Figure (17): Stream Lines Indicating Average Flow Speed in the Model with Various Nose shapes, Measured at Mid-Depth and at the Flow Surface Level, at a Flow Rate of 78 Liters per Second.

Conducting experimental and numerical studies to analyze theimpact of the base nose shape on flow hydraulics in PKW weirusing FLOW-3D

FLOW-3D를 사용하여 PKW 둑의 흐름 수력학에 대한 베이스 노즈 모양의 영향을 분석하기 위한 실험 및 수치 연구 수행

Behshad Mardasi 1
Rasoul Ilkhanipour Zeynali 2
Majid Heydari 3

Abstract

Weirs are essential structures used to manage excess water flow from behind dams to downstream areas. Enhancing discharge efficiency often involves extending the effective length of Piano Key Weirs (PKW) in dams or regulating flow within irrigation and drainage networks. This study employed both numerical and laboratory investigations to assess the impact of different base nose shapes installed beneath the outlet keys and varying Input to output key width ratios (Wi/Wo) on discharges ranging from 5 to 80 liters per second. Furthermore, the study aimed to achieve research objectives and compare the performance of Piano Key Weirs with Ogee Weir. For numerical simulation, the optimal number of cells for meshing was determined, and an appropriate turbulence model was selected. The results indicated that the numerical model accurately simulated the laboratory sample with a high degree of precision. Moreover, the numerical model closely approximated PKW for all parameters Q, H, and Cd compared to the laboratory sample. The findings revealed that in laboratory models with a maximum discharge area of 80 liters per second, the weir with Wi/Wo=1.2 and a flow head value of 285 mm exhibited the lowest value, whereas the weir with Wi/Wo=0.71 and a flow head value of 305 mm showed the highest, attributed to the higher discharge in the input-output ratio. Additionally, as the ratio of flow head to weir height H/P increased, the discharge coefficient Cd decreased. Comparing the flow conditions in weirs with different base nose shapes, it was observed that the weir with a spindle nose shape (PKW1.2S) outperformed the PKW with a flat (PKW1.2), semi-cylindrical (PKW1.2CL) and triangular base nose (PKW1.2TR). The results emphasized that models featuring semi-cylindrical and flat noses exhibited notable flow deviation and abrupt disruption upon impact with the nose. However, this effect was significantly reduced in models equipped with triangular and spindle-shaped noses. Also, the coefficient of discharge in PKW1.2S and PKW1.2TR weirs, compared to the PKW1.20 weir, increased by 27% and 20%, respectively.

웨어는 댐 뒤에서 하류 지역으로의 과도한 물 흐름을 관리하는 데 사용되는 필수 구조물입니다. 배출 효율을 높이는 데에는 댐의 피아노 키 위어(PKW) 유효 길이를 연장하거나 관개 및 배수 네트워크 내 흐름을 조절하는 것이 포함됩니다.

이 연구에서는 콘센트 키 아래에 설치된 다양한 베이스 노즈 모양과 초당 5~80리터 범위의 배출에 대한 다양한 입력 대 출력 키 너비 비율(Wi/Wo)의 영향을 평가하기 위해 수치 및 실험실 조사를 모두 사용했습니다. 또한 본 연구에서는 연구 목적을 달성하고 Piano Key Weir와 Ogee Weir의 성능을 비교하는 것을 목표로 했습니다.

수치 시뮬레이션을 위해 메시 생성을 위한 최적의 셀 수를 결정하고 적절한 난류 모델을 선택했습니다. 결과는 수치 모델이 높은 정밀도로 실험실 샘플을 정확하게 시뮬레이션했음을 나타냅니다. 더욱이, 수치 모델은 실험실 샘플과 비교하여 모든 매개변수 Q, H 및 Cd에 대해 PKW에 매우 근접했습니다.

연구 결과, 최대 배출 면적이 초당 80리터인 실험실 모델에서는 Wi/Wo=1.2, 플로우 헤드 값이 285mm인 웨어가 가장 낮은 값을 나타냈고, Wi/Wo=0.71 및 a인 웨어는 가장 낮은 값을 나타냈습니다. 플로우 헤드 값은 305mm로 가장 높은 것으로 나타났는데, 이는 입출력 비율의 높은 토출량에 기인합니다. 또한, 웨어 높이에 대한 유수두 비율 H/P가 증가함에 따라 유출계수 Cd는 감소하였다.

베이스 노즈 모양이 다른 웨어의 흐름 조건을 비교해 보면, 스핀들 노즈 모양(PKW1.2S)의 웨어가 평면(PKW1.2), 반원통형(PKW1.2CL) 및 삼각형 모양의 PKW보다 성능이 우수한 것으로 관찰되었습니다. 베이스 노즈(PKW1.2TR) 결과는 반원통형 및 편평한 노즈를 특징으로 하는 모델이 노즈에 충격을 가할 때 눈에 띄는 흐름 편차와 급격한 중단을 나타냄을 강조했습니다.

그러나 삼각형 및 방추형 노즈를 장착한 모델에서는 이러한 효과가 크게 감소했습니다. 또한 PKW1.20보에 비해 PKW1.2S보와 PKW1.2TR보의 유출계수는 각각 27%, 20% 증가하였다.

Keywords

Piano Key Weir, Base Nose Shape, Flow Hydraulics, Numerical Model, Triangular
Nose Shape, Flat Nose Shape, Semi-Cylindrical Nose Shape, Spindle Nose Shape

Figure (17): Stream Lines Indicating Average Flow Speed in the Model with Various Nose shapes, Measured at Mid-Depth and at the Flow Surface Level, at a Flow Rate of 78 Liters per Second.
Figure (17): Stream Lines Indicating Average Flow Speed in the Model with Various Nose shapes, Measured at Mid-Depth and at the Flow Surface Level, at a Flow Rate of 78 Liters per Second.

Reference

  1. Chow, V.T. (1959). “Open channel hydraulics.” McGraw-Hill Book Company, New York,
    NY.
  2. Ouamane, A., and Lempérière, F. (2006). “Design of a new economic shape of weir.” Proc.,
    Intl. Symp. on Dams in the Societies of the 21st Century, 463-470, Barcelona, Spain.
  3. Crookston, B. M., Anderson, A., Shearin-Feimster, L., and Tullis, B. P. (2014). “Mitigation
    investigation of flow-induced vibrations at a rehabilitated spillway.” Proc., 5th IAHR Intl.
    Symp. on Hydraulic Structures, Univ. of Queensland Brisbane, Brisbane, Australia.
  4. Machiels, O. (2012). “Experimental study of the hydraulic behaviour of Piano Key Weirs.”
    Ph.D. Dissertation, Faculty of Applied Science, University of Liège, Liège, Belgium.
  5. Blanc, P., and Lempérière, F. (2001). “Labyrinth spillways have a promising future.” Intl. J.
    of Hydropower and Dams, 8(4), 129-131.
  6. Muslu, Y. (2001). “Numerical analysis for lateral weir flow.” J. of Irrigation and Drainage
    Eng., ASCE, 127, 246.
  7. Erpicum, S., Machiels, O., Dewals, B., Pirotton, M., and Archambeau, P. (2012).
    “Numerical and physical hydraulic modeling of Piano Key Weirs.” Proc., ASIA 2012 – 4th
    Intl. Conf. on Water Resources and Renewable Energy Development in Asia, Chiang Mai,
    Thailand.
  8. Tullis, J.P., Amanian, N., and Waldron, D. (1995). “Design of Labyrinth Spillways.” J. of
    Hydraulic Eng., ASCE, 121.
  9. Lux, F.L., and Hinchcliff, D. (1985). “Design and construction of labyrinth spillways.”
    Proc., 15th Intl. Congress on Large Dams, ICOLD, Vol. 4, 249-274, Paris, France.
  10. Erpicum, S., Laugier, F., Ho to Khanh, M., & Pfister, M. (2017). Labyrinth and Piano Key
    Weirs III–PKW 2017. CRC Press, Boca Raton, FL.
  11. Kabiri-Samani, A., and Javaheri, A. (2012). “Discharge coefficient for free and submerged flow over Piano Key weirs.” Hydraulic Research J., 50(1), 114-120.
  12. Hien, T.C., Son, H.T., and Khanh, M.H.T. (2006). “Results of some piano Key weirs
    hydraulic model tests in Vietnam.” Proc., 22nd ICOLD Congress, CIGB/ICOLD,
    Barcelona, Spain.
  13. Laugier, F., Lochu, A., Gille, C., Leite Ribeiro, M., and Boillat, J-L. (2009). “Design and
    construction of a labyrinth PKW spillway at St-Marc Dam.” Hydropower and Dams J.,
    15(5), 100-107.
  14. Cicero, G.M., Menon, J.M., Luck, M., and Pinchard, T. (2011). “Experimental study of side
    and scale effects on hydraulic performances of a Piano Key Weir.” In: Erpicum, S., Laugier,
    F., Boillat, J-L, Pirotton, M., Reverchon, B., and Schleiss, A-J (Eds.), Labyrinth and Piano
    Key Weirs, 167-172, CRC Press, London.
  15. Pralong, J., Vermeulen, J., Blancher, B., Laugier, F., Erpicum, S., Machiels, O., Pirotton,
    M., Boillat, J.L, Leite Ribeiro, M., and Schleiss, A.J. (2011). “A naming convention for the
    piano key weirs geometrical parameters.” In: Erpicum, S., Laugier, F., Boillat, J-L, Pirotton,
    M., Reverchon, B., and Schleiss, A-J (Eds.), Labyrinth and Piano Key Weirs, 271-278,
    CRC Press, London.
  16. Denys, F. J. M., and Basson, G. R. (2018). “Transient hydrodynamics of Piano Key Weirs.”
    Proc., 7th IAHR Intl. Symp. on Hydraulic Structures, ISHS2018, 518-527,
    DigitalCommons@USU, Logan, UT.
  17. Anderson, A., and Tullis, B. P. (2018). “Finite crest length weir nappe oscillation.” J. of
    Hydraulic Eng., ASCE, 144(6), 04018020. https://doi.org/10.1061/(ASCE)HY.1943-
    7900.0001461
  18. Erpicum, S., Laugier, F., Boillat, J.-L., Pirotton, M., Reverchon, B., and Schleiss, A. J.
    (2011). “Labyrinth and Piano Key Weirs–PKW 2011.” CRC Press, Boca Raton, FL.
  19. Aydin, C.M., and Emiroglu, M.E. (2011). “Determination of capacity of labyrinth side weir
    by CFD.” Flow Measurement and Instrumentation, 29, 1-8.
  20. Cicero, G.M., Delisle, J.R., Lefebvre, V., and Vermeulen, J. (2013). “Experimental and
    numerical study of the hydraulic performance of a trapezoidal PKW.” Proc., Intl. Workshop
    on Labyrinths and Piano Key Weirs PKW II 2013, 265-272, CRC Press.
  21. Anderson, R. M. (2011). “Piano Key Weir Head Discharge Relationships.” Master’s Thesis,
    Utah State University, Logan, Utah.
  22. Crookston, B.M., Anderson, R.M., and Tullis, B.P. (2018). “Free-flow discharge estimation
    method for Piano Key weir geometries.” J. of Hydro-environment Research, 19, 160-167

Lab-on-a-Chip 시스템의 혈류 역학에 대한 검토: 엔지니어링 관점

Review on Blood Flow Dynamics in Lab-on-a-Chip Systems: An Engineering Perspective

  • Bin-Jie Lai
  • Li-Tao Zhu
  • Zhe Chen*
  • Bo Ouyang*
  • , and 
  • Zheng-Hong Luo*

Abstract

다양한 수송 메커니즘 하에서, “LOC(lab-on-a-chip)” 시스템에서 유동 전단 속도 조건과 밀접한 관련이 있는 혈류 역학은 다양한 수송 현상을 초래하는 것으로 밝혀졌습니다.

본 연구는 적혈구의 동적 혈액 점도 및 탄성 거동과 같은 점탄성 특성의 역할을 통해 LOC 시스템의 혈류 패턴을 조사합니다. 모세관 및 전기삼투압의 주요 매개변수를 통해 LOC 시스템의 혈액 수송 현상에 대한 연구는 실험적, 이론적 및 수많은 수치적 접근 방식을 통해 제공됩니다.

전기 삼투압 점탄성 흐름에 의해 유발되는 교란은 특히 향후 연구 기회를 위해 혈액 및 기타 점탄성 유체를 취급하는 LOC 장치의 혼합 및 분리 기능 향상에 논의되고 적용됩니다. 또한, 본 연구는 보다 정확하고 단순화된 혈류 모델에 대한 요구와 전기역학 효과 하에서 점탄성 유체 흐름에 대한 수치 연구에 대한 강조와 같은 LOC 시스템 하에서 혈류 역학의 수치 모델링의 문제를 식별합니다.

전기역학 현상을 연구하는 동안 제타 전위 조건에 대한 보다 실용적인 가정도 강조됩니다. 본 연구는 모세관 및 전기삼투압에 의해 구동되는 미세유체 시스템의 혈류 역학에 대한 포괄적이고 학제적인 관점을 제공하는 것을 목표로 한다.

KEYWORDS: 

1. Introduction

1.1. Microfluidic Flow in Lab-on-a-Chip (LOC) Systems

Over the past several decades, the ability to control and utilize fluid flow patterns at microscales has gained considerable interest across a myriad of scientific and engineering disciplines, leading to growing interest in scientific research of microfluidics. 

(1) Microfluidics, an interdisciplinary field that straddles physics, engineering, and biotechnology, is dedicated to the behavior, precise control, and manipulation of fluids geometrically constrained to a small, typically submillimeter, scale. 

(2) The engineering community has increasingly focused on microfluidics, exploring different driving forces to enhance working fluid transport, with the aim of accurately and efficiently describing, controlling, designing, and applying microfluidic flow principles and transport phenomena, particularly for miniaturized applications. 

(3) This attention has chiefly been fueled by the potential to revolutionize diagnostic and therapeutic techniques in the biomedical and pharmaceutical sectorsUnder various driving forces in microfluidic flows, intriguing transport phenomena have bolstered confidence in sustainable and efficient applications in fields such as pharmaceutical, biochemical, and environmental science. The “lab-on-a-chip” (LOC) system harnesses microfluidic flow to enable fluid processing and the execution of laboratory tasks on a chip-sized scale. LOC systems have played a vital role in the miniaturization of laboratory operations such as mixing, chemical reaction, separation, flow control, and detection on small devices, where a wide variety of fluids is adapted. Biological fluid flow like blood and other viscoelastic fluids are notably studied among the many working fluids commonly utilized by LOC systems, owing to the optimization in small fluid sample volumed, rapid response times, precise control, and easy manipulation of flow patterns offered by the system under various driving forces. 

(4)The driving forces in blood flow can be categorized as passive or active transport mechanisms and, in some cases, both. Under various transport mechanisms, the unique design of microchannels enables different functionalities in driving, mixing, separating, and diagnosing blood and drug delivery in the blood. 

(5) Understanding and manipulating these driving forces are crucial for optimizing the performance of a LOC system. Such knowledge presents the opportunity to achieve higher efficiency and reliability in addressing cellular level challenges in medical diagnostics, forensic studies, cancer detection, and other fundamental research areas, for applications of point-of-care (POC) devices. 

(6)

1.2. Engineering Approach of Microfluidic Transport Phenomena in LOC Systems

Different transport mechanisms exhibit unique properties at submillimeter length scales in microfluidic devices, leading to significant transport phenomena that differ from those of macroscale flows. An in-depth understanding of these unique transport phenomena under microfluidic systems is often required in fluidic mechanics to fully harness the potential functionality of a LOC system to obtain systematically designed and precisely controlled transport of microfluids under their respective driving force. Fluid mechanics is considered a vital component in chemical engineering, enabling the analysis of fluid behaviors in various unit designs, ranging from large-scale reactors to separation units. Transport phenomena in fluid mechanics provide a conceptual framework for analytically and descriptively explaining why and how experimental results and physiological phenomena occur. The Navier–Stokes (N–S) equation, along with other governing equations, is often adapted to accurately describe fluid dynamics by accounting for pressure, surface properties, velocity, and temperature variations over space and time. In addition, limiting factors and nonidealities for these governing equations should be considered to impose corrections for empirical consistency before physical models are assembled for more accurate controls and efficiency. Microfluidic flow systems often deviate from ideal conditions, requiring adjustments to the standard governing equations. These deviations could arise from factors such as viscous effects, surface interactions, and non-Newtonian fluid properties from different microfluid types and geometrical layouts of microchannels. Addressing these nonidealities supports the refining of theoretical models and prediction accuracy for microfluidic flow behaviors.

The analytical calculation of coupled nonlinear governing equations, which describes the material and energy balances of systems under ideal conditions, often requires considerable computational efforts. However, advancements in computation capabilities, cost reduction, and improved accuracy have made numerical simulations using different numerical and modeling methods a powerful tool for effectively solving these complex coupled equations and modeling various transport phenomena. Computational fluid dynamics (CFD) is a numerical technique used to investigate the spatial and temporal distribution of various flow parameters. It serves as a critical approach to provide insights and reasoning for decision-making regarding the optimal designs involving fluid dynamics, even prior to complex physical model prototyping and experimental procedures. The integration of experimental data, theoretical analysis, and reliable numerical simulations from CFD enables systematic variation of analytical parameters through quantitative analysis, where adjustment to delivery of blood flow and other working fluids in LOC systems can be achieved.

Numerical methods such as the Finite-Difference Method (FDM), Finite-Element-Method (FEM), and Finite-Volume Method (FVM) are heavily employed in CFD and offer diverse approaches to achieve discretization of Eulerian flow equations through filling a mesh of the flow domain. A more in-depth review of numerical methods in CFD and its application for blood flow simulation is provided in Section 2.2.2.

1.3. Scope of the Review

In this Review, we explore and characterize the blood flow phenomena within the LOC systems, utilizing both physiological and engineering modeling approaches. Similar approaches will be taken to discuss capillary-driven flow and electric-osmotic flow (EOF) under electrokinetic phenomena as a passive and active transport scheme, respectively, for blood transport in LOC systems. Such an analysis aims to bridge the gap between physical (experimental) and engineering (analytical) perspectives in studying and manipulating blood flow delivery by different driving forces in LOC systems. Moreover, the Review hopes to benefit the interests of not only blood flow control in LOC devices but also the transport of viscoelastic fluids, which are less studied in the literature compared to that of Newtonian fluids, in LOC systems.

Section 2 examines the complex interplay between viscoelastic properties of blood and blood flow patterns under shear flow in LOC systems, while engineering numerical modeling approaches for blood flow are presented for assistance. Sections 3 and 4 look into the theoretical principles, numerical governing equations, and modeling methodologies for capillary driven flow and EOF in LOC systems as well as their impact on blood flow dynamics through the quantification of key parameters of the two driving forces. Section 5 concludes the characterized blood flow transport processes in LOC systems under these two forces. Additionally, prospective areas of research in improving the functionality of LOC devices employing blood and other viscoelastic fluids and potentially justifying mechanisms underlying microfluidic flow patterns outside of LOC systems are presented. Finally, the challenges encountered in the numerical studies of blood flow under LOC systems are acknowledged, paving the way for further research.

2. Blood Flow Phenomena

ARTICLE SECTIONS

Jump To


2.1. Physiological Blood Flow Behavior

Blood, an essential physiological fluid in the human body, serves the vital role of transporting oxygen and nutrients throughout the body. Additionally, blood is responsible for suspending various blood cells including erythrocytes (red blood cells or RBCs), leukocytes (white blood cells), and thrombocytes (blood platelets) in a plasma medium.Among the cells mentioned above, red blood cells (RBCs) comprise approximately 40–45% of the volume of healthy blood. 

(7) An RBC possesses an inherent elastic property with a biconcave shape of an average diameter of 8 μm and a thickness of 2 μm. This biconcave shape maximizes the surface-to-volume ratio, allowing RBCs to endure significant distortion while maintaining their functionality. 

(8,9) Additionally, the biconcave shape optimizes gas exchange, facilitating efficient uptake of oxygen due to the increased surface area. The inherent elasticity of RBCs allows them to undergo substantial distortion from their original biconcave shape and exhibits high flexibility, particularly in narrow channels.RBC deformability enables the cell to deform from a biconcave shape to a parachute-like configuration, despite minor differences in RBC shape dynamics under shear flow between initial cell locations. As shown in Figure 1(a), RBCs initiating with different resting shapes and orientations displaying display a similar deformation pattern 

(10) in terms of its shape. Shear flow induces an inward bending of the cell at the rear position of the rim to the final bending position, 

(11) resulting in an alignment toward the same position of the flow direction.

Figure 1. Images of varying deformation of RBCs and different dynamic blood flow behaviors. (a) The deforming shape behavior of RBCs at four different initiating positions under the same experimental conditions of a flow from left to right, (10) (b) RBC aggregation, (13) (c) CFL region. (18) Reproduced with permission from ref (10). Copyright 2011 Elsevier. Reproduced with permission from ref (13). Copyright 2022 The Authors, under the terms of the Creative Commons (CC BY 4.0) License https://creativecommons.org/licenses/by/4.0/. Reproduced with permission from ref (18). Copyright 2019 Elsevier.

The flexible property of RBCs enables them to navigate through narrow capillaries and traverse a complex network of blood vessels. The deformability of RBCs depends on various factors, including the channel geometry, RBC concentration, and the elastic properties of the RBC membrane. 

(12) Both flexibility and deformability are vital in the process of oxygen exchange among blood and tissues throughout the body, allowing cells to flow in vessels even smaller than the original cell size prior to deforming.As RBCs serve as major components in blood, their collective dynamics also hugely affect blood rheology. RBCs exhibit an aggregation phenomenon due to cell to cell interactions, such as adhesion forces, among populated cells, inducing unique blood flow patterns and rheological behaviors in microfluidic systems. For blood flow in large vessels between a diameter of 1 and 3 cm, where shear rates are not high, a constant viscosity and Newtonian behavior for blood can be assumed. However, under low shear rate conditions (0.1 s

–1) in smaller vessels such as the arteries and venules, which are within a diameter of 0.2 mm to 1 cm, blood exhibits non-Newtonian properties, such as shear-thinning viscosity and viscoelasticity due to RBC aggregation and deformability. The nonlinear viscoelastic property of blood gives rise to a complex relationship between viscosity and shear rate, primarily influenced by the highly elastic behavior of RBCs. A wide range of research on the transient behavior of the RBC shape and aggregation characteristics under varied flow circumstances has been conducted, aiming to obtain a better understanding of the interaction between blood flow shear forces from confined flows.

For a better understanding of the unique blood flow structures and rheological behaviors in microfluidic systems, some blood flow patterns are introduced in the following section.

2.1.1. RBC Aggregation

RBC aggregation is a vital phenomenon to be considered when designing LOC devices due to its impact on the viscosity of the bulk flow. Under conditions of low shear rate, such as in stagnant or low flow rate regions, RBCs tend to aggregate, forming structures known as rouleaux, resembling stacks of coins as shown in Figure 1(b). 

(13) The aggregation of RBCs increases the viscosity at the aggregated region, 

(14) hence slowing down the overall blood flow. However, when exposed to high shear rates, RBC aggregates disaggregate. As shear rates continue to increase, RBCs tend to deform, elongating and aligning themselves with the direction of the flow. 

(15) Such a dynamic shift in behavior from the cells in response to the shear rate forms the basis of the viscoelastic properties observed in whole blood. In essence, the viscosity of the blood varies according to the shear rate conditions, which are related to the velocity gradient of the system. It is significant to take the intricate relationship between shear rate conditions and the change of blood viscosity due to RBC aggregation into account since various flow driving conditions may induce varied effects on the degree of aggregation.

2.1.2. Fåhræus-Lindqvist Effect

The Fåhræus–Lindqvist (FL) effect describes the gradual decrease in the apparent viscosity of blood as the channel diameter decreases. 

(16) This effect is attributed to the migration of RBCs toward the central region in the microchannel, where the flow rate is higher, due to the presence of higher pressure and asymmetric distribution of shear forces. This migration of RBCs, typically observed at blood vessels less than 0.3 mm, toward the higher flow rate region contributes to the change in blood viscosity, which becomes dependent on the channel size. Simultaneously, the increase of the RBC concentration in the central region of the microchannel results in the formation of a less viscous region close to the microchannel wall. This region called the Cell-Free Layer (CFL), is primarily composed of plasma. 

(17) The combination of the FL effect and the following CFL formation provides a unique phenomenon that is often utilized in passive and active plasma separation mechanisms, involving branched and constriction channels for various applications in plasma separation using microfluidic systems.

2.1.3. Cell-Free Layer Formation

In microfluidic blood flow, RBCs form aggregates at the microchannel core and result in a region that is mostly devoid of RBCs near the microchannel walls, as shown in Figure 1(c). 

(18) The region is known as the cell-free layer (CFL). The CFL region is often known to possess a lower viscosity compared to other regions within the blood flow due to the lower viscosity value of plasma when compared to that of the aggregated RBCs. Therefore, a thicker CFL region composed of plasma correlates to a reduced apparent whole blood viscosity. 

(19) A thicker CFL region is often established following the RBC aggregation at the microchannel core under conditions of decreasing the tube diameter. Apart from the dependence on the RBC concentration in the microchannel core, the CFL thickness is also affected by the volume concentration of RBCs, or hematocrit, in whole blood, as well as the deformability of RBCs. Given the influence CFL thickness has on blood flow rheological parameters such as blood flow rate, which is strongly dependent on whole blood viscosity, investigating CFL thickness under shear flow is crucial for LOC systems accounting for blood flow.

2.1.4. Plasma Skimming in Bifurcation Networks

The uneven arrangement of RBCs in bifurcating microchannels, commonly termed skimming bifurcation, arises from the axial migration of RBCs within flowing streams. This uneven distribution contributes to variations in viscosity across differing sizes of bifurcating channels but offers a stabilizing effect. Notably, higher flow rates in microchannels are associated with increased hematocrit levels, resulting in higher viscosity compared with those with lower flow rates. Parametric investigations on bifurcation angle, 

(20) thickness of the CFL, 

(21) and RBC dynamics, including aggregation and deformation, 

(22) may alter the varying viscosity of blood and its flow behavior within microchannels.

2.2. Modeling on Blood Flow Dynamics

2.2.1. Blood Properties and Mathematical Models of Blood Rheology

Under different shear rate conditions in blood flow, the elastic characteristics and dynamic changes of the RBC induce a complex velocity and stress relationship, resulting in the incompatibility of blood flow characterization through standard presumptions of constant viscosity used for Newtonian fluid flow. Blood flow is categorized as a viscoelastic non-Newtonian fluid flow where constitutive equations governing this type of flow take into consideration the nonlinear viscometric properties of blood. To mathematically characterize the evolving blood viscosity and the relationship between the elasticity of RBC and the shear blood flow, respectively, across space and time of the system, a stress tensor (τ) defined by constitutive models is often coupled in the Navier–Stokes equation to account for the collective impact of the constant dynamic viscosity (η) and the elasticity from RBCs on blood flow.The dynamic viscosity of blood is heavily dependent on the shear stress applied to the cell and various parameters from the blood such as hematocrit value, plasma viscosity, mechanical properties of the RBC membrane, and red blood cell aggregation rate. The apparent blood viscosity is considered convenient for the characterization of the relationship between the evolving blood viscosity and shear rate, which can be defined by Casson’s law, as shown in eq 1.

𝜇=𝜏0𝛾˙+2𝜂𝜏0𝛾˙⎯⎯⎯⎯⎯⎯⎯√+𝜂�=�0�˙+2��0�˙+�

(1)where τ

0 is the yield stress–stress required to initiate blood flow motion, η is the Casson rheological constant, and γ̇ is the shear rate. The value of Casson’s law parameters under blood with normal hematocrit level can be defined as τ

0 = 0.0056 Pa and η = 0.0035 Pa·s. 

(23) With the known property of blood and Casson’s law parameters, an approximation can be made to the dynamic viscosity under various flow condition domains. The Power Law model is often employed to characterize the dynamic viscosity in relation to the shear rate, since precise solutions exist for specific geometries and flow circumstances, acting as a fundamental standard for definition. The Carreau and Carreau–Yasuda models can be advantageous over the Power Law model due to their ability to evaluate the dynamic viscosity at low to zero shear rate conditions. However, none of the above-mentioned models consider the memory or other elastic behavior of blood and its RBCs. Some other commonly used mathematical models and their constants for the non-Newtonian viscosity property characterization of blood are listed in Table 1 below. 

(24−26)Table 1. Comparison of Various Non-Newtonian Models for Blood Viscosity 

(24−26)

ModelNon-Newtonian ViscosityParameters
Power Law(2)n = 0.61, k = 0.42
Carreau(3)μ0 = 0.056 Pa·s, μ = 0.00345 Pa·s, λ = 3.1736 s, m = 2.406, a = 0.254
Walburn–Schneck(4)C1 = 0.000797 Pa·s, C2 = 0.0608 Pa·s, C3 = 0.00499, C4 = 14.585 g–1, TPMA = 25 g/L
Carreau–Yasuda(5)μ0 = 0.056 Pa·s, μ = 0.00345 Pa·s, λ = 1.902 s, n = 0.22, a = 1.25
Quemada(6)μp = 0.0012 Pa·s, k = 2.07, k0 = 4.33, γ̇c = 1.88 s–1

The blood rheology is commonly known to be influenced by two key physiological factors, namely, the hematocrit value (H

t) and the fibrinogen concentration (c

f), with an average value of 42% and 0.252 gd·L

–1, respectively. Particularly in low shear conditions, the presence of varying fibrinogen concentrations affects the tendency for aggregation and rouleaux formation, while the occurrence of aggregation is contingent upon specific levels of hematocrit. 

(27) The study from Apostolidis et al. 

(28) modifies the Casson model through emphasizing its reliance on hematocrit and fibrinogen concentration parameter values, owing to the extensive knowledge of the two physiological blood parameters.The viscoelastic response of blood is heavily dependent on the elasticity of the RBC, which is defined by the relationship between the deformation and stress relaxation from RBCs under a specific location of shear flow as a function of the velocity field. The stress tensor is usually characterized by constitutive equations such as the Upper-Convected Maxwell Model 

(29) and the Oldroyd-B model 

(30) to track the molecule effects under shear from different driving forces. The prominent non-Newtonian features, such as shear thinning and yield stress, have played a vital role in the characterization of blood rheology, particularly with respect to the evaluation of yield stress under low shear conditions. The nature of stress measurement in blood, typically on the order of 1 mPa, is challenging due to its low magnitude. The occurrence of the CFL complicates the measurement further due to the significant decrease in apparent viscosity near the wall over time and a consequential disparity in viscosity compared to the bulk region.In addition to shear thinning viscosity and yield stress, the formation of aggregation (rouleaux) from RBCs under low shear rates also contributes to the viscoelasticity under transient flow 

(31) and thixotropy 

(32) of whole blood. Given the difficulty in evaluating viscoelastic behavior of blood under low strain magnitudes and limitations in generalized Newtonian models, the utilization of viscoelastic models is advocated to encompass elasticity and delineate non-shear components within the stress tensor. Extending from the Oldroyd-B model, Anand et al. 

(33) developed a viscoelastic model framework for adapting elasticity within blood samples and predicting non-shear stress components. However, to also address the thixotropic effects, the model developed by Horner et al. 

(34) serves as a more comprehensive approach than the viscoelastic model from Anand et al. Thixotropy 

(32) typically occurs from the structural change of the rouleaux, where low shear rate conditions induce rouleaux formation. Correspondingly, elasticity increases, while elasticity is more representative of the isolated RBCs, under high shear rate conditions. The model of Horner et al. 

(34) considers the contribution of rouleaux to shear stress, taking into account factors such as the characteristic time for Brownian aggregation, shear-induced aggregation, and shear-induced breakage. Subsequent advancements in the model from Horner et al. often revolve around refining the three aforementioned key terms for a more substantial characterization of rouleaux dynamics. Notably, this has led to the recently developed mHAWB model 

(35) and other model iterations to enhance the accuracy of elastic and viscoelastic contributions to blood rheology, including the recently improved model suggested by Armstrong et al. 

(36)

2.2.2. Numerical Methods (FDM, FEM, FVM)

Numerical simulation has become increasingly more significant in analyzing the geometry, boundary layers of flow, and nonlinearity of hyperbolic viscoelastic flow constitutive equations. CFD is a powerful and efficient tool utilizing numerical methods to solve the governing hydrodynamic equations, such as the Navier–Stokes (N–S) equation, continuity equation, and energy conservation equation, for qualitative evaluation of fluid motion dynamics under different parameters. CFD overcomes the challenge of analytically solving nonlinear forms of differential equations by employing numerical methods such as the Finite-Difference Method (FDM), Finite-Element Method (FEM), and Finite-Volume Method (FVM) to discretize and solve the partial differential equations (PDEs), allowing for qualitative reproduction of transport phenomena and experimental observations. Different numerical methods are chosen to cope with various transport systems for optimization of the accuracy of the result and control of error during the discretization process.FDM is a straightforward approach to discretizing PDEs, replacing the continuum representation of equations with a set of finite-difference equations, which is typically applied to structured grids for efficient implementation in CFD programs. 

(37) However, FDM is often limited to simple geometries such as rectangular or block-shaped geometries and struggles with curved boundaries. In contrast, FEM divides the fluid domain into small finite grids or elements, approximating PDEs through a local description of physics. 

(38) All elements contribute to a large, sparse matrix solver. However, FEM may not always provide accurate results for systems involving significant deformation and aggregation of particles like RBCs due to large distortion of grids. 

(39) FVM evaluates PDEs following the conservation laws and discretizes the selected flow domain into small but finite size control volumes, with each grid at the center of a finite volume. 

(40) The divergence theorem allows the conversion of volume integrals of PDEs with divergence terms into surface integrals of surface fluxes across cell boundaries. Due to its conservation property, FVM offers efficient outcomes when dealing with PDEs that embody mass, momentum, and energy conservation principles. Furthermore, widely accessible software packages like the OpenFOAM toolbox 

(41) include a viscoelastic solver, making it an attractive option for viscoelastic fluid flow modeling. 

(42)

2.2.3. Modeling Methods of Blood Flow Dynamics

The complexity in the blood flow simulation arises from deformability and aggregation that RBCs exhibit during their interaction with neighboring cells under different shear rate conditions induced by blood flow. Numerical models coupled with simulation programs have been applied as a groundbreaking method to predict such unique rheological behavior exhibited by RBCs and whole blood. The conventional approach of a single-phase flow simulation is often applied to blood flow simulations within large vessels possessing a moderate shear rate. However, such a method assumes the properties of plasma, RBCs and other cellular components to be evenly distributed as average density and viscosity in blood, resulting in the inability to simulate the mechanical dynamics, such as RBC aggregation under high-shear flow field, inherent in RBCs. To accurately describe the asymmetric distribution of RBC and blood flow, multiphase flow simulation, where numerical simulations of blood flows are often modeled as two immiscible phases, RBCs and blood plasma, is proposed. A common assumption is that RBCs exhibit non-Newtonian behavior while the plasma is treated as a continuous Newtonian phase.Numerous multiphase numerical models have been proposed to simulate the influence of RBCs on blood flow dynamics by different assumptions. In large-scale simulations (above the millimeter range), continuum-based methods are wildly used due to their lower computational demands. 

(43) Eulerian multiphase flow simulations offer the solution of a set of conservation equations for each separate phase and couple the phases through common pressure and interphase exchange coefficients. Xu et al. 

(44) utilized the combined finite-discrete element method (FDEM) to replicate the dynamic behavior and distortion of RBCs subjected to fluidic forces, utilizing the Johnson–Kendall–Roberts model 

(45) to define the adhesive forces of cell-to-cell interactions. The iterative direct-forcing immersed boundary method (IBM) is commonly employed in simulations of the fluid–cell interface of blood. This method effectively captures the intricacies of the thin and flexible RBC membranes within various external flow fields. 

(46) The study by Xu et al. 

(44) also adopts this approach to bridge the fluid dynamics and RBC deformation through IBM. Yoon and You utilized the Maxwell model to define the viscosity of the RBC membrane. 

(47) It was discovered that the Maxwell model could represent the stress relaxation and unloading processes of the cell. Furthermore, the reduced flexibility of an RBC under particular situations such as infection is specified, which was unattainable by the Kelvin–Voigt model 

(48) when compared to the Maxwell model in the literature. The Yeoh hyperplastic material model was also adapted to predict the nonlinear elasticity property of RBCs with FEM employed to discretize the RBC membrane using shell-type elements. Gracka et al. 

(49) developed a numerical CFD model with a finite-volume parallel solver for multiphase blood flow simulation, where an updated Maxwell viscoelasticity model and a Discrete Phase Model are adopted. In the study, the adapted IBM, based on unstructured grids, simulates the flow behavior and shape change of the RBCs through fluid-structure coupling. It was found that the hybrid Euler–Lagrange (E–L) approach 

(50) for the development of the multiphase model offered better results in the simulated CFL region in the microchannels.To study the dynamics of individual behaviors of RBCs and the consequent non-Newtonian blood flow, cell-shape-resolved computational models are often adapted. The use of the boundary integral method has become prevalent in minimizing computational expenses, particularly in the exclusive determination of fluid velocity on the surfaces of RBCs, incorporating the option of employing IBM or particle-based techniques. The cell-shaped-resolved method has enabled an examination of cell to cell interactions within complex ambient or pulsatile flow conditions 

(51) surrounding RBC membranes. Recently, Rydquist et al. 

(52) have looked to integrate statistical information from macroscale simulations to obtain a comprehensive overview of RBC behavior within the immediate proximity of the flow through introduction of respective models characterizing membrane shape definition, tension, bending stresses of RBC membranes.At a macroscopic scale, continuum models have conventionally been adapted for assessing blood flow dynamics through the application of elasticity theory and fluid dynamics. However, particle-based methods are known for their simplicity and adaptability in modeling complex multiscale fluid structures. Meshless methods, such as the boundary element method (BEM), smoothed particle hydrodynamics (SPH), and dissipative particle dynamics (DPD), are often used in particle-based characterization of RBCs and the surrounding fluid. By representing the fluid as discrete particles, meshless methods provide insights into the status and movement of the multiphase fluid. These methods allow for the investigation of cellular structures and microscopic interactions that affect blood rheology. Non-confronting mesh methods like IBM can also be used to couple a fluid solver such as FEM, FVM, or the Lattice Boltzmann Method (LBM) through membrane representation of RBCs. In comparison to conventional CFD methods, LBM has been viewed as a favorable numerical approach for solving the N–S equations and the simulation of multiphase flows. LBM exhibits the notable advantage of being amenable to high-performance parallel computing environments due to its inherently local dynamics. In contrast to DPD and SPH where RBC membranes are modeled as physically interconnected particles, LBM employs the IBM to account for the deformation dynamics of RBCs 

(53,54) under shear flows in complex channel geometries. 

(54,55) However, it is essential to acknowledge that the utilization of LBM in simulating RBC flows often entails a significant computational overhead, being a primary challenge in this context. Krüger et al. 

(56) proposed utilizing LBM as a fluid solver, IBM to couple the fluid and FEM to compute the response of membranes to deformation under immersed fluids. This approach decouples the fluid and membranes but necessitates significant computational effort due to the requirements of both meshes and particles.Despite the accuracy of current blood flow models, simulating complex conditions remains challenging because of the high computational load and cost. Balachandran Nair et al. 

(57) suggested a reduced order model of RBC under the framework of DEM, where the RBC is represented by overlapping constituent rigid spheres. The Morse potential force is adapted to account for the RBC aggregation exhibited by cell to cell interactions among RBCs at different distances. Based upon the IBM, the reduced-order RBC model is adapted to simulate blood flow transport for validation under both single and multiple RBCs with a resolved CFD-DEM solver. 

(58) In the resolved CFD-DEM model, particle sizes are larger than the grid size for a more accurate computation of the surrounding flow field. A continuous forcing approach is taken to describe the momentum source of the governing equation prior to discretization, which is different from a Direct Forcing Method (DFM). 

(59) As no body-conforming moving mesh is required, the continuous forcing approach offers lower complexity and reduced cost when compared to the DFM. Piquet et al. 

(60) highlighted the high complexity of the DFM due to its reliance on calculating an additional immersed boundary flux for the velocity field to ensure its divergence-free condition.The fluid–structure interaction (FSI) method has been advocated to connect the dynamic interplay of RBC membranes and fluid plasma within blood flow such as the coupling of continuum–particle interactions. However, such methodology is generally adapted for anatomical configurations such as arteries 

(61,62) and capillaries, 

(63) where both the structural components and the fluid domain undergo substantial deformation due to the moving boundaries. Due to the scope of the Review being blood flow simulation within microchannels of LOC devices without deformable boundaries, the Review of the FSI method will not be further carried out.In general, three numerical methods are broadly used: mesh-based, particle-based, and hybrid mesh–particle techniques, based on the spatial scale and the fundamental numerical approach, mesh-based methods tend to neglect the effects of individual particles, assuming a continuum and being efficient in terms of time and cost. However, the particle-based approach highlights more of the microscopic and mesoscopic level, where the influence of individual RBCs is considered. A review from Freund et al. 

(64) addressed the three numerical methodologies and their respective modeling approaches of RBC dynamics. Given the complex mechanics and the diverse levels of study concerning numerical simulations of blood and cellular flow, a broad spectrum of numerical methods for blood has been subjected to extensive review. 

(64−70) Ye at al. 

(65) offered an extensive review of the application of the DPD, SPH, and LBM for numerical simulations of RBC, while Rathnayaka et al. 

(67) conducted a review of the particle-based numerical modeling for liquid marbles through drawing parallels to the transport of RBCs in microchannels. A comparative analysis between conventional CFD methods and particle-based approaches for cellular and blood flow dynamic simulation can be found under the review by Arabghahestani et al. 

(66) Literature by Li et al. 

(68) and Beris et al. 

(69) offer an overview of both continuum-based models at micro/macroscales and multiscale particle-based models encompassing various length and temporal dimensions. Furthermore, these reviews deliberate upon the potential of coupling continuum-particle methods for blood plasma and RBC modeling. Arciero et al. 

(70) investigated various modeling approaches encompassing cellular interactions, such as cell to cell or plasma interactions and the individual cellular phases. A concise overview of the reviews is provided in Table 2 for reference.

Table 2. List of Reviews for Numerical Approaches Employed in Blood Flow Simulation

ReferenceNumerical methods
Li et al. (2013) (68)Continuum-based modeling (BIM), particle-based modeling (LBM, LB-FE, SPH, DPD)
Freund (2014) (64)RBC dynamic modeling (continuum-based modeling, complementary discrete microstructure modeling), blood flow dynamic modeling (FDM, IBM, LBM, particle-mesh methods, coupled boundary integral and mesh-based methods, DPD)
Ye et al. (2016) (65)DPD, SPH, LBM, coupled IBM-Smoothed DPD
Arciero et al. (2017) (70)LBM, IBM, DPD, conventional CFD Methods (FDM, FVM, FEM)
Arabghahestani et al. (2019) (66)Particle-based methods (LBM, DPD, direct simulation Monte Carlo, molecular dynamics), SPH, conventional CFD methods (FDM, FVM, FEM)
Beris et al. (2021) (69)DPD, smoothed DPD, IBM, LBM, BIM
Rathnayaka (2022) (67)SPH, CG, LBM

3. Capillary Driven Blood Flow in LOC Systems

ARTICLE SECTIONS

Jump To


3.1. Capillary Driven Flow Phenomena

Capillary driven (CD) flow is a pivotal mechanism in passive microfluidic flow systems 

(9) such as the blood circulation system and LOC systems. 

(71) CD flow is essentially the movement of a liquid to flow against drag forces, where the capillary effect exerts a force on the liquid at the borders, causing a liquid–air meniscus to flow despite gravity or other drag forces. A capillary pressure drops across the liquid–air interface with surface tension in the capillary radius and contact angle. The capillary effect depends heavily on the interaction between the different properties of surface materials. Different values of contact angles can be manipulated and obtained under varying levels of surface wettability treatments to manipulate the surface properties, resulting in different CD blood delivery rates for medical diagnostic device microchannels. CD flow techniques are appealing for many LOC devices, because they require no external energy. However, due to the passive property of liquid propulsion by capillary forces and the long-term instability of surface treatments on channel walls, the adaptability of CD flow in geometrically complex LOC devices may be limited.

3.2. Theoretical and Numerical Modeling of Capillary Driven Blood Flow

3.2.1. Theoretical Basis and Assumptions of Microfluidic Flow

The study of transport phenomena regarding either blood flow driven by capillary forces or externally applied forces under microfluid systems all demands a comprehensive recognition of the significant differences in flow dynamics between microscale and macroscale. The fundamental assumptions and principles behind fluid transport at the microscale are discussed in this section. Such a comprehension will lay the groundwork for the following analysis of the theoretical basis of capillary forces and their role in blood transport in LOC systems.

At the macroscale, fluid dynamics are often strongly influenced by gravity due to considerable fluid mass. However, the high surface to volume ratio at the microscale shifts the balance toward surface forces (e.g., surface tension and viscous forces), much larger than the inertial force. This difference gives rise to transport phenomena unique to microscale fluid transport, such as the prevalence of laminar flow due to a very low Reynolds number (generally lower than 1). Moreover, the fluid in a microfluidic system is often assumed to be incompressible due to the small flow velocity, indicating constant fluid density in both space and time.Microfluidic flow behaviors are governed by the fundamental principles of mass and momentum conservation, which are encapsulated in the continuity equation and the Navier–Stokes (N–S) equation. The continuity equation describes the conservation of mass, while the N–S equation captures the spatial and temporal variations in velocity, pressure, and other physical parameters. Under the assumption of the negligible influence of gravity in microfluidic systems, the continuity equation and the Eulerian representation of the incompressible N–S equation can be expressed as follows:

∇·𝐮⇀=0∇·�⇀=0

(7)

−∇𝑝+𝜇∇2𝐮⇀+∇·𝝉⇀−𝐅⇀=0−∇�+�∇2�⇀+∇·�⇀−�⇀=0

(8)Here, p is the pressure, u is the fluid viscosity, 

𝝉⇀�⇀ represents the stress tensor, and F is the body force exerted by external forces if present.

3.2.2. Theoretical Basis and Modeling of Capillary Force in LOC Systems

The capillary force is often the major driving force to manipulate and transport blood without an externally applied force in LOC systems. Forces induced by the capillary effect impact the free surface of fluids and are represented not directly in the Navier–Stokes equations but through the pressure boundary conditions of the pressure term p. For hydrophilic surfaces, the liquid generally induces a contact angle between 0° and 30°, encouraging the spread and attraction of fluid under a positive cos θ condition. For this condition, the pressure drop becomes positive and generates a spontaneous flow forward. A hydrophobic solid surface repels the fluid, inducing minimal contact. Generally, hydrophobic solids exhibit a contact angle larger than 90°, inducing a negative value of cos θ. Such a value will result in a negative pressure drop and a flow in the opposite direction. The induced contact angle is often utilized to measure the wall exposure of various surface treatments on channel walls where different wettability gradients and surface tension effects for CD flows are established. Contact angles between different interfaces are obtainable through standard values or experimental methods for reference. 

(72)For the characterization of the induced force by the capillary effect, the Young–Laplace (Y–L) equation 

(73) is widely employed. In the equation, the capillary is considered a pressure boundary condition between the two interphases. Through the Y–L equation, the capillary pressure force can be determined, and subsequently, the continuity and momentum balance equations can be solved to obtain the blood filling rate. Kim et al. 

(74) studied the effects of concentration and exposure time of a nonionic surfactant, Silwet L-77, on the performance of a polydimethylsiloxane (PDMS) microchannel in terms of plasma and blood self-separation. The study characterized the capillary pressure force by incorporating the Y–L equation and further evaluated the effects of the changing contact angle due to different levels of applied channel wall surface treatments. The expression of the Y–L equation utilized by Kim et al. 

(74) is as follows:

𝑃=−𝜎(cos𝜃b+cos𝜃tℎ+cos𝜃l+cos𝜃r𝑤)�=−�(cos⁡�b+cos⁡�tℎ+cos⁡�l+cos⁡�r�)

(9)where σ is the surface tension of the liquid and θ

bθ

tθ

l, and θ

r are the contact angle values between the liquid and the bottom, top, left, and right walls, respectively. A numerical simulation through Coventor software is performed to evaluate the dynamic changes in the filling rate within the microchannel. The simulation results for the blood filling rate in the microchannel are expressed at a specific time stamp, shown in Figure 2. The results portray an increasing instantaneous filling rate of blood in the microchannel following the decrease in contact angle induced by a higher concentration of the nonionic surfactant treated to the microchannel wall.

Figure 2. Numerical simulation of filling rate of capillary driven blood flow under various contact angle conditions at a specific timestamp. (74) Reproduced with permission from ref (74). Copyright 2010 Elsevier.

When in contact with hydrophilic or hydrophobic surfaces, blood forms a meniscus with a contact angle due to surface tension. The Lucas–Washburn (L–W) equation 

(75) is one of the pioneering theoretical definitions for the position of the meniscus over time. In addition, the L–W equation provides the possibility for research to obtain the velocity of the blood formed meniscus through the derivation of the meniscus position. The L–W equation 

(75) can be shown below:

𝐿(𝑡)=𝑅𝜎cos(𝜃)𝑡2𝜇⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯√�(�)=��⁡cos(�)�2�

(10)Here L(t) represents the distance of the liquid driven by the capillary forces. However, the generalized L–W equation solely assumes the constant physical properties from a Newtonian fluid rather than considering the non-Newtonian fluid behavior of blood. Cito et al. 

(76) constructed an enhanced version of the L–W equation incorporating the power law to consider the RBC aggregation and the FL effect. The non-Newtonian fluid apparent viscosity under the Power Law model is defined as

𝜇=𝑘·(𝛾˙)𝑛−1�=�·(�˙)�−1

(11)where γ̇ is the strain rate tensor defined as 

𝛾˙=12𝛾˙𝑖𝑗𝛾˙𝑗𝑖⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯√�˙=12�˙���˙��. The stress tensor term τ is computed as τ = μγ̇

ij. The updated L–W equation by Cito 

(76) is expressed as

𝐿(𝑡)=𝑅[(𝑛+13𝑛+1)(𝜎cos(𝜃)𝑅𝑘)1/𝑛𝑡]𝑛/𝑛+1�(�)=�[(�+13�+1)(�⁡cos(�)��)1/��]�/�+1

(12)where k is the flow consistency index and n is the power law index, respectively. The power law index, from the Power Law model, characterizes the extent of the non-Newtonian behavior of blood. Both the consistency and power law index rely on blood properties such as hematocrit, the appearance of the FL effect, the formation of RBC aggregates, etc. The updated L–W equation computes the location and velocity of blood flow caused by capillary forces at specified time points within the LOC devices, taking into account the effects of blood flow characteristics such as RBC aggregation and the FL effect on dynamic blood viscosity.Apart from the blood flow behaviors triggered by inherent blood properties, unique flow conditions driven by capillary forces that are portrayed under different microchannel geometries also hold crucial implications for CD blood delivery. Berthier et al. 

(77) studied the spontaneous Concus–Finn condition, the condition to initiate the spontaneous capillary flow within a V-groove microchannel, as shown in Figure 3(a) both experimentally and numerically. Through experimental studies, the spontaneous Concus–Finn filament development of capillary driven blood flow is observed, as shown in Figure 3(b), while the dynamic development of blood flow is numerically simulated through CFD simulation.

Figure 3. (a) Sketch of the cross-section of Berthier’s V-groove microchannel, (b) experimental view of blood in the V-groove microchannel, (78) (c) illustration of the dynamic change of the extension of filament from FLOW 3D under capillary flow at three increasing time intervals. (78) Reproduced with permission from ref (78). Copyright 2014 Elsevier.

Berthier et al. 

(77) characterized the contact angle needed for the initiation of the capillary driving force at a zero-inlet pressure, through the half-angle (α) of the V-groove geometry layout, and its relation to the Concus–Finn filament as shown below:

𝜃<𝜋2−𝛼sin𝛼1+2(ℎ2/𝑤)sin𝛼<cos𝜃{�<�2−�sin⁡�1+2(ℎ2/�)⁡sin⁡�<cos⁡�

(13)Three possible regimes were concluded based on the contact angle value for the initiation of flow and development of Concus–Finn filament:

𝜃>𝜃1𝜃1>𝜃>𝜃0𝜃0no SCFSCF without a Concus−Finn filamentSCF without a Concus−Finn filament{�>�1no SCF�1>�>�0SCF without a Concus−Finn filament�0SCF without a Concus−Finn filament

(14)Under Newton’s Law, the force balance with low Reynolds and Capillary numbers results in the neglect of inertial terms. The force balance between the capillary forces and the viscous force induced by the channel wall is proposed to derive the analytical fluid velocity. This relation between the two forces offers insights into the average flow velocity and the penetration distance function dependent on time. The apparent blood viscosity is defined by Berthier et al. 

(78) through Casson’s law, 

(23) given in eq 1. The research used the FLOW-3D program from Flow Science Inc. software, which solves transient, free-surface problems using the FDM in multiple dimensions. The Volume of Fluid (VOF) method 

(79) is utilized to locate and track the dynamic extension of filament throughout the advancing interface within the channel ahead of the main flow at three progressing time stamps, as depicted in Figure 3(c).

4. Electro-osmotic Flow (EOF) in LOC Systems

ARTICLE SECTIONS

Jump To


The utilization of external forces, such as electric fields, has significantly broadened the possibility of manipulating microfluidic flow in LOC systems. 

(80) Externally applied electric field forces induce a fluid flow from the movement of ions in fluid terms as the “electro-osmotic flow” (EOF).Unique transport phenomena, such as enhanced flow velocity and flow instability, induced by non-Newtonian fluids, particularly viscoelastic fluids, under EOF, have sparked considerable interest in microfluidic devices with simple or complicated geometries within channels. 

(81) However, compared to the study of Newtonian fluids and even other electro-osmotic viscoelastic fluid flows, the literature focusing on the theoretical and numerical modeling of electro-osmotic blood flow is limited due to the complexity of blood properties. Consequently, to obtain a more comprehensive understanding of the complex blood flow behavior under EOF, theoretical and numerical studies of the transport phenomena in the EOF section will be based on the studies of different viscoelastic fluids under EOF rather than that of blood specifically. Despite this limitation, we believe these studies offer valuable insights that can help understand the complex behavior of blood flow under EOF.

4.1. EOF Phenomena

Electro-osmotic flow occurs at the interface between the microchannel wall and bulk phase solution. When in contact with the bulk phase, solution ions are absorbed or dissociated at the solid–liquid interface, resulting in the formation of a charge layer, as shown in Figure 4. This charged channel surface wall interacts with both negative and positive ions in the bulk sample, causing repulsion and attraction forces to create a thin layer of immobilized counterions, known as the Stern layer. The induced electric potential from the wall gradually decreases with an increase in the distance from the wall. The Stern layer potential, commonly termed the zeta potential, controls the intensity of the electrostatic interactions between mobile counterions and, consequently, the drag force from the applied electric field. Next to the Stern layer is the diffuse mobile layer, mainly composed of a mobile counterion. These two layers constitute the “electrical double layer” (EDL), the thickness of which is directly proportional to the ionic strength (concentration) of the bulk fluid. The relationship between the two parameters is characterized by a Debye length (λ

D), expressed as

𝜆𝐷=𝜖𝑘B𝑇2(𝑍𝑒)2𝑐0⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯√��=��B�2(��)2�0

(15)where ϵ is the permittivity of the electrolyte solution, k

B is the Boltzmann constant, T is the electron temperature, Z is the integer valence number, e is the elementary charge, and c

0 is the ionic density.

Figure 4. Schematic diagram of an electro-osmotic flow in a microchannel with negative surface charge. (82) Reproduced with permission from ref (82). Copyright 2012 Woodhead Publishing.

When an electric field is applied perpendicular to the EDL, viscous drag is generated due to the movement of excess ions in the EDL. Electro-osmotic forces can be attributed to the externally applied electric potential (ϕ) and the zeta potential, the system wall induced potential by charged walls (ψ). As illustrated in Figure 4, the majority of ions in the bulk phase have a uniform velocity profile, except for a shear rate condition confined within an extremely thin Stern layer. Therefore, EOF displays a unique characteristic of a “near flat” or plug flow velocity profile, different from the parabolic flow typically induced by pressure-driven microfluidic flow (Hagen–Poiseuille flow). The plug-shaped velocity profile of the EOF possesses a high shear rate above the Stern layer.Overall, the EOF velocity magnitude is typically proportional to the Debye Length (λ

D), zeta potential, and magnitude of the externally applied electric field, while a more viscous liquid reduces the EOF velocity.

4.2. Modeling on Electro-osmotic Viscoelastic Fluid Flow

4.2.1. Theoretical Basis of EOF Mechanisms

The EOF of an incompressible viscoelastic fluid is commonly governed by the continuity and incompressible N–S equations, as shown in eqs 7 and 8, where the stress tensor and the electrostatic force term are coupled. The electro-osmotic body force term F, representing the body force exerted by the externally applied electric force, is defined as 

𝐹⇀=𝑝𝐸𝐸⇀�⇀=���⇀, where ρ

E and 

𝐸⇀�⇀ are the net electric charge density and the applied external electric field, respectively.Numerous models are established to theoretically study the externally applied electric potential and the system wall induced potential by charged walls. The following Laplace equation, expressed as eq 16, is generally adapted and solved to calculate the externally applied potential (ϕ).

∇2𝜙=0∇2�=0

(16)Ion diffusion under applied electric fields, together with mass transport resulting from convection and diffusion, transports ionic solutions in bulk flow under electrokinetic processes. The Nernst–Planck equation can describe these transport methods, including convection, diffusion, and electro-diffusion. Therefore, the Nernst–Planck equation is used to determine the distribution of the ions within the electrolyte. The electric potential induced by the charged channel walls follows the Poisson–Nernst–Plank (PNP) equation, which can be written as eq 17.

∇·[𝐷𝑖∇𝑛𝑖−𝑢⇀𝑛𝑖+𝑛𝑖𝐷𝑖𝑧𝑖𝑒𝑘𝑏𝑇∇(𝜙+𝜓)]=0∇·[��∇��−�⇀��+����������∇(�+�)]=0

(17)where D

in

i, and z

i are the diffusion coefficient, ionic concentration, and ionic valence of the ionic species I, respectively. However, due to the high nonlinearity and numerical stiffness introduced by different lengths and time scales from the PNP equations, the Poisson–Boltzmann (PB) model is often considered the major simplified method of the PNP equation to characterize the potential distribution of the EDL region in microchannels. In the PB model, it is assumed that the ionic species in the fluid follow the Boltzmann distribution. This model is typically valid for steady-state problems where charge transport can be considered negligible, the EDLs do not overlap with each other, and the intrinsic potentials are low. It provides a simplified representation of the potential distribution in the EDL region. The PB equation governing the EDL electric potential distribution is described as

∇2𝜓=(2𝑒𝑧𝑛0𝜀𝜀0)sinh(𝑧𝑒𝜓𝑘b𝑇)∇2�=(2���0��0)⁡sinh(����b�)

(18)where n

0 is the ion bulk concentration, z is the ionic valence, and ε

0 is the electric permittivity in the vacuum. Under low electric potential conditions, an even further simplified model to illustrate the EOF phenomena is the Debye–Hückel (DH) model. The DH model is derived by obtaining a charge density term by expanding the exponential term of the Boltzmann equation in a Taylor series.

4.2.2. EOF Modeling for Viscoelastic Fluids

Many studies through numerical modeling were performed to obtain a deeper understanding of the effect exhibited by externally applied electric fields on viscoelastic flow in microchannels under various geometrical designs. Bello et al. 

(83) found that methylcellulose solution, a non-Newtonian polymer solution, resulted in stronger electro-osmotic mobility in experiments when compared to the predictions by the Helmholtz–Smoluchowski equation, which is commonly used to define the velocity of EOF of a Newtonian fluid. Being one of the pioneers to identify the discrepancies between the EOF of Newtonian and non-Newtonian fluids, Bello et al. attributed such discrepancies to the presence of a very high shear rate in the EDL, resulting in a change in the orientation of the polymer molecules. Park and Lee 

(84) utilized the FVM to solve the PB equation for the characterization of the electric field induced force. In the study, the concept of fractional calculus for the Oldroyd-B model was adapted to illustrate the elastic and memory effects of viscoelastic fluids in a straight microchannel They observed that fluid elasticity and increased ratio of viscoelastic fluid contribution to overall fluid viscosity had a significant impact on the volumetric flow rate and sensitivity of velocity to electric field strength compared to Newtonian fluids. Afonso et al. 

(85) derived an analytical expression for EOF of viscoelastic fluid between parallel plates using the DH model to account for a zeta potential condition below 25 mV. The study established the understanding of the electro-osmotic viscoelastic fluid flow under low zeta potential conditions. Apart from the electrokinetic forces, pressure forces can also be coupled with EOF to generate a unique fluid flow behavior within the microchannel. Sousa et al. 

(86) analytically studied the flow of a standard viscoelastic solution by combining the pressure gradient force with an externally applied electric force. It was found that, at a near wall skimming layer and the outer layer away from the wall, macromolecules migrating away from surface walls in viscoelastic fluids are observed. In the study, the Phan-Thien Tanner (PTT) constitutive model is utilized to characterize the viscoelastic properties of the solution. The approach is found to be valid when the EDL is much thinner than the skimming layer under an enhanced flow rate. Zhao and Yang 

(87) solved the PB equation and Carreau model for the characterization of the EOF mechanism and non-Newtonian fluid respectively through the FEM. The numerical results depict that, different from the EOF of Newtonian fluids, non-Newtonian fluids led to an increase of electro-osmotic mobility for shear thinning fluids but the opposite for shear thickening fluids.Like other fluid transport driving forces, EOF within unique geometrical layouts also portrays unique transport phenomena. Pimenta and Alves 

(88) utilized the FVM to perform numerical simulations of the EOF of viscoelastic fluids considering the PB equation and the Oldroyd-B model, in a cross-slot and flow-focusing microdevices. It was found that electroelastic instabilities are formed due to the development of large stresses inside the EDL with streamlined curvature at geometry corners. Bezerra et al. 

(89) used the FDM to numerically analyze the vortex formation and flow instability from an electro-osmotic non-Newtonian fluid flow in a microchannel with a nozzle geometry and parallel wall geometry setting. The PNP equation is utilized to characterize the charge motion in the EOF and the PTT model for non-Newtonian flow characterization. A constriction geometry is commonly utilized in blood flow adapted in LOC systems due to the change in blood flow behavior under narrow dimensions in a microchannel. Ji et al. 

(90) recently studied the EOF of viscoelastic fluid in a constriction microchannel connected by two relatively big reservoirs on both ends (as seen in Figure 5) filled with the polyacrylamide polymer solution, a viscoelastic fluid, and an incompressible monovalent binary electrolyte solution KCl.

Figure 5. Schematic diagram of a negatively charged constriction microchannel connected to two reservoirs at both ends. An electro-osmotic flow is induced in the system by the induced potential difference between the anode and cathode. (90) Reproduced with permission from ref (90). Copyright 2021 The Authors, under the terms of the Creative Commons (CC BY 4.0) License https://creativecommons.org/licenses/by/4.0/.

In studying the EOF of viscoelastic fluids, the Oldroyd-B model is often utilized to characterize the polymeric stress tensor and the deformation rate of the fluid. The Oldroyd-B model is expressed as follows:

𝜏=𝜂p𝜆(𝐜−𝐈)�=�p�(�−�)

(19)where η

p, λ, c, and I represent the polymer dynamic viscosity, polymer relaxation time, symmetric conformation tensor of the polymer molecules, and the identity matrix, respectively.A log-conformation tensor approach is taken to prevent convergence difficulty induced by the viscoelastic properties. The conformation tensor (c) in the polymeric stress tensor term is redefined by a new tensor (Θ) based on the natural logarithm of the c. The new tensor is defined as

Θ=ln(𝐜)=𝐑ln(𝚲)𝐑Θ=ln(�)=�⁡ln(�)�

(20)in which Λ is the diagonal matrix and R is the orthogonal matrix.Under the new conformation tensor, the induced EOF of a viscoelastic fluid is governed by the continuity and N–S equations adapting the Oldroyd-B model, which is expressed as

∂𝚯∂𝑡+𝐮·∇𝚯=𝛀Θ−ΘΩ+2𝐁+1𝜆(eΘ−𝐈)∂�∂�+�·∇�=�Θ−ΘΩ+2�+1�(eΘ−�)

(21)where Ω and B represent the anti-symmetric matrix and the symmetric traceless matrix of the decomposition of the velocity gradient tensor ∇u, respectively. The conformation tensor can be recovered by c = exp(Θ). The PB model and Laplace equation are utilized to characterize the charged channel wall induced potential and the externally applied potential.The governing equations are numerically solved through the FVM by RheoTool, 

(42) an open-source viscoelastic EOF solver on the OpenFOAM platform. A SIMPLEC (Semi-Implicit Method for Pressure Linked Equations-Consistent) algorithm was applied to solve the velocity-pressure coupling. The pressure field and velocity field were computed by the PCG (Preconditioned Conjugate Gradient) solver and the PBiCG (Preconditioned Biconjugate Gradient) solver, respectively.Ranging magnitudes of an applied electric field or fluid concentration induce both different streamlines and velocity magnitudes at various locations and times of the microchannel. In the study performed by Ji et al., 

(90) notable fluctuation of streamlines and vortex formation is formed at the upper stream entrance of the constriction as shown in Figure 6(a) and (b), respectively, due to the increase of electrokinetic effect, which is seen as a result of the increase in polymeric stress (τ

xx). 

(90) The contraction geometry enhances the EOF velocity within the constriction channel under high E

app condition (600 V/cm). Such phenomena can be attributed to the dependence of electro-osmotic viscoelastic fluid flow on the system wall surface and bulk fluid properties. 

(91)

Figure 6. Schematic diagram of vortex formation and streamlines of EOF depicting flow instability at (a) 1.71 s and (b) 1.75 s. Spatial distribution of the elastic normal stress at (c) high Eapp condition. Streamline of an electro-osmotic flow under Eapp of 600 V/cm (90) for (d) non-Newtonian and (e) Newtonian fluid through a constriction geometry. Reproduced with permission from ref (90). Copyright 2021 The Authors, under the terms of the Creative Commons (CC BY 4.0) License https://creativecommons.org/licenses/by/4.0/.

As elastic normal stress exceeds the local shear stress, flow instability and vortex formation occur. The induced elastic stress under EOF not only enhances the instability of the flow but often generates an irregular secondary flow leading to strong disturbance. 

(92) It is also vital to consider the effect of the constriction layout of microchannels on the alteration of the field strength within the system. The contraction geometry enhances a larger electric field strength compared with other locations of the channel outside the constriction region, resulting in a higher velocity gradient and stronger extension on the polymer within the viscoelastic solution. Following the high shear flow condition, a higher magnitude of stretch for polymer molecules in viscoelastic fluids exhibits larger elastic stresses and enhancement of vortex formation at the region. 

(93)As shown in Figure 6(c), significant elastic normal stress occurs at the inlet of the constriction microchannel. Such occurrence of a polymeric flow can be attributed to the dominating elongational flow, giving rise to high deformation of the polymers within the viscoelastic fluid flow, resulting in higher elastic stress from the polymers. Such phenomena at the entrance result in the difference in velocity streamline as circled in Figure 6(d) compared to that of the Newtonian fluid at the constriction entrance in Figure 6(e). 

(90) The difference between the Newtonian and polymer solution at the exit, as circled in Figure 6(d) and (e), can be attributed to the extrudate swell effect of polymers 

(94) within the viscoelastic fluid flow. The extrudate swell effect illustrates that, as polymers emerge from the constriction exit, they tend to contract in the flow direction and grow in the normal direction, resulting in an extrudate diameter greater than the channel size. The deformation of polymers within the polymeric flow at both the entrance and exit of the contraction channel facilitates the change in shear stress conditions of the flow, leading to the alteration in streamlines of flows for each region.

4.3. EOF Applications in LOC Systems

4.3.1. Mixing in LOC Systems

Rather than relying on the micromixing controlled by molecular diffusion under low Reynolds number conditions, active mixers actively leverage convective instability and vortex formation induced by electro-osmotic flows from alternating current (AC) or direct current (DC) electric fields. Such adaptation is recognized as significant breakthroughs for promotion of fluid mixing in chemical and biological applications such as drug delivery, medical diagnostics, chemical synthesis, and so on. 

(95)Many researchers proposed novel designs of electro-osmosis micromixers coupled with numerical simulations in conjunction with experimental findings to increase their understanding of the role of flow instability and vortex formation in the mixing process under electrokinetic phenomena. Matsubara and Narumi 

(96) numerically modeled the mixing process in a microchannel with four electrodes on each side of the microchannel wall, which generated a disruption through unstable electro-osmotic vortices. It was found that particle mixing was sensitive to both the convection effect induced by the main and secondary vortex within the micromixer and the change in oscillation frequency caused by the supplied AC voltage when the Reynolds number was varied. Qaderi et al. 

(97) adapted the PNP equation to numerically study the effect of the geometry and zeta potential configuration of the microchannel on the mixing process with a combined electro-osmotic pressure driven flow. It was reported that the application of heterogeneous zeta potential configuration enhances the mixing efficiency by around 23% while the height of the hurdles increases the mixing efficiency at most 48.1%. Cho et al. 

(98) utilized the PB model and Laplace equation to numerically simulate the electro-osmotic non-Newtonian fluid mixing process within a wavy and block layout of microchannel walls. The Power Law model is adapted to describe the fluid rheological characteristic. It was found that shear-thinning fluids possess a higher volumetric flow rate, which could result in poorer mixing efficiency compared to that of Newtonian fluids. Numerous studies have revealed that flow instability and vortex generation, in particular secondary vortices produced by barriers or greater magnitudes of heterogeneous zeta potential distribution, enhance mixing by increasing bulk flow velocity and reducing flow distance.To better understand the mechanism of disturbance formed in the system due to externally applied forces, known as electrokinetic instability, literature often utilize the Rayleigh (Ra) number, 

(1) as described below:

𝑅𝑎𝑣=𝑢ev𝑢eo=(𝛾−1𝛾+1)2𝑊𝛿2𝐸el2𝐻2𝜁𝛿Ra�=�ev�eo=(�−1�+1)2��2�el2�2��

(22)where γ is the conductivity ratio of the two streams and can be written as 

𝛾=𝜎el,H𝜎el,L�=�el,H�el,L. The Ra number characterizes the ratio between electroviscous and electro-osmotic flow. A high Ra

v value often results in good mixing. It is evident that fluid properties such as the conductivity (σ) of the two streams play a key role in the formation of disturbances to enhance mixing in microsystems. At the same time, electrokinetic parameters like the zeta potential (ζ) in the Ra number is critical in the characterization of electro-osmotic velocity and a slip boundary condition at the microchannel wall.To understand the mixing result along the channel, the concentration field can be defined and simulated under the assumption of steady state conditions and constant diffusion coefficient for each of the working fluid within the system through the convection–diffusion equation as below:

∂𝑐𝒊∂𝑡+∇⇀(𝑐𝑖𝑢⇀−𝐷𝑖∇⇀𝑐𝒊)=0∂��∂�+∇⇀(���⇀−��∇⇀��)=0

(23)where c

i is the species concentration of species i and D

i is the diffusion coefficient of the corresponding species.The standard deviation of concentration (σ

sd) can be adapted to evaluate the mixing quality of the system. 

(97) The standard deviation for concentration at a specific portion of the channel may be calculated using the equation below:

𝜎sd=∫10(𝐶∗(𝑦∗)−𝐶m)2d𝑦∗∫10d𝑦∗⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯�sd=∫01(�*(�*)−�m)2d�*∫01d�*

(24)where C*(y*) and C

m are the non-dimensional concentration profile and the mean concentration at the portion, respectively. C* is the non-dimensional concentration and can be calculated as 

𝐶∗=𝐶𝐶ref�*=��ref, where C

ref is the reference concentration defined as the bulk solution concentration. The mean concentration profile can be calculated as 

𝐶m=∫10(𝐶∗(𝑦∗)d𝑦∗∫10d𝑦∗�m=∫01(�*(�*)d�*∫01d�*. With the standard deviation of concentration, the mixing efficiency 

(97) can then be calculated as below:

𝜀𝑥=1−𝜎sd𝜎sd,0��=1−�sd�sd,0

(25)where σ

sd,0 is the standard derivation of the case of no mixing. The value of the mixing efficiency is typically utilized in conjunction with the simulated flow field and concentration field to explore the effect of geometrical and electrokinetic parameters on the optimization of the mixing results.

5. Summary

ARTICLE SECTIONS

Jump To


5.1. Conclusion

Viscoelastic fluids such as blood flow in LOC systems are an essential topic to proceed with diagnostic analysis and research through microdevices in the biomedical and pharmaceutical industries. The complex blood flow behavior is tightly controlled by the viscoelastic characteristics of blood such as the dynamic viscosity and the elastic property of RBCs under various shear rate conditions. Furthermore, the flow behaviors under varied driving forces promote an array of microfluidic transport phenomena that are critical to the management of blood flow and other adapted viscoelastic fluids in LOC systems. This review addressed the blood flow phenomena, the complicated interplay between shear rate and blood flow behaviors, and their numerical modeling under LOC systems through the lens of the viscoelasticity characteristic. Furthermore, a theoretical understanding of capillary forces and externally applied electric forces leads to an in-depth investigation of the relationship between blood flow patterns and the key parameters of the two driving forces, the latter of which is introduced through the lens of viscoelastic fluids, coupling numerical modeling to improve the knowledge of blood flow manipulation in LOC systems. The flow disturbances triggered by the EOF of viscoelastic fluids and their impact on blood flow patterns have been deeply investigated due to their important role and applications in LOC devices. Continuous advancements of various numerical modeling methods with experimental findings through more efficient and less computationally heavy methods have served as an encouraging sign of establishing more accurate illustrations of the mechanisms for multiphase blood and other viscoelastic fluid flow transport phenomena driven by various forces. Such progress is fundamental for the manipulation of unique transport phenomena, such as the generated disturbances, to optimize functionalities offered by microdevices in LOC systems.

The following section will provide further insights into the employment of studied blood transport phenomena to improve the functionality of micro devices adapting LOC technology. A discussion of the novel roles that external driving forces play in microfluidic flow behaviors is also provided. Limitations in the computational modeling of blood flow and electrokinetic phenomena in LOC systems will also be emphasized, which may provide valuable insights for future research endeavors. These discussions aim to provide guidance and opportunities for new paths in the ongoing development of LOC devices that adapt blood flow.

5.2. Future Directions

5.2.1. Electro-osmosis Mixing in LOC Systems

Despite substantial research, mixing results through flow instability and vortex formation phenomena induced by electro-osmotic mixing still deviate from the effective mixing results offered by chaotic mixing results such as those seen in turbulent flows. However, recent discoveries of a mixing phenomenon that is generally observed under turbulent flows are found within electro-osmosis micromixers under low Reynolds number conditions. Zhao 

(99) experimentally discovered a rapid mixing process in an AC applied micromixer, where the power spectrum of concentration under an applied voltage of 20 V

p-p induces a −5/3 slope within a frequency range. This value of the slope is considered as the O–C spectrum in macroflows, which is often visible under relatively high Re conditions, such as the Taylor microscale Reynolds number Re > 500 in turbulent flows. 

(100) However, the Re value in the studied system is less than 1 at the specific location and applied voltage. A secondary flow is also suggested to occur close to microchannel walls, being attributed to the increase of convective instability within the system.Despite the experimental phenomenon proposed by Zhao et al., 

(99) the range of effects induced by vital parameters of an EOF mixing system on the enhanced mixing results and mechanisms of disturbance generated by the turbulent-like flow instability is not further characterized. Such a gap in knowledge may hinder the adaptability and commercialization of the discovery of micromixers. One of the parameters for further evaluation is the conductivity gradient of the fluid flow. A relatively strong conductivity gradient (5000:1) was adopted in the system due to the conductive properties of the two fluids. The high conductivity gradients may contribute to the relatively large Rayleigh number and differences in EDL layer thickness, resulting in an unusual disturbance in laminar flow conditions and enhanced mixing results. However, high conductivity gradients are not always achievable by the working fluids due to diverse fluid properties. The reliance on turbulent-like phenomena and rapid mixing results in a large conductivity gradient should be established to prevent the limited application of fluids for the mixing system. In addition, the proposed system utilizes distinct zeta potential distributions at the top and bottom walls due to their difference in material choices, which may be attributed to the flow instability phenomena. Further studies should be made on varying zeta potential magnitude and distribution to evaluate their effect on the slip boundary conditions of the flow and the large shear rate condition close to the channel wall of EOF. Such a study can potentially offer an optimized condition in zeta potential magnitude through material choices and geometrical layout of the zeta potential for better mixing results and manipulation of mixing fluid dynamics. The two vital parameters mentioned above can be varied with the aid of numerical simulation to understand the effect of parameters on the interaction between electro-osmotic forces and electroviscous forces. At the same time, the relationship of developed streamlines of the simulated velocity and concentration field, following their relationship with the mixing results, under the impact of these key parameters can foster more insight into the range of impact that the two parameters have on the proposed phenomena and the microfluidic dynamic principles of disturbances.

In addition, many of the current investigations of electrokinetic mixers commonly emphasize the fluid dynamics of mixing for Newtonian fluids, while the utilization of biofluids, primarily viscoelastic fluids such as blood, and their distinctive response under shear forces in these novel mixing processes of LOC systems are significantly less studied. To develop more compatible microdevice designs and efficient mixing outcomes for the biomedical industry, it is necessary to fill the knowledge gaps in the literature on electro-osmotic mixing for biofluids, where properties of elasticity, dynamic viscosity, and intricate relationship with shear flow from the fluid are further considered.

5.2.2. Electro-osmosis Separation in LOC Systems

Particle separation in LOC devices, particularly in biological research and diagnostics, is another area where disturbances may play a significant role in optimization. 

(101) Plasma analysis in LOC systems under precise control of blood flow phenomena and blood/plasma separation procedures can detect vital information about infectious diseases from particular antibodies and foreign nucleic acids for medical treatments, diagnostics, and research, 

(102) offering more efficient results and simple operating procedures compared to that of the traditional centrifugation method for blood and plasma separation. However, the adaptability of LOC devices for blood and plasma separation is often hindered by microchannel clogging, where flow velocity and plasma yield from LOC devices is reduced due to occasional RBC migration and aggregation at the filtration entrance of microdevices. 

(103)It is important to note that the EOF induces flow instability close to microchannel walls, which may provide further solutions to clogging for the separation process of the LOC systems. Mohammadi et al. 

(104) offered an anti-clogging effect of RBCs at the blood and plasma separating device filtration entry, adjacent to the surface wall, through RBC disaggregation under high shear rate conditions generated by a forward and reverse EOF direction.

Further theoretical and numerical research can be conducted to characterize the effect of high shear rate conditions near microchannel walls toward the detachment of binding blood cells on surfaces and the reversibility of aggregation. Through numerical modeling with varying electrokinetic parameters to induce different degrees of disturbances or shear conditions at channel walls, it may be possible to optimize and better understand the process of disrupting the forces that bind cells to surface walls and aggregated cells at filtration pores. RBCs that migrate close to microchannel walls are often attracted by the adhesion force between the RBC and the solid surface originating from the van der Waals forces. Following RBC migration and attachment by adhesive forces adjacent to the microchannel walls as shown in Figure 7, the increase in viscosity at the region causes a lower shear condition and encourages RBC aggregation (cell–cell interaction), which clogs filtering pores or microchannels and reduces flow velocity at filtration region. Both the impact that shear forces and disturbances may induce on cell binding forces with surface walls and other cells leading to aggregation may suggest further characterization. Kinetic parameters such as activation energy and the rate-determining step for cell binding composition attachment and detachment should be considered for modeling the dynamics of RBCs and blood flows under external forces in LOC separation devices.

Figure 7. Schematic representations of clogging at a microchannel pore following the sequence of RBC migration, cell attachment to channel walls, and aggregation. (105) Reproduced with permission from ref (105). Copyright 2018 The Authors under the terms of the Creative Commons (CC BY 4.0) License https://creativecommons.org/licenses/by/4.0/.

5.2.3. Relationship between External Forces and Microfluidic Systems

In blood flow, a thicker CFL suggests a lower blood viscosity, suggesting a complex relationship between shear stress and shear rate, affecting the blood viscosity and blood flow. Despite some experimental and numerical studies on electro-osmotic non-Newtonian fluid flow, limited literature has performed an in-depth investigation of the role that applied electric forces and other external forces could play in the process of CFL formation. Additional studies on how shear rates from external forces affect CFL formation and microfluidic flow dynamics can shed light on the mechanism of the contribution induced by external driving forces to the development of a separate phase of layer, similar to CFL, close to the microchannel walls and distinct from the surrounding fluid within the system, then influencing microfluidic flow dynamics.One of the mechanisms of phenomena to be explored is the formation of the Exclusion Zone (EZ) region following a “Self-Induced Flow” (SIF) phenomenon discovered by Li and Pollack, 

(106) as shown in Figure 8(a) and (b), respectively. A spontaneous sustained axial flow is observed when hydrophilic materials are immersed in water, resulting in the buildup of a negative layer of charges, defined as the EZ, after water molecules absorb infrared radiation (IR) energy and break down into H and OH

+.

Figure 8. Schematic representations of (a) the Exclusion Zone region and (b) the Self Induced Flow through visualization of microsphere movement within a microchannel. (106) Reproduced with permission from ref (106). Copyright 2020 The Authors under the terms of the Creative Commons (CC BY 4.0) License https://creativecommons.org/licenses/by/4.0/.

Despite the finding of such a phenomenon, the specific mechanism and role of IR energy have yet to be defined for the process of EZ development. To further develop an understanding of the role of IR energy in such phenomena, a feasible study may be seen through the lens of the relationships between external forces and microfluidic flow. In the phenomena, the increase of SIF velocity under a rise of IR radiation resonant characteristics is shown in the participation of the external electric field near the microchannel walls under electro-osmotic viscoelastic fluid flow systems. The buildup of negative charges at the hydrophilic surfaces in EZ is analogous to the mechanism of electrical double layer formation. Indeed, research has initiated the exploration of the core mechanisms for EZ formation through the lens of the electrokinetic phenomena. 

(107) Such a similarity of the role of IR energy and the transport phenomena of SIF with electrokinetic phenomena paves the way for the definition of the unknown SIF phenomena and EZ formation. Furthermore, Li and Pollack 

(106) suggest whether CFL formation might contribute to a SIF of blood using solely IR radiation, a commonly available source of energy in nature, as an external driving force. The proposition may be proven feasible with the presence of the CFL region next to the negatively charged hydrophilic endothelial glycocalyx layer, coating the luminal side of blood vessels. 

(108) Further research can dive into the resonating characteristics between the formation of the CFL region next to the hydrophilic endothelial glycocalyx layer and that of the EZ formation close to hydrophilic microchannel walls. Indeed, an increase in IR energy is known to rapidly accelerate EZ formation and SIF velocity, depicting similarity to the increase in the magnitude of electric field forces and greater shear rates at microchannel walls affecting CFL formation and EOF velocity. Such correlation depicts a future direction in whether SIF blood flow can be observed and characterized theoretically further through the lens of the relationship between blood flow and shear forces exhibited by external energy.

The intricate link between the CFL and external forces, more specifically the externally applied electric field, can receive further attention to provide a more complete framework for the mechanisms between IR radiation and EZ formation. Such characterization may also contribute to a greater comprehension of the role IR can play in CFL formation next to the endothelial glycocalyx layer as well as its role as a driving force to propel blood flow, similar to the SIF, but without the commonly assumed pressure force from heart contraction as a source of driving force.

5.3. Challenges

Although there have been significant improvements in blood flow modeling under LOC systems over the past decade, there are still notable constraints that may require special attention for numerical simulation applications to benefit the adaptability of the designs and functionalities of LOC devices. Several points that require special attention are mentioned below:

1.The majority of CFD models operate under the relationship between the viscoelasticity of blood and the shear rate conditions of flow. The relative effect exhibited by the presence of highly populated RBCs in whole blood and their forces amongst the cells themselves under complex flows often remains unclearly defined. Furthermore, the full range of cell populations in whole blood requires a much more computational load for numerical modeling. Therefore, a vital goal for future research is to evaluate a reduced modeling method where the impact of cell–cell interaction on the viscoelastic property of blood is considered.
2.Current computational methods on hemodynamics rely on continuum models based upon non-Newtonian rheology at the macroscale rather than at molecular and cellular levels. Careful considerations should be made for the development of a constructive framework for the physical and temporal scales of micro/nanoscale systems to evaluate the intricate relationship between fluid driving forces, dynamic viscosity, and elasticity.
3.Viscoelastic fluids under the impact of externally applied electric forces often deviate from the assumptions of no-slip boundary conditions due to the unique flow conditions induced by externally applied forces. Furthermore, the mechanism of vortex formation and viscoelastic flow instability at laminar flow conditions should be better defined through the lens of the microfluidic flow phenomenon to optimize the prediction of viscoelastic flow across different geometrical layouts. Mathematical models and numerical methods are needed to better predict such disturbance caused by external forces and the viscoelasticity of fluids at such a small scale.
4.Under practical situations, zeta potential distribution at channel walls frequently deviates from the common assumption of a constant distribution because of manufacturing faults or inherent surface charges prior to the introduction of electrokinetic influence. These discrepancies frequently lead to inconsistent surface potential distribution, such as excess positive ions at relatively more negatively charged walls. Accordingly, unpredicted vortex formation and flow instability may occur. Therefore, careful consideration should be given to these discrepancies and how they could trigger the transport process and unexpected results of a microdevice.

Author Information

ARTICLE SECTIONS

Jump To


  • Corresponding Authors
    • Zhe Chen – Department of Chemical Engineering, School of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai 200240, P. R. China;  Email: zaccooky@sjtu.edu.cn
    • Bo Ouyang – Department of Chemical Engineering, School of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai 200240, P. R. China;  Email: bouy93@sjtu.edu.cn
    • Zheng-Hong Luo – Department of Chemical Engineering, School of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai 200240, P. R. China;  Orcidhttps://orcid.org/0000-0001-9011-6020; Email: luozh@sjtu.edu.cn
  • Authors
    • Bin-Jie Lai – Department of Chemical Engineering, School of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai 200240, P. R. China;  Orcidhttps://orcid.org/0009-0002-8133-5381
    • Li-Tao Zhu – Department of Chemical Engineering, School of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai 200240, P. R. China;  Orcidhttps://orcid.org/0000-0001-6514-8864
  • NotesThe authors declare no competing financial interest.

Acknowledgments

ARTICLE SECTIONS

Jump To


This work was supported by the National Natural Science Foundation of China (No. 22238005) and the Postdoctoral Research Foundation of China (No. GZC20231576).

Vocabulary

ARTICLE SECTIONS

Jump To


Microfluidicsthe field of technological and scientific study that investigates fluid flow in channels with dimensions between 1 and 1000 μm
Lab-on-a-Chip Technologythe field of research and technological development aimed at integrating the micro/nanofluidic characteristics to conduct laboratory processes on handheld devices
Computational Fluid Dynamics (CFD)the method utilizing computational abilities to predict physical fluid flow behaviors mathematically through solving the governing equations of corresponding fluid flows
Shear Ratethe rate of change in velocity where one layer of fluid moves past the adjacent layer
Viscoelasticitythe property holding both elasticity and viscosity characteristics relying on the magnitude of applied shear stress and time-dependent strain
Electro-osmosisthe flow of fluid under an applied electric field when charged solid surface is in contact with the bulk fluid
Vortexthe rotating motion of a fluid revolving an axis line

References

ARTICLE SECTIONS

Jump To


This article references 108 other publications.

  1. 1Neethirajan, S.; Kobayashi, I.; Nakajima, M.; Wu, D.; Nandagopal, S.; Lin, F. Microfluidics for food, agriculture and biosystems industries. Lab Chip 201111 (9), 1574– 1586,  DOI: 10.1039/c0lc00230eViewGoogle Scholar
  2. 2Whitesides, G. M. The origins and the future of microfluidics. Nature 2006442 (7101), 368– 373,  DOI: 10.1038/nature05058ViewGoogle Scholar
  3. 3Burklund, A.; Tadimety, A.; Nie, Y.; Hao, N.; Zhang, J. X. J. Chapter One – Advances in diagnostic microfluidics; Elsevier, 2020; DOI:  DOI: 10.1016/bs.acc.2019.08.001 .ViewGoogle Scholar
  4. 4Abdulbari, H. A. Chapter 12 – Lab-on-a-chip for analysis of blood. In Nanotechnology for Hematology, Blood Transfusion, and Artificial Blood; Denizli, A., Nguyen, T. A., Rajan, M., Alam, M. F., Rahman, K., Eds.; Elsevier, 2022; pp 265– 283.ViewGoogle Scholar
  5. 5Vladisavljević, G. T.; Khalid, N.; Neves, M. A.; Kuroiwa, T.; Nakajima, M.; Uemura, K.; Ichikawa, S.; Kobayashi, I. Industrial lab-on-a-chip: Design, applications and scale-up for drug discovery and delivery. Advanced Drug Delivery Reviews 201365 (11), 1626– 1663,  DOI: 10.1016/j.addr.2013.07.017ViewGoogle Scholar
  6. 6Kersaudy-Kerhoas, M.; Dhariwal, R.; Desmulliez, M. P. Y.; Jouvet, L. Hydrodynamic blood plasma separation in microfluidic channels. Microfluid. Nanofluid. 20108 (1), 105– 114,  DOI: 10.1007/s10404-009-0450-5ViewGoogle Scholar
  7. 7Popel, A. S.; Johnson, P. C. Microcirculation and Hemorheology. Annu. Rev. Fluid Mech. 200537 (1), 43– 69,  DOI: 10.1146/annurev.fluid.37.042604.133933ViewGoogle Scholar
  8. 8Fedosov, D. A.; Peltomäki, M.; Gompper, G. Deformation and dynamics of red blood cells in flow through cylindrical microchannels. Soft Matter 201410 (24), 4258– 4267,  DOI: 10.1039/C4SM00248BViewGoogle Scholar
  9. 9Chakraborty, S. Dynamics of capillary flow of blood into a microfluidic channel. Lab Chip 20055 (4), 421– 430,  DOI: 10.1039/b414566fViewGoogle Scholar
  10. 10Tomaiuolo, G.; Guido, S. Start-up shape dynamics of red blood cells in microcapillary flow. Microvascular Research 201182 (1), 35– 41,  DOI: 10.1016/j.mvr.2011.03.004ViewGoogle Scholar
  11. 11Sherwood, J. M.; Dusting, J.; Kaliviotis, E.; Balabani, S. The effect of red blood cell aggregation on velocity and cell-depleted layer characteristics of blood in a bifurcating microchannel. Biomicrofluidics 20126 (2), 24119,  DOI: 10.1063/1.4717755ViewGoogle Scholar
  12. 12Nader, E.; Skinner, S.; Romana, M.; Fort, R.; Lemonne, N.; Guillot, N.; Gauthier, A.; Antoine-Jonville, S.; Renoux, C.; Hardy-Dessources, M.-D. Blood Rheology: Key Parameters, Impact on Blood Flow, Role in Sickle Cell Disease and Effects of Exercise. Frontiers in Physiology 201910, 01329,  DOI: 10.3389/fphys.2019.01329ViewGoogle Scholar
  13. 13Trejo-Soto, C.; Lázaro, G. R.; Pagonabarraga, I.; Hernández-Machado, A. Microfluidics Approach to the Mechanical Properties of Red Blood Cell Membrane and Their Effect on Blood Rheology. Membranes 202212 (2), 217,  DOI: 10.3390/membranes12020217ViewGoogle Scholar
  14. 14Wagner, C.; Steffen, P.; Svetina, S. Aggregation of red blood cells: From rouleaux to clot formation. Comptes Rendus Physique 201314 (6), 459– 469,  DOI: 10.1016/j.crhy.2013.04.004ViewGoogle Scholar
  15. 15Kim, H.; Zhbanov, A.; Yang, S. Microfluidic Systems for Blood and Blood Cell Characterization. Biosensors 202313 (1), 13,  DOI: 10.3390/bios13010013ViewGoogle Scholar
  16. 16Fåhræus, R.; Lindqvist, T. THE VISCOSITY OF THE BLOOD IN NARROW CAPILLARY TUBES. American Journal of Physiology-Legacy Content 193196 (3), 562– 568,  DOI: 10.1152/ajplegacy.1931.96.3.562ViewGoogle Scholar
  17. 17Ascolese, M.; Farina, A.; Fasano, A. The Fåhræus-Lindqvist effect in small blood vessels: how does it help the heart?. J. Biol. Phys. 201945 (4), 379– 394,  DOI: 10.1007/s10867-019-09534-4ViewGoogle Scholar
  18. 18Bento, D.; Fernandes, C. S.; Miranda, J. M.; Lima, R. In vitro blood flow visualizations and cell-free layer (CFL) measurements in a microchannel network. Experimental Thermal and Fluid Science 2019109, 109847,  DOI: 10.1016/j.expthermflusci.2019.109847ViewGoogle Scholar
  19. 19Namgung, B.; Ong, P. K.; Wong, Y. H.; Lim, D.; Chun, K. J.; Kim, S. A comparative study of histogram-based thresholding methods for the determination of cell-free layer width in small blood vessels. Physiological Measurement 201031 (9), N61,  DOI: 10.1088/0967-3334/31/9/N01ViewGoogle Scholar
  20. 20Hymel, S. J.; Lan, H.; Fujioka, H.; Khismatullin, D. B. Cell trapping in Y-junction microchannels: A numerical study of the bifurcation angle effect in inertial microfluidics. Phys. Fluids (1994) 201931 (8), 082003,  DOI: 10.1063/1.5113516ViewGoogle Scholar
  21. 21Li, X.; Popel, A. S.; Karniadakis, G. E. Blood-plasma separation in Y-shaped bifurcating microfluidic channels: a dissipative particle dynamics simulation study. Phys. Biol. 20129 (2), 026010,  DOI: 10.1088/1478-3975/9/2/026010ViewGoogle Scholar
  22. 22Yin, X.; Thomas, T.; Zhang, J. Multiple red blood cell flows through microvascular bifurcations: Cell free layer, cell trajectory, and hematocrit separation. Microvascular Research 201389, 47– 56,  DOI: 10.1016/j.mvr.2013.05.002ViewGoogle Scholar
  23. 23Shibeshi, S. S.; Collins, W. E. The Rheology of Blood Flow in a Branched Arterial System. Appl. Rheol 200515 (6), 398– 405,  DOI: 10.1515/arh-2005-0020ViewGoogle Scholar
  24. 24Sequeira, A.; Janela, J. An Overview of Some Mathematical Models of Blood Rheology. In A Portrait of State-of-the-Art Research at the Technical University of Lisbon; Pereira, M. S., Ed.; Springer Netherlands: Dordrecht, 2007; pp 65– 87.ViewGoogle Scholar
  25. 25Walburn, F. J.; Schneck, D. J. A constitutive equation for whole human blood. Biorheology 197613, 201– 210,  DOI: 10.3233/BIR-1976-13307ViewGoogle Scholar
  26. 26Quemada, D. A rheological model for studying the hematocrit dependence of red cell-red cell and red cell-protein interactions in blood. Biorheology 198118, 501– 516,  DOI: 10.3233/BIR-1981-183-615ViewGoogle Scholar
  27. 27Varchanis, S.; Dimakopoulos, Y.; Wagner, C.; Tsamopoulos, J. How viscoelastic is human blood plasma?. Soft Matter 201814 (21), 4238– 4251,  DOI: 10.1039/C8SM00061AViewGoogle Scholar
  28. 28Apostolidis, A. J.; Moyer, A. P.; Beris, A. N. Non-Newtonian effects in simulations of coronary arterial blood flow. J. Non-Newtonian Fluid Mech. 2016233, 155– 165,  DOI: 10.1016/j.jnnfm.2016.03.008ViewGoogle Scholar
  29. 29Luo, X. Y.; Kuang, Z. B. A study on the constitutive equation of blood. J. Biomech. 199225 (8), 929– 934,  DOI: 10.1016/0021-9290(92)90233-QViewGoogle Scholar
  30. 30Oldroyd, J. G.; Wilson, A. H. On the formulation of rheological equations of state. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 1950200 (1063), 523– 541,  DOI: 10.1098/rspa.1950.0035ViewGoogle Scholar
  31. 31Prado, G.; Farutin, A.; Misbah, C.; Bureau, L. Viscoelastic transient of confined red blood cells. Biophys J. 2015108 (9), 2126– 2136,  DOI: 10.1016/j.bpj.2015.03.046ViewGoogle Scholar
  32. 32Huang, C. R.; Pan, W. D.; Chen, H. Q.; Copley, A. L. Thixotropic properties of whole blood from healthy human subjects. Biorheology 198724 (6), 795– 801,  DOI: 10.3233/BIR-1987-24630ViewGoogle Scholar
  33. 33Anand, M.; Kwack, J.; Masud, A. A new generalized Oldroyd-B model for blood flow in complex geometries. International Journal of Engineering Science 201372, 78– 88,  DOI: 10.1016/j.ijengsci.2013.06.009ViewGoogle Scholar
  34. 34Horner, J. S.; Armstrong, M. J.; Wagner, N. J.; Beris, A. N. Investigation of blood rheology under steady and unidirectional large amplitude oscillatory shear. J. Rheol. 201862 (2), 577– 591,  DOI: 10.1122/1.5017623ViewGoogle Scholar
  35. 35Horner, J. S.; Armstrong, M. J.; Wagner, N. J.; Beris, A. N. Measurements of human blood viscoelasticity and thixotropy under steady and transient shear and constitutive modeling thereof. J. Rheol. 201963 (5), 799– 813,  DOI: 10.1122/1.5108737ViewGoogle Scholar
  36. 36Armstrong, M.; Tussing, J. A methodology for adding thixotropy to Oldroyd-8 family of viscoelastic models for characterization of human blood. Phys. Fluids 202032 (9), 094111,  DOI: 10.1063/5.0022501ViewGoogle Scholar
  37. 37Crank, J.; Nicolson, P. A practical method for numerical evaluation of solutions of partial differential equations of the heat-conduction type. Mathematical Proceedings of the Cambridge Philosophical Society 194743 (1), 50– 67,  DOI: 10.1017/S0305004100023197ViewGoogle Scholar
  38. 38Clough, R. W. Original formulation of the finite element method. Finite Elements in Analysis and Design 19907 (2), 89– 101,  DOI: 10.1016/0168-874X(90)90001-UViewGoogle Scholar
  39. 39Liu, W. K.; Liu, Y.; Farrell, D.; Zhang, L.; Wang, X. S.; Fukui, Y.; Patankar, N.; Zhang, Y.; Bajaj, C.; Lee, J.Immersed finite element method and its applications to biological systems. Computer Methods in Applied Mechanics and Engineering 2006195 (13), 1722– 1749,  DOI: 10.1016/j.cma.2005.05.049ViewGoogle Scholar
  40. 40Lopes, D.; Agujetas, R.; Puga, H.; Teixeira, J.; Lima, R.; Alejo, J. P.; Ferrera, C. Analysis of finite element and finite volume methods for fluid-structure interaction simulation of blood flow in a real stenosed artery. International Journal of Mechanical Sciences 2021207, 106650,  DOI: 10.1016/j.ijmecsci.2021.106650ViewGoogle Scholar
  41. 41Favero, J. L.; Secchi, A. R.; Cardozo, N. S. M.; Jasak, H. Viscoelastic flow analysis using the software OpenFOAM and differential constitutive equations. J. Non-Newtonian Fluid Mech. 2010165 (23), 1625– 1636,  DOI: 10.1016/j.jnnfm.2010.08.010ViewGoogle Scholar
  42. 42Pimenta, F.; Alves, M. A. Stabilization of an open-source finite-volume solver for viscoelastic fluid flows. J. Non-Newtonian Fluid Mech. 2017239, 85– 104,  DOI: 10.1016/j.jnnfm.2016.12.002ViewGoogle Scholar
  43. 43Chee, C. Y.; Lee, H. P.; Lu, C. Using 3D fluid-structure interaction model to analyse the biomechanical properties of erythrocyte. Phys. Lett. A 2008372 (9), 1357– 1362,  DOI: 10.1016/j.physleta.2007.09.067ViewGoogle Scholar
  44. 44Xu, D.; Kaliviotis, E.; Munjiza, A.; Avital, E.; Ji, C.; Williams, J. Large scale simulation of red blood cell aggregation in shear flows. J. Biomech. 201346 (11), 1810– 1817,  DOI: 10.1016/j.jbiomech.2013.05.010ViewGoogle Scholar
  45. 45Johnson, K. L.; Kendall, K.; Roberts, A. Surface energy and the contact of elastic solids. Proceedings of the royal society of London. A. mathematical and physical sciences 1971324 (1558), 301– 313,  DOI: 10.1098/rspa.1971.0141ViewGoogle Scholar
  46. 46Shi, L.; Pan, T.-W.; Glowinski, R. Deformation of a single red blood cell in bounded Poiseuille flows. Phys. Rev. E 201285 (1), 016307,  DOI: 10.1103/PhysRevE.85.016307ViewGoogle Scholar
  47. 47Yoon, D.; You, D. Continuum modeling of deformation and aggregation of red blood cells. J. Biomech. 201649 (11), 2267– 2279,  DOI: 10.1016/j.jbiomech.2015.11.027ViewGoogle Scholar
  48. 48Mainardi, F.; Spada, G. Creep, relaxation and viscosity properties for basic fractional models in rheology. European Physical Journal Special Topics 2011193 (1), 133– 160,  DOI: 10.1140/epjst/e2011-01387-1ViewGoogle Scholar
  49. 49Gracka, M.; Lima, R.; Miranda, J. M.; Student, S.; Melka, B.; Ostrowski, Z. Red blood cells tracking and cell-free layer formation in a microchannel with hyperbolic contraction: A CFD model validation. Computer Methods and Programs in Biomedicine 2022226, 107117,  DOI: 10.1016/j.cmpb.2022.107117ViewGoogle Scholar
  50. 50Aryan, H.; Beigzadeh, B.; Siavashi, M. Euler-Lagrange numerical simulation of improved magnetic drug delivery in a three-dimensional CT-based carotid artery bifurcation. Computer Methods and Programs in Biomedicine 2022219, 106778,  DOI: 10.1016/j.cmpb.2022.106778ViewGoogle Scholar
  51. 51Czaja, B.; Závodszky, G.; Azizi Tarksalooyeh, V.; Hoekstra, A. G. Cell-resolved blood flow simulations of saccular aneurysms: effects of pulsatility and aspect ratio. J. R Soc. Interface 201815 (146), 20180485,  DOI: 10.1098/rsif.2018.0485ViewGoogle Scholar
  52. 52Rydquist, G.; Esmaily, M. A cell-resolved, Lagrangian solver for modeling red blood cell dynamics in macroscale flows. J. Comput. Phys. 2022461, 111204,  DOI: 10.1016/j.jcp.2022.111204ViewGoogle Scholar
  53. 53Dadvand, A.; Baghalnezhad, M.; Mirzaee, I.; Khoo, B. C.; Ghoreishi, S. An immersed boundary-lattice Boltzmann approach to study the dynamics of elastic membranes in viscous shear flows. Journal of Computational Science 20145 (5), 709– 718,  DOI: 10.1016/j.jocs.2014.06.006ViewGoogle Scholar
  54. 54Krüger, T.; Holmes, D.; Coveney, P. V. Deformability-based red blood cell separation in deterministic lateral displacement devices─A simulation study. Biomicrofluidics 20148 (5), 054114,  DOI: 10.1063/1.4897913ViewGoogle Scholar
  55. 55Takeishi, N.; Ito, H.; Kaneko, M.; Wada, S. Deformation of a Red Blood Cell in a Narrow Rectangular Microchannel. Micromachines 201910 (3), 199,  DOI: 10.3390/mi10030199ViewGoogle Scholar
  56. 56Krüger, T.; Varnik, F.; Raabe, D. Efficient and accurate simulations of deformable particles immersed in a fluid using a combined immersed boundary lattice Boltzmann finite element method. Computers & Mathematics with Applications 201161 (12), 3485– 3505,  DOI: 10.1016/j.camwa.2010.03.057ViewGoogle Scholar
  57. 57Balachandran Nair, A. N.; Pirker, S.; Umundum, T.; Saeedipour, M. A reduced-order model for deformable particles with application in bio-microfluidics. Computational Particle Mechanics 20207 (3), 593– 601,  DOI: 10.1007/s40571-019-00283-8ViewGoogle Scholar
  58. 58Balachandran Nair, A. N.; Pirker, S.; Saeedipour, M. Resolved CFD-DEM simulation of blood flow with a reduced-order RBC model. Computational Particle Mechanics 20229 (4), 759– 774,  DOI: 10.1007/s40571-021-00441-xViewGoogle Scholar
  59. 59Mittal, R.; Iaccarino, G. IMMERSED BOUNDARY METHODS. Annu. Rev. Fluid Mech. 200537 (1), 239– 261,  DOI: 10.1146/annurev.fluid.37.061903.175743ViewGoogle Scholar
  60. 60Piquet, A.; Roussel, O.; Hadjadj, A. A comparative study of Brinkman penalization and direct-forcing immersed boundary methods for compressible viscous flows. Computers & Fluids 2016136, 272– 284,  DOI: 10.1016/j.compfluid.2016.06.001ViewGoogle Scholar
  61. 61Akerkouch, L.; Le, T. B. A Hybrid Continuum-Particle Approach for Fluid-Structure Interaction Simulation of Red Blood Cells in Fluid Flows. Fluids 20216 (4), 139,  DOI: 10.3390/fluids6040139ViewGoogle Scholar
  62. 62Barker, A. T.; Cai, X.-C. Scalable parallel methods for monolithic coupling in fluid-structure interaction with application to blood flow modeling. J. Comput. Phys. 2010229 (3), 642– 659,  DOI: 10.1016/j.jcp.2009.10.001ViewGoogle Scholar
  63. 63Cetin, A.; Sahin, M. A monolithic fluid-structure interaction framework applied to red blood cells. International Journal for Numerical Methods in Biomedical Engineering 201935 (2), e3171  DOI: 10.1002/cnm.3171ViewGoogle Scholar
  64. 64Freund, J. B. Numerical Simulation of Flowing Blood Cells. Annu. Rev. Fluid Mech. 201446 (1), 67– 95,  DOI: 10.1146/annurev-fluid-010313-141349ViewGoogle Scholar
  65. 65Ye, T.; Phan-Thien, N.; Lim, C. T. Particle-based simulations of red blood cells─A review. J. Biomech. 201649 (11), 2255– 2266,  DOI: 10.1016/j.jbiomech.2015.11.050ViewGoogle Scholar
  66. 66Arabghahestani, M.; Poozesh, S.; Akafuah, N. K. Advances in Computational Fluid Mechanics in Cellular Flow Manipulation: A Review. Applied Sciences 20199 (19), 4041,  DOI: 10.3390/app9194041ViewGoogle Scholar
  67. 67Rathnayaka, C. M.; From, C. S.; Geekiyanage, N. M.; Gu, Y. T.; Nguyen, N. T.; Sauret, E. Particle-Based Numerical Modelling of Liquid Marbles: Recent Advances and Future Perspectives. Archives of Computational Methods in Engineering 202229 (5), 3021– 3039,  DOI: 10.1007/s11831-021-09683-7ViewGoogle Scholar
  68. 68Li, X.; Vlahovska, P. M.; Karniadakis, G. E. Continuum- and particle-based modeling of shapes and dynamics of red blood cells in health and disease. Soft Matter 20139 (1), 28– 37,  DOI: 10.1039/C2SM26891DViewGoogle Scholar
  69. 69Beris, A. N.; Horner, J. S.; Jariwala, S.; Armstrong, M. J.; Wagner, N. J. Recent advances in blood rheology: a review. Soft Matter 202117 (47), 10591– 10613,  DOI: 10.1039/D1SM01212FViewGoogle Scholar
  70. 70Arciero, J.; Causin, P.; Malgaroli, F. Mathematical methods for modeling the microcirculation. AIMS Biophysics 20174 (3), 362– 399,  DOI: 10.3934/biophy.2017.3.362ViewGoogle Scholar
  71. 71Maria, M. S.; Chandra, T. S.; Sen, A. K. Capillary flow-driven blood plasma separation and on-chip analyte detection in microfluidic devices. Microfluid. Nanofluid. 201721 (4), 72,  DOI: 10.1007/s10404-017-1907-6ViewGoogle Scholar
  72. 72Huhtamäki, T.; Tian, X.; Korhonen, J. T.; Ras, R. H. A. Surface-wetting characterization using contact-angle measurements. Nat. Protoc. 201813 (7), 1521– 1538,  DOI: 10.1038/s41596-018-0003-zViewGoogle Scholar
  73. 73Young, T., III. An essay on the cohesion of fluids. Philosophical Transactions of the Royal Society of London 180595, 65– 87,  DOI: 10.1098/rstl.1805.0005ViewGoogle Scholar
  74. 74Kim, Y. C.; Kim, S.-H.; Kim, D.; Park, S.-J.; Park, J.-K. Plasma extraction in a capillary-driven microfluidic device using surfactant-added poly(dimethylsiloxane). Sens. Actuators, B 2010145 (2), 861– 868,  DOI: 10.1016/j.snb.2010.01.017ViewGoogle Scholar
  75. 75Washburn, E. W. The Dynamics of Capillary Flow. Physical Review 192117 (3), 273– 283,  DOI: 10.1103/PhysRev.17.273ViewGoogle Scholar
  76. 76Cito, S.; Ahn, Y. C.; Pallares, J.; Duarte, R. M.; Chen, Z.; Madou, M.; Katakis, I. Visualization and measurement of capillary-driven blood flow using spectral domain optical coherence tomography. Microfluid Nanofluidics 201213 (2), 227– 237,  DOI: 10.1007/s10404-012-0950-6ViewGoogle Scholar
  77. 77Berthier, E.; Dostie, A. M.; Lee, U. N.; Berthier, J.; Theberge, A. B. Open Microfluidic Capillary Systems. Anal Chem. 201991 (14), 8739– 8750,  DOI: 10.1021/acs.analchem.9b01429ViewGoogle Scholar
  78. 78Berthier, J.; Brakke, K. A.; Furlani, E. P.; Karampelas, I. H.; Poher, V.; Gosselin, D.; Cubizolles, M.; Pouteau, P. Whole blood spontaneous capillary flow in narrow V-groove microchannels. Sens. Actuators, B 2015206, 258– 267,  DOI: 10.1016/j.snb.2014.09.040ViewGoogle Scholar
  79. 79Hirt, C. W.; Nichols, B. D. Volume of fluid (VOF) method for the dynamics of free boundaries. J. Comput. Phys. 198139 (1), 201– 225,  DOI: 10.1016/0021-9991(81)90145-5ViewGoogle Scholar
  80. 80Chen, J.-L.; Shih, W.-H.; Hsieh, W.-H. AC electro-osmotic micromixer using a face-to-face, asymmetric pair of planar electrodes. Sens. Actuators, B 2013188, 11– 21,  DOI: 10.1016/j.snb.2013.07.012ViewGoogle Scholar
  81. 81Zhao, C.; Yang, C. Electrokinetics of non-Newtonian fluids: A review. Advances in Colloid and Interface Science 2013201-202, 94– 108,  DOI: 10.1016/j.cis.2013.09.001ViewGoogle Scholar
  82. 82Oh, K. W. 6 – Lab-on-chip (LOC) devices and microfluidics for biomedical applications. In MEMS for Biomedical Applications; Bhansali, S., Vasudev, A., Eds.; Woodhead Publishing, 2012; pp 150– 171.ViewGoogle Scholar
  83. 83Bello, M. S.; De Besi, P.; Rezzonico, R.; Righetti, P. G.; Casiraghi, E. Electroosmosis of polymer solutions in fused silica capillaries. ELECTROPHORESIS 199415 (1), 623– 626,  DOI: 10.1002/elps.1150150186ViewGoogle Scholar
  84. 84Park, H. M.; Lee, W. M. Effect of viscoelasticity on the flow pattern and the volumetric flow rate in electroosmotic flows through a microchannel. Lab Chip 20088 (7), 1163– 1170,  DOI: 10.1039/b800185eViewGoogle Scholar
  85. 85Afonso, A. M.; Alves, M. A.; Pinho, F. T. Analytical solution of mixed electro-osmotic/pressure driven flows of viscoelastic fluids in microchannels. J. Non-Newtonian Fluid Mech. 2009159 (1), 50– 63,  DOI: 10.1016/j.jnnfm.2009.01.006ViewGoogle Scholar
  86. 86Sousa, J. J.; Afonso, A. M.; Pinho, F. T.; Alves, M. A. Effect of the skimming layer on electro-osmotic─Poiseuille flows of viscoelastic fluids. Microfluid. Nanofluid. 201110 (1), 107– 122,  DOI: 10.1007/s10404-010-0651-yViewGoogle Scholar
  87. 87Zhao, C.; Yang, C. Electro-osmotic mobility of non-Newtonian fluids. Biomicrofluidics 20115 (1), 014110,  DOI: 10.1063/1.3571278ViewGoogle Scholar
  88. 88Pimenta, F.; Alves, M. A. Electro-elastic instabilities in cross-shaped microchannels. J. Non-Newtonian Fluid Mech. 2018259, 61– 77,  DOI: 10.1016/j.jnnfm.2018.04.004ViewGoogle Scholar
  89. 89Bezerra, W. S.; Castelo, A.; Afonso, A. M. Numerical Study of Electro-Osmotic Fluid Flow and Vortex Formation. Micromachines (Basel) 201910 (12), 796,  DOI: 10.3390/mi10120796ViewGoogle Scholar
  90. 90Ji, J.; Qian, S.; Liu, Z. Electroosmotic Flow of Viscoelastic Fluid through a Constriction Microchannel. Micromachines (Basel) 202112 (4), 417,  DOI: 10.3390/mi12040417ViewGoogle Scholar
  91. 91Zhao, C.; Yang, C. Exact solutions for electro-osmotic flow of viscoelastic fluids in rectangular micro-channels. Applied Mathematics and Computation 2009211 (2), 502– 509,  DOI: 10.1016/j.amc.2009.01.068ViewGoogle Scholar
  92. 92Gerum, R.; Mirzahossein, E.; Eroles, M.; Elsterer, J.; Mainka, A.; Bauer, A.; Sonntag, S.; Winterl, A.; Bartl, J.; Fischer, L. Viscoelastic properties of suspended cells measured with shear flow deformation cytometry. Elife 202211, e78823,  DOI: 10.7554/eLife.78823ViewGoogle Scholar
  93. 93Sadek, S. H.; Pinho, F. T.; Alves, M. A. Electro-elastic flow instabilities of viscoelastic fluids in contraction/expansion micro-geometries. J. Non-Newtonian Fluid Mech. 2020283, 104293,  DOI: 10.1016/j.jnnfm.2020.104293ViewGoogle Scholar
  94. 94Spanjaards, M.; Peters, G.; Hulsen, M.; Anderson, P. Numerical Study of the Effect of Thixotropy on Extrudate Swell. Polymers 202113 (24), 4383,  DOI: 10.3390/polym13244383ViewGoogle Scholar
  95. 95Rashidi, S.; Bafekr, H.; Valipour, M. S.; Esfahani, J. A. A review on the application, simulation, and experiment of the electrokinetic mixers. Chemical Engineering and Processing – Process Intensification 2018126, 108– 122,  DOI: 10.1016/j.cep.2018.02.021ViewGoogle Scholar
  96. 96Matsubara, K.; Narumi, T. Microfluidic mixing using unsteady electroosmotic vortices produced by a staggered array of electrodes. Chemical Engineering Journal 2016288, 638– 647,  DOI: 10.1016/j.cej.2015.12.013ViewGoogle Scholar
  97. 97Qaderi, A.; Jamaati, J.; Bahiraei, M. CFD simulation of combined electroosmotic-pressure driven micro-mixing in a microchannel equipped with triangular hurdle and zeta-potential heterogeneity. Chemical Engineering Science 2019199, 463– 477,  DOI: 10.1016/j.ces.2019.01.034ViewGoogle Scholar
  98. 98Cho, C.-C.; Chen, C.-L.; Chen, C. o.-K. Mixing enhancement in crisscross micromixer using aperiodic electrokinetic perturbing flows. International Journal of Heat and Mass Transfer 201255 (11), 2926– 2933,  DOI: 10.1016/j.ijheatmasstransfer.2012.02.006ViewGoogle Scholar
  99. 99Zhao, W.; Yang, F.; Wang, K.; Bai, J.; Wang, G. Rapid mixing by turbulent-like electrokinetic microflow. Chemical Engineering Science 2017165, 113– 121,  DOI: 10.1016/j.ces.2017.02.027ViewGoogle Scholar
  100. 100Tran, T.; Chakraborty, P.; Guttenberg, N.; Prescott, A.; Kellay, H.; Goldburg, W.; Goldenfeld, N.; Gioia, G. Macroscopic effects of the spectral structure in turbulent flows. Nat. Phys. 20106 (6), 438– 441,  DOI: 10.1038/nphys1674ViewGoogle Scholar
  101. 101Toner, M.; Irimia, D. Blood-on-a-chip. Annu. Rev. Biomed Eng. 20057, 77– 103,  DOI: 10.1146/annurev.bioeng.7.011205.135108ViewGoogle Scholar
  102. 102Maria, M. S.; Rakesh, P. E.; Chandra, T. S.; Sen, A. K. Capillary flow of blood in a microchannel with differential wetting for blood plasma separation and on-chip glucose detection. Biomicrofluidics 201610 (5), 054108,  DOI: 10.1063/1.4962874ViewGoogle Scholar
  103. 103Tripathi, S.; Varun Kumar, Y. V. B.; Prabhakar, A.; Joshi, S. S.; Agrawal, A. Passive blood plasma separation at the microscale: a review of design principles and microdevices. Journal of Micromechanics and Microengineering 201525 (8), 083001,  DOI: 10.1088/0960-1317/25/8/083001ViewGoogle Scholar
  104. 104Mohammadi, M.; Madadi, H.; Casals-Terré, J. Microfluidic point-of-care blood panel based on a novel technique: Reversible electroosmotic flow. Biomicrofluidics 20159 (5), 054106,  DOI: 10.1063/1.4930865ViewGoogle Scholar
  105. 105Kang, D. H.; Kim, K.; Kim, Y. J. An anti-clogging method for improving the performance and lifespan of blood plasma separation devices in real-time and continuous microfluidic systems. Sci. Rep 20188 (1), 17015,  DOI: 10.1038/s41598-018-35235-4ViewGoogle Scholar
  106. 106Li, Z.; Pollack, G. H. Surface-induced flow: A natural microscopic engine using infrared energy as fuel. Science Advances 20206 (19), eaba0941  DOI: 10.1126/sciadv.aba0941ViewGoogle Scholar
  107. 107Mercado-Uribe, H.; Guevara-Pantoja, F. J.; García-Muñoz, W.; García-Maldonado, J. S.; Méndez-Alcaraz, J. M.; Ruiz-Suárez, J. C. On the evolution of the exclusion zone produced by hydrophilic surfaces: A contracted description. J. Chem. Phys. 2021154 (19), 194902,  DOI: 10.1063/5.0043084ViewGoogle Scholar
  108. 108Yalcin, O.; Jani, V. P.; Johnson, P. C.; Cabrales, P. Implications Enzymatic Degradation of the Endothelial Glycocalyx on the Microvascular Hemodynamics and the Arteriolar Red Cell Free Layer of the Rat Cremaster Muscle. Front Physiol 20189, 168,  DOI: 10.3389/fphys.2018.00168ViewGoogle Scholar
Figure 11. Sketch of scour mechanism around USAF under random waves.

Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves

by Ruigeng Hu 1,Hongjun Liu 2,Hao Leng 1,Peng Yu 3 andXiuhai Wang 1,2,*

1College of Environmental Science and Engineering, Ocean University of China, Qingdao 266000, China

2Key Lab of Marine Environment and Ecology (Ocean University of China), Ministry of Education, Qingdao 266000, China

3Qingdao Geo-Engineering Survering Institute, Qingdao 266100, China

*Author to whom correspondence should be addressed.

J. Mar. Sci. Eng. 20219(8), 886; https://doi.org/10.3390/jmse9080886

Received: 6 July 2021 / Revised: 8 August 2021 / Accepted: 13 August 2021 / Published: 17 August 2021

(This article belongs to the Section Ocean Engineering)

Download 

Abstract

A series of numerical simulation were conducted to study the local scour around umbrella suction anchor foundation (USAF) under random waves. In this study, the validation was carried out firstly to verify the accuracy of the present model. Furthermore, the scour evolution and scour mechanism were analyzed respectively. In addition, two revised models were proposed to predict the equilibrium scour depth Seq around USAF. At last, a parametric study was carried out to study the effects of the Froude number Fr and Euler number Eu for the Seq. The results indicate that the present numerical model is accurate and reasonable for depicting the scour morphology under random waves. The revised Raaijmakers’s model shows good agreement with the simulating results of the present study when KCs,p < 8. The predicting results of the revised stochastic model are the most favorable for n = 10 when KCrms,a < 4. The higher Fr and Eu both lead to the more intensive horseshoe vortex and larger Seq.

Keywords: 

scournumerical investigationrandom wavesequilibrium scour depthKC number

1. Introduction

The rapid expansion of cities tends to cause social and economic problems, such as environmental pollution and traffic jam. As a kind of clean energy, offshore wind power has developed rapidly in recent years. The foundation of offshore wind turbine (OWT) supports the upper tower, and suffers the cyclic loading induced by waves, tides and winds, which exerts a vital influence on the OWT system. The types of OWT foundation include the fixed and floating foundation, and the fixed foundation was used usually for nearshore wind turbine. After the construction of fixed foundation, the hydrodynamic field changes in the vicinity of the foundation, leading to the horseshoe vortex formation and streamline compression at the upside and sides of foundation respectively [1,2,3,4]. As a result, the neighboring soil would be carried away by the shear stress induced by vortex, and the scour hole would emerge in the vicinity of foundation. The scour holes increase the cantilever length, and weaken the lateral bearing capacity of foundation [5,6,7,8,9]. Moreover, the natural frequency of OWT system increases with the increase of cantilever length, causing the resonance occurs when the system natural frequency equals the wave or wind frequency [10,11,12]. Given that, an innovative foundation called umbrella suction anchor foundation (USAF) has been designed for nearshore wind power. The previous studies indicated the USAF was characterized by the favorable lateral bearing capacity with the low cost [6,13,14]. The close-up of USAF is shown in Figure 1, and it includes six parts: 1-interal buckets, 2-external skirt, 3-anchor ring, 4-anchor branch, 5-supporting rod, 6-telescopic hook. The detailed description and application method of USAF can be found in reference [13].

Jmse 09 00886 g001 550

Figure 1. The close-up of umbrella suction anchor foundation (USAF).

Numerical and experimental investigations of scour around OWT foundation under steady currents and waves have been extensively studied by many researchers [1,2,15,16,17,18,19,20,21,22,23,24]. The seabed scour can be classified as two types according to Shields parameter θ, i.e., clear bed scour (θ < θcr) or live bed scour (θ > θcr). Due to the set of foundation, the adverse hydraulic pressure gradient exists at upstream foundation edges, resulting in the streamline separation between boundary layer flow and seabed. The separating boundary layer ascended at upstream anchor edges and developed into the horseshoe vortex. Then, the horseshoe vortex moved downstream gradually along the periphery of the anchor, and the vortex shed off continually at the lee-side of the anchor, i.e., wake vortex. The core of wake vortex is a negative pressure center, liking a vacuum cleaner. Hence, the soil particles were swirled into the negative pressure core and carried away by wake vortexes. At the same time, the onset of scour at rear side occurred. Finally, the wake vortex became downflow when the turbulence energy could not support the survival of wake vortex. According to Tavouktsoglou et al. [25], the scale of pile wall boundary layer is proportional to 1/ln(Rd) (Rd is pile Reynolds), which means the turbulence intensity induced by the flow-structure interaction would decrease with Rd increases, but the effects of Rd can be neglected only if the flow around the foundation is fully turbulent [26]. According to previous studies [1,15,27,28,29,30,31,32], the scour development around pile foundation under waves was significantly influenced by Shields parameter θ and KC number simultaneously (calculated by Equation (1)). Sand ripples widely existed around pile under waves in the case of live bed scour, and the scour morphology is related with θ and KC. Compared with θKC has a greater influence on the scour morphology [21,27,28]. The influence mechanism of KC on the scour around the pile is reflected in two aspects: the horseshoe vortex at upstream and wake vortex shedding at downstream.

KC=UwmTD��=�wm��(1)

where, Uwm is the maximum velocity of the undisturbed wave-induced oscillatory flow at the sea bottom above the wave boundary layer, T is wave period, and D is pile diameter.

There are two prerequisites to satisfy the formation of horseshoe vortex at upstream pile edges: (1) the incoming flow boundary layer with sufficient thickness and (2) the magnitude of upstream adverse pressure gradient making the boundary layer separating [1,15,16,18,20]. The smaller KC results the lower adverse pressure gradient, and the boundary layer cannot separate, herein, there is almost no horseshoe vortex emerging at upside of pile. Sumer et al. [1,15] carried out several sets of wave flume experiments under regular and irregular waves respectively, and the experiment results show that there is no horseshoe vortex when KC is less than 6. While the scale and lifespan of horseshoe vortex increase evidently with the increase of KC when KC is larger than 6. Moreover, the wake vortex contributes to the scour at lee-side of pile. Similar with the case of horseshoe vortex, there is no wake vortex when KC is less than 6. The wake vortex is mainly responsible for scour around pile when KC is greater than 6 and less than O(100), while horseshoe vortex controls scour nearly when KC is greater than O(100).

Sumer et al. [1] found that the equilibrium scour depth was nil around pile when KC was less than 6 under regular waves for live bed scour, while the equilibrium scour depth increased with the increase of KC. Based on that, Sumer proposed an equilibrium scour depth predicting equation (Equation (2)). Carreiras et al. [33] revised Sumer’s equation with m = 0.06 for nonlinear waves. Different with the findings of Sumer et al. [1] and Carreiras et al. [33], Corvaro et al. [21] found the scour still occurred for KC ≈ 4, and proposed the revised equilibrium scour depth predicting equation (Equation (3)) for KC > 4.

Rudolph and Bos [2] conducted a series of wave flume experiments to investigate the scour depth around monopile under waves only, waves and currents combined respectively, indicting KC was one of key parameters in influencing equilibrium scour depth, and proposed the equilibrium scour depth predicting equation (Equation (4)) for low KC (1 < KC < 10). Through analyzing the extensive data from published literatures, Raaijmakers and Rudolph [34] developed the equilibrium scour depth predicting equation (Equation (5)) for low KC, which was suitable for waves only, waves and currents combined. Khalfin [35] carried out several sets of wave flume experiments to study scour development around monopile, and proposed the equilibrium scour depth predicting equation (Equation (6)) for low KC (0.1 < KC < 3.5). Different with above equations, the Khalfin’s equation considers the Shields parameter θ and KC number simultaneously in predicting equilibrium scour depth. The flow reversal occurred under through in one wave period, so sand particles would be carried away from lee-side of pile to upside, resulting in sand particles backfilled into the upstream scour hole [20,29]. Considering the backfilling effects, Zanke et al. [36] proposed the equilibrium scour depth predicting equation (Equation (7)) around pile by theoretical analysis, and the equation is suitable for the whole range of KC number under regular waves and currents combined.

S/D=1.3(1−exp([−m(KC−6)])�/�=1.3(1−exp(−�(��−6))(2)

where, m = 0.03 for linear waves.

S/D=1.3(1−exp([−0.02(KC−4)])�/�=1.3(1−exp(−0.02(��−4))(3)

S/D=1.3γKwaveKhw�/�=1.3��wave�ℎw(4)

where, γ is safety factor, depending on design process, typically γ = 1.5, Kwave is correction factor considering wave action, Khw is correction factor considering water depth.

S/D=1.5[tanh(hwD)]KwaveKhw�/�=1.5tanh(ℎw�)�wave�ℎw(5)

where, hw is water depth.

S/D=0.0753(θθcr−−−√−0.5)0.69KC0.68�/�=0.0753(��cr−0.5)0.69��0.68(6)

where, θ is shields parameter, θcr is critical shields parameter.

S/D=2.5(1−0.5u/uc)xrelxrel=xeff/(1+xeff)xeff=0.03(1−0.35ucr/u)(KC−6)⎫⎭⎬⎪⎪�/�=2.5(1−0.5�/��)��������=����/(1+����)����=0.03(1−0.35�cr/�)(��−6)(7)

where, u is near-bed orbital velocity amplitude, uc is critical velocity corresponding the onset of sediment motion.

S/D=1.3{1−exp[−0.03(KC2lnn+36)1/2−6]}�/�=1.31−exp−0.03(��2ln�+36)1/2−6(8)

where, n is the 1/n’th highest wave for random waves

For predicting equilibrium scour depth under irregular waves, i.e., random waves, Sumer and Fredsøe [16] found it’s suitable to take Equation (2) to predict equilibrium scour depth around pile under random waves with the root-mean-square (RMS) value of near-bed orbital velocity amplitude Um and peak wave period TP to calculate KC. Khalfin [35] recommended the RMS wave height Hrms and peak wave period TP were used to calculate KC for Equation (6). References [37,38,39,40] developed a series of stochastic theoretical models to predict equilibrium scour depth around pile under random waves, nonlinear random waves plus currents respectively. The stochastic approach thought the 1/n’th highest wave were responsible for scour in vicinity of pile under random waves, and the KC was calculated in Equation (8) with Um and mean zero-crossing wave period Tz. The results calculated by Equation (8) agree well with experimental values of Sumer and Fredsøe [16] if the 1/10′th highest wave was used. To author’s knowledge, the stochastic approach proposed by Myrhaug and Rue [37] is the only theoretical model to predict equilibrium scour depth around pile under random waves for the whole range of KC number in published documents. Other methods of predicting scour depth under random waves are mainly originated from the equation for regular waves-only, waves and currents combined, which are limited to the large KC number, such as KC > 6 for Equation (2) and KC > 4 for Equation (3) respectively. However, situations with relatively low KC number (KC < 4) often occur in reality, for example, monopile or suction anchor for OWT foundations in ocean environment. Moreover, local scour around OWT foundations under random waves has not yet been investigated fully. Therefore, further study are still needed in the aspect of scour around OWT foundations with low KC number under random waves. Given that, this study presents the scour sediment model around umbrella suction anchor foundation (USAF) under random waves. In this study, a comparison of equilibrium scour depth around USAF between this present numerical models and the previous theoretical models and experimental results was presented firstly. Then, this study gave a comprehensive analysis for the scour mechanisms around USAF. After that, two revised models were proposed according to the model of Raaijmakers and Rudolph [34] and the stochastic model developed by Myrhaug and Rue [37] respectively to predict the equilibrium scour depth. Finally, a parametric study was conducted to study the effects of the Froude number (Fr) and Euler number (Eu) to equilibrium scour depth respectively.

2. Numerical Method

2.1. Governing Equations of Flow

The following equations adopted in present model are already available in Flow 3D software. The authors used these theoretical equations to simulate scour in random waves without modification. The incompressible viscous fluid motion satisfies the Reynolds-averaged Navier-Stokes (RANS) equation, so the present numerical model solves RANS equations:

∂u∂t+1VF(uAx∂u∂x+vAy∂u∂y+wAz∂u∂z)=−1ρf∂p∂x+Gx+fx∂�∂�+1��(���∂�∂�+���∂�∂�+���∂�∂�)=−1�f∂�∂�+��+��(9)

∂v∂t+1VF(uAx∂v∂x+vAy∂v∂y+wAz∂v∂z)=−1ρf∂p∂y+Gy+fy∂�∂�+1��(���∂�∂�+���∂�∂�+���∂�∂�)=−1�f∂�∂�+��+��(10)

∂w∂t+1VF(uAx∂w∂x+vAy∂w∂y+wAz∂w∂z)=−1ρf∂p∂z+Gz+fz∂�∂�+1��(���∂�∂�+���∂�∂�+���∂�∂�)=−1�f∂�∂�+��+��(11)

where, VF is the volume fraction; uv, and w are the velocity components in xyz direction respectively with Cartesian coordinates; Ai is the area fraction; ρf is the fluid density, fi is the viscous fluid acceleration, Gi is the fluid body acceleration (i = xyz).

2.2. Turbulent Model

The turbulence closure is available by the turbulent model, such as one-equation, the one-equation k-ε model, the standard k-ε model, RNG k-ε turbulent model and large eddy simulation (LES) model. The LES model requires very fine mesh grid, so the computational time is large, which hinders the LES model application in engineering. The RNG k-ε model can reduce computational time greatly with high accuracy in the near-wall region. Furthermore, the RNG k-ε model computes the maximum turbulent mixing length dynamically in simulating sediment scour model. Therefore, the RNG k-ε model was adopted to study the scour around anchor under random waves [41,42].

∂kT∂T+1VF(uAx∂kT∂x+vAy∂kT∂y+wAz∂kT∂z)=PT+GT+DiffkT−εkT∂��∂�+1��(���∂��∂�+���∂��∂�+���∂��∂�)=��+��+������−���(12)

∂εT∂T+1VF(uAx∂εT∂x+vAy∂εT∂y+wAz∂εT∂z)=CDIS1εTkT(PT+CDIS3GT)+Diffε−CDIS2ε2TkT∂��∂�+1��(���∂��∂�+���∂��∂�+���∂��∂�)=����1����(��+����3��)+�����−����2��2��(13)

where, kT is specific kinetic energy involved with turbulent velocity, GT is the turbulent energy generated by buoyancy; εT is the turbulent energy dissipating rate, PT is the turbulent energy, Diffε and DiffkT are diffusion terms associated with VFAiCDIS1CDIS2 and CDIS3 are dimensionless parameters, and CDIS1CDIS3 have default values of 1.42, 0.2 respectively. CDIS2 can be obtained from PT and kT.

2.3. Sediment Scour Model

The sand particles may suffer four processes under waves, i.e., entrainment, bed load transport, suspended load transport, and deposition, so the sediment scour model should depict the above processes efficiently. In present numerical simulation, the sediment scour model includes the following aspects:

2.3.1. Entrainment and Deposition

The combination of entrainment and deposition determines the net scour rate of seabed in present sediment scour model. The entrainment lift velocity of sand particles was calculated as [43]:

ulift,i=αinsd0.3∗(θ−θcr)1.5∥g∥di(ρi−ρf)ρf−−−−−−−−−−−−√�lift,i=�����*0.3(�−�cr)1.5���(��−�f)�f(14)

where, αi is the entrainment parameter, ns is the outward point perpendicular to the seabed, d* is the dimensionless diameter of sand particles, which was calculated by Equation (15), θcr is the critical Shields parameter, g is the gravity acceleration, di is the diameter of sand particles, ρi is the density of seabed species.

d∗=di(∥g∥ρf(ρi−ρf)μ2f)1/3�*=��(��f(��−�f)�f2)1/3(15)

where μf is the fluid dynamic viscosity.

In Equation (14), the entrainment parameter αi confirms the rate at which sediment erodes when the given shear stress is larger than the critical shear stress, and the recommended value 0.018 was adopted according to the experimental data of Mastbergen and Von den Berg [43]. ns is the outward pointing normal to the seabed interface, and ns = (0,0,1) according to the Cartesian coordinates used in present numerical model.

The shields parameter was obtained from the following equation:

θ=U2f,m(ρi/ρf−1)gd50�=�f,m2(��/�f−1)��50(16)

where, Uf,m is the maximum value of the near-bed friction velocity; d50 is the median diameter of sand particles. The detailed calculation procedure of θ was available in Soulsby [44].

The critical shields parameter θcr was obtained from the Equation (17) [44]

θcr=0.31+1.2d∗+0.055[1−exp(−0.02d∗)]�cr=0.31+1.2�*+0.0551−exp(−0.02�*)(17)

The sand particles begin to deposit on seabed when the turbulence energy weaken and cann’t support the particles suspending. The setting velocity of the particles was calculated from the following equation [44]:

usettling,i=νfdi[(10.362+1.049d3∗)0.5−10.36]�settling,�=�f��(10.362+1.049�*3)0.5−10.36(18)

where νf is the fluid kinematic viscosity.

2.3.2. Bed Load Transport

This is called bed load transport when the sand particles roll or bounce over the seabed and always have contact with seabed. The bed load transport velocity was computed by [45]:

ubedload,i=qb,iδicb,ifb�bedload,�=�b,����b,��b(19)

where, qb,i is the bed load transport rate, which was obtained from Equation (20), δi is the bed load thickness, which was calculated by Equation (21), cb,i is the volume fraction of sand i in the multiple species, fb is the critical packing fraction of the seabed.

qb,i=8[∥g∥(ρi−ρfρf)d3i]1/2�b,�=8�(��−�f�f)��31/2(20)

δi=0.3d0.7∗(θθcr−1)0.5di��=0.3�*0.7(��cr−1)0.5��(21)

2.3.3. Suspended Load Transport

Through the following transport equation, the suspended sediment concentration could be acquired.

∂Cs,i∂t+∇(us,iCs,i)=∇∇(DfCs,i)∂�s,�∂�+∇(�s,��s,�)=∇∇(�f�s,�)(22)

where, Cs,i is the suspended sand particles mass concentration of sand i in the multiple species, us,i is the sand particles velocity of sand iDf is the diffusivity.

The velocity of sand i in the multiple species could be obtained from the following equation:

us,i=u¯¯+usettling,ics,i�s,�=�¯+�settling,��s,�(23)

where, u¯�¯ is the velocity of mixed fluid-particles, which can be calculated by the RANS equation with turbulence model, cs,i is the suspended sand particles volume concentration, which was computed from Equation (24).

cs,i=Cs,iρi�s,�=�s,���(24)

3. Model Setup

The seabed-USAF-wave three-dimensional scour numerical model was built using Flow-3D software. As shown in Figure 2, the model includes sandy seabed, USAF model, sea water, two baffles and porous media. The dimensions of USAF are shown in Table 1. The sandy bed (210 m in length, 30 m in width and 11 m in height) is made up of uniform fine sand with median diameter d50 = 0.041 cm. The USAF model includes upper steel tube with the length of 20 m, which was installed in the middle of seabed. The location of USAF is positioned at 140 m from the upstream inflow boundary and 70 m from the downstream outflow boundary. Two baffles were installed at two ends of seabed. In order to eliminate the wave reflection basically, the porous media was set at the outflow side on the seabed.

Jmse 09 00886 g002 550

Figure 2. (a) The sketch of seabed-USAF-wave three-dimensional model; (b) boundary condation:Wv-wave boundary, S-symmetric boundary, O-outflow boundary; (c) USAF model.

Table 1. Numerical simulating cases.

Table

3.1. Mesh Geometric Dimensions

In the simulation of the scour under the random waves, the model includes the umbrella suction anchor foundation, seabed and fluid. As shown in Figure 3, the model mesh includes global mesh grid and nested mesh grid, and the total number of grids is 1,812,000. The basic procedure for building mesh grid consists of two steps. Step 1: Divide the global mesh using regular hexahedron with size of 0.6 × 0.6. The global mesh area is cubic box, embracing the seabed and whole fluid volume, and the dimensions are 210 m in length, 30 m in width and 32 m in height. The details of determining the grid size can see the following mesh sensitivity section. Step 2: Set nested fine mesh grid in vicinity of the USAF with size of 0.3 × 0.3 so as to shorten the computation cost and improve the calculation accuracy. The encryption range is −15 m to 15 m in x direction, −15 m to 15 m in y direction and 0 m to 32 m in z direction, respectively. In order to accurately capture the free-surface dynamics, such as the fluid-air interface, the volume of fluid (VOF) method was adopted for tracking the free water surface. One specific algorithm called FAVORTM (Fractional Area/Volume Obstacle Representation) was used to define the fractional face areas and fractional volumes of the cells which are open to fluid flow.

Jmse 09 00886 g003 550

Figure 3. The sketch of mesh grid.

3.2. Boundary Conditions

As shown in Figure 2, the initial fluid length is 210 m as long as seabed. A wave boundary was specified at the upstream offshore end. The details of determining the random wave spectrum can see the following wave parameters section. The outflow boundary was set at the downstream onshore end. The symmetry boundary was used at the top and two sides of the model. The symmetric boundaries were the better strategy to improve the computation efficiency and save the calculation cost [46]. At the seabed bottom, the wall boundary was adopted, which means the u = v = w= 0. Besides, the upper steel tube of USAF was set as no-slip condition.

3.3. Wave Parameters

The random waves with JONSWAP wave spectrum were used for all simulations as realistic representation of offshore conditions. The unidirectional JONSWAP frequency spectrum was described as [47]:

S(ω)=αg2ω5exp[−54(ωpω)4]γexp[−(ω−ωp)22σ2ω2p]�(�)=��2�5exp−54(�p�)4�exp−(�−�p)22�2�p2(25)

where, α is wave energy scale parameter, which is calculated by Equation (26), ω is frequency, ωp is wave spectrum peak frequency, which can be obtained from Equation (27). γ is wave spectrum peak enhancement factor, in this study γ = 3.3. σ is spectral width factor, σ equals 0.07 for ω ≤ ωp and 0.09 for ω > ωp respectively.

α=0.0076(gXU2)−0.22�=0.0076(���2)−0.22(26)

ωp=22(gU)(gXU2)−0.33�p=22(��)(���2)−0.33(27)

where, X is fetch length, U is average wind velocity at 10 m height from mean sea level.

In present numerical model, the input key parameters include X and U for wave boundary with JONSWAP wave spectrum. The objective wave height and period are available by different combinations of X and U. In this study, we designed 9 cases with different wave heights, periods and water depths for simulating scour around USAF under random waves (see Table 2). For random waves, the wave steepness ε and Ursell number Ur were acquired form Equations (28) and (29) respectively

ε=2πgHsT2a�=2���s�a2(28)

Ur=Hsk2h3w�r=�s�2ℎw3(29)

where, Hs is significant wave height, Ta is average wave period, k is wave number, hw is water depth. The Shield parameter θ satisfies θ > θcr for all simulations in current study, indicating the live bed scour prevails.

Table 2. Numerical simulating cases.

Table

3.4. Mesh Sensitivity

In this section, a mesh sensitivity analysis was conducted to investigate the influence of mesh grid size to results and make sure the calculation is mesh size independent and converged. Three mesh grid size were chosen: Mesh 1—global mesh grid size of 0.75 × 0.75, nested fine mesh grid size of 0.4 × 0.4, and total number of grids 1,724,000, Mesh 2—global mesh grid size of 0.6 × 0.6, nested fine mesh grid size of 0.3 × 0.3, and total number of grids 1,812,000, Mesh 3—global mesh grid size of 0.4 × 0.4, nested fine mesh grid size of 0.2 × 0.2, and total number of grids 1,932,000. The near-bed shear velocity U* is an important factor for influencing scour process [1,15], so U* at the position of (4,0,11.12) was evaluated under three mesh sizes. As the Figure 4 shown, the maximum error of shear velocity ∆U*1,2 is about 39.8% between the mesh 1 and mesh 2, and 4.8% between the mesh 2 and mesh 3. According to the mesh sensitivity criterion adopted by Pang et al. [48], it’s reasonable to think the results are mesh size independent and converged with mesh 2. Additionally, the present model was built according to prototype size, and the mesh size used in present model is larger than the mesh size adopted by Higueira et al. [49] and Corvaro et al. [50]. If we choose the smallest cell size, it will take too much time. For example, the simulation with Mesh3 required about 260 h by using a computer with Intel Xeon Scalable Gold 4214 CPU @24 Cores, 2.2 GHz and 64.00 GB RAM. Therefore, in this case, considering calculation accuracy and computation efficiency, the mesh 2 was chosen for all the simulation in this study.

Jmse 09 00886 g004 550

Figure 4. Comparison of near-bed shear velocity U* with different mesh grid size.

The nested mesh block was adopted for seabed in vicinity of the USAF, which was overlapped with the global mesh block. When two mesh blocks overlap each other, the governing equations are by default solved on the mesh block with smaller average cell size (i.e., higher grid resolution). It is should be noted that the Flow 3D software used the moving mesh captures the scour evolution and automatically adjusts the time step size to be as large as possible without exceeding any of the stability limits, affecting accuracy, or unduly increasing the effort required to enforce the continuity condition [51].

3.5. Model Validation

In order to verify the reliability of the present model, the results of present study were compared with the experimental data of Khosronejad et al. [52]. The experiment was conducted in an open channel with a slender vertical pile under unidirectional currents. The comparison of scour development between the present results and the experimental results is shown in Figure 5. The Figure 5 reveals that the present results agree well with the experimental data of Khosronejad et al. [52]. In the first stage, the scour depth increases rapidly. After that, the scour depth achieves a maximum value gradually. The equilibrium scour depth calculated by the present model is basically corresponding with the experimental results of Khosronejad et al. [52], although scour depth in the present model is slightly larger than the experimental results at initial stage.

Jmse 09 00886 g005 550

Figure 5. Comparison of time evolution of scour between the present study and Khosronejad et al. [52], Petersen et al. [17].

Secondly, another comparison was further conducted between the results of present study and the experimental data of Petersen et al. [17]. The experiment was carried out in a flume with a circular vertical pile in combined waves and current. Figure 4 shows a comparison of time evolution of scour depth between the simulating and the experimental results. As Figure 5 indicates, the scour depth in this study has good overall agreement with the experimental results proposed in Petersen et al. [17]. The equilibrium scour depth calculated by the present model is 0.399 m, which equals to the experimental value basically. Overall, the above verifications prove the present model is accurate and capable in dealing with sediment scour under waves.

In addition, in order to calibrate and validate the present model for hydrodynamic parameters, the comparison of water surface elevation was carried out with laboratory experiments conducted by Stahlmann [53] for wave gauge No. 3. The Figure 6 depicts the surface wave profiles between experiments and numerical model results. The comparison indicates that there is a good agreement between the model results and experimental values, especially the locations of wave crest and trough. Comparison of the surface elevation instructs the present model has an acceptable relative error, and the model is a calibrated in terms of the hydrodynamic parameters.

Jmse 09 00886 g006 550

Figure 6. Comparison of surface elevation between the present study and Stahlmann [53].

Finally, another comparison was conducted for equilibrium scour depth or maximum scour depth under random waves with the experimental data of Sumer and Fredsøe [16] and Schendel et al. [22]. The Figure 7 shows the comparison between the numerical results and experimental data of Run01, Run05, Run21 and Run22 in Sumer and Fredsøe [16] and test A05 and A09 in Schendel et al. [22]. As shown in Figure 7, the equilibrium scour depth or maximum scour depth distributed within the ±30 error lines basically, meaning the reliability and accuracy of present model for predicting equilibrium scour depth around foundation in random waves. However, compared with the experimental values, the present model overestimated the equilibrium scour depth generally. Given that, a calibration for scour depth was carried out by multiplying the mean reduced coefficient 0.85 in following section.

Jmse 09 00886 g007 550

Figure 7. Comparison of equilibrium (or maximum) scour depth between the present study and Sumer and Fredsøe [16], Schendel et al. [22].

Through the various examination for hydrodynamic and morphology parameters, it can be concluded that the present model is a validated and calibrated model for scour under random waves. Thus, the present numerical model would be utilized for scour simulation around foundation under random waves.

4. Numerical Results and Discussions

4.1. Scour Evolution

Figure 8 displays the scour evolution for case 1–9. As shown in Figure 8a, the scour depth increased rapidly at the initial stage, and then slowed down at the transition stage, which attributes to the backfilling occurred in scour holes under live bed scour condition, resulting in the net scour decreasing. Finally, the scour reached the equilibrium state when the amount of sediment backfilling equaled to that of scouring in the scour holes, i.e., the net scour transport rate was nil. Sumer and Fredsøe [16] proposed the following formula for the scour development under waves

St=Seq(1−exp(−t/Tc))�t=�eq(1−exp(−�/�c))(30)

where Tc is time scale of scour process.

Jmse 09 00886 g008 550

Figure 8. Time evolution of scour for case 1–9: (a) Case 1–5; (b) Case 6–9.

The computing time is 3600 s and the scour development curves in Figure 8 kept fluctuating, meaning it’s still not in equilibrium scour stage in these cases. According to Sumer and Fredsøe [16], the equilibrium scour depth can be acquired by fitting the data with Equation (30). From Figure 8, it can be seen that the scour evolution obtained from Equation (30) is consistent with the present study basically at initial stage, but the scour depth predicted by Equation (30) developed slightly faster than the simulating results and the Equation (30) overestimated the scour depth to some extent. Overall, the whole tendency of the results calculated by Equation (30) agrees well with the simulating results of the present study, which means the Equation (30) is applicable to depict the scour evolution around USAF under random waves.

4.2. Scour Mechanism under Random Waves

The scour morphology and scour evolution around USAF are similar under random waves in case 1~9. Taking case 7 as an example, the scour morphology is shown in Figure 9.

Jmse 09 00886 g009 550

Figure 9. Scour morphology under different times for case 7.

From Figure 9, at the initial stage (t < 1200 s), the scour occurred at upstream foundation edges between neighboring anchor branches. The maximum scour depth appeared at the lee-side of the USAF. Correspondingly, the sediments deposited at the periphery of the USAF, and the location of the maximum accretion depth was positioned at an angle of about 45° symmetrically with respect to the wave propagating direction in the lee-side of the USAF. After that, when t > 2400 s, the location of the maximum scour depth shifted to the upside of the USAF at an angle of about 45° with respect to the wave propagating direction.

According to previous studies [1,15,16,19,30,31], the horseshoe vortex, streamline compression and wake vortex shedding were responsible for scour around foundation. The Figure 10 displays the distribution of flow velocity in vicinity of foundation, which reflects the evolving processes of horseshoe vertex.

Jmse 09 00886 g010a 550
Jmse 09 00886 g010b 550

Figure 10. Velocity profile around USAF: (a) Flow runup and down stream at upstream anchor edges; (b) Horseshoe vortex at upstream anchor edges; (c) Flow reversal during wave through stage at lee side.

As shown in Figure 10, the inflow tripped to the upstream edges of the USAF and it was blocked by the upper tube of USAF. Then, the downflow formed the horizontal axis clockwise vortex and rolled on the seabed bypassing the tube, that is, the horseshoe vortex (Figure 11). The Figure 12 displays the turbulence intensity around the tube on the seabed. From Figure 12, it can be seen that the turbulence intensity was high-intensity with respect to the region of horseshoe vortex. This phenomenon occurred because of drastic water flow momentum exchanging in the horseshoe vortex. As a result, it created the prominent shear stress on the seabed, causing the local scour at the upstream edges of USAF. Besides, the horseshoe vortex moved downstream gradually along the periphery of the tube and the wake vortex shed off continually at the lee-side of the USAF, i.e., wake vortex.

Jmse 09 00886 g011 550

Figure 11. Sketch of scour mechanism around USAF under random waves.

Jmse 09 00886 g012 550

Figure 12. Turbulence intensity: (a) Turbulence intensity of horseshoe vortex; (b) Turbulence intensity of wake vortex; (c) Turbulence intensity of accretion area.

The core of wake vortex is a negative pressure center, liking a vacuum cleaner [11,42]. Hence, the soil particles were swirled into the negative pressure core and carried away by wake vortex. At the same time, the onset of scour at rear side occurred. Finally, the wake vortex became downflow at the downside of USAF. As is shown in Figure 12, the turbulence intensity was low where the downflow occurred at lee-side, which means the turbulence energy may not be able to support the survival of wake vortex, leading to accretion happening. As mentioned in previous section, the formation of horseshoe vortex was dependent with adverse pressure gradient at upside of foundation. As shown in Figure 13, the evaluated range of pressure distribution is −15 m to 15 m in x direction. The t = 450 s and t = 1800 s indicate that the wave crest and trough arrived at the upside and lee-side of the foundation respectively, and the t = 350 s was neither the wave crest nor trough. The adverse gradient pressure reached the maximum value at t = 450 s corresponding to the wave crest phase. In this case, it’s helpful for the wave boundary separating fully from seabed, which leads to the formation of horseshoe vortex with high turbulence intensity. Therefore, the horseshoe vortex is responsible for the local scour between neighboring anchor branches at upside of USAF. What’s more, due to the combination of the horseshoe vortex and streamline compression, the maximum scour depth occurred at the upside of the USAF with an angle of about 45° corresponding to the wave propagating direction. This is consistent with the findings of Pang et al. [48] and Sumer et al. [1,15] in case of regular waves. At the wave trough phase (t = 1800 s), the pressure gradient became positive at upstream USAF edges, which hindered the separating of wave boundary from seabed. In the meantime, the flow reversal occurred (Figure 10) and the adverse gradient pressure appeared at downstream USAF edges, but the magnitude of adverse gradient pressure at lee-side was lower than the upstream gradient pressure under wave crest. In this way, the intensity of horseshoe vortex behind the USAF under wave trough was low, which explains the difference of scour depth at upstream and downstream, i.e., the scour asymmetry. In other words, the scour asymmetry at upside and downside of USAF was attributed to wave asymmetry for random waves, and the phenomenon became more evident for nonlinear waves [21]. Briefly speaking, the vortex system at wave crest phase was mainly related to the scour process around USAF under random waves.

Jmse 09 00886 g013 550

Figure 13. Pressure distribution around USAF.

4.3. Equilibrium Scour Depth

The KC number is a key parameter for horseshoe vortex emerging and evolving under waves. According to Equation (1), when pile diameter D is fixed, the KC depends on the maximum near-bed velocity Uwm and wave period T. For random waves, the Uwm can be denoted by the root-mean-square (RMS) value of near-bed velocity amplitude Uwm,rms or the significant value of near-bed velocity amplitude Uwm,s. The Uwm,rms and Uwm,s for all simulating cases of the present study are listed in Table 3 and Table 4. The T can be denoted by the mean up zero-crossing wave period Ta, peak wave period Tp, significant wave period Ts, the maximum wave period Tm, 1/10′th highest wave period Tn = 1/10 and 1/5′th highest wave period Tn = 1/5 for random waves, so the different combinations of Uwm and T will acquire different KC. The Table 3 and Table 4 list 12 types of KC, for example, the KCrms,s was calculated by Uwm,rms and Ts. Sumer and Fredsøe [16] conducted a series of wave flume experiments to investigate the scour depth around monopile under random waves, and found the equilibrium scour depth predicting equation (Equation (2)) for regular waves was applicable for random waves with KCrms,p. It should be noted that the Equation (2) is only suitable for KC > 6 under regular waves or KCrms,p > 6 under random waves.

Table 3. Uwm,rms and KC for case 1~9.

Table

Table 4. Uwm,s and KC for case 1~9.

Table

Raaijmakers and Rudolph [34] proposed the equilibrium scour depth predicting model (Equation (5)) around pile under waves, which is suitable for low KC. The format of Equation (5) is similar with the formula proposed by Breusers [54], which can predict the equilibrium scour depth around pile at different scour stages. In order to verify the applicability of Raaijmakers’s model for predicting the equilibrium scour depth around USAF under random waves, a validation of the equilibrium scour depth Seq between the present study and Raaijmakers’s equation was conducted. The position where the scour depth Seq was evaluated is the location of the maximum scour depth, and it was depicted in Figure 14. The Figure 15 displays the comparison of Seq with different KC between the present study and Raaijmakers’s model.

Jmse 09 00886 g014 550

Figure 14. Sketch of the position where the Seq was evaluated.

Jmse 09 00886 g015a 550
Jmse 09 00886 g015b 550

Figure 15. Comparison of the equilibrium scour depth between the present model and the model of Raaijmakers and Rudolph [34]: (aKCrms,sKCrms,a; (bKCrms,pKCrms,m; (cKCrms,n = 1/10KCrms,n = 1/5; (dKCs,sKCs,a; (eKCs,pKCs,m; (fKCs,n = 1/10KCs,n = 1/5.

As shown in Figure 15, there is an error in predicting Seq between the present study and Raaijmakers’s model, and Raaijmakers’s model underestimates the results generally. Although the error exists, the varying trend of Seq with KC obtained from Raaijmakers’s model is consistent with the present study basically. What’s more, the error is minimum and the Raaijmakers’s model is of relatively high accuracy for predicting scour around USAF under random waves by using KCs,p. Based on this, a further revision was made to eliminate the error as much as possible, i.e., add the deviation value ∆S/D in the Raaijmakers’s model. The revised equilibrium scour depth predicting equation based on Raaijmakers’s model can be written as

S′eq/D=1.95[tanh(hD)](1−exp(−0.012KCs,p))+ΔS/D�eq′/�=1.95tanh(ℎ�)(1−exp(−0.012��s,p))+∆�/�(31)

As the Figure 16 shown, through trial-calculation, when ∆S/D = 0.05, the results calculated by Equation (31) show good agreement with the simulating results of the present study. The maximum error is about 18.2% and the engineering requirements have been met basically. In order to further verify the accuracy of the revised model for large KC (KCs,p > 4) under random waves, a validation between the revised model and the previous experimental results [21]. The experiment was conducted in a flume (50 m in length, 1.0 m in width and 1.3 m in height) with a slender vertical pile (D = 0.1 m) under random waves. The seabed is composed of 0.13 m deep layer of sand with d50 = 0.6 mm and the water depth is 0.5 m for all tests. The significant wave height is 0.12~0.21 m and the KCs,p is 5.52~11.38. The comparison between the predicting results by Equation (31) and the experimental results of Corvaro et al. [21] is shown in Figure 17. From Figure 17, the experimental data evenly distributes around the predicted results and the prediction accuracy is favorable when KCs,p < 8. However, the gap between the predicting results and experimental data becomes large and the Equation (31) overestimates the equilibrium scour depth to some extent when KCs,p > 8.

Jmse 09 00886 g016 550

Figure 16. Comparison of Seq between the simulating results and the predicting values by Equation (31).

Jmse 09 00886 g017 550

Figure 17. Comparison of Seq/D between the Experimental results of Corvaro et al. [21] and the predicting values by Equation (31).

In ocean environment, the waves are composed of a train of sinusoidal waves with different frequencies and amplitudes. The energy of constituent waves with very large and very small frequencies is relatively low, and the energy of waves is mainly concentrated in a certain range of moderate frequencies. Myrhaug and Rue [37] thought the 1/n’th highest wave was responsible for scour and proposed the stochastic model to predict the equilibrium scour depth around pile under random waves for full range of KC. Noteworthy is that the KC was denoted by KCrms,a in the stochastic model. To verify the application of the stochastic model for predicting scour depth around USAF, a validation between the simulating results of present study and predicting results by the stochastic model with n = 2,3,5,10,20,500 was carried out respectively.

As shown in Figure 18, compared with the simulating results, the stochastic model underestimates the equilibrium scour depth around USAF generally. Although the error exists, the varying trend of Seq with KCrms,a obtained from the stochastic model is consistent with the present study basically. What’s more, the gap between the predicting values by stochastic model and the simulating results decreases with the increase of n, but for large n, for example n = 500, the varying trend diverges between the predicting values and simulating results, meaning it’s not feasible only by increasing n in stochastic model to predict the equilibrium scour depth around USAF.

Jmse 09 00886 g018 550

Figure 18. Comparison of Seq between the simulating results and the predicting values by Equation (8).

The Figure 19 lists the deviation value ∆Seq/D′ between the predicting values and simulating results with different KCrms,a and n. Then, fitted the relationship between the ∆S′and n under different KCrms,a, and the fitting curve can be written by Equation (32). The revised stochastic model (Equation (33)) can be acquired by adding ∆Seq/D′ to Equation (8).

ΔSeq/D=0.052*exp(−n/6.566)+0.068∆�eq/�=0.052*exp(−�/6.566)+0.068(32)

S′eq¯/D=S′eq/D+0.052*exp(−n/6.566)+0.068�eq′¯/�=�eq′/�+0.052*exp(−�/6.566)+0.068(33)

Jmse 09 00886 g019 550

Figure 19. The fitting line between ∆S′and n.

The comparison between the predicting results by Equation (33) and the simulating results of present study is shown in Figure 20. According to the Figure 20, the varying trend of Seq with KCrms,a obtained from the stochastic model is consistent with the present study basically. Compared with predicting results by the stochastic model, the results calculated by Equation (33) is favorable. Moreover, comparison with simulating results indicates that the predicting results are the most favorable for n = 10, which is consistent with the findings of Myrhaug and Rue [37] for equilibrium scour depth predicting around slender pile in case of random waves.

Jmse 09 00886 g020 550

Figure 20. Comparison of Seq between the simulating results and the predicting values by Equation (33).

In order to further verify the accuracy of the Equation (33) for large KC (KCrms,a > 4) under random waves, a validation was conducted between the Equation (33) and the previous experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21]. The details of experiments conducted by Corvaro et al. [21] were described in above section. Sumer and Fredsøe [16] investigated the local scour around pile under random waves. The experiments were conducted in a wave basin with a slender vertical pile (D = 0.032, 0.055 m). The seabed is composed of 0.14 m deep layer of sand with d50 = 0.2 mm and the water depth was maintained at 0.5 m. The JONSWAP wave spectrum was used and the KCrms,a was 5.29~16.95. The comparison between the predicting results by Equation (33) and the experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21] are shown in Figure 21. From Figure 21, contrary to the case of low KCrms,a (KCrms,a < 4), the error between the predicting values and experimental results increases with decreasing of n for KCrms,a > 4. Therefore, the predicting results are the most favorable for n = 2 when KCrms,a > 4.

Jmse 09 00886 g021 550

Figure 21. Comparison of Seq between the experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21] and the predicting values by Equation (33).

Noteworthy is that the present model was built according to prototype size, so the errors between the numerical results and experimental data of References [16,21] may be attribute to the scale effects. In laboratory experiments on scouring process, it is typically impossible to ensure a rigorous similarity of all physical parameters between the model and prototype structure, leading to the scale effects in the laboratory experiments. To avoid a cohesive behaviour, the bed material was not scaled geometrically according to model scale. As a consequence, the relatively large-scaled sediments sizes may result in the overestimation of bed load transport and underestimation of suspended load transport compared with field conditions. What’s more, the disproportional scaled sediment presumably lead to the difference of bed roughness between the model and prototype, and thus large influences for wave boundary layer on the seabed and scour process. Besides, according to Corvaro et al. [21] and Schendel et al. [55], the pile Reynolds numbers and Froude numbers both affect the scour depth for the condition of non fully developed turbulent flow in laboratory experiments.

4.4. Parametric Study

4.4.1. Influence of Froude Number

As described above, the set of foundation leads to the adverse pressure gradient appearing at upstream, leading to the wave boundary layer separating from seabed, then horseshoe vortex formatting and the horseshoe vortex are mainly responsible for scour around foundation (see Figure 22). The Froude number Fr is the key parameter to influence the scale and intensity of horseshoe vortex. The Fr under waves can be calculated by the following formula [42]

Fr=UwgD−−−√�r=�w��(34)

where Uw is the mean water particle velocity during 1/4 cycle of wave oscillation, obtained from the following formula. Noteworthy is that the root-mean-square (RMS) value of near-bed velocity amplitude Uwm,rms is used for calculating Uwm.

Uw=1T/4∫0T/4Uwmsin(t/T)dt=2πUwm�w=1�/4∫0�/4�wmsin(�/�)��=2��wm(35)

Jmse 09 00886 g022 550

Figure 22. Sketch of flow field at upstream USAF edges.

Tavouktsoglou et al. [25] proposed the following formula between Fr and the vertical location of the stagnation y

yh∝Fer�ℎ∝�r�(36)

where e is constant.

The Figure 23 displays the relationship between Seq/D and Fr of the present study. In order to compare with the simulating results, the experimental data of Corvaro et al. [21] was also depicted in Figure 23. As shown in Figure 23, the equilibrium scour depth appears a logarithmic increase as Fr increases and approaches the mathematical asymptotic value, which is also consistent with the experimental results of Corvaro et al. [21]. According to Figure 24, the adverse pressure gradient pressure at upstream USAF edges increases with the increase of Fr, which is benefit for the wave boundary layer separating from seabed, resulting in the high-intensity horseshoe vortex, hence, causing intensive scour around USAF. Based on the previous study of Tavouktsoglou et al. [25] for scour around pile under currents, the high Fr leads to the stagnation point is closer to the mean sea level for shallow water, causing the stronger downflow kinetic energy. As mentioned in previous section, the energy of downflow at upstream makes up the energy of the subsequent horseshoe vortex, so the stronger downflow kinetic energy results in the more intensive horseshoe vortex. Therefore, the higher Fr leads to the more intensive horseshoe vortex by influencing the position of stagnation point y presumably. Qi and Gao [19] carried out a series of flume tests to investigate the scour around pile under regular waves, and proposed the fitting formula between Seq/D and Fr as following

lg(Seq/D)=Aexp(B/Fr)+Clg(�eq/�)=�exp(�/�r)+�(37)

where AB and C are constant.

Jmse 09 00886 g023 550

Figure 23. The fitting curve between Seq/D and Fr.

Jmse 09 00886 g024 550

Figure 24. Sketch of adverse pressure gradient at upstream USAF edges.

Took the Equation (37) to fit the simulating results with A = −0.002, B = 0.686 and C = −0.808, and the results are shown in Figure 23. From Figure 23, the simulating results evenly distribute around the Equation (37) and the varying trend of Seq/D and Fr in present study is consistent with Equation (37) basically, meaning the Equation (37) is applicable to express the relationship of Seq/D with Fr around USAF under random waves.

4.4.2. Influence of Euler Number

The Euler number Eu is the influencing factor for the hydrodynamic field around foundation. The Eu under waves can be calculated by the following formula. The Eu can be represented by the Equation (38) for uniform cylinders [25]. The root-mean-square (RMS) value of near-bed velocity amplitude Um,rms is used for calculating Um.

Eu=U2mgD�u=�m2��(38)

where Um is depth-averaged flow velocity.

The Figure 25 displays the relationship between Seq/D and Eu of the present study. In order to compare with the simulating results, the experimental data of Sumer and Fredsøe [16] and Corvaro et al. [21] were also plotted in Figure 25. As shown in Figure 25, similar with the varying trend of Seq/D and Fr, the equilibrium scour depth appears a logarithmic increase as Eu increases and approaches the mathematical asymptotic value, which is also consistent with the experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21]. According to Figure 24, the adverse pressure gradient pressure at upstream USAF edges increases with the increasing of Eu, which is benefit for the wave boundary layer separating from seabed, inducing the high-intensity horseshoe vortex, hence, causing intensive scour around USAF.

Jmse 09 00886 g025 550

Figure 25. The fitting curve between Seq/D and Eu.

Therefore, the variation of Fr and Eu reflect the magnitude of adverse pressure gradient pressure at upstream. Given that, the Equation (37) also was used to fit the simulating results with A = 8.875, B = 0.078 and C = −9.601, and the results are shown in Figure 25. From Figure 25, the simulating results evenly distribute around the Equation (37) and the varying trend of Seq/D and Eu in present study is consistent with Equation (37) basically, meaning the Equation (37) is also applicable to express the relationship of Seq/D with Eu around USAF under random waves. Additionally, according to the above description of Fr, it can be inferred that the higher Fr and Eu both lead to the more intensive horseshoe vortex by influencing the position of stagnation point y presumably.

5. Conclusions

A series of numerical models were established to investigate the local scour around umbrella suction anchor foundation (USAF) under random waves. The numerical model was validated for hydrodynamic and morphology parameters by comparing with the experimental data of Khosronejad et al. [52], Petersen et al. [17], Sumer and Fredsøe [16] and Schendel et al. [22]. Based on the simulating results, the scour evolution and scour mechanisms around USAF under random waves were analyzed respectively. Two revised models were proposed according to the model of Raaijmakers and Rudolph [34] and the stochastic model developed by Myrhaug and Rue [37] to predict the equilibrium scour depth around USAF under random waves. Finally, a parametric study was carried out with the present model to study the effects of the Froude number Fr and Euler number Eu to the equilibrium scour depth around USAF under random waves. The main conclusions can be described as follows.(1)

The packed sediment scour model and the RNG k−ε turbulence model were used to simulate the sand particles transport processes and the flow field around UASF respectively. The scour evolution obtained by the present model agrees well with the experimental results of Khosronejad et al. [52], Petersen et al. [17], Sumer and Fredsøe [16] and Schendel et al. [22], which indicates that the present model is accurate and reasonable for depicting the scour morphology around UASF under random waves.(2)

The vortex system at wave crest phase is mainly related to the scour process around USAF under random waves. The maximum scour depth appeared at the lee-side of the USAF at the initial stage (t < 1200 s). Subsequently, when t > 2400 s, the location of the maximum scour depth shifted to the upside of the USAF at an angle of about 45° with respect to the wave propagating direction.(3)

The error is negligible and the Raaijmakers’s model is of relatively high accuracy for predicting scour around USAF under random waves when KC is calculated by KCs,p. Given that, a further revision model (Equation (31)) was proposed according to Raaijmakers’s model to predict the equilibrium scour depth around USAF under random waves and it shows good agreement with the simulating results of the present study when KCs,p < 8.(4)

Another further revision model (Equation (33)) was proposed according to the stochastic model established by Myrhaug and Rue [37] to predict the equilibrium scour depth around USAF under random waves, and the predicting results are the most favorable for n = 10 when KCrms,a < 4. However, contrary to the case of low KCrms,a, the predicting results are the most favorable for n = 2 when KCrms,a > 4 by the comparison with experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21].(5)

The same formula (Equation (37)) is applicable to express the relationship of Seq/D with Eu or Fr, and it can be inferred that the higher Fr and Eu both lead to the more intensive horseshoe vortex and larger Seq.

Author Contributions

Conceptualization, H.L. (Hongjun Liu); Data curation, R.H. and P.Y.; Formal analysis, X.W. and H.L. (Hao Leng); Funding acquisition, X.W.; Writing—original draft, R.H. and P.Y.; Writing—review & editing, X.W. and H.L. (Hao Leng); The final manuscript has been approved by all the authors. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Fundamental Research Funds for the Central Universities (grant number 202061027) and the National Natural Science Foundation of China (grant number 41572247).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data presented in this study are available on request from the corresponding author.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Sumer, B.M.; Fredsøe, J.; Christiansen, N. Scour Around Vertical Pile in Waves. J. Waterw. Port. Coast. Ocean Eng. 1992118, 15–31. [Google Scholar] [CrossRef]
  2. Rudolph, D.; Bos, K. Scour around a monopile under combined wave-current conditions and low KC-numbers. In Proceedings of the 6th International Conference on Scour and Erosion, Amsterdam, The Netherlands, 1–3 November 2006; pp. 582–588. [Google Scholar]
  3. Nielsen, A.W.; Liu, X.; Sumer, B.M.; Fredsøe, J. Flow and bed shear stresses in scour protections around a pile in a current. Coast. Eng. 201372, 20–38. [Google Scholar] [CrossRef]
  4. Ahmad, N.; Bihs, H.; Myrhaug, D.; Kamath, A.; Arntsen, Ø.A. Three-dimensional numerical modelling of wave-induced scour around piles in a side-by-side arrangement. Coast. Eng. 2018138, 132–151. [Google Scholar] [CrossRef]
  5. Li, H.; Ong, M.C.; Leira, B.J.; Myrhaug, D. Effects of Soil Profile Variation and Scour on Structural Response of an Offshore Monopile Wind Turbine. J. Offshore Mech. Arct. Eng. 2018140, 042001. [Google Scholar] [CrossRef]
  6. Li, H.; Liu, H.; Liu, S. Dynamic analysis of umbrella suction anchor foundation embedded in seabed for offshore wind turbines. Géoméch. Energy Environ. 201710, 12–20. [Google Scholar] [CrossRef]
  7. Fazeres-Ferradosa, T.; Rosa-Santos, P.; Taveira-Pinto, F.; Vanem, E.; Carvalho, H.; Correia, J.A.F.D.O. Editorial: Advanced research on offshore structures and foundation design: Part 1. Proc. Inst. Civ. Eng. Marit. Eng. 2019172, 118–123. [Google Scholar] [CrossRef]
  8. Chavez, C.E.A.; Stratigaki, V.; Wu, M.; Troch, P.; Schendel, A.; Welzel, M.; Villanueva, R.; Schlurmann, T.; De Vos, L.; Kisacik, D.; et al. Large-Scale Experiments to Improve Monopile Scour Protection Design Adapted to Climate Change—The PROTEUS Project. Energies 201912, 1709. [Google Scholar] [CrossRef][Green Version]
  9. Wu, M.; De Vos, L.; Chavez, C.E.A.; Stratigaki, V.; Fazeres-Ferradosa, T.; Rosa-Santos, P.; Taveira-Pinto, F.; Troch, P. Large Scale Experimental Study of the Scour Protection Damage Around a Monopile Foundation Under Combined Wave and Current Conditions. J. Mar. Sci. Eng. 20208, 417. [Google Scholar] [CrossRef]
  10. Sørensen, S.P.H.; Ibsen, L.B. Assessment of foundation design for offshore monopiles unprotected against scour. Ocean Eng. 201363, 17–25. [Google Scholar] [CrossRef]
  11. Prendergast, L.; Gavin, K.; Doherty, P. An investigation into the effect of scour on the natural frequency of an offshore wind turbine. Ocean Eng. 2015101, 1–11. [Google Scholar] [CrossRef][Green Version]
  12. Fazeres-Ferradosa, T.; Chambel, J.; Taveira-Pinto, F.; Rosa-Santos, P.; Taveira-Pinto, F.; Giannini, G.; Haerens, P. Scour Protections for Offshore Foundations of Marine Energy Harvesting Technologies: A Review. J. Mar. Sci. Eng. 20219, 297. [Google Scholar] [CrossRef]
  13. Yang, Q.; Yu, P.; Liu, Y.; Liu, H.; Zhang, P.; Wang, Q. Scour characteristics of an offshore umbrella suction anchor foundation under the combined actions of waves and currents. Ocean Eng. 2020202, 106701. [Google Scholar] [CrossRef]
  14. Yu, P.; Hu, R.; Yang, J.; Liu, H. Numerical investigation of local scour around USAF with different hydraulic conditions under currents and waves. Ocean Eng. 2020213, 107696. [Google Scholar] [CrossRef]
  15. Sumer, B.M.; Christiansen, N.; Fredsøe, J. The horseshoe vortex and vortex shedding around a vertical wall-mounted cylinder exposed to waves. J. Fluid Mech. 1997332, 41–70. [Google Scholar] [CrossRef]
  16. Sumer, B.M.; Fredsøe, J. Scour around Pile in Combined Waves and Current. J. Hydraul. Eng. 2001127, 403–411. [Google Scholar] [CrossRef]
  17. Petersen, T.U.; Sumer, B.M.; Fredsøe, J. Time scale of scour around a pile in combined waves and current. In Proceedings of the 6th International Conference on Scour and Erosion, Paris, France, 27–31 August 2012. [Google Scholar]
  18. Petersen, T.U.; Sumer, B.M.; Fredsøe, J.; Raaijmakers, T.C.; Schouten, J.-J. Edge scour at scour protections around piles in the marine environment—Laboratory and field investigation. Coast. Eng. 2015106, 42–72. [Google Scholar] [CrossRef]
  19. Qi, W.; Gao, F. Equilibrium scour depth at offshore monopile foundation in combined waves and current. Sci. China Ser. E Technol. Sci. 201457, 1030–1039. [Google Scholar] [CrossRef][Green Version]
  20. Larsen, B.E.; Fuhrman, D.R.; Baykal, C.; Sumer, B.M. Tsunami-induced scour around monopile foundations. Coast. Eng. 2017129, 36–49. [Google Scholar] [CrossRef][Green Version]
  21. Corvaro, S.; Marini, F.; Mancinelli, A.; Lorenzoni, C.; Brocchini, M. Hydro- and Morpho-dynamics Induced by a Vertical Slender Pile under Regular and Random Waves. J. Waterw. Port. Coast. Ocean Eng. 2018144, 04018018. [Google Scholar] [CrossRef]
  22. Schendel, A.; Welzel, M.; Schlurmann, T.; Hsu, T.-W. Scour around a monopile induced by directionally spread irregular waves in combination with oblique currents. Coast. Eng. 2020161, 103751. [Google Scholar] [CrossRef]
  23. Fazeres-Ferradosa, T.; Taveira-Pinto, F.; Romão, X.; Reis, M.; das Neves, L. Reliability assessment of offshore dynamic scour protections using copulas. Wind. Eng. 201843, 506–538. [Google Scholar] [CrossRef]
  24. Fazeres-Ferradosa, T.; Welzel, M.; Schendel, A.; Baelus, L.; Santos, P.R.; Pinto, F.T. Extended characterization of damage in rubble mound scour protections. Coast. Eng. 2020158, 103671. [Google Scholar] [CrossRef]
  25. Tavouktsoglou, N.S.; Harris, J.M.; Simons, R.R.; Whitehouse, R.J.S. Equilibrium Scour-Depth Prediction around Cylindrical Structures. J. Waterw. Port. Coast. Ocean Eng. 2017143, 04017017. [Google Scholar] [CrossRef][Green Version]
  26. Ettema, R.; Melville, B.; Barkdoll, B. Scale Effect in Pier-Scour Experiments. J. Hydraul. Eng. 1998124, 639–642. [Google Scholar] [CrossRef]
  27. Umeda, S. Scour Regime and Scour Depth around a Pile in Waves. J. Coast. Res. Spec. Issue 201164, 845–849. [Google Scholar]
  28. Umeda, S. Scour process around monopiles during various phases of sea storms. J. Coast. Res. 2013165, 1599–1604. [Google Scholar] [CrossRef]
  29. Baykal, C.; Sumer, B.; Fuhrman, D.R.; Jacobsen, N.; Fredsøe, J. Numerical simulation of scour and backfilling processes around a circular pile in waves. Coast. Eng. 2017122, 87–107. [Google Scholar] [CrossRef][Green Version]
  30. Miles, J.; Martin, T.; Goddard, L. Current and wave effects around windfarm monopile foundations. Coast. Eng. 2017121, 167–178. [Google Scholar] [CrossRef][Green Version]
  31. Miozzi, M.; Corvaro, S.; Pereira, F.A.; Brocchini, M. Wave-induced morphodynamics and sediment transport around a slender vertical cylinder. Adv. Water Resour. 2019129, 263–280. [Google Scholar] [CrossRef]
  32. Yu, T.; Zhang, Y.; Zhang, S.; Shi, Z.; Chen, X.; Xu, Y.; Tang, Y. Experimental study on scour around a composite bucket foundation due to waves and current. Ocean Eng. 2019189, 106302. [Google Scholar] [CrossRef]
  33. Carreiras, J.; Larroudé, P.; Seabra-Santos, F.; Mory, M. Wave Scour Around Piles. In Proceedings of the Coastal Engineering 2000, American Society of Civil Engineers (ASCE), Sydney, Australia, 16–21 July 2000; pp. 1860–1870. [Google Scholar]
  34. Raaijmakers, T.; Rudolph, D. Time-dependent scour development under combined current and waves conditions—Laboratory experiments with online monitoring technique. In Proceedings of the 4th International Conference on Scour and Erosion, Tokyo, Japan, 5–7 November 2008; pp. 152–161. [Google Scholar]
  35. Khalfin, I.S. Modeling and calculation of bed score around large-diameter vertical cylinder under wave action. Water Resour. 200734, 357. [Google Scholar] [CrossRef][Green Version]
  36. Zanke, U.C.; Hsu, T.-W.; Roland, A.; Link, O.; Diab, R. Equilibrium scour depths around piles in noncohesive sediments under currents and waves. Coast. Eng. 201158, 986–991. [Google Scholar] [CrossRef]
  37. Myrhaug, D.; Rue, H. Scour below pipelines and around vertical piles in random waves. Coast. Eng. 200348, 227–242. [Google Scholar] [CrossRef]
  38. Myrhaug, D.; Ong, M.C.; Føien, H.; Gjengedal, C.; Leira, B.J. Scour below pipelines and around vertical piles due to second-order random waves plus a current. Ocean Eng. 200936, 605–616. [Google Scholar] [CrossRef]
  39. Myrhaug, D.; Ong, M.C. Random wave-induced onshore scour characteristics around submerged breakwaters using a stochastic method. Ocean Eng. 201037, 1233–1238. [Google Scholar] [CrossRef]
  40. Ong, M.C.; Myrhaug, D.; Hesten, P. Scour around vertical piles due to long-crested and short-crested nonlinear random waves plus a current. Coast. Eng. 201373, 106–114. [Google Scholar] [CrossRef]
  41. Yakhot, V.; Orszag, S.A. Renormalization group analysis of turbulence. I. Basic theory. J. Sci. Comput. 19861, 3–51. [Google Scholar] [CrossRef]
  42. Yakhot, V.; Smith, L.M. The renormalization group, the e-expansion and derivation of turbulence models. J. Sci. Comput. 19927, 35–61. [Google Scholar] [CrossRef]
  43. Mastbergen, D.R.; Berg, J.V.D. Breaching in fine sands and the generation of sustained turbidity currents in submarine canyons. Sedimentology 200350, 625–637. [Google Scholar] [CrossRef]
  44. Soulsby, R. Dynamics of Marine Sands; Thomas Telford Ltd.: London, UK, 1998. [Google Scholar] [CrossRef]
  45. Van Rijn, L.C. Sediment Transport, Part I: Bed Load Transport. J. Hydraul. Eng. 1984110, 1431–1456. [Google Scholar] [CrossRef][Green Version]
  46. Zhang, Q.; Zhou, X.-L.; Wang, J.-H. Numerical investigation of local scour around three adjacent piles with different arrangements under current. Ocean Eng. 2017142, 625–638. [Google Scholar] [CrossRef]
  47. Yu, Y.X.; Liu, S.X. Random Wave and Its Applications to Engineering, 4th ed.; Dalian University of Technology Press: Dalian, China, 2011. [Google Scholar]
  48. Pang, A.; Skote, M.; Lim, S.; Gullman-Strand, J.; Morgan, N. A numerical approach for determining equilibrium scour depth around a mono-pile due to steady currents. Appl. Ocean Res. 201657, 114–124. [Google Scholar] [CrossRef]
  49. Higuera, P.; Lara, J.L.; Losada, I.J. Three-dimensional interaction of waves and porous coastal structures using Open-FOAM®. Part I: Formulation and validation. Coast. Eng. 201483, 243–258. [Google Scholar] [CrossRef]
  50. Corvaro, S.; Crivellini, A.; Marini, F.; Cimarelli, A.; Capitanelli, L.; Mancinelli, A. Experimental and Numerical Analysis of the Hydrodynamics around a Vertical Cylinder in Waves. J. Mar. Sci. Eng. 20197, 453. [Google Scholar] [CrossRef][Green Version]
  51. Flow3D User Manual, version 11.0.3; Flow Science, Inc.: Santa Fe, NM, USA, 2013.
  52. Khosronejad, A.; Kang, S.; Sotiropoulos, F. Experimental and computational investigation of local scour around bridge piers. Adv. Water Resour. 201237, 73–85. [Google Scholar] [CrossRef]
  53. Stahlmann, A. Experimental and Numerical Modeling of Scour at Foundation Structures for Offshore Wind Turbines. Ph.D. Thesis, Franzius-Institute for Hydraulic, Estuarine and Coastal Engineering, Leibniz Universität Hannover, Hannover, Germany, 2013. [Google Scholar]
  54. Breusers, H.N.C.; Nicollet, G.; Shen, H. Local Scour Around Cylindrical Piers. J. Hydraul. Res. 197715, 211–252. [Google Scholar] [CrossRef]
  55. Schendel, A.; Hildebrandt, A.; Goseberg, N.; Schlurmann, T. Processes and evolution of scour around a monopile induced by tidal currents. Coast. Eng. 2018139, 65–84. [Google Scholar] [CrossRef]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

      

MDPI and ACS Style

Hu, R.; Liu, H.; Leng, H.; Yu, P.; Wang, X. Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves. J. Mar. Sci. Eng. 20219, 886. https://doi.org/10.3390/jmse9080886

AMA Style

Hu R, Liu H, Leng H, Yu P, Wang X. Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves. Journal of Marine Science and Engineering. 2021; 9(8):886. https://doi.org/10.3390/jmse9080886Chicago/Turabian Style

Hu, Ruigeng, Hongjun Liu, Hao Leng, Peng Yu, and Xiuhai Wang. 2021. “Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves” Journal of Marine Science and Engineering 9, no. 8: 886. https://doi.org/10.3390/jmse9080886

Find Other Styles

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

For more information on the journal statistics, click here.

Multiple requests from the same IP address are counted as one view.

Fig. 1 Oscillation of a free surface due to the step reduction of gravity acceleration from kzi ≈ 9.81 to kz ≈ 0

Reorientation of Cryogenic Fluids Upon Step Reduction of Gravity

단계적 중력 감소 시 극저온 유체의 방향 전환

Malte Stief∗, Jens Gerstmann∗∗, and Michael E. Dreyer∗∗∗
ZARM, Center of Applied Space Technology and Microgravity, University of Bremen, Am Fallturm, D-28359 Bremen
Experiments to observe the surface oscillation of cryogenic liquids have been performed with liquid nitrogen inside a 50 mm
diameter right circular cylinder. The surface oscillation is driven by the capillary force that becomes dominant after a sudden
reduction of the gravity acceleration acting on the liquid. The experiments show differences from the speculated behavior and
enables one to observe new features.

Introduction and motivation

최근 몇 년 동안 Bremen의 낙하탑에서 중력의 단계적 감소 시 방향 재지향 거동과 표면 진동을 조사하기 위해 수많은 실험이 수행되었습니다[1]. 이 실험의 원리는 그림 1에 나와 있습니다.

그림 1의 왼쪽에 표시된 것처럼 오른쪽 원형 원통형 용기에 테스트 액체를 레벨 h0까지 채웁니다. 처음에 액체는 정지 상태이며 중앙에서 평평한 인터페이스를 형성합니다.

초기 중력 가속도 kzi ≈ 9.81 [m/s2]와 결과적으로 높은 BOND 수(Bo = ρkziR2/σ)로 인해 실린더의 대칭축에서. 낙하탑에서 실험 캡슐의 방출에 의해 확립된 μ-중력 환경 kz ≈ 0 [m/s2]로의 갑작스러운 전환과 함께 자유 표면은 진동 운동으로 새로운 평형 구성을 찾기 시작합니다(그림의 오른쪽) 1). 이러한 움직임은 그림 1의 중앙에 스케치되어 있습니다.

표면 진동의 구동력은 접착력과 결합된 표면 장력이며, 댐핑은 액체의 점도에 의해 제어됩니다. 위치가 zw인 벽에서 접촉선의 이동은 접촉각 γ에 의해 제어됩니다. 접촉각이 작은 액체용 γ ≈ 0◦

In recent years numerous experiments have been carried out to investigate the reorientation behavior and surface oscillations upon step reduction of gravity at the drop tower in Bremen [1]. The principals of these experiments are shown in figure 1. A right circular cylindrical container is filled up to the level h0 with the test liquid, as shown on the left of figure 1. Initially the liquid is quiescent and forms a flat interface at the center, in the symmetry axis of the cylinder, due to the initial gravity acceleration kzi ≈ 9.81 [m/s2] and the resulting high BOND number (Bo = ρkziR2/σ). With the sudden transition to the µ-gravity environment kz ≈ 0 [m/s2], which is established by the release of the experiment capsular in the drop tower, the free surface is initiated to search its new equilibrium configuration (right side of figure 1) with an oscillatory motion. These movements are sketched in the center of figure 1. The driving force for the surface oscillation is the surface tension in combination with the adhesion force where the damping is controlled by the viscosity of the liquid. The movement of the contact line at the wall, with its position zw, is governed by the contact angle γ. For liquids with small contact angle γ ≈ 0◦

Fig. 1 Oscillation of a free surface due to the step reduction of gravity acceleration from kzi ≈ 9.81 to kz ≈ 0
Fig. 1 Oscillation of a free surface due to the step reduction of gravity acceleration from kzi ≈ 9.81 to kz ≈ 0
Fig. 2 Experiment picture-series showing the oscillation of the free surface at different times for a 50 mm diameter cylinder.
Fig. 2 Experiment picture-series showing the oscillation of the free surface at different times for a 50 mm diameter cylinder.

References

[1] M. Michaelis, Kapillarinduzierte Schwingungen freier Fl¨ussigkeitsoberfl¨achen, Dissertation Universit¨at Bremen, Fortschritt-Berichte
Nr. 454 (VDI Verlag, D¨usseldorf, 2003).

Figure 5. Schematic view of flap and support structure [32]

Design Optimization of Ocean Renewable Energy Converter Using a Combined Bi-level Metaheuristic Approach

결합된 Bi-level 메타휴리스틱 접근법을 사용한 해양 재생 에너지 변환기의 설계 최적화

Erfan Amini a1, Mahdieh Nasiri b1, Navid Salami Pargoo a, Zahra Mozhgani c, Danial Golbaz d, Mehrdad Baniesmaeil e, Meysam Majidi Nezhad f, Mehdi Neshat gj, Davide Astiaso Garcia h, Georgios Sylaios i

Abstract

In recent years, there has been an increasing interest in renewable energies in view of the fact that fossil fuels are the leading cause of catastrophic environmental consequences. Ocean wave energy is a renewable energy source that is particularly prevalent in coastal areas. Since many countries have tremendous potential to extract this type of energy, a number of researchers have sought to determine certain effective factors on wave converters’ performance, with a primary emphasis on ambient factors. In this study, we used metaheuristic optimization methods to investigate the effects of geometric factors on the performance of an Oscillating Surge Wave Energy Converter (OSWEC), in addition to the effects of hydrodynamic parameters. To do so, we used CATIA software to model different geometries which were then inserted into a numerical model developed in Flow3D software. A Ribed-surface design of the converter’s flap is also introduced in this study to maximize wave-converter interaction. Besides, a Bi-level Hill Climbing Multi-Verse Optimization (HCMVO) method was also developed for this application. The results showed that the converter performs better with greater wave heights, flap freeboard heights, and shorter wave periods. Additionally, the added ribs led to more wave-converter interaction and better performance, while the distance between the flap and flume bed negatively impacted the performance. Finally, tracking the changes in the five-dimensional objective function revealed the optimum value for each parameter in all scenarios. This is achieved by the newly developed optimization algorithm, which is much faster than other existing cutting-edge metaheuristic approaches.

Keywords

Wave Energy Converter

OSWEC

Hydrodynamic Effects

Geometric Design

Metaheuristic Optimization

Multi-Verse Optimizer

1Introduction

The increase in energy demand, the limitations of fossil fuels, as well as environmental crises, such as air pollution and global warming, are the leading causes of calling more attention to harvesting renewable energy recently [1][2][3]. While still in its infancy, ocean wave energy has neither reached commercial maturity nor technological convergence. In recent decades, remarkable progress has been made in the marine energy domain, which is still in the early stage of development, to improve the technology performance level (TPL) [4][5]and technology readiness level (TRL) of wave energy converters (WECs). This has been achieved using novel modeling techniques [6][7][8][9][10][11][12][13][14] to gain the following advantages [15]: (i) As a source of sustainable energy, it contributes to the mix of energy resources that leads to greater diversity and attractiveness for coastal cities and suppliers. [16] (ii) Since wave energy can be exploited offshore and does not require any land, in-land site selection would be less expensive and undesirable visual effects would be reduced. [17] (iii) When the best layout and location of offshore site are taken into account, permanent generation of energy will be feasible (as opposed to using solar energy, for example, which is time-dependent) [18].

In general, the energy conversion process can be divided into three stages in a WEC device, including primary, secondary, and tertiary stages [19][20]. In the first stage of energy conversion, which is the subject of this study, the wave power is converted to mechanical power by wave-structure interaction (WSI) between ocean waves and structures. Moreover, the mechanical power is transferred into electricity in the second stage, in which mechanical structures are coupled with power take-off systems (PTO). At this stage, optimal control strategies are useful to tune the system dynamics to maximize power output [10][13][12]. Furthermore, the tertiary energy conversion stage revolves around transferring the non-standard AC power into direct current (DC) power for energy storage or standard AC power for grid integration [21][22]. We discuss only the first stage regardless of the secondary and tertiary stages. While Page 1 of 16 WECs include several categories and technologies such as terminators, point absorbers, and attenuators [15][23], we focus on oscillating surge wave energy converters (OSWECs) in this paper due to its high capacity for industrialization [24].

Over the past two decades, a number of studies have been conducted to understand how OSWECs’ structures and interactions between ocean waves and flaps affect converters performance. Henry et al.’s experiment on oscillating surge wave energy converters is considered as one of the most influential pieces of research [25], which demonstrated how the performance of oscillating surge wave energy converters (OSWECs) is affected by seven different factors, including wave period, wave power, flap’s relative density, water depth, free-board of the flap, the gap between the tubes, gap underneath the flap, and flap width. These parameters were assessed in their two models in order to estimate the absorbed energy from incoming waves [26][27]. In addition, Folly et al. investigated the impact of water depth on the OSWECs performance analytically, numerically, and experimentally. According to this and further similar studies, the average annual incident wave power is significantly reduced by water depth. Based on the experimental results, both the surge wave force and the power capture of OSWECs increase in shallow water [28][29]. Following this, Sarkar et al. found that under such circumstances, the device that is located near the coast performs much better than those in the open ocean [30]. On the other hand, other studies are showing that the size of the converter, including height and width, is relatively independent of the location (within similar depth) [31]. Subsequently, Schmitt et al. studied OSWECs numerically and experimentally. In fact, for the simulation of OSWEC, OpenFOAM was used to test the applicability of Reynolds-averaged Navier-Stokes (RANS) solvers. Then, the experimental model reproduced the numerical results with satisfying accuracy [32]. In another influential study, Wang et al. numerically assessed the effect of OSWEC’s width on their performance. According to their findings, as converter width increases, its efficiency decreases in short wave periods while increases in long wave periods [33]. One of the main challenges in the analysis of the OSWEC is the coupled effect of hydrodynamic and geometric variables. As a result, numerous cutting-edge geometry studies have been performed in recent years in order to find the optimal structure that maximizes power output and minimizes costs. Garcia et al. reviewed hull geometry optimization studies in the literature in [19]. In addition, Guo and Ringwood surveyed geometric optimization methods to improve the hydrodynamic performance of OSWECs at the primary stage [14]. Besides, they classified the hull geometry of OSWECs based on Figure 1. Subsequently, Whittaker et al. proposed a different design of OSWEC called Oyster2. There have been three examples of different geometries of oysters with different water depths. Based on its water depth, they determined the width and height of the converter. They also found that in the constant wave period the less the converter’s width, the less power captures the converter has [34]. Afterward, O’Boyle et al. investigated a type of OSWEC called Oyster 800. They compared the experimental and numerical models with the prototype model. In order to precisely reproduce the shape, mass distribution, and buoyancy properties of the prototype, a 40th-scale experimental model has been designed. Overall, all the models were fairly accurate according to the results [35].

Inclusive analysis of recent research avenues in the area of flap geometry has revealed that the interaction-based designs of such converters are emerging as a novel approach. An initiative workflow is designed in the current study to maximizing the wave energy extrication by such systems. To begin with, a sensitivity analysis plays its role of determining the best hydrodynamic values for installing the converter’s flap. Then, all flap dimensions and characteristics come into play to finalize the primary model. Following, interactive designs is proposed to increase the influence of incident waves on the body by adding ribs on both sides of the flap as a novel design. Finally, a new bi-level metaheuristic method is proposed to consider the effects of simultaneous changes in ribs properties and other design parameters. We hope this novel approach will be utilized to make big-scale projects less costly and justifiable. The efficiency of the method is also compared with four well known metaheuristic algorithms and out weight them for this application.

This paper is organized as follows. First, the research methodology is introduced by providing details about the numerical model implementation. To that end, we first introduced the primary model’s geometry and software details. That primary model is later verified with a benchmark study with regard to the flap angle of rotation and water surface elevation. Then, governing equations and performance criteria are presented. In the third part of the paper, we discuss the model’s sensitivity to lower and upper parts width (we proposed a two cross-sectional design for the flap), bottom elevation, and freeboard. Finally, the novel optimization approach is introduced in the final part and compared with four recent metaheuristic algorithms.

2. Numerical Methods

In this section, after a brief introduction of the numerical software, Flow3D, boundary conditions are defined. Afterwards, the numerical model implementation, along with primary model properties are described. Finally, governing equations, as part of numerical process, are discussed.

2.1Model Setup

FLOW-3D is a powerful and comprehensive CFD simulation platform for studying fluid dynamics. This software has several modules to solve many complex engineering problems. In addition, modeling complex flows is simple and effective using FLOW-3D’s robust meshing capabilities [36]. Interaction between fluid and moving objects might alter the computational range. Dynamic meshes are used in our modeling to take these changes into account. At each time step, the computational node positions change in order to adapt the meshing area to the moving object. In addition, to choose mesh dimensions, some factors are taken into account such as computational accuracy, computational time, and stability. The final grid size is selected based on the detailed procedure provided in [37]. To that end, we performed grid-independence testing on a CFD model using three different mesh grid sizes of 0.01, 0.015, and 0.02 meters. The problem geometry and boundary conditions were defined the same, and simulations were run on all three grids under the same conditions. The predicted values of the relevant variable, such as velocity, was compared between the grids. The convergence behavior of the numerical solution was analyzed by calculating the relative L2 norm error between two consecutive grids. Based on the results obtained, it was found that the grid size of 0.02 meters showed the least error, indicating that it provided the most accurate and reliable solution among the three grids. Therefore, the grid size of 0.02 meters was selected as the optimal spatial resolution for the mesh grid.

In this work, the flume dimensions are 10 meters long, 0.1 meters wide, and 2.2 meters high, which are shown in figure2. In addition, input waves with linear characteristics have a height of 0.1 meters and a period of 1.4 seconds. Among the linear wave methods included in this software, RNGk-ε and k- ε are appropriate for turbulence model. The research of Lopez et al. shows that RNGk- ε provides the most accurate simulation of turbulence in OSWECs [21]. We use CATIA software to create the flap primary model and other innovative designs for this project. The flap measures 0.1 m x 0.65 m x 0.360 m in x, y and z directions, respectively. In Figure 3, the primary model of flap and its dimensions are shown. In this simulation, five boundaries have been defined, including 1. Inlet, 2. Outlet, 3. Converter flap, 4. Bed flume, and 5. Water surface, which are shown in figure 2. Besides, to avoid wave reflection in inlet and outlet zones, Flow3D is capable of defining some areas as damping zones, the length of which has to be one to one and a half times the wavelength. Therefore, in the model, this length is considered equal to 2 meters. Furthermore, there is no slip in all the boundaries. In other words, at every single time step, the fluid velocity is zero on the bed flume, while it is equal to the flap velocity on the converter flap. According to the wave theory defined in the software, at the inlet boundary, the water velocity is called from the wave speed to be fed into the model.

2.2Verification

In the current study, we utilize the Schmitt experimental model as a benchmark for verification, which was developed at the Queen’s University of Belfast. The experiments were conducted on the flap of the converter, its rotation, and its interaction with the water surface. Thus, the details of the experiments are presented below based up on the experimental setup’s description [38]. In the experiment, the laboratory flume has a length of 20m and a width of 4.58m. Besides, in order to avoid incident wave reflection, a wave absorption source is devised at the end of the left flume. The flume bed, also, includes two parts with different slops. The flap position and dimensions of the flume can be seen in Figure4. In addition, a wave-maker with 6 paddles is installed at one end. At the opposite end, there is a beach with wire meshes. Additionally, there are 6 indicators to extract the water level elevation. In the flap model, there are three components: the fixed support structure, the hinge, and the flap. The flap measures 0.1m x 0.65m x 0.341m in x, y and z directions, respectively. In Figure5, the details are given [32]. The support structure consists of a 15 mm thick stainless steel base plate measuring 1m by 1.4m, which is screwed onto the bottom of the tank. The hinge is supported by three bearing blocks. There is a foam centerpiece on the front and back of the flap which is sandwiched between two PVC plates. Enabling changes of the flap, three metal fittings link the flap to the hinge. Moreover, in this experiment, the selected wave is generated based on sea wave data at scale 1:40. The wave height and the wave period are equal to 0.038 (m) and 2.0625 (s), respectively, which are tantamount to a wave with a period of 13 (s) and a height of 1.5 (m).

Two distinct graphs illustrate the numerical and experi-mental study results. Figure6 and Figure7 are denoting the angle of rotation of flap and surface elevation in computational and experimental models, respectively. The two figures roughly represent that the numerical and experimental models are a good match. However, for the purpose of verifying the match, we calculated the correlation coefficient (C) and root mean square error (RMSE). According to Figure6, correlation coefficient and RMSE are 0.998 and 0.003, respectively, and in Figure7 correlation coefficient and RMSE are respectively 0.999 and 0.001. Accordingly, there is a good match between the numerical and empirical models. It is worth mentioning that the small differences between the numerical and experimental outputs may be due to the error of the measuring devices and the calibration of the data collection devices.

Including continuity equation and momentum conserva- tion for incompressible fluid are given as [32][39]:(1)

where P represents the pressure, g denotes gravitational acceleration, u represents fluid velocity, and Di is damping coefficient. Likewise, the model uses the same equation. to calculate the fluid velocity in other directions as well. Considering the turbulence, we use the two-equation model of RNGK- ε. These equations are:

(3)��t(��)+����(����)=����[�eff�������]+��-��and(4)���(��)+����(����)=����[�eff�������]+�1�∗����-��2��2�Where �2� and �1� are constants. In addition, �� and �� represent the turbulent Prandtl number of � and k, respectively.

�� also denote the production of turbulent kinetic energy of k under the effect of velocity gradient, which is calculated as follows:(5)��=�eff[�����+�����]�����(6)�eff=�+��(7)�eff=�+��where � is molecular viscosity,�� represents turbulence viscosity, k denotes kinetic energy, and ∊∊ is energy dissipation rate. The values of constant coefficients in the two-equation RNGK ∊-∊ model is as shown in the Table 1 [40].Table 2.

Table 1. Constant coefficients in RNGK- model

Factors�0�1�2������
Quantity0.0124.381.421.681.391.390.084

Table 2. Flap properties

Joint height (m)0.476
Height of the center of mass (m)0.53
Weight (Kg)10.77

It is worth mentioning that the volume of fluid method is used to separate water and air phases in this software [41]. Below is the equation of this method [40].(8)����+����(���)=0where α and 1 − α are portion of water phase and air phase, respectively. As a weighting factor, each fluid phase portion is used to determine the mixture properties. Finally, using the following equations, we calculate the efficiency of converters [42][34][43]:(9)�=14|�|2�+�2+(�+�a)2(�n2-�2)2where �� represents natural frequency, I denotes the inertia of OSWEC, Ia is the added inertia, F is the complex wave force, and B denotes the hydrodynamic damping coefficient. Afterward, the capture factor of the converter is calculated by [44]:(10)��=�1/2��2����gw where �� represents the capture factor, which is the total efficiency of device per unit length of the wave crest at each time step [15], �� represent the dimensional amplitude of the incident wave, w is the flap’s width, and Cg is the group velocity of the incident wave, as below:(11)��=��0·121+2�0ℎsinh2�0ℎwhere �0 denotes the wave number, h is water depth, and H is the height of incident waves.

According to previous sections ∊,����-∊ modeling is used for all models simulated in this section. For this purpose, the empty boundary condition is used for flume walls. In order to preventing wave reflection at the inlet and outlet of the flume, the length of wave absorption is set to be at least one incident wavelength. In addition, the structured mesh is chosen, and the mesh dimensions are selected in two distinct directions. In each model, all grids have a length of 2 (cm) and a height of 1 (cm). Afterwards, as an input of the software for all of the models, we define the time step as 0.001 (s). Moreover, the run time of every simulation is 30 (s). As mentioned before, our primary model is Schmitt model, and the flap properties is given in table2. For all simulations, the flume measures 15 meters in length and 0.65 meters in width, and water depth is equal to 0.335 (m). The flap is also located 7 meters from the flume’s inlet.

Finally, in order to compare the results, the capture factor is calculated for each simulation and compared to the primary model. It is worth mentioning that capture factor refers to the ratio of absorbed wave energy to the input wave energy.

According to primary model simulation and due to the decreasing horizontal velocity with depth, the wave crest has the highest velocity. Considering the fact that the wave’s orbital velocity causes the flap to move, the contact between the upper edge of the flap and the incident wave can enhance its performance. Additionally, the numerical model shows that the dynamic pressure decreases as depth increases, and the hydrostatic pressure increases as depth increases.

To determine the OSWEC design, it is imperative to understand the correlation between the capture factor, wave period, and wave height. Therefore, as it is shown in Figure8, we plot the change in capture factor over the variations in wave period and wave height in 3D and 2D. In this diagram, the first axis features changes in wave period, the second axis displays changes in wave height, and the third axis depicts changes in capture factor. According to our wave properties in the numerical model, the wave period and wave height range from 2 to 14 seconds and 2 to 8 meters, respectively. This is due to the fact that the flap does not oscillate if the wave height is less than 2 (m), and it does not reverse if the wave height is more than 8 (m). In addition, with wave periods more than 14 (s), the wavelength would be so long that it would violate the deep-water conditions, and with wave periods less than 2 (s), the flap would not oscillate properly due to the shortness of wavelength. The results of simulation are shown in Figure 8. As it can be perceived from Figure 8, in a constant wave period, the capture factor is in direct proportion to the wave height. It is because of the fact that waves with more height have more energy to rotate the flap. Besides, in a constant wave height, the capture factor increases when the wave period increases, until a given wave period value. However, the capture factor falls after this point. These results are expected since the flap’s angular displacement is not high in lower wave periods, while the oscillating motion of that is not fast enough to activate the power take-off system in very high wave periods.

As is shown in Figure 9, we plot the change in capture factor over the variations in wave period (s) and water depth (m) in 3D. As it can be seen in this diagram, the first axis features changes in water depth (m), the second axis depicts the wave period (s), and the third axis displays OSWEC’s capture factor. The wave period ranges from 0 to 10 seconds based on our wave properties, which have been adopted from Schmitt’s model, while water depth ranges from 0 to 0.5 meters according to the flume and flap dimensions and laboratory limitations. According to Figure9, for any specific water depth, the capture factor increases in a varying rate when the wave period increases, until a given wave period value. However, the capture factor falls steadily after this point. In fact, the maximum capture factor occurs when the wave period is around 6 seconds. This trend is expected since, in a specific water depth, the flap cannot oscillate properly when the wavelength is too short. As the wave period increases, the flap can oscillate more easily, and consequently its capture factor increases. However, the capture factor drops in higher wave periods because the wavelength is too large to move the flap. Furthermore, in a constant wave period, by changing the water depth, the capture factor does not alter. In other words, the capture factor does not depend on the water depth when it is around its maximum value.

3Sensitivity Analysis

Based on previous studies, in addition to the flap design, the location of the flap relative to the water surface (freeboard) and its elevation relative to the flume bed (flap bottom elevation) play a significant role in extracting energy from the wave energy converter. This study measures the sensitivity of the model to various parameters related to the flap design including upper part width of the flap, lower part width of the flap, the freeboard, and the flap bottom elevation. Moreover, as a novel idea, we propose that the flap widths differ in the lower and upper parts. In Figure10, as an example, a flap with an upper thickness of 100 (mm) and a lower thickness of 50 (mm) and a flap with an upper thickness of 50 (mm) and a lower thickness of 100 (mm) are shown. The influence of such discrepancy between the widths of the upper and lower parts on the interaction between the wave and the flap, or in other words on the capture factor, is evaluated. To do so, other parameters are remained constant, such as the freeboard, the distance between the flap and the flume bed, and the wave properties.

In Figure11, models are simulated with distinct upper and lower widths. As it is clear in this figure, the first axis depicts the lower part width of the flap, the second axis indicates the upper part width of the flap, and the colors represent the capture factor values. Additionally, in order to consider a sufficient range of change, the flap thickness varies from half to double the value of the primary model for each part.

According to this study, the greater the discrepancy in these two parts, the lower the capture factor. It is on account of the fact that when the lower part of the flap is thicker than the upper part, and this thickness difference in these two parts is extremely conspicuous, the inertia against the motion is significant at zero degrees of rotation. Consequently, it is difficult to move the flap, which results in a low capture factor. Similarly, when the upper part of the flap is thicker than the lower part, and this thickness difference in these two parts is exceedingly noticeable, the inertia is so great that the flap can not reverse at the maximum degree of rotation. As the results indicate, the discrepancy can enhance the performance of the converter if the difference between these two parts is around 20%. As it is depicted in the Figure11, the capture factor reaches its own maximum amount, when the lower part thickness is from 5 to 6 (cm), and the upper part thickness is between 6 and 7 (cm). Consequently, as a result of this discrepancy, less material will be used, and therefore there will be less cost.

As illustrated in Figure12, this study examines the effects of freeboard (level difference between the flap top and water surface) and the flap bottom elevation (the distance between the flume bed and flap bottom) on the converter performance. In this diagram, the first axis demonstrates the freeboard and the second axis on the left side displays the flap bottom elevation, while the colors indicate the capture factor. In addition, the feasible range of freeboard is between -15 to 15 (cm) due to the limitation of the numerical model, so that we can take the wave slamming and the overtopping into consideration. Additionally, based on the Schmitt model and its scaled model of 1:40 of the base height, the flap bottom should be at least 9 (cm) high. Since the effect of surface waves is distributed over the depth of the flume, it is imperative to maintain a reasonable flap height exposed to incoming waves. Thus, the maximum flap bottom elevation is limited to 19 (cm). As the Figure12 pictures, at constant negative values of the freeboard, the capture factor is in inverse proportion with the flap bottom elevation, although slightly.

Furthermore, at constant positive values of the freeboard, the capture factor fluctuates as the flap bottom elevation decreases while it maintains an overall increasing trend. This is on account of the fact that increasing the flap bottom elevation creates turbulence flow behind the flap, which encumbers its rotation, as well as the fact that the flap surface has less interaction with the incoming waves. Furthermore, while keeping the flap bottom elevation constant, the capture factor increases by raising the freeboard. This is due to the fact that there is overtopping with adverse impacts on the converter performance when the freeboard is negative and the flap is under the water surface. Besides, increasing the freeboard makes the wave slam more vigorously, which improves the converter performance.

Adding ribs to the flap surface, as shown in Figure13, is a novel idea that is investigated in the next section. To achieve an optimized design for the proposed geometry of the flap, we determine the optimal number and dimensions of ribs based on the flap properties as our decision variables in the optimization process. As an example, Figure13 illustrates a flap with 3 ribs on each side with specific dimensions.

Figure14 shows the flow velocity field around the flap jointed to the flume bed. During the oscillation of the flap, the pressure on the upper and lower surfaces of the flap changes dynamically due to the changing angle of attack and the resulting change in the direction of fluid flow. As the flap moves upwards, the pressure on the upper surface decreases, and the pressure on the lower surface increases. Conversely, as the flap moves downwards, the pressure on the upper surface increases, and the pressure on the lower surface decreases. This results in a cyclic pressure variation around the flap. Under certain conditions, the pressure field around the flap can exhibit significant variations in magnitude and direction, forming vortices and other flow structures. These flow structures can affect the performance of the OSWEC by altering the lift and drag forces acting on the flap.

4Design Optimization

We consider optimizing the design parameters of the flap of converter using a nature-based swarm optimization method, that fall in the category of metaheuristic algorithms [45]. Accordingly, we choose four state-of-the-art algorithms to perform an optimization study. Then, based on their performances to achieve the highest capture factor, one of them will be chosen to be combined with the Hill Climb algorithm to carry out a local search. Therefore, in the remainder of this section, we discuss the search process of each algorithm and visualize their performance and convergence curve as they try to find the best values for decision variables.

4.1. Metaheuristic Approaches

As the first considered algorithm, the Gray Wolf Optimizer (GWO) algorithm simulates the natural leadership and hunting performance of gray wolves which tend to live in colonies. Hunters must obey the alpha wolf, the leader, who is responsible for hunting. Then, the beta wolf is at the second level of the gray wolf hierarchy. A subordinate of alpha wolf, beta stands under the command of the alpha. At the next level in this hierarchy, there are the delta wolves. They are subordinate to the alpha and beta wolves. This category of wolves includes scouts, sentinels, elders, hunters, and caretakers. In this ranking, omega wolves are at the bottom, having the lowest level and obeying all other wolves. They are also allowed to eat the prey just after others have eaten. Despite the fact that they seem less important than others, they are really central to the pack survival. Since, it has been shown that without omega wolves, the entire pack would experience some problems like fighting, violence, and frustration. In this simulation, there are three primary steps of hunting including searching, surrounding, and finally attacking the prey. Mathematically model of gray wolves’ hunting technique and their social hierarchy are applied in determined by optimization. this study. As mentioned before, gray wolves can locate their prey and surround them. The alpha wolf also leads the hunt. Assuming that the alpha, beta, and delta have more knowledge about prey locations, we can mathematically simulate gray wolf hunting behavior. Hence, in addition to saving the top three best solutions obtained so far, we compel the rest of the search agents (also the omegas) to adjust their positions based on the best search agent. Encircling behavior can be mathematically modeled by the following equations: [46].(12)�→=|�→·��→(�)-�→(�)|(13)�→(�+1)=��→(�)-�→·�→(14)�→=2.�2→(15)�→=2�→·�1→-�→Where �→indicates the position vector of gray wolf, ��→ defines the vector of prey, t indicates the current iteration, and �→and �→are coefficient vectors. To force the search agent to diverge from the prey, we use �→ with random values greater than 1 or less than -1. In addition, C→ contains random values in the range [0,2], and �→ 1 and �2→ are random vectors in [0,1]. The second considered technique is the Moth Flame Optimizer (MFO) algorithm. This method revolves around the moths’ navigation mechanism, which is realized by positioning themselves and maintaining a fixed angle relative to the moon while flying. This effective mechanism helps moths to fly in a straight path. However, when the source of light is artificial, maintaining an angle with the light leads to a spiral flying path towards the source that causes the moth’s death [47]. In MFO algorithm, moths and flames are both solutions. The moths are actual search agents that fly in hyper-dimensional space by changing their position vectors, and the flames are considered pins that moths drop when searching the search space [48]. The problem’s variables are the position of moths in the space. Each moth searches around a flame and updates it in case of finding a better solution. The fitness value is the return value of each moth’s fitness (objective) function. The position vector of each moth is passed to the fitness function, and the output of the fitness function is assigned to the corresponding moth. With this mechanism, a moth never loses its best solution [49]. Some attributes of this algorithm are as follows:

  • •It takes different values to converge moth in any point around the flame.
  • •Distance to the flame is lowered to be eventually minimized.
  • •When the position gets closer to the flame, the updated positions around the flame become more frequent.

As another method, the Multi-Verse Optimizer is based on a multiverse theory which proposes there are other universes besides the one in which we all live. According to this theory, there are more than one big bang in the universe, and each big bang leads to the birth of a new universe [50]. Multi-Verse Optimizer (MVO) is mainly inspired by three phenomena in cosmology: white holes, black holes, and wormholes. A white hole has never been observed in our universe, but physicists believe the big bang could be considered a white hole [51]. Black holes, which behave completely in contrast to white holes, attract everything including light beams with their extremely high gravitational force [52]. In the multiverse theory, wormholes are time and space tunnels that allow objects to move instantly between any two corners of a universe (or even simultaneously from one universe to another) [53]. Based on these three concepts, mathematical models are designed to perform exploration, exploitation, and local search, respectively. The concept of white and black holes is implied as an exploration phase, while the concept of wormholes is considered as an exploitation phase by MVO. Additionally, each solution is analogous to a universe, and each variable in the solution represents an object in that universe. Furthermore, each solution is assigned an inflation rate, and the time is used instead of iterations. Following are the universe rules in MVO:

  • •The possibility of having white hole increases with the inflation rate.
  • •The possibility of having black hole decreases with the inflation rate.
  • •Objects tend to pass through black holes more frequently in universes with lower inflation rates.
  • •Regardless of inflation rate, wormholes may cause objects in universes to move randomly towards the best universe. [54]

Modeling the white/black hole tunnels and exchanging objects of universes mathematically was accomplished by using the roulette wheel mechanism. With every iteration, the universes are sorted according to their inflation rates, then, based on the roulette wheel, the one with the white hole is selected as the local extremum solution. This is accomplished through the following steps:

Assume that

(16)���=����1<��(��)����1≥��(��)

Where ��� represents the jth parameter of the ith universe, Ui indicates the ith universe, NI(Ui) is normalized inflation rate of the ith universe, r1 is a random number in [0,1], and j xk shows the jth parameter of the kth universe selected by a roulette wheel selection mechanism [54]. It is assumed that wormhole tunnels always exist between a universe and the best universe formed so far. This mechanism is as follows:(17)���=if�2<���:��+���×((���-���)×�4+���)�3<0.5��-���×((���-���)×�4+���)�3≥0.5����:���where Xj indicates the jth parameter of the best universe formed so far, TDR and WEP are coefficients, where Xj indicates the jth parameter of the best universelbjshows the lower bound of the jth variable, ubj is the upper bound of the jth variable, and r2, r3, and r4 are random numbers in [1][54].

Finally, one of the newest optimization algorithms is WOA. The WOA algorithm simulates the movement of prey and the whale’s discipline when looking for their prey. Among several species, Humpback whales have a specific method of hunting [55]. Humpback whales can recognize the location of prey and encircle it before hunting. The optimal design position in the search space is not known a priori, and the WOA algorithm assumes that the best candidate solution is either the target prey or close to the optimum. This foraging behavior is called the bubble-net feeding method. Two maneuvers are associated with bubbles: upward spirals and double loops. A unique behavior exhibited only by humpback whales is bubble-net feeding. In fact, The WOA algorithm starts with a set of random solutions. At each iteration, search agents update their positions for either a randomly chosen search agent or the best solution obtained so far [56][55]. When the best search agent is determined, the other search agents will attempt to update their positions toward that agent. It is important to note that humpback whales swim around their prey simultaneously in a circular, shrinking circle and along a spiral-shaped path. By using a mathematical model, the spiral bubble-net feeding maneuver is optimized. The following equation represents this behavior:(18)�→(�+1)=�′→·�bl·cos(2��)+�∗→(�)

Where:(19)�′→=|�∗→(�)-�→(�)|

X→(t+ 1) indicates the distance of the it h whale to the prey (best solution obtained so far),� is a constant for defining the shape of the logarithmic spiral, l is a random number in [−1, 1], and dot (.) is an element-by-element multiplication [55].

Comparing the four above-mentioned methods, simulations are run with 10 search agents for 400 iterations. In Figure 15, there are 20 plots the optimal values of different parameters in optimization algorithms. The five parameters of this study are freeboard, bottom elevations, number of ribs on the converter, rib thickness, and rib Height. The optimal value for each was found by optimization algorithms, naming WOA, MVO, MFO, and GWO. By looking through the first row, the freeboard parameter converges to its maximum possible value in the optimization process of GWO after 300 iterations. Similarly, MFO finds the same result as GWO. In contrast, the freeboard converges to its minimum possible value in MVO optimizing process, which indicates positioning the converter under the water. Furthermore, WOA found the optimal value of freeboard as around 0.02 after almost 200 iterations. In the second row, the bottom elevation is found at almost 0.11 (m) in all algorithms; however, the curves follow different trends in each algorithm. The third row shows the number of ribs, where results immediately reveal that it should be over 4. All algorithms coincide at 5 ribs as the optimal number in this process. The fourth row displays the trends of algorithms to find optimal rib thickness. MFO finds the optimal value early and sets it to around 0.022, while others find the same value in higher iterations. Finally, regarding the rib height, MVO, MFO, and GWO state that the optimal value is 0.06 meters, but WOA did not find a higher value than 0.039.

4.2. HCMVO Bi-level Approach

Despite several strong search characteristics of MVO and its high performance in various optimization problems, it suffers from a few deficiencies in local and global search mechanisms. For instance, it is trapped in the local optimum when wormholes stochastically generate many solutions near the best universe achieved throughout iterations, especially in solving complex multimodal problems with high dimensions [57]. Furthermore, MVO needs to be modified by an escaping strategy from the local optima to enhance the global search abilities. To address these shortages, we propose a fast and effective meta-algorithm (HCMVO) to combine MVO with a Random-restart hill-climbing local search. This meta-algorithm uses MVO on the upper level to develop global tracking and provide a range of feasible and proper solutions. The hill-climbing algorithm is designed to develop a comprehensive neighborhood search around the best-found solution proposed by the upper-level (MVO) when MVO is faced with a stagnation issue or falling into a local optimum. The performance threshold is formulated as follows.(20)Δ����THD=∑�=1�����TH��-����TH��-1�where BestTHDis the best-found solution per generation, andM is related to the domain of iterations to compute the average performance of MVO. If the proposed best solution by the local search is better than the initial one, the global best of MVO will be updated. HCMVO iteratively runs hill climbing when the performance of MVO goes down, each time with an initial condition to prepare for escaping such undesirable situations. In order to get a better balance between exploration and exploitation, the search step size linearly decreases as follows:(21)��=��-����Ma�iter��+1where iter and Maxiter are the current iteration and maximum number of evaluation, respectively. �� stands for the step size of the neighborhood search. Meanwhile, this strategy can improve the convergence rate of MVO compared with other algorithms.

Algorithm 1 shows the technical details of the proposed optimization method (HCMVO). The initial solution includes freeboard (�), bottom elevation (�), number of ribs (Nr), rib thickness (�), and rib height(�).

5. Conclusion

The high trend of diminishing worldwide energy resources has entailed a great crisis upon vulnerable societies. To withstand this effect, developing renewable energy technologies can open doors to a more reliable means, among which the wave energy converters will help the coastal residents and infrastructure. This paper set out to determine the optimized design for such devices that leads to the highest possible power output. The main goal of this research was to demonstrate the best design for an oscillating surge wave energy converter using a novel metaheuristic optimization algorithm. In this regard, the methodology was devised such that it argued the effects of influential parameters, including wave characteristics, WEC design, and interaction criteria.

To begin with, a numerical model was developed in Flow 3D software to simulate the response of the flap of a wave energy converter to incoming waves, followed by a validation study based upon a well-reputed experimental study to verify the accuracy of the model. Secondly, the hydrodynamics of the flap was investigated by incorporating the turbulence. The effect of depth, wave height, and wave period are also investigated in this part. The influence of two novel ideas on increasing the wave-converter interaction was then assessed: i) designing a flap with different widths in the upper and lower part, and ii) adding ribs on the surface of the flap. Finally, four trending single-objective metaheuristic optimization methods

Empty CellAlgorithm 1: Hill Climb Multiverse Optimization
01:procedure HCMVO
02:�=30,�=5▹���������������������������������
03:�=〈F1,B1,N,R,H1〉,…〈FN,B2,N,R,HN〉⇒lb1N⩽�⩽ubN
04:Initialize parameters�ER,�DR,�EP,Best�,���ite��▹Wormhole existence probability (WEP)
05:��=����(��)
06:��=Normalize the inflation rate��
07:for iter in[1,⋯,���iter]do
08:for�in[1,⋯,�]do
09:Update�EP,�DR,Black����Index=�
10:for���[1,⋯,�]��
11:�1=����()
12:if�1≤��(��)then
13:White HoleIndex=Roulette�heelSelection(-��)
14:�(Black HoleIndex,�)=��(White HoleIndex,�)
15:end if
16:�2=����([0,�])
17:if�2≤�EPthen
18:�3=����(),�4=����()
19:if�3<0.5then
20:�1=((��(�)-��(�))�4+��(�))
21:�(�,�)=Best�(�)+�DR�
22:else
23:�(�,�)=Best�(�)-�DR�
24:end if
25:end if
26:end for
27:end for
28:�HD=����([�1,�2,⋯,�Np])
29:Bes�TH�itr=����HD
30:ΔBestTHD=∑�=1�BestTII��-BestTII��-1�
31:ifΔBestTHD<��then▹Perform hill climbing local search
32:BestTHD=����-�lim��������THD
33:end if
34:end for
35:return�,BestTHD▹Final configuration
36:end procedure

The implementation details of the hill-climbing algorithm applied in HCMPA can be seen in Algorithm 2. One of the critical parameters isg, which denotes the resolution of the neighborhood search around the proposed global best by MVO. If we set a small step size for hill-climbing, the convergence speed will be decreased. On the other hand, a large step size reinforces the exploration ability. Still, it may reduce the exploitation ability and in return increase the act of jumping from a global optimum or surfaces with high-potential solutions. Per each decision variable, the neighborhood search evaluates two different direct searches, incremental or decremental. After assessing the generated solutions, the best candidate will be selected to iterate the search algorithm. It is noted that the hill-climbing algorithm should not be applied in the initial iteration of the optimization process due to the immense tendency for converging to local optima. Meanwhile, for optimizing largescale problems, hill-climbing is not an appropriate selection. In order to improve understanding of the proposed hybrid optimization algorithm’s steps, the flowchart of HCMVO is designed and can be seen in Figure 16.

Figure 17 shows the observed capture factor (which is the absorbed energy with respect to the available energy) by each optimization algorithm from iterations 1 to 400. The algorithms use ten search agents in their modified codes to find the optimal solutions. While GWO and MFO remain roughly constant after iterations 54 and 40, the other three algorithms keep improving the capture factor. In this case, HCMVO and MVO worked very well in the optimizing process with a capture factor obtained by the former as 0.594 and by the latter as 0.593. MFO almost found its highest value before the iteration 50, which means the exploration part of the algorithm works out well. Similarly, HCMVO does the same. However, it keeps finding the better solution during the optimization process until the last iteration, indicating the strong exploitation part of the algorithm. GWO reveals a weakness in exploration and exploitation because not only does it evoke the least capture factor value, but also the curve remains almost unchanged throughout 350 iterations.

Figure 18 illustrates complex interactions between the five optimization parameters and the capture factor for HCMVO (a), MPA (b), and MFO (c) algorithms. The first interesting observation is that there is a high level of nonlinear relationships among the setting parameters that can make a multi-modal search space. The dark blue lines represent the best-found configuration throughout the optimisation process. Based on both HCMVO (a) and MVO (b), we can infer that the dark blue lines concentrate in a specific range, showing the high convergence ability of both HCMVO and MVO. However, MFO (c) could not find the exact optimal range of the decision variables, and the best-found solutions per generation distribute mostly all around the search space.

Empty CellAlgorithm 1: Hill Climb Multiverse Optimization
01:procedure HCMVO
02:Initialization
03:Initialize the constraints��1�,��1�
04:�1�=Mi�1�+���1�/�▹Compute the step size,�is search resolution
05:So�1=〈�,�,�,�,�〉▹���������������
06:�������1=����So�1▹���������ℎ���������
07:Main loop
08:for iter≤���ita=do
09:���=���±��
10:while�≤���(Sol1)do
11:���=���+�,▹����ℎ���ℎ��������ℎ
12:fitness��iter=�������
13:t = t+1
14:end while
15:〈�����,������max〉=����������
16:���itev=���Inde�max▹�������ℎ�������������������������������ℎ�������
17:��=��-����Max��+1▹�����������������
18:end for
19:return���iter,����
20:end procedure

were utilized to illuminate the optimum values of the design parameters, and the best method was chosen to develop a new algorithm that performs both local and global search methods.

The correlation between hydrodynamic parameters and the capture factor of the converter was supported by the results. For any given water depth, the capture factor increases as the wave period increases, until a certain wave period value (6 seconds) is reached, after which the capture factor gradually decreases. It is expected since the flap cannot oscillate effectively when the wavelength is too short for a certain water depth. Conversely, when the wavelength is too long, the capture factor decreases. Furthermore, under a constant wave period, increasing the water depth does not affect the capture factor. Regarding the sensitivity analysis, the study found that increasing the flap bottom elevation causes turbulence flow behind the flap and limitation of rotation, which leads to less interaction with the incoming waves. Furthermore, while keeping the flap bottom elevation constant, increasing the freeboard improves the capture factor. Overtopping happens when the freeboard is negative and the flap is below the water surface, which has a detrimental influence on converter performance. Furthermore, raising the freeboard causes the wave impact to become more violent, which increases converter performance.

In the last part, we discussed the search process of each algorithm and visualized their performance and convergence curves as they try to find the best values for decision variables. Among the four selected metaheuristic algorithms, the Multi-verse Optimizer proved to be the most effective in achieving the best answer in terms of the WEC capture factor. However, the MVO needed modifications regarding its escape approach from the local optima in order to improve its global search capabilities. To overcome these constraints, we presented a fast and efficient meta-algorithm (HCMVO) that combines MVO with a Random-restart hill-climbing local search. On a higher level, this meta-algorithm employed MVO to generate global tracking and present a range of possible and appropriate solutions. Taken together, the results demonstrated that there is a significant degree of nonlinearity among the setup parameters that might result in a multimodal search space. Since MVO was faced with a stagnation issue or fell into a local optimum, we constructed a complete neighborhood search around the best-found solution offered by the upper level. In sum, the newly-developed algorithm proved to be highly effective for the problem compared to other similar optimization methods. The strength of the current findings may encourage future investigation on design optimization of wave energy converters using developed geometry as well as the novel approach.

CRediT authorship contribution statement

Erfan Amini: Conceptualization, Methodology, Validation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Mahdieh Nasiri: Conceptualization, Methodology, Validation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Navid Salami Pargoo: Writing – original draft, Writing – review & editing. Zahra Mozhgani: Conceptualization, Methodology. Danial Golbaz: Writing – original draft. Mehrdad Baniesmaeil: Writing – original draft. Meysam Majidi Nezhad: . Mehdi Neshat: Supervision, Conceptualization, Writing – original draft, Writing – review & editing, Visualization. Davide Astiaso Garcia: Supervision. Georgios Sylaios: Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

This research has been carried out within ILIAD (Inte-grated Digital Framework for Comprehensive Maritime Data and Information Services) project that received funding from the European Union’s H2020 programme.

Data availability

Data will be made available on request.

References

Figure 15. Velocity distribution of impinging jet on a wall under different Reynolds numbers.

Hydraulic Characteristics of Continuous Submerged Jet Impinging on a Wall by Using Numerical Simulation and PIV Experiment

by Hongbo Mi 1,2, Chuan Wang 1,3, Xuanwen Jia 3,*, Bo Hu 2, Hongliang Wang 4, Hui Wang 3 and Yong Zhu 5

1College of Mechatronics Engineering, Hainan Vocational University of Science and Technology, Haikou 571126, China

2Department of Energy and Power Engineering, Tsinghua University, Beijing 100084, China

3College of Hydraulic Science and Engineering, Yangzhou University, Yangzhou 225009, China

4School of Aerospace and Mechanical Engineering/Flight College, Changzhou Institute of Technology, Changzhou 213032, China

5National Research Center of Pumps, Jiangsu University, Zhenjiang 212013, China

*Author to whom correspondence should be addressed.Sustainability202315(6), 5159; https://doi.org/10.3390/su15065159

Received: 30 January 2023 / Revised: 4 March 2023 / Accepted: 10 March 2023 / Published: 14 March 2023(This article belongs to the Special Issue Advanced Technologies of Renewable Energy and Water Management for Sustainable Environment

Abstract

Due to their high efficiency, low heat loss and associated sustainability advantages, impinging jets have been used extensively in marine engineering, geotechnical engineering and other engineering practices. In this paper, the flow structure and impact characteristics of impinging jets with different Reynolds numbers and impact distances are systematically studied by Flow-3D based on PIV experiments. In the study, the relevant state parameters of the jets are dimensionlessly treated, obtaining not only the linear relationship between the length of the potential nucleation zone and the impinging distance, but also the linear relationship between the axial velocity and the axial distance in the impinging zone. In addition, after the jet impinges on the flat plate, the vortex action range caused by the wall-attached flow of the jet gradually decreases inward with the increase of the impinging distance. By examining the effect of Reynolds number Re on the hydraulic characteristics of the submerged impact jet, it can be found that the structure of the continuous submerged impact jet is relatively independent of the Reynolds number. At the same time, the final simulation results demonstrate the applicability of the linear relationship between the length of the potential core region and the impact distance. This study provides methodological guidance and theoretical support for relevant engineering practice and subsequent research on impinging jets, which has strong theoretical and practical significance.

Keywords: 

PIVFlow-3Dimpinging jethydraulic characteristicsimpinging distance

Sustainability 15 05159 g001 550

Figure 1. Geometric model.

Sustainability 15 05159 g002 550

Figure 2. Model grid schematic.

Sustainability 15 05159 g003 550

Figure 3. (a) Schematic diagram of the experimental setup; (b) PIV images of vertical impinging jets with velocity fields.

Sustainability 15 05159 g004 550

Figure 4. (a) Velocity distribution verification at the outlet of the jet pipe; (b) Distribution of flow angle in the mid-axis of the jet [39].

Sustainability 15 05159 g005 550

Figure 5. Along-range distribution of the dimensionless axial velocity of the jet at different impact distances.Figure 6 shows the variation of H

Sustainability 15 05159 g006 550

Figure 6. Relationship between the distribution of potential core region and the impact height H/D.

Sustainability 15 05159 g007 550

Figure 7. The relationship between the potential core length 

Sustainability 15 05159 g008 550

Figure 8. Along-range distribution of the flow angle φ of the jet at different impact distances.

Sustainability 15 05159 g009 550

Figure 9. Velocity distribution along the axis of the jet at different impinging regions.

Sustainability 15 05159 g010 550

Figure 10. The absolute value distribution of slope under different impact distances.

Sustainability 15 05159 g011a 550
Sustainability 15 05159 g011b 550

Figure 11. Velocity distribution of impinging jet on wall under different impinging distances.

Sustainability 15 05159 g012 550

Figure 12. Along-range distribution of the dimensionless axial velocity of the jet at different Reynolds numbers.

Sustainability 15 05159 g013 550

Figure 13. Along-range distribution of the flow angle φ of the jet at different Reynolds numbers.

Sustainability 15 05159 g014 550

Figure 14. Velocity distribution along the jet axis at different Reynolds numbers.

Sustainability 15 05159 g015 550

Figure 15. Velocity distribution of impinging jet on a wall under different Reynolds numbers.

References

  1. Zhang, J.; Li, Y.; Zhang, Y.; Yang, F.; Liang, C.; Tan, S. Using a high-pressure water jet-assisted tunnel boring machine to break rock. Adv. Mech. Eng. 202012, 1687814020962290. [Google Scholar] [CrossRef]
  2. Shi, X.; Zhang, G.; Xu, G.; Ma, Y.; Wu, X. Inactivating Microorganism on Medical Instrument Using Plasma Jet. High Volt. Eng. 200935, 632–635. [Google Scholar]
  3. Gao, Y.; Han, P.; Wang, F.; Cao, J.; Zhang, S. Study on the Characteristics of Water Jet Breaking Coal Rock in a Drilling Hole. Sustainability 202214, 8258. [Google Scholar] [CrossRef]
  4. Xu, W.; Wang, C.; Zhang, L.; Ge, J.; Zhang, D.; Gao, Z. Numerical study of continuous jet impinging on a rotating wall based on Wray—Agarwal turbulence model. J. Braz. Soc. Mech. Sci. Eng. 202244, 433. [Google Scholar] [CrossRef]
  5. Hu, B.; Wang, C.; Wang, H.; Yu, Q.; Liu, J.; Zhu, Y.; Ge, J.; Chen, X.; Yang, Y. Numerical Simulation Study of the Horizontal Submerged Jet Based on the Wray—Agarwal Turbulence Model. J. Mar. Sci. Eng. 202210, 1217. [Google Scholar] [CrossRef]
  6. Dahiya, A.K.; Bhuyan, B.K.; Kumar, S. Perspective study of abrasive water jet machining of composites—A review. J. Mech. Sci. Technol. 202236, 213–224. [Google Scholar] [CrossRef]
  7. Abushanab, W.S.; Moustafa, E.B.; Harish, M.; Shanmugan, S.; Elsheikh, A.H. Experimental investigation on surface characteristics of Ti6Al4V alloy during abrasive water jet machining process. Alex. Eng. J. 202261, 7529–7539. [Google Scholar] [CrossRef]
  8. Hu, B.; Wang, H.; Liu, J.; Zhu, Y.; Wang, C.; Ge, J.; Zhang, Y. A numerical study of a submerged water jet impinging on a stationary wall. J. Mar. Sci. Eng. 202210, 228. [Google Scholar] [CrossRef]
  9. Peng, J.; Shen, H.; Xie, W.; Zhai, S.; Xi, G. Influence of flow fluctuation characteristics on flow and heat transfer in different regions. J. Drain. Irrig. Mach. Eng. 202240, 826–833. [Google Scholar]
  10. Zhai, S.; Xie, F.; Yin, G.; Xi, G. Effect of gap ratio on vortex-induced vibration characteristics of different blunt bodies near-wall. J. Drain. Irrig. Mach. Eng. 202139, 1132–1138. [Google Scholar]
  11. Lin, W.; Zhou, Y.; Wang, L.; Tao, L. PIV experiment and numerical simulation of trailing vortex structure of improved INTER-MIG impeller. J. Drain. Irrig. Mach. Eng. 202139, 158–164. [Google Scholar]
  12. Han, B.; Yao, Z.; Tang, R.; Xu, H. On the supersonic impinging jet by laser Doppler velocimetry. Exp. Meas. Fluid Mech. 200216, 99–103. [Google Scholar]
  13. Darisse, A.; Lemay, J.; Benaissa, A. LDV measurements of well converged third order moments in the far field of a free turbulent round jet. Exp. Therm. Fluid Sci. 201344, 825–833. [Google Scholar] [CrossRef]
  14. Kumar, S.; Kumar, A. Effect of initial conditions on mean flow characteristics of a three dimensional turbulent wall jet. Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 2021235, 6177–6190. [Google Scholar] [CrossRef]
  15. Tao, D.; Zhang, R.; Ying, C. Development and application of the pollutant diffusion testing apparatus based on the image analysis. J. Saf. Environ. 201616, 247–251. [Google Scholar]
  16. Seo, H.; Kim, K.C. Experimental study on flow and turbulence characteristics of bubbly jet with low void fraction. Int. J. Multiph. Flow 2021142, 103738. [Google Scholar] [CrossRef]
  17. Wen, Q.; Sha, J.; Liu, Y. TR-PIV measurement of the turbulent submerged jet and POB analysis of the dynamic structure. J. Exp. Fluid Mech. 20144, 16–24. [Google Scholar]
  18. Yang, Y.; Zhou, L.; Shi, W.; He, Z.; Han, Y.; Xiao, Y. Interstage difference of pressure pulsation in a three-stage electrical submersible pump. J. Petrol. Sci. Eng. 2021196, 107653. [Google Scholar] [CrossRef]
  19. Tang, S.; Zhu, Y.; Yuan, S. An improved convolutional neural network with an adaptable learning rate towards multi-signal fault diagnosis of hydraulic piston pump. Adv. Eng. Inform. 202150, 101406. [Google Scholar] [CrossRef]
  20. Han, Y.; Song, X.; Li, K.; Yan, X. Hybrid modeling for submergence depth of the pumping well using stochastic configuration networks with random sampling. J. Petrol. Sci. Eng. 2022208, 109423. [Google Scholar] [CrossRef]
  21. Tang, S.; Zhu, Y.; Yuan, S. A novel adaptive convolutional neural network for fault diagnosis of hydraulic piston pump with acoustic images. Adv. Eng. Inform. 202252, 101554. [Google Scholar] [CrossRef]
  22. Long, J.; Song, X.; Shi, J.; Chen, J. Optimization and CFD Analysis on Nozzle Exit Position of Two-phase Ejector. J. Refrig. 202243, 39–45. [Google Scholar]
  23. Ni, Q.; Ruan, W. Optimization design of desilting jet pump parameters based on response surface model. J. Ship Mech. 202226, 365–374. [Google Scholar]
  24. Zhang, K.; Zhu, X.; Ren, X.; Qiu, Q.; Shen, S. Numerical investigation on the effect of nozzle position for design of high performance ejector. Appl. Therm. Eng. 2017126, 594–601. [Google Scholar] [CrossRef]
  25. Fu, W.; Liu, Z.; Li, Y.; Wu, H.; Tang, Y. Numerical study for the influences of primary steam nozzle distance and mixing chamber throat diameter on steam ejector performance. Int. J. Therm. Sci. 2018132, 509–516. [Google Scholar] [CrossRef]
  26. Lucas, C.; Rusche, H.; Schroeder, A.; Koehler, J. Numerical investigation of a two-phase CO2 ejector. Int. J. Refrigeration 201443, 154–166. [Google Scholar] [CrossRef]
  27. Ma, X.; Zhu, T.; Fu, Y.; Yan, Y.; Chen, W. Numerical simulation of rock breaking by abrasive water jet. J. Coast. Res. 201993, 274–283. [Google Scholar] [CrossRef]
  28. He, L.; Liu, Y.; Shen, K.; Yang, X.; Ba, Q.; Xiong, W. Numerical research on the dynamic rock-breaking process of impact drilling with multi-nozzle water jets. J. Pet. Sci. Eng. 2021207, 109145. [Google Scholar] [CrossRef]
  29. Yu, Z.; Wang, Z.; Lei, C.; Zhou, Y.; Qiu, X. Numerical Simulation on Internal Flow Field of a Self-excited Oscillation Pulsed Jet Nozzle with Back-flow. Mech. Sci. Technol. Aerosp. Eng. 202241, 998–1002. [Google Scholar]
  30. Huang, J.; Ni, F.; Gu, L. Numerical method of FLOW-3D for sediment erosion simulation. China Harb. Eng. 201939, 6–11. [Google Scholar]
  31. Al Shaikhli, H.I.; Khassaf, S.I. Using of flow 3d as CFD materials approach in waves generation. Mater. Today Proc. 202249, 2907–2911. [Google Scholar] [CrossRef]
  32. Kosaj, R.; Alboresha, R.S.; Sulaiman, S.O. Comparison Between Numerical Flow3d Software and Laboratory Data, For Sediment Incipient Motion. IOP Conf. Ser. Earth Environ. Sci. 2022961, 012031. [Google Scholar] [CrossRef]
  33. Du, C.; Liu, X.; Zhang, J.; Wang, B.; Chen, X.; Yu, X. Long-distance water hammer protection of pipeline after pump being first lowered and then rasied. J. Drain. Irrig. Mach. Eng. 202240, 1248–1253, 1267. [Google Scholar]
  34. Gao, F.; Li, X.; Gao, Q. Experiment and numerical simulation on hydraulic characteristics of novel trapezoidal measuring weir. J. Drain. Irrig. Mach. Eng. 202240, 1104–1111. [Google Scholar]
  35. Tu, A.; Nie, X.; Li, Y.; Li, H. Experimental and simulation study on water infiltration characteristics of layered red soil. J. Drain. Irrig. Mach. Eng. 202139, 1243–1249. [Google Scholar]
  36. Chen, J.; Zeng, B.; Liu, L.; Tao, K.; Zhao, H.; Zhang, C.; Zhang, J.; Li, D. Investigating the anchorage performance of full-grouted anchor bolts with a modified numerical simulation method. Eng. Fail. Anal. 2022141, 106640. [Google Scholar] [CrossRef]
  37. Hu, B.; Yao, Y.; Wang, M.; Wang, C.; Liu, Y. Flow and Performance of the Disk Cavity of a Marine Gas Turbine at Varying Nozzle Pressure and Low Rotation Speeds: A Numerical Investigation. Machines 202311, 68. [Google Scholar] [CrossRef]
  38. Yao, J.; Wang, X.; Zhang, S.; Xu, S.; Jin, B.; Ding, S. Orthogonal test of important parameters affecting hydraulic performance of negative pressure feedback jet sprinkler. J. Drain. Irrig. Mach. Eng. 202139, 966–972. [Google Scholar]
  39. Wang, C.; Wang, X.; Shi, W.; Lu, W.; Tan, S.K.; Zhou, L. Experimental investigation on impingement of a submerged circular water jet at varying impinging angles and Reynolds numbers. Exp. Therm. Fluid Sci. 201789, 189–198. [Google Scholar] [CrossRef]
  40. Speziale, C.G.; Thangam, S. Analysis of an RNG based turbulence model for separated flows. Int. J. Eng. Sci. 199230, 1379–1388. [Google Scholar] [CrossRef]
  41. El Hassan, M.; Assoum, H.H.; Sobolik, V.; Vétel, J.; Abed-Meraim, K.; Garon, A.; Sakout, A. Experimental investigation of the wall shear stress and the vortex dynamics in a circular impinging jet. Exp. Fluids 201252, 1475–1489. [Google Scholar] [CrossRef]
  42. Fairweather, M.; Hargrave, G. Experimental investigation of an axisymmetric, impinging turbulent jet. 1. Velocity field. Exp. Fluids 200233, 464–471. [Google Scholar] [CrossRef]
  43. Ashforth-Frost, S.; Jambunathan, K. Effect of nozzle geometry and semi-confinement on the potential core of a turbulent axisymmetric free jet. Int. Commun. Heat Mass Transf. 199623, 155–162. [Google Scholar] [CrossRef]
  44. Chen, M.; Huang, H.; Wang, D.; Lv, S.; Chen, Y. PIV tests for flow characteristics of impinging jet in a semi-closed circular pipe. J. Vib. Shock 202140, 90–97, 113. [Google Scholar]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

      

MDPI and ACS Style

Mi, H.; Wang, C.; Jia, X.; Hu, B.; Wang, H.; Wang, H.; Zhu, Y. Hydraulic Characteristics of Continuous Submerged Jet Impinging on a Wall by Using Numerical Simulation and PIV Experiment. Sustainability 202315, 5159. https://doi.org/10.3390/su15065159

AMA Style

Mi H, Wang C, Jia X, Hu B, Wang H, Wang H, Zhu Y. Hydraulic Characteristics of Continuous Submerged Jet Impinging on a Wall by Using Numerical Simulation and PIV Experiment. Sustainability. 2023; 15(6):5159. https://doi.org/10.3390/su15065159Chicago/Turabian Style

Mi, Hongbo, Chuan Wang, Xuanwen Jia, Bo Hu, Hongliang Wang, Hui Wang, and Yong Zhu. 2023. “Hydraulic Characteristics of Continuous Submerged Jet Impinging on a Wall by Using Numerical Simulation and PIV Experiment” Sustainability 15, no. 6: 5159. https://doi.org/10.3390/su15065159

Figure 2. Different PKW Types.

A review of Piano Key Weir as a superior alternative for dam rehabilitation

댐 복구를 위한 우수한 대안으로서의 Piano Key Weir에 대한 검토

Amiya Abhash &

K. K. Pandey

Pages 541-551 | Received 03 Mar 2020, Accepted 07 May 2020, Published online: 21 May 2020

ABSTRACT

Dams fall in ‘installations containing dangerous forces’ because of their massive impact on the environment and civilian life and property as per International humanitarian law. As such, it becomes vital for hydraulic engineers to refurbish various solutions for dam rehabilitation. This paper presents a review of a new type of weir installation called Piano Key Weir (PKW), which is becoming popular around the world for its higher spillway capacity both for existing and new dam spillway installations. This paper reviews the geometry along with structural integrity, discharging capacity, economic aspects, aeration requirements, sediment transport and erosion aspects of Piano Key Weir (PKW) as compared with other traditional spillway structures and alternatives from literature. The comparison with other alternatives shows PKW to be an excellent alternative for dam risk mitigation owing to its high spillway capabilities and economy, along with its use in both existing and new hydraulic structures.

댐은 국제 인도법에 따라 환경과 민간인 생활 및 재산에 막대한 영향을 미치기 때문에 ‘위험한 힘을 포함하는 시설물’에 속합니다. 따라서 유압 엔지니어는 댐 복구를 위한 다양한 솔루션을 재정비해야 합니다.

이 백서에서는 PKW(Piano Key Weir)라는 새로운 유형의 둑 설치에 대한 검토를 제공합니다. PKW는 기존 및 신규 댐 방수로 설치 모두에서 더 높은 방수로 용량으로 전 세계적으로 인기를 얻고 있습니다.

이 백서에서는 구조적 무결성, 배출 용량, 경제적 측면, 폭기 요구 사항, 퇴적물 운반 및 PKW(Piano Key Weir)의 침식 측면과 함께 다른 전통적인 여수로 구조 및 문헌의 대안과 비교하여 기하학을 검토합니다.

다른 대안과의 비교는 PKW가 높은 여수로 기능과 경제성으로 인해 댐 위험 완화를 위한 탁월한 대안이며 기존 및 새로운 수력 구조물 모두에 사용됨을 보여줍니다.

KEYWORDS: 

Figure 2. Different PKW Types.
Figure 2. Different PKW Types.

References

  • Anderson, R., and Tullis, B. (2011). Influence of Piano Key Weir geometry on discharge. Proc. Int. Conf. Labyrinth and Piano Key Weirs Liège B. Liège, Belgium. [Crossref][Google Scholar]
  • Anderson, R., and Tullis, B. (2012a). “Piano key weir hydraulics and labyrinth weir comparison”. J. Irrig. Drain. Eng., 139(3), 246–253. doi:https://doi.org/10.1061/(ASCE)IR.1943-4774.0000530 [Crossref][Web of Science ®][Google Scholar]
  • Anderson, R., and Tullis, B. (2012b). “Piano key weir: Reservoir versus channel application”. J. Irrig. Drain. Eng., 138(8), 773–776. doi:https://doi.org/10.1061/(ASCE)IR.1943-4774.0000464 [Crossref][Web of Science ®][Google Scholar]
  • Anderson, R.M. 2011. Piano key weir head discharge relationships, M.S. Thesis, Utah State University, Logan, Utah. [Google Scholar]
  • Bashiri, H., Dewals, B., Pirotton, M., Archambeau, P., and Erpicum, S. (2016). “Towards a new design equation for piano key weirs discharge capacity.” Proc. of the 6th International Symposium on Hydraulic Structures. Portland, USA. [Google Scholar]
  • Bianucci, S.P., Sordo Ward, Á.F., Pérez Díaz, J.I., García-Palacios, J.H., Mediero Orduña, L.J., and Garrote de Marcos, L. (2013). “Risk-based methodology for parameter calibration of a reservoir flood control model”. Natl. Hazard Earth Syst. Sci., 13(4), 965–981. doi:https://doi.org/10.5194/nhess-13-965-2013 [Crossref][Web of Science ®][Google Scholar]
  • Blancher, B., Montarros, F., and Laugier, F. (2011). Hydraulic comparison between Piano Key Weirs and labyrinth spillways. Proc. Int. Conf. Labyrinth and Piano Key Weirs Liège B. Liège, Belgium. [Crossref][Google Scholar]
  • Botha, A., Fitz, I., Moore, A., Mulder, F., and Van Deventer, N. 2013. “Application of the Piano Key Weir spillway in the Republic of South Africa”. Labyrinth and Piano Key Weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs, Chatou, Paris, France, 20–22, 185. [Crossref][Google Scholar]
  • Chahartaghi, M.K., Nazari, S., and Shooshtari, M.M. 2019. “Experimental and numerical simulation of arced trapezoidal Piano Key Weirs”. Flow Meas. Instrum., 68, 101576. doi:https://doi.org/10.1016/j.flowmeasinst.2019.101576 [Crossref][Web of Science ®][Google Scholar]
  • Chi Hien, T., Thanh Son, H., and Ho Ta Khanh, M. (2006). Results of some ‘piano keys’ weir hydraulic model tests in Vietnam. Proc., 22nd Int. Congress of Large Dams, Question 87, Response 39, International Commission on Large Dams (ICOLD). Barcelona, Spain. [Google Scholar]
  • Cicero, G., Barcouda, M., Luck, M., and Vettori, E. (2011). Study of a piano key morning glory to increase the spillway capacity of the Bage dam. Proc. Int. Conf. Labyrinth Piano Key Weirs-PKW2011, Taylor & Francis, London. [Crossref][Google Scholar]
  • Cicero, G., De Miranda, D., and Luck, M. (2012). “Assessment of the code Wolf 1D PKW for predicting the hydraulic behaviour of PK-Weirs.” Congrès SHF-33èmes journées de l’hydraulique “Grands aménagements hydrauliques 2012”, Paris, France. [Google Scholar]
  • Cicero, G., and Delisle, J. (2013). “Discharge characteristics of Piano Key weirs under submerged flow”. Labyrinth and Piano Key Weirs II–PKW 2013, 101–109. [Crossref][Google Scholar]
  • Cicero, G., Delisle, J., Lefebvre, V., and Vermeulen, J. (2013). “Experimental and numerical study of the hydraulic performance of a trapezoidal Piano Key weir.” Labyrinth and Piano Key Weirs II: Proceedings of the Second International Workshop on Labyrinth and Piano key weirs, Chatou, Paris, France, 20–22, 265. [Crossref][Google Scholar]
  • Cicéro, G., Guene, C., Luck, M., Pinchard, T., Lochu, A., and Brousse, P. (2010). “Experimental optimization of a Piano Key Weir to increase the spillway capacity of the Malarce dam.” 1st IAHR European Congress, Edinbourgh, Mai 4–6, 2010. [Google Scholar]
  • Crookston, B., Anderson, R., and Tullis, B. (2018). “Free-flow discharge estimation method for Piano Key weir geometries.” J. Hydro. Environ. Res., 19, 160–167. doi:https://doi.org/10.1016/j.jher.2017.10.003 [Crossref][Web of Science ®][Google Scholar]
  • Das Singhal, G., and Sharma, N. 2011. “Rehabilitation of Sawara Kuddu Hydroelectric Project–Model studies of Piano Key Weir in India”. Proc. Int. Workshop on Labyrinths and Piano Key Weirs PKW 2011. Taylor & Francis, London. [Crossref][Google Scholar]
  • Denys, F., Basson, G., and Strasheim, J. (2017). Fluid Structure Interaction of Piano Key Weirs. Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), Feb 22–24, 2017, CRC Press, Qui Nhon, Vietnam. [Crossref][Google Scholar]
  • Eichenberger, P. (2013). “The first commercial piano key weir in Switzerland.” Labyrinth and Piano Key Weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, 227. [Crossref][Google Scholar]
  • Erpicum, S., Laugier, F., Pfister, M., Pirotton, M., Cicero, G.-M., and Schleiss, A.J. 2013. Labyrinth and Piano Key Weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, CRC Press. [Crossref][Google Scholar]
  • Erpicum, S., Machiels, O., Dewals, B., Pirotton, M., and Archambeau, P. (2012). “Numerical and physical hydraulic modelling of Piano Key Weirs.” Proceedings of the 4th Int. Conf. on Water Resources and Renewable Energy Development in Asia. Chiang Mai, Thailande. [Google Scholar]
  • Erpicum, S., Nagel, V., and Laugier, F. (2011). “Piano Key Weir design study at Raviege dam”. Labyrinth and Piano Key Weirs–PKW 2011, 43–50. [Crossref][Google Scholar]
  • Ervine, D., and Elsawy, E. (1975). “The effect of a falling nappe on river aeration.” Proc. 16th IAHR Congress, Sao Paulo, Brazil. [Google Scholar]
  • Falvey, H.T. 1980. “Air-water flow in hydraulic structures”. NASA STI/Recon Technical Report N, 81. [Google Scholar]
  • Gabriel-Martin, I., Sordo-Ward, A., Garrote, L., and Castillo, L.G. (2017). “Influence of initial reservoir level and gate failure in dam safety analysis. Stochastic approach.” J. Hydrol., 550, 669–684. doi:https://doi.org/10.1016/j.jhydrol.2017.05.032 [Crossref][Web of Science ®][Google Scholar]
  • Gebhardt, M., Herbst, J., Merkel, J., and Belzner, F. (2019). “Sedimentation at labyrinth weirs–an experimental study of the self-cleaning process”. J. Hydraulic Res., 57(4), 579–590. doi:https://doi.org/10.1080/00221686.2018.1494053 [Taylor & Francis Online][Web of Science ®][Google Scholar]
  • Hu, H., Qian, Z., Yang, W., Hou, D., and Du, L. (2018). “Numerical study of characteristics and discharge capacity of piano key weirs.” Flow Meas. Instrum., 62, 27–32. doi:https://doi.org/10.1016/j.flowmeasinst.2018.05.004 [Crossref][Web of Science ®][Google Scholar]
  • Javaheri, A., and Kabiri-Samani, A. (2012). “Threshold submergence of flow over PK weirs”. Int. J. Civil Geol. Eng., 6, 46–49. [Google Scholar]
  • Jayatillake, H., and Perera, K. (2013). “Design of a Piano-Key Weir for Giritale Dam spillway in Sri Lanka.” Labyrinth and Piano Key Weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, 151. [Crossref][Google Scholar]
  • Jayatillake, H., and Perera, K. (2017). “Adoption of a type D Piano Key Weir spillway with tapered noses at Rambawa Tank, Sri Lanka.” Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), Feb 22–24, 2017, CRC Press, Qui Nhon, Vietnam. [Crossref][Google Scholar]
  • Jüstrich, S., Pfister, M., and Schleiss, A.J. (2016). “Mobile riverbed scour downstream of a Piano Key weir”. J. Hydraulic Eng., 142(11), 04016043. doi:https://doi.org/10.1061/(ASCE)HY.1943-7900.0001189 [Crossref][Google Scholar]
  • Kabiri-Samani, A., and Javaheri, A. (2012). “Discharge coefficients for free and submerged flow over Piano Key weirs”. J. Hydraulic Res., 50(1), 114–120. doi:https://doi.org/10.1080/00221686.2011.647888 [Taylor & Francis Online][Web of Science ®][Google Scholar]
  • Karimi, M., Attari, J., Saneie, M., and Jalili Ghazizadeh, M.R. (2018). “Side weir flow characteristics: comparison of piano key, labyrinth, and linear types”. J. Hydraulic Eng., 144(12), 04018075. doi:https://doi.org/10.1061/(ASCE)HY.1943-7900.0001539 [Crossref][Google Scholar]
  • Karimi, M., Attari, J., Saneie, M., and Jalili-Ghazizadeh, M. (2017). “Experimental study of discharge coefficient of a piano key side weir.” Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017). Proceedings of the Third International Workshop on Labyrinth and Piano key weirs 2017, Qui Nhon, Vietnam, 22–24. [Crossref][Google Scholar]
  • Khanh, M.H.T. (2013). “The Piano Key Weirs: 15 years of Research & Development–Prospect.” Labyrinth and piano key weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, 3. [Crossref][Google Scholar]
  • Khanh, M.H.T. (2017). “History and development of Piano Key Weirs in Vietnam from 2004 to 2016.” Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), Feb 22–24, 2017, CRC Press, Qui Nhon, Vietnam. [Google Scholar]
  • Khanh, M.H.T., Hien, T.C., and Hai, N.T. (2011). “Main results of the PK weir model tests in Vietnam (2004 to 2010).” Labyrinth and Piano Key Weirs, 191. Liège, Belgium. [Crossref][Google Scholar]
  • Khassaf, S.I., Aziz, L.J., and Elkatib, Z.A. (2016). “Hydraulic behavior of piano key weir type B under free flow conditions”. Int. J. Sci. Technol. Res., 5(3), 158–163. [Google Scholar]
  • Khassaf, S.I., and Al-Baghdadi, M.B. (2015). “Experimental study of non-rectangular piano key weir discharge coefficient”. J. Homepage, 6(5), 425–436. [Google Scholar]
  • Khassaf, S.I., and Al-Baghdadi, M.B.N. (2018). “Experimental investigation of submerged flow over piano key weir”. Int. J. Energy Environ., 9(3), 249–260. [Google Scholar]
  • Kwon, -H.-H., and Moon, Y.-I. (2006). “Improvement of overtopping risk evaluations using probabilistic concepts for existing dams”. Stochastic Environ. Res. Risk Assess., 20(4), 223. doi:https://doi.org/10.1007/s00477-005-0017-2 [Crossref][Web of Science ®][Google Scholar]
  • Laugier, F. (2007). “Design and construction of the first Piano Key Weir spillway at Goulours dam”. Int. J. Hydropower Dams, 14(5), 94. [Google Scholar]
  • Laugier, F., Lochu, A., Gille, C., Leite Ribeiro, M., and Boillat, J.-L. (2009). “Design and construction of a labyrinth PKW spillway at Saint-Marc dam, France”. Hydropower Dams, 16(LCH–ARTICLE–2009–023), 100–107. [Google Scholar]
  • Laugier, F., Pralong, J., and Blancher, B. (2011). “Influence of structural thickness of sidewalls on PKW spillway discharge capacity.” Proc. Intl Workshop on Labyrinths and Piano Key Weirs PKW 2011. Liège, Belgium. [Crossref][Google Scholar]
  • Le Blanc, M., Spinazzola, U., and Kocahan, H. (2011). “Labyrinth fusegate applications on free overflow spillways–Overview of recent projects.” Labyrinth and Piano Key Weirs, 261, Liège, Belgium. [Crossref][Google Scholar]
  • Leite Ribeiro, M., Bieri, M., Boillat, J.-L., Schleiss, A., Delorme, F., and Laugier, F. (2009). “Hydraulic capacity improvement of existing spillways–design of a piano key weirs.” Proc. (on CD) of the 23rd Congress of the Int. Commission on Large Dams CIGB-ICOLD. Brasilia, Brazil. [Google Scholar]
  • Leite Ribeiro, M., Bieri, M., Boillat, J.-L., Schleiss, A., Singhal, G., and Sharma, N. (2011). “Discharge capacity of piano key weirs”. J. Hydraulic Eng., 138(2), 199–203. doi:https://doi.org/10.1061/(ASCE)HY.1943-7900.0000490 [Crossref][Google Scholar]
  • Lempérière, F., and Ouamane, A. (2003). “The Piano Keys weir: a new cost-effective solution for spillways”. Int. J. Hydropower Dams, 10(5), 144–149. [Google Scholar]
  • Lempérière, F., and Vigny, J. (2011). “General comments on labyrinth and Piano Keys Weirs–The future”. Labyrinth and Piano Key weirs–PKW 2011, 289–294. [Crossref][Google Scholar]
  • Lempérière, F., Vigny, J., and Ouamane, A. (2011). General comments on Labyrinth and Piano Key Weirs: The past and present. Proc. Intl. Conf. Labyrinth and Piano Key Weirs, Liège B. Liège, Belgium. [Crossref][Google Scholar]
  • Lewin, J., Ballard, G., and Bowles, D.S. (2003). “Spillway gate reliability in the context of overall dam failure risk.” USSD Annual Lecture, Charleston, South Carolina. [Google Scholar]
  • Lodomez, M., Pirotton, M., Dewals, B., Archambeau, P., and Erpicum, S. (2017). “Could piano key weirs be subject to nappe oscillations?” Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), Feb 22–24, 2017, CRC Press, Qui Nhon, Vietnam [Crossref][Google Scholar]
  • Machiels, O., Erpicum, S., Archambeau, P., Dewals, B., and Pirotton, M. (2009). “Large scale experimental study of piano key weirs.” Proc. 33rd IAHR Congress: Water Engineering for a Sustainable Environment, IAHR. Vancouver, Canada [Google Scholar]
  • Machiels, O., Erpicum, S., Archambeau, P., Dewals, B., and Pirotton, M. (2011a). “Piano Key Weir preliminary design method–Application to a new dam project.” Proc. Int. Conf. Labyrinth and Piano Key Weirs Liège B. Liège, Belgium. [Crossref][Google Scholar]
  • Machiels, O., Erpicum, S., Dewals, B., Archambeau, P., and Pirotton, M. (2010). “Piano Key Weirs: The experimental study of an efficient solution for rehabilitation”. WIT Trans. Ecol., 133, 95–106. [Crossref][Google Scholar]
  • Machiels, O., Erpicum, S., Dewals, B.J., Archambeau, P., and Pirotton, M. (2011b). “Experimental observation of flow characteristics over a Piano Key Weir”. J Hydraulic Res, 49(3), 359–366. doi:https://doi.org/10.1080/00221686.2011.567761 [Taylor & Francis Online][Web of Science ®][Google Scholar]
  • Machiels, O., Pirotton, M., Pierre, A., Dewals, B., and Erpicum, S. (2014). “Experimental parametric study and design of Piano Key Weirs”. J. Hydraulic Res., 52(3), 326–335. doi:https://doi.org/10.1080/00221686.2013.875070 [Taylor & Francis Online][Web of Science ®][Google Scholar]
  • Mehboudi, A., Attari, J., and Hosseini, S. (2016). “Experimental study of discharge coefficient for trapezoidal piano key weirs.” Flow Meas. Instrum., 50, 65–72. doi:https://doi.org/10.1016/j.flowmeasinst.2016.06.005 [Crossref][Web of Science ®][Google Scholar]
  • Micovic, Z., Hartford, D.N., Schaefer, M.G., and Barker, B.L. (2016). “A non-traditional approach to the analysis of flood hazard for dams”. Stochastic Environ. Res. Risk Assess., 30(2), 559–581. doi:https://doi.org/10.1007/s00477-015-1052-2 [Crossref][Web of Science ®][Google Scholar]
  • Monjezi, R., Heidarnejad, M., Masjedi, A., Purmohammadi, M.H., and Kamanbedast, A. (2018). “Laboratory investigation of the discharge coefficient of flow in arced labyrinth weirs with triangular plans.” Flow Meas. Instrum., 64, 64–70. doi:https://doi.org/10.1016/j.flowmeasinst.2018.10.011 [Crossref][Web of Science ®][Google Scholar]
  • Noseda, M., Stojnic, I., Pfister, M., and Schleiss, A.J. (2019). “Upstream Erosion and sediment passage at piano key weirs”. J. Hydraulic Eng., 145(8), 04019029. doi:https://doi.org/10.1061/(ASCE)HY.1943-7900.0001616 [Crossref][Google Scholar]
  • Oertel, M. (2015). “Discharge coefficients of piano key weirs from experimental and numerical modelS.” E= proceedings of the 36th IAHR world congress. 28 June – 3 July, The Hague, The Netherlands. [Google Scholar]
  • Ouamane, A. (2011). Nine years of study of the Piano Key Weir in the university laboratory of Biskra “lessons and reflections”. Proc. Int. Conf. Labyrinth Piano Key Weirs-PKW2011, Taylor & Francis, London. [Crossref][Google Scholar]
  • Ouamane, A., Debabeche, M., Lempérière, F., and Vigny, J. (2017). Twenty years of research in Biskra University for Labyrinths and Piano Key Weirs and associated fuse plugs. Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), Feb 22–24, 2017, CRC Press, Qui Nhon, Vietnam. [Crossref][Google Scholar]
  • Ouamane, A., and Lempérière, F. (2006). Design of a new economic shape of weir. Proc. Int. Symp. on Dams in the Societies of the 21st Century. Barcelona, Spain. [Crossref][Google Scholar]
  • Patev, R., and Putcha, C. (2005). “Development of fault trees for risk assessment of dam gates and associated operating equipment”. Int. J. Modell. Simul., 25(3), 190–201. doi:https://doi.org/10.1080/02286203.2005.11442336 [Taylor & Francis Online][Google Scholar]
  • Paxson, G., Tullis, B., and Hertel, D. 2013. “Comparison of Piano Key Weirs with labyrinth and gated spillways: Hydraulics, cost, constructability and operations”. Labyrinth and Piano Key Weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, 123–130. [Crossref][Google Scholar]
  • Pfister, M., Capobianco, D., Tullis, B., and Schleiss, A.J. (2013). “Debris-blocking sensitivity of piano key weirs under reservoir-type approach flow”. J. Hydraulic Eng., 139(11), 1134–1141. doi:https://doi.org/10.1061/(ASCE)HY.1943-7900.0000780 [Crossref][Google Scholar]
  • Phillips, M., and Lesleighter, E. 2013. “Piano Key Weir spillway: Upgrade option for a major dam”. Labyrinth and Piano Key Weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, 159–168. [Crossref][Google Scholar]
  • Pinchard, T., Boutet, J., and Cicero, G. (2011). “Spillway capacity upgrade at Malarce dam: design of an additional Piano Key Weir spillway.” Proc. Int. Workshop on Labyrinths and Piano Key Weirs PKW. Liège, Belgium. [Crossref][Google Scholar]
  • Pralong, J., J. Vermeulen, B. Blancher, F. Laugier, S. Erpicum, O. Machiels, M. Pirotton, J.-L. Boillat, M. Leite Ribeiro and A. Schleiss (2011). “A naming convention for the piano key weirs geometrical parameters.” Labyrinth and piano key weirs, 271–278. [Crossref][Google Scholar]
  • Ribeiro, M.L., Boillat, J.-L., Schleiss, A., Laugier, F., and Albalat, C. (2007). “Rehabilitation of St-Marc dam.” Experimental optimization of a piano key weir. Proc. of 32nd Congress of IAHR, Vince, Italy. [Google Scholar]
  • Ribeiro, M.L., Pfister, M., and Schleiss, A.J. (2013). “Overview of Piano Key weir prototypes and scientific model investigations”. Labyrinth and Piano Key Weirs II, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, 273. [Crossref][Google Scholar]
  • Ribeiro, M.L., Pfister, M., Schleiss, A.J., and Boillat, J.-L. (2012). “Hydraulic design of A-type piano key weirs”. J. Hydraulic Res., 50(4), 400–408. doi:https://doi.org/10.1080/00221686.2012.695041 [Taylor & Francis Online][Web of Science ®][Google Scholar]
  • Ribi, J., Spahni, B., Dorthe, D., and Pfister, M. (2017). Piano Key Weir as overflow on sedimentation basin of wastewater treatment plant. Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), Feb 22–24, 2017, CRC Press, Qui Nhon, Vietnam [Crossref][Google Scholar]
  • Schleiss, A. (2011). “From labyrinth to piano key weirs: a historical review.” Proc. Int. Conf. Labyrinth and Piano Key Weirs Liège B. Liège, Belgium. [Crossref][Google Scholar]
  • Sharma, N., and Tiwari, H. (2013). “Experimental study on vertical velocity and submergence depth near Piano Key Weir.” Labyrinth and Piano Key Weirs II-PKW, Proceedings of the Second International Workshop on Labyrinth and Piano key weirs 2013, Chatou, Paris, France, 20–22, 93–100. [Crossref][Google Scholar]
  • Tiwari, H. (2016). Experimental Study of Turbulence Characteristics Near Piano Key Weir. PhD, Indian Institute of Technology Roorkee. [Google Scholar]
  • Tiwari, H., and Sharma, N. 2017. “Empirical and Mathematical Modeling of Head and Discharge Over Piano Key Weir”. Development of Water Resources in India. Springer, Cham. 341–354. https://doi.org/10.1007/978-3-319-55125-8_29 [Crossref][Google Scholar]
  • Valley, P., and Blancher, B. (2017). Construction and testing of two Piano Key Weirs at Charmines dam. Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), Feb 22–24, 2017, CRC Press, Qui Nhon, Vietnam. [Crossref][Google Scholar]
  • Vermeulen, J., Lassus, C., and Pinchard, T. (2017). Design of a Piano Key Weir aeration network. Labyrinth and Piano Key Weirs III: Proceedings of the 3rd International Workshop on Labyrinth and Piano Key Weirs (PKW 2017), February 22- 24,2017, Qui Nhon, Vietnam, CRC Press. [Crossref][Google Scholar]
  • Vermeulen, J., Laugier, F., Faramond, L., and Gille, C. (2011). “Lessons learnt from design and construction of EDF first Piano Key Weirs”. Labyrinth and Piano Key weirs-PKW 2011, 215–224. [Crossref][Google Scholar]

이종 금속 인터커넥트의 펄스 레이저 용접을 위한 가공 매개변수 최적화

Optimization of processing parameters for pulsed laser welding of dissimilar metal interconnects

본 논문은 독자의 편의를 위해 기계번역된 내용이어서 자세한 내용은 원문을 참고하시기 바랍니다.

NguyenThi TienaYu-LungLoabM.Mohsin RazaaCheng-YenChencChi-PinChiuc

aNational Cheng Kung University, Department of Mechanical Engineering, Tainan, Taiwan

bNational Cheng Kung University, Academy of Innovative Semiconductor and Sustainable Manufacturing, Tainan, Taiwan

cJum-bo Co., Ltd, Xinshi District, Tainan, Taiwan

Abstract

워블 전략이 포함된 펄스 레이저 용접(PLW) 방법을 사용하여 알루미늄 및 구리 이종 랩 조인트의 제조를 위한 최적의 가공 매개변수에 대해 실험 및 수치 조사가 수행됩니다. 피크 레이저 출력과 접선 용접 속도의 대표적인 조합 43개를 선택하기 위해 원형 패킹 설계 알고리즘이 먼저 사용됩니다.

선택한 매개변수는 PLW 프로세스의 전산유체역학(CFD) 모델에 제공되어 용융 풀 형상(즉, 인터페이스 폭 및 침투 깊이) 및 구리 농도를 예측합니다. 시뮬레이션 결과는 설계 공간 내에서 PLW 매개변수의 모든 조합에 대한 용융 풀 형상 및 구리 농도를 예측하기 위해 3개의 대리 모델을 교육하는 데 사용됩니다.

마지막으로, 대체 모델을 사용하여 구성된 처리 맵은 용융 영역에 균열이나 기공이 없고 향상된 기계적 및 전기적 특성이 있는 이종 조인트를 생성하는 PLW 매개변수를 결정하기 위해 세 가지 품질 기준에 따라 필터링됩니다.

제안된 최적화 접근법의 타당성은 최적의 용접 매개변수를 사용하여 생성된 실험 샘플의 전단 강도, 금속간 화합물(IMC) 형성 및 전기 접촉 저항을 평가하여 입증됩니다.

결과는 최적의 매개변수가 1209N의 높은 전단 강도와 86µΩ의 낮은 전기 접촉 저항을 생성함을 확인합니다. 또한 용융 영역에는 균열 및 기공과 같은 결함이 없습니다.

An experimental and numerical investigation is performed into the optimal processing parameters for the fabrication of aluminum and copper dissimilar lap joints using a pulsed laser welding (PLW) method with a wobble strategy. A circle packing design algorithm is first employed to select 43 representative combinations of the peak laser power and tangential welding speed. The selected parameters are then supplied to a computational fluidic dynamics (CFD) model of the PLW process to predict the melt pool geometry (i.e., interface width and penetration depth) and copper concentration. The simulation results are used to train three surrogate models to predict the melt pool geometry and copper concentration for any combination of the PLW parameters within the design space. Finally, the processing maps constructed using the surrogate models are filtered in accordance with three quality criteria to determine the PLW parameters that produce dissimilar joints with no cracks or pores in the fusion zone and enhanced mechanical and electrical properties. The validity of the proposed optimization approach is demonstrated by evaluating the shear strength, intermetallic compound (IMC) formation, and electrical contact resistance of experimental samples produced using the optimal welding parameters. The results confirm that the optimal parameters yield a high shear strength of 1209 N and a low electrical contact resistance of 86 µΩ. Moreover, the fusion zone is free of defects, such as cracks and pores.

Fig. 1. Schematic illustration of Al-Cu lap-joint arrangement
Fig. 1. Schematic illustration of Al-Cu lap-joint arrangement
Fig. 2. Machine setup (MFQS-150W_1500W
Fig. 2. Machine setup (MFQS-150W_1500W
Fig. 5. Lap-shear mechanical tests: (a) experimental setup and specimen dimensions, and (b) two different failures of lap-joint welding.
N. Thi Tien et al.
Fig. 5. Lap-shear mechanical tests: (a) experimental setup and specimen dimensions, and (b) two different failures of lap-joint welding. N. Thi Tien et al.
Fig. 9. Simulation and experimental results for melt pool profile. (a) Simulation results for melt pool cross-section, and (b) OM image of melt pool cross-section.
(Note that laser processing parameter of 830 W and 565 mm/s is chosen.).
Fig. 9. Simulation and experimental results for melt pool profile. (a) Simulation results for melt pool cross-section, and (b) OM image of melt pool cross-section. (Note that laser processing parameter of 830 W and 565 mm/s is chosen.).

References

[1]

G. Santos

Road transport and CO2 emissions: What are the challenges?

Transport Policy, 59 (2017), pp. 71-74

ArticleDownload PDFView Record in ScopusGoogle Scholar[2]

A. Das, D. Li, D. Williams, D. Greenwood

Joining technologies for automotive battery systems manufacturing

World Electric Veh. J., 9 (2) (2018), p. 22 View PDF

CrossRefGoogle Scholar[3]

M. Zwicker, M. Moghadam, W. Zhang, C. Nielsen

Automotive battery pack manufacturing–a review of battery to tab joining

J. Adv. Joining Process., 1 (2020), Article 100017

ArticleDownload PDFView Record in ScopusGoogle Scholar[4]

T. Mai, A. Spowage

Characterisation of dissimilar joints in laser welding of steel–kovar, copper–steel and copper–aluminium

Mater. Sci. Eng. A, 374 (1–2) (2004), pp. 224-233

ArticleDownload PDFView Record in ScopusGoogle Scholar[5]

S.S. Lee, T.H. Kim, S.J. Hu, W.W. Cai, J. Li, J.A. Abell

Characterization of joint quality in ultrasonic welding of battery tabs

International Manufacturing Science and Engineering Conference, vol. 54990, American Society of Mechanical Engineers (2012), pp. 249-261

Google Scholar[6]

Y. Zhou, P. Gorman, W. Tan, K. Ely

Weldability of thin sheet metals during small-scale resistance spot welding using an alternating-current power supply

J. Electron. Mater., 29 (9) (2000), pp. 1090-1099 View PDF

CrossRefView Record in ScopusGoogle Scholar[7]

S. Katayama

Handbook of laser welding technologies

Elsevier (2013)

Google Scholar[8]

A. Sadeghian, N. Iqbal

A review on dissimilar laser welding of steel-copper, steel-aluminum, aluminum-copper, and steel-nickel for electric vehicle battery manufacturing

Opt. Laser Technol., 146 (2022), Article 107595

ArticleDownload PDFView Record in ScopusGoogle Scholar[9]

M.J. Brand, P.A. Schmidt, M.F. Zaeh, A. Jossen

Welding techniques for battery cells and resulting electrical contact resistances

J. Storage Mater., 1 (2015), pp. 7-14

ArticleDownload PDFView Record in ScopusGoogle Scholar[10]

M. Jarwitz, F. Fetzer, R. Weber, T. Graf

Weld seam geometry and electrical resistance of laser-welded, aluminum-copper dissimilar joints produced with spatial beam oscillation

Metals, 8 (7) (2018), p. 510 View PDF

CrossRefView Record in ScopusGoogle Scholar[11]S. Smith, J. Blackburn, M. Gittos, P. de Bono, and P. Hilton, “Welding of dissimilar metallic materials using a scanned laser beam,” in International Congress on Applications of Lasers & Electro-Optics, 2013, vol. 2013, no. 1: Laser Institute of America, pp. 493-502.

Google Scholar[12]

P. Schmitz, J.B. Habedank, M.F. Zaeh

Spike laser welding for the electrical connection of cylindrical lithium-ion batteries

J. Laser Appl., 30 (1) (2018), Article 012004 View PDF

CrossRefView Record in ScopusGoogle Scholar[13]

P. Kah, C. Vimalraj, J. Martikainen, R. Suoranta

Factors influencing Al-Cu weld properties by intermetallic compound formation

Int. J. Mech. Mater. Eng., 10 (1) (2015), pp. 1-13

Google Scholar[14]

Z. Lei, X. Zhang, J. Liu, P. Li

Interfacial microstructure and reaction mechanism with various weld fillers on laser welding-brazing of Al/Cu lap joint

J. Manuf. Process., 67 (2021), pp. 226-240

ArticleDownload PDFView Record in ScopusGoogle Scholar[15]

T. Solchenbach, P. Plapper

Mechanical characteristics of laser braze-welded aluminium–copper connections

Opt. Laser Technol., 54 (2013), pp. 249-256

ArticleDownload PDFView Record in ScopusGoogle Scholar[16]

T. Solchenbach, P. Plapper, W. Cai

Electrical performance of laser braze-welded aluminum–copper interconnects

J. Manuf. Process., 16 (2) (2014), pp. 183-189

ArticleDownload PDFView Record in ScopusGoogle Scholar[17]

S.J. Lee, H. Nakamura, Y. Kawahito, S. Katayama

Effect of welding speed on microstructural and mechanical properties of laser lap weld joints in dissimilar Al and Cu sheets

Sci. Technol. Weld. Join., 19 (2) (2014), pp. 111-118

Google Scholar[18]

Z. Xue, S. Hu, D. Zuo, W. Cai, D. Lee, K.-A. Elijah Jr

Molten pool characterization of laser lap welded copper and aluminum

J. Phys. D Appl. Phys., 46 (49) (2013), Article 495501 View PDF

CrossRefView Record in ScopusGoogle Scholar[19]

S. Zhao, G. Yu, X. He, Y. Zhang, W. Ning

Numerical simulation and experimental investigation of laser overlap welding of Ti6Al4V and 42CrMo

J. Mater. Process. Technol., 211 (3) (2011), pp. 530-537

ArticleDownload PDFView Record in ScopusGoogle Scholar[20]

W. Huang, H. Wang, T. Rinker, W. Tan

Investigation of metal mixing in laser keyhole welding of dissimilar metals

Mater. Des., 195 (2020), Article 109056

ArticleDownload PDFView Record in ScopusGoogle Scholar[21]

E. Kaiser, G. Ambrosy, E. Papastathopoulos

Welding strategies for joining copper and aluminum by fast oscillating, high quality laser beam

High-Power Laser Materials Processing: Applications, Diagnostics, and Systems IX, vol. 11273, International Society for Optics and Photonics (2020), p. 112730C

View Record in ScopusGoogle Scholar[22]

V. Dimatteo, A. Ascari, A. Fortunato

Continuous laser welding with spatial beam oscillation of dissimilar thin sheet materials (Al-Cu and Cu-Al): Process optimization and characterization

J. Manuf. Process., 44 (2019), pp. 158-165

ArticleDownload PDFView Record in ScopusGoogle Scholar[23]

V. Dimatteo, A. Ascari, E. Liverani, A. Fortunato

Experimental investigation on the effect of spot diameter on continuous-wave laser welding of copper and aluminum thin sheets for battery manufacturing

Opt. Laser Technol., 145 (2022), Article 107495

ArticleDownload PDFView Record in ScopusGoogle Scholar[24]

D. Wu, X. Hua, F. Li, L. Huang

Understanding of spatter formation in fiber laser welding of 5083 aluminum alloy

Int. J. Heat Mass Transf., 113 (2017), pp. 730-740

ArticleDownload PDFView Record in ScopusGoogle Scholar[25]

R. Ducharme, K. Williams, P. Kapadia, J. Dowden, B. Steen, M. Glowacki

The laser welding of thin metal sheets: an integrated keyhole and weld pool model with supporting experiments

J. Phys. D Appl. Phys., 27 (8) (1994), p. 1619 View PDF

CrossRefView Record in ScopusGoogle Scholar[26]

C.W. Hirt, B.D. Nichols

Volume of fluid (VOF) method for the dynamics of free boundaries

J. Comput. Phys., 39 (1) (1981), pp. 201-225

ArticleDownload PDFGoogle Scholar[27]

W. Piekarska, M. Kubiak

Three-dimensional model for numerical analysis of thermal phenomena in laser–arc hybrid welding process

Int. J. Heat Mass Transf., 54 (23–24) (2011), pp. 4966-4974

ArticleDownload PDFView Record in ScopusGoogle Scholar[28]J. Zhou, H.-L. Tsai, and P.-C. Wang, “Transport phenomena and keyhole dynamics during pulsed laser welding,” 2006.

Google Scholar[29]

D. Harrison, D. Yan, S. Blairs

The surface tension of liquid copper

J. Chem. Thermodyn., 9 (12) (1977), pp. 1111-1119

ArticleDownload PDFView Record in ScopusGoogle Scholar[30]

M. Leitner, T. Leitner, A. Schmon, K. Aziz, G. Pottlacher

Thermophysical properties of liquid aluminum

Metall. Mater. Trans. A, 48 (6) (2017), pp. 3036-3045 View PDF

This article is free to access.

CrossRefView Record in ScopusGoogle Scholar[31]

H.-C. Tran, Y.-L. Lo

Systematic approach for determining optimal processing parameters to produce parts with high density in selective laser melting process

Int. J. Adv. Manuf. Technol., 105 (10) (2019), pp. 4443-4460 View PDF

CrossRefView Record in ScopusGoogle Scholar[32]A. Ascari, A. Fortunato, E. Liverani, and A. Lutey, “Application of different pulsed laser sources to dissimilar welding of Cu and Al alloys,” in Proceedings of Lasers in Manufacturing Conference (LIM), 2019.

Google Scholar[33]

A. Fortunato, A. Ascari

Laser welding of thin copper and aluminum sheets: feasibility and challenges in continuous-wave welding of dissimilar metals

Lasers in Manufacturing and Materials Processing, 6 (2) (2019), pp. 136-157 View PDF

CrossRefView Record in ScopusGoogle Scholar[34]

A. Boucherit, M.-N. Avettand-Fènoël, R. Taillard

Effect of a Zn interlayer on dissimilar FSSW of Al and Cu

Mater. Des., 124 (2017), pp. 87-99

ArticleDownload PDFView Record in ScopusGoogle Scholar[35]

N. Kumar, I. Masters, A. Das

In-depth evaluation of laser-welded similar and dissimilar material tab-to-busbar electrical interconnects for electric vehicle battery pack

J. Manuf. Process., 70 (2021), pp. 78-96

ArticleDownload PDFView Record in ScopusGoogle Scholar[36]

M. Abbasi, A.K. Taheri, M. Salehi

Growth rate of intermetallic compounds in Al/Cu bimetal produced by cold roll welding process

J. Alloy. Compd., 319 (1–2) (2001), pp. 233-241

ArticleDownload PDFGoogle Scholar[37]

D. Zuo, S. Hu, J. Shen, Z. Xue

Intermediate layer characterization and fracture behavior of laser-welded copper/aluminum metal joints

Mater. Des., 58 (2014), pp. 357-362

ArticleDownload PDFView Record in ScopusGoogle Scholar[38]

S. Yan, Y. Shi

Influence of Ni interlayer on microstructure and mechanical properties of laser welded joint of Al/Cu bimetal

J. Manuf. Process., 59 (2020), pp. 343-354

ArticleDownload PDFView Record in ScopusGoogle Scholar[39]

S. Yan, Y. Shi

Influence of laser power on microstructure and mechanical property of laser-welded Al/Cu dissimilar lap joints

J. Manuf. Process., 45 (2019), pp. 312-321

ArticleDownload PDFView Record in ScopusGoogle Scholar

Extratropical cyclone damage to the seawall in Dawlish, UK: eyewitness accounts, sea level analysis and numerical modelling

영국 Dawlish의 방파제에 대한 온대 저기압 피해: 목격자 설명, 해수면 분석 및 수치 모델링

Extratropical cyclone damage to the seawall in Dawlish, UK: eyewitness accounts, sea level analysis and numerical modelling

Natural Hazards (2022)Cite this article

Abstract

2014년 2월 영국 해협(영국)과 특히 Dawlish에 영향을 미친 온대 저기압 폭풍 사슬은 남서부 지역과 영국의 나머지 지역을 연결하는 주요 철도에 심각한 피해를 입혔습니다.

이 사건으로 라인이 두 달 동안 폐쇄되어 5천만 파운드의 피해와 12억 파운드의 경제적 손실이 발생했습니다. 이 연구에서는 폭풍의 파괴력을 해독하기 위해 목격자 계정을 수집하고 해수면 데이터를 분석하며 수치 모델링을 수행합니다.

우리의 분석에 따르면 이벤트의 재난 관리는 성공적이고 효율적이었으며 폭풍 전과 도중에 인명과 재산을 구하기 위해 즉각적인 조치를 취했습니다. 파도 부이 분석에 따르면 주기가 4–8, 8–12 및 20–25초인 복잡한 삼중 봉우리 바다 상태가 존재하는 반면, 조위계 기록에 따르면 최대 0.8m의 상당한 파도와 최대 1.5m의 파도 성분이 나타났습니다.

이벤트에서 가능한 기여 요인으로 결합된 진폭. 최대 286 KN의 상당한 임펄스 파동이 손상의 시작 원인일 가능성이 가장 높았습니다. 수직 벽의 반사는 파동 진폭의 보강 간섭을 일으켜 파고가 증가하고 최대 16.1m3/s/m(벽의 미터 너비당)의 상당한 오버탑핑을 초래했습니다.

이 정보와 우리의 공학적 판단을 통해 우리는 이 사고 동안 다중 위험 계단식 실패의 가장 가능성 있는 순서는 다음과 같다고 결론을 내립니다. 조적 파괴로 이어지는 파도 충격력, 충전물 손실 및 연속적인 조수에 따른 구조물 파괴.

The February 2014 extratropical cyclonic storm chain, which impacted the English Channel (UK) and Dawlish in particular, caused significant damage to the main railway connecting the south-west region to the rest of the UK. The incident caused the line to be closed for two months, £50 million of damage and an estimated £1.2bn of economic loss. In this study, we collate eyewitness accounts, analyse sea level data and conduct numerical modelling in order to decipher the destructive forces of the storm. Our analysis reveals that the disaster management of the event was successful and efficient with immediate actions taken to save lives and property before and during the storm. Wave buoy analysis showed that a complex triple peak sea state with periods at 4–8, 8–12 and 20–25 s was present, while tide gauge records indicated that significant surge of up to 0.8 m and wave components of up to 1.5 m amplitude combined as likely contributing factors in the event. Significant impulsive wave force of up to 286 KN was the most likely initiating cause of the damage. Reflections off the vertical wall caused constructive interference of the wave amplitudes that led to increased wave height and significant overtopping of up to 16.1 m3/s/m (per metre width of wall). With this information and our engineering judgement, we conclude that the most probable sequence of multi-hazard cascading failure during this incident was: wave impact force leading to masonry failure, loss of infill and failure of the structure following successive tides.

Introduction

The progress of climate change and increasing sea levels has started to have wide ranging effects on critical engineering infrastructure (Shakou et al. 2019). The meteorological effects of increased atmospheric instability linked to warming seas mean we may be experiencing more frequent extreme storm events and more frequent series or chains of events, as well as an increase in the force of these events, a phenomenon called storminess (Mölter et al. 2016; Feser et al. 2014). Features of more extreme weather events in extratropical latitudes (30°–60°, north and south of the equator) include increased gusting winds, more frequent storm squalls, increased prolonged precipitation and rapid changes in atmospheric pressure and more frequent and significant storm surges (Dacre and Pinto 2020). A recent example of these events impacting the UK with simultaneous significant damage to coastal infrastructure was the extratropical cyclonic storm chain of winter 2013/2014 (Masselink et al. 2016; Adams and Heidarzadeh 2021). The cluster of storms had a profound effect on both coastal and inland infrastructure, bringing widespread flooding events and large insurance claims (RMS 2014).

The extreme storms of February 2014, which had a catastrophic effect on the seawall of the south Devon stretch of the UK’s south-west mainline, caused a two-month closure of the line and significant disruption to the local and regional economy (Fig. 1b) (Network Rail 2014; Dawson et al. 2016; Adams and Heidarzadeh 2021). Restoration costs were £35 m, and economic effects to the south-west region of England were estimated up to £1.2bn (Peninsula Rail Taskforce 2016). Adams and Heidarzadeh (2021) investigated the disparate cascading failure mechanisms which played a part in the failure of the railway through Dawlish and attempted to put these in the context of the historical records of infrastructure damage on the line. Subsequent severe storms in 2016 in the region have continued to cause damage and disruption to the line in the years since 2014 (Met Office 2016). Following the events of 2014, Network Rail Footnote1 who owns the network has undertaken a resilience study. As a result, it has proposed a £400 m refurbishment of the civil engineering assets that support the railway (Fig. 1) (Network Rail 2014). The new seawall structure (Fig. 1a,c), which is constructed of pre-cast concrete sections, encases the existing Brunel seawall (named after the project lead engineer, Isambard Kingdom Brunel) and has been improved with piled reinforced concrete foundations. It is now over 2 m taller to increase the available crest freeboard and incorporates wave return features to minimise wave overtopping. The project aims to increase both the resilience of the assets to extreme weather events as well as maintain or improve amenity value of the coastline for residents and visitors.

figure 1
Fig. 1

In this work, we return to the Brunel seawall and the damage it sustained during the 2014 storms which affected the assets on the evening of the 4th and daytime of the 5th of February and eventually resulted in a prolonged closure of the line. The motivation for this research is to analyse and model the damage made to the seawall and explain the damage mechanisms in order to improve the resilience of many similar coastal structures in the UK and worldwide. The innovation of this work is the multidisciplinary approach that we take comprising a combination of analysis of eyewitness accounts (social science), sea level and wave data analysis (physical science) as well as numerical modelling and engineering judgement (engineering sciences). We investigate the contemporary wave climate and sea levels by interrogating the real-time tide gauge and wave buoys installed along the south-west coast of the English Channel. We then model a typical masonry seawall (Fig. 2), applying the computational fluid dynamics package FLOW3D-Hydro,Footnote2 to quantify the magnitude of impact forces that the seawall would have experienced leading to its failure. We triangulate this information to determine the probable sequence of failures that led to the disaster in 2014.

figure 2
Fig. 2

Data and methods

Our data comprise eyewitness accounts, sea level records from coastal tide gauges and offshore wave buoys as well as structural details of the seawall. As for methodology, we analyse eyewitness data, process and investigate sea level records through Fourier transform and conduct numerical simulations using the Flow3D-Hydro package (Flow Science 2022). Details of the data and methodology are provided in the following.

Eyewitness data

The scale of damage to the seawall and its effects led the local community to document the first-hand accounts of those most closely affected by the storms including residents, local businesses, emergency responders, politicians and engineering contractors involved in the post-storm restoration work. These records now form a permanent exhibition in the local museum in DawlishFootnote3, and some of these accounts have been transcribed into a DVD account of the disaster (Dawlish Museum 2015). We have gathered data from the Dawlish Museum, national and international news reports, social media tweets and videos. Table 1 provides a summary of the eyewitness accounts. Overall, 26 entries have been collected around the time of the incident. Our analysis of the eyewitness data is provided in the third column of Table 1 and is expanded in Sect. 3.Table 1 Eyewitness accounts of damage to the Dawlish railway due to the February 2014 storm and our interpretations

Full size table

Sea level data and wave environment

Our sea level data are a collection of three tide gauge stations (Newlyn, Devonport and Swanage Pier—Fig. 5a) owned and operated by the UK National Tide and Sea Level FacilityFootnote4 for the Environment Agency and four offshore wave buoys (Dawlish, West Bay, Torbay and Chesil Beach—Fig. 6a). The tide gauge sites are all fitted with POL-EKO (www.pol-eko.com.pl) data loggers. Newlyn has a Munro float gauge with one full tide and one mid-tide pneumatic bubbler system. Devonport has a three-channel data pneumatic bubbler system, and Swanage Pier consists of a pneumatic gauge. Each has a sampling interval of 15 min, except for Swanage Pier which has a sampling interval of 10 min. The tide gauges are located within the port areas, whereas the offshore wave buoys are situated approximately 2—3.3 km from the coast at water depths of 10–15 m. The wave buoys are all Datawell Wavemaker Mk III unitsFootnote5 and come with sampling interval of 0.78 s. The buoys have a maximum saturation amplitude of 20.5 m for recording the incident waves which implies that every wave larger than this threshold will be recorded at 20.5 m. The data are provided by the British Oceanographic Data CentreFootnote6 for tide gauges and the Channel Coastal ObservatoryFootnote7 for wave buoys.

Sea level analysis

The sea level data underwent quality control to remove outliers and spikes as well as gaps in data (e.g. Heidarzadeh et al. 2022; Heidarzadeh and Satake 2015). We processed the time series of the sea level data using the Matlab signal processing tool (MathWorks 2018). For calculations of the tidal signals, we applied the tidal package TIDALFIT (Grinsted 2008), which is based on fitting tidal harmonics to the observed sea level data. To calculate the surge signals, we applied a 30-min moving average filter to the de-tided data in order to remove all wind, swell and infra-gravity waves from the time series. Based on the surge analysis and the variations of the surge component before the time period of the incident, an error margin of approximately ± 10 cm is identified for our surge analysis. Spectral analysis of the wave buoy data is performed using the fast Fourier transform (FFT) of Matlab package (Mathworks 2018).

Numerical modelling

Numerical modelling of wave-structure interaction is conducted using the computational fluid dynamics package Flow3D-Hydro version 1.1 (Flow Science 2022). Flow3D-Hydro solves the transient Navier–Stokes equations of conservation of mass and momentum using a finite difference method and on Eulerian and Lagrangian frameworks (Flow Science 2022). The aforementioned governing equations are:

∇.u=0∇.u=0

(1)

∂u∂t+u.∇u=−∇Pρ+υ∇2u+g∂u∂t+u.∇u=−∇Pρ+υ∇2u+g

(2)

where uu is the velocity vector, PP is the pressure, ρρ is the water density, υυ is the kinematic viscosity and gg is the gravitational acceleration. A Fractional Area/Volume Obstacle Representation (FAVOR) is adapted in Flow3D-Hydro, which applies solid boundaries within the Eulerian grid and calculates the fraction of areas and volume in partially blocked volume in order to compute flows on corresponding boundaries (Hirt and Nichols 1981). We validated the numerical modelling through comparing the results with Sainflou’s analytical equation for the design of vertical seawalls (Sainflou 1928; Ackhurst 2020), which is as follows:

pd=ρgHcoshk(d+z)coshkdcosσtpd=ρgHcoshk(d+z)coshkdcosσt

(3)

where pdpd is the hydrodynamic pressure, ρρ is the water density, gg is the gravitational acceleration, HH is the wave height, dd is the water depth, kk is the wavenumber, zz is the difference in still water level and mean water level, σσ is the angular frequency and tt is the time. The Sainflou’s equation (Eq. 3) is used to calculate the dynamic pressure from wave action, which is combined with static pressure on the seawall.

Using Flow3D-Hydro, a model of the Dawlish seawall was made with a computational domain which is 250.0 m in length, 15.0 m in height and 0.375 m in width (Fig. 3a). The computational domain was discretised using a single uniform grid with a mesh size of 0.125 m. The model has a wave boundary at the left side of the domain (x-min), an outflow boundary on the right side (x-max), a symmetry boundary at the bottom (z-min) and a wall boundary at the top (z-max). A wall boundary implies that water or waves are unable to pass through the boundary, whereas a symmetry boundary means that the two edges of the boundary are identical and therefore there is no flow through it. The water is considered incompressible in our model. For volume of fluid advection for the wave boundary (i.e. the left-side boundary) in our simulations, we utilised the “Split Lagrangian Method”, which guarantees the best accuracy (Flow Science, 2022).

figure 3
Fig. 3

The stability of the numerical scheme is controlled and maintained through checking the Courant number (CC) as given in the following:

C=VΔtΔxC=VΔtΔx

(4)

where VV is the velocity of the flow, ΔtΔt is the time step and ΔxΔx is the spatial step (i.e. grid size). For stability and convergence of the numerical simulations, the Courant number must be sufficiently below one (Courant et al. 1928). This is maintained by a careful adjustment of the ΔxΔx and ΔtΔt selections. Flow3D-Hydro applies a dynamic Courant number, meaning the program adjusts the value of time step (ΔtΔt) during the simulations to achieve a balance between accuracy of results and speed of simulation. In our simulation, the time step was in the range ΔtΔt = 0.0051—0.051 s.

In order to achieve the most efficient mesh resolution, we varied cell size for five values of ΔxΔx = 0.1 m, 0.125 m, 0.15 m, 0.175 m and 0.20 m. Simulations were performed for all mesh sizes, and the results were compared in terms of convergence, stability and speed of simulation (Fig. 3). A linear wave with an amplitude of 1.5 m and a period of 6 s was used for these optimisation simulations. We considered wave time histories at two gauges A and B and recorded the waves from simulations using different mesh sizes (Fig. 3). Although the results are close (Fig. 3), some limited deviations are observed for larger mesh sizes of 0.20 m and 0.175 m. We therefore selected mesh size of 0.125 m as the optimum, giving an extra safety margin as a conservative solution.

The pressure from the incident waves on the vertical wall is validated in our model by comparing them with the analytical equation of Sainflou (1928), Eq. (3), which is one of the most common set of equations for design of coastal structures (Fig. 4). The model was tested by running a linear wave of period 6 s and wave amplitude of 1.5 m against the wall, with a still water level of 4.5 m. It can be seen that the model results are very close to those from analytical equations of Sainflou (1928), indicating that our numerical model is accurately modelling the wave-structure interaction (Fig. 4).

figure 4
Fig. 4

Eyewitness account analysis

Contemporary reporting of the 4th and 5th February 2014 storms by the main national news outlets in the UK highlights the extreme nature of the events and the significant damage and disruption they were likely to have on the communities of the south-west of England. In interviews, this was reinforced by Network Rail engineers who, even at this early stage, were forecasting remedial engineering works to last for at least 6 weeks. One week later, following subsequent storms the cascading nature of the events was obvious. Multiple breaches of the seawall had taken place with up to 35 separate landslide events and significant damage to parapet walls along the coastal route also were reported. Residents of the area reported extreme effects of the storm, one likening it to an earthquake and reporting water ingress through doors windows and even through vertical chimneys (Table 1). This suggests extreme wave overtopping volumes and large wave impact forces. One resident described the structural effects as: “the house was jumping up and down on its footings”.

Disaster management plans were quickly and effectively put into action by the local council, police service and National Rail. A major incident was declared, and decisions regarding evacuation of the residents under threat were taken around 2100 h on the night of 4th February when reports of initial damage to the seawall were received (Table 1). Local hotels were asked to provide short-term refuge to residents while local leisure facilities were prepared to accept residents later that evening. Initial repair work to the railway line was hampered by successive high spring tides and storms in the following days although significant progress was still made when weather conditions permitted (Table 1).

Sea level observations and spectral analysis

The results of surge and wave analyses are presented in Figs. 5 and 6. A surge height of up to 0.8 m was recorded in the examined tide gauge stations (Fig. 5b-d). Two main episodes of high surge heights are identified: the first surge started on 3rd February 2014 at 03:00 (UTC) and lasted until 4th of February 2014 at 00:00; the second event occurred in the period 4th February 2014 15:00 to 5th February 2014 at 17:00 (Fig. 5b-d). These data imply surge durations of 21 h and 26 h for the first and the second events, respectively. Based on the surge data in Fig. 5, we note that the storm event of early February 2014 and the associated surges was a relatively powerful one, which impacted at least 230 km of the south coast of England, from Land’s End to Weymouth, with large surge heights.

figure 5
Fig. 5
figure 6
Fig. 6

Based on wave buoy records, the maximum recorded amplitudes are at least 20.5 m in Dawlish and West Bay, 1.9 m in Tor Bay and 4.9 m in Chesil (Fig. 6a-b). The buoys at Tor Bay and Chesil recorded dual peak period bands of 4–8 and 8–12 s, whereas at Dawlish and West Bay registered triple peak period bands at 4–8, 8–12 and 20–25 s (Fig. 6c, d). It is important to note that the long-period waves at 20–25 s occur with short durations (approximately 2 min) while the waves at the other two bands of 4–8 and 8–12 s appear to be present at all times during the storm event.

The wave component at the period band of 4–8 s can be most likely attributed to normal coastal waves while the one at 8–12 s, which is longer, is most likely the swell component of the storm. Regarding the third component of the waves with long period of 20 -25 s, which occurs with short durations of 2 min, there are two hypotheses; it is either the result of a local (port and harbour) and regional (the Lyme Bay) oscillations (eg. Rabinovich 1997; Heidarzadeh and Satake 2014; Wang et al. 1992), or due to an abnormally long swell. To test the first hypothesis, we consider various water bodies such as Lyme Bay (approximate dimensions of 70 km × 20 km with an average water depth of 30 m; Fig. 6), several local bays (approximate dimensions of 3.6 km × 0.6 km with an average water depth of 6 m) and harbours (approximate dimensions of 0.5 km × 0.5 km with an average water depth of 4 m). Their water depths are based on the online Marine navigation website.Footnote8 According to Rabinovich (2010), the oscillation modes of a semi-enclosed rectangle basin are given by the following equation:

Tmn=2gd−−√[(m2L)2+(nW)2]−1/2Tmn=2gd[(m2L)2+(nW)2]−1/2

(5)

where TmnTmn is the oscillation period, gg is the gravitational acceleration, dd is the water depth, LL is the length of the basin, WW is the width of the basin, m=1,2,3,…m=1,2,3,… and n=0,1,2,3,…n=0,1,2,3,…; mm and nn are the counters of the different modes. Applying Eq. (5) to the aforementioned water bodies results in oscillation modes of at least 5 min, which is far longer than the observed period of 20–25 s. Therefore, we rule out the first hypothesis and infer that the long period of 20–25 s is most likely a long swell wave coming from distant sources. As discussed by Rabinovich (1997) and Wang et al. (2022), comparison between sea level spectra before and after the incident is a useful method to distinguish the spectrum of the weather event. A visual inspection of Fig. 6 reveals that the forcing at the period band of 20–25 s is non-existent before the incident.

Numerical simulations of wave loading and overtopping

Based on the results of sea level data analyses in the previous section (Fig. 6), we use a dual peak wave spectrum with peak periods of 10.0 s and 25.0 s for numerical simulations because such a wave would be comprised of the most energetic signals of the storm. For variations of water depth (2.0–4.0 m), coastal wave amplitude (0.5–1.5 m) (Fig. 7) and storm surge height (0.5–0.8 m) (Fig. 5), we developed 20 scenarios (Scn) which we used in numerical simulations (Table 2). Data during the incident indicated that water depth was up to the crest level of the seawall (approximately 4 m water depth); therefore, we varied water depth from 2 to 4 m in our simulation scenarios. Regarding wave amplitudes, we referred to the variations at a nearby tide gauge station (West Bay) which showed wave amplitude up to 1.2 m (Fig. 7). Therefore, wave amplitude was varied from 0.5 m to 1.5 m by considering a factor a safety of 25% for the maximum wave amplitude. As for the storm surge component, time series of storm surges calculated at three coastal stations adjacent to Dawlish showed that it was in the range of 0.5 m to 0.8 m (Fig. 5). These 20 scenarios would help to study uncertainties associated with wave amplitudes and pressures. Figure 8 shows snapshots of wave propagation and impacts on the seawall at different times.

figure 7
Fig. 7

Table 2 The 20 scenarios considered for numerical simulations in this study

Full size table

figure 8
Fig. 8

Results of wave amplitude simulations

Large wave amplitudes can induce significant wave forcing on the structure and cause overtopping of the seawall, which could eventually cascade to other hazards such as erosion of the backfill and scour (Adams and Heidarzadeh, 2021). The first 10 scenarios of our modelling efforts are for the same incident wave amplitudes of 0.5 m, which occur at different water depths (2.0–4.0 m) and storm surge heights (0.5–0.8 m) (Table 2 and Fig. 9). This is because we aim at studying the impacts of effective water depth (deff—the sum of mean sea level and surge height) on the time histories of wave amplitudes as the storm evolves. As seen in Fig. 9a, by decreasing effective water depth, wave amplitude increases. For example, for Scn-1 with effective depth of 4.5 m, the maximum amplitude of the first wave is 1.6 m, whereas it is 2.9 m for Scn-2 with effective depth of 3.5 m. However, due to intensive reflections and interferences of the waves in front of the vertical seawall, such a relationship is barely seen for the second and the third wave peaks. It is important to note that the later peaks (second or third) produce the largest waves rather than the first wave. Extraordinary wave amplifications are seen for the Scn-2 (deff = 3.5 m) and Scn-7 (deff = 3.3 m), where the corresponding wave amplitudes are 4.5 m and 3.7 m, respectively. This may indicate that the effective water depth of deff = 3.3–3.5 m is possibly a critical water depth for this structure resulting in maximum wave amplitudes under similar storms. In the second wave impact, the combined wave height (i.e. the wave amplitude plus the effective water depth), which is ultimately an indicator of wave overtopping, shows that the largest wave heights are generated by Scn-2, 7 and 8 (Fig. 9a) with effective water depths of 3.5 m, 3.3 m and 3.8 m and combined heights of 8.0 m, 7.0 m and 6.9 m (Fig. 9b). Since the height of seawall is 5.4 m, the combined wave heights for Scn-2, 7 and 8 are greater than the crest height of the seawall by 2.6 m, 1.6 m and 1.5 m, respectively, which indicates wave overtopping.

figure 9
Fig. 9

For scenarios 11–20 (Fig. 10), with incident wave amplitudes of 1.5 m (Table 2), the largest wave amplitudes are produced by Scn-17 (deff = 3.3 m), Scn-13 (deff = 2.5 m) and Scn-12 (deff = 3.5 m), which are 5.6 m, 5.1 m and 4.5 m. The maximum combined wave heights belong to Scn-11 (deff = 4.5 m) and Scn-17 (deff = 3.3 m), with combined wave heights of 9.0 m and 8.9 m (Fig. 10b), which are greater than the crest height of the seawall by 4.6 m and 3.5 m, respectively.

figure 10
Fig. 10

Our simulations for all 20 scenarios reveal that the first wave is not always the largest and wave interactions, reflections and interferences play major roles in amplifying the waves in front of the seawall. This is primarily because the wall is fully vertical and therefore has a reflection coefficient of close to one (i.e. full reflection). Simulations show that the combined wave height is up to 4.6 m higher than the crest height of the wall, implying that severe overtopping would be expected.

Results of wave loading calculations

The pressure calculations for scenarios 1–10 are given in Fig. 11 and those of scenarios 11–20 in Fig. 12. The total pressure distribution in Figs. 1112 mostly follows a triangular shape with maximum pressure at the seafloor as expected from the Sainflou (1928) design equations. These pressure plots comprise both static (due to mean sea level in front of the wall) and dynamic (combined effects of surge and wave) pressures. For incident wave amplitudes of 0.5 m (Fig. 11), the maximum wave pressure varies in the range of 35–63 kPa. At the sea surface, it is in the range of 4–20 kPa (Fig. 11). For some scenarios (Scn-2 and 7), the pressure distribution deviates from a triangular shape and shows larger pressures at the top, which is attributed to the wave impacts and partial breaking at the sea surface. This adds an additional triangle-shaped pressure distribution at the sea surface elevation consistent with the design procedure developed by Goda (2000) for braking waves. The maximum force on the seawall due to scenarios 1–10, which is calculated by integrating the maximum pressure distribution over the wave-facing surface of the seawall, is in the range of 92–190 KN (Table 2).

figure 11
Fig. 11
figure 12
Fig. 12

For scenarios 11–20, with incident wave amplitude of 1.5 m, wave pressures of 45–78 kPa and 7–120 kPa, for  the bottom and top of the wall, respectively, were observed (Fig. 12). Most of the plots show a triangular pressure distribution, except for Scn-11 and 15. A significant increase in wave impact pressure is seen for Scn-15 at the top of the structure, where a maximum pressure of approximately 120 kPa is produced while other scenarios give a pressure of 7–32 kPa for the sea surface. In other words, the pressure from Scn-15 is approximately four times larger than the other scenarios. Such a significant increase of the pressure at the top is most likely attributed to the breaking wave impact loads as detailed by Goda (2000) and Cuomo et al. (2010). The wave simulation snapshots in Fig. 8 show that the wave breaks before reaching the wall. The maximum force due to scenarios 11–20 is 120–286 KN.

The breaking wave impacts peaking at 286 KN in our simulations suggest destabilisation of the upper masonry blocks, probably by grout malfunction. This significant impact force initiated the failure of the seawall which in turn caused extensive ballast erosion. Wave impact damage was proposed by Adams and Heidarzadeh (2021) as one of the primary mechanisms in the 2014 Dawlish disaster. In the multi-hazard risk model proposed by these authors, damage mechanism III (failure pathway 5 in Adams and Heidarzadeh, 2021) was characterised by wave impact force causing damage to the masonry elements, leading to failure of the upper sections of the seawall and loss of infill material. As blocks were removed, access to the track bed was increased for inbound waves allowing infill material from behind the seawall to be fluidised and subsequently removed by backwash. The loss of infill material critically compromised the stability of the seawall and directly led to structural failure. In parallel, significant wave overtopping (discussed in the next section) led to ballast washout and cascaded, in combination with masonry damage, to catastrophic failure of the wall and suspension of the rails in mid-air (Fig. 1b), leaving the railway inoperable for two months.

Wave Overtopping

The two most important factors contributing to the 2014 Dawlish railway catastrophe were wave impact forces and overtopping. Figure 13 gives the instantaneous overtopping rates for different scenarios, which experienced overtopping. It can be seen that the overtopping rates range from 0.5 m3/s/m to 16.1 m3/s/m (Fig. 13). Time histories of the wave overtopping rates show that the phenomenon occurs intermittently, and each time lasts 1.0–7.0 s. It is clear that the longer the overtopping time, the larger the volume of the water poured on the structure. The largest wave overtopping rates of 16.1 m3/s/m and 14.4 m3/s/m belong to Scn-20 and 11, respectively. These are the two scenarios that also give the largest combined wave heights (Fig. 10b).

figure 13
Fig. 13

The cumulative overtopping curves (Figs. 1415) show the total water volume overtopped the structure during the entire simulation time. This is an important hazard factor as it determines the level of soil saturation, water pore pressure in the soil and soil erosion (Van der Meer et al. 2018). The maximum volume belongs to Scn-20, which is 65.0 m3/m (m-cubed of water per metre length of the wall). The overtopping volumes are 42.7 m3/m for Scn-11 and 28.8 m3/m for Scn-19. The overtopping volume is in the range of 0.7–65.0 m3/m for all scenarios.

figure 14
Fig. 14
figure 15
Fig. 15

For comparison, we compare our modelling results with those estimated using empirical equations. For the case of the Dawlish seawall, we apply the equation proposed by Van Der Meer et al. (2018) to estimate wave overtopping rates, based on a set of decision criteria which are the influence of foreshore, vertical wall, possible breaking waves and low freeboard:

qgH3m−−−−√=0.0155(Hmhs)12e(−2.2RcHm)qgHm3=0.0155(Hmhs)12e(−2.2RcHm)

(6)

where qq is the mean overtopping rate per metre length of the seawall (m3/s/m), gg is the acceleration due to gravity, HmHm is the incident wave height at the toe of the structure, RcRc is the wall crest height above mean sea level, hshs is the deep-water significant wave height and e(x)e(x) is the exponential function. It is noted that Eq. (6) is valid for 0.1<RcHm<1.350.1<RcHm<1.35. For the case of the Dawlish seawall and considering the scenarios with larger incident wave amplitude of 1.5 m (hshs= 1.5 m), the incident wave height at the toe of the structure is HmHm = 2.2—5.6 m, and the wall crest height above mean sea level is RcRc = 0.6–2.9 m. As a result, Eq. (6) gives mean overtopping rates up to approximately 2.9 m3/s/m. A visual inspection of simulated overtopping rates in Fig. 13 for Scn 11–20 shows that the mean value of the simulated overtopping rates (Fig. 13) is close to estimates using Eq. (6).

Discussion and conclusions

We applied a combination of eyewitness account analysis, sea level data analysis and numerical modelling in combination with our engineering judgement to explain the damage to the Dawlish railway seawall in February 2014. Main findings are:

  • Eyewitness data analysis showed that the extreme nature of the event was well forecasted in the hours prior to the storm impact; however, the magnitude of the risks to the structures was not well understood. Multiple hazards were activated simultaneously, and the effects cascaded to amplify the damage. Disaster management was effective, exemplified by the establishment of an emergency rendezvous point and temporary evacuation centre during the storm, indicating a high level of hazard awareness and preparedness.
  • Based on sea level data analysis, we identified triple peak period bands at 4–8, 8–12 and 20–25 s in the sea level data. Storm surge heights and wave oscillations were up to 0.8 m and 1.5 m, respectively.
  • Based on the numerical simulations of 20 scenarios with different water depths, incident wave amplitudes, surge heights and peak periods, we found that the wave oscillations at the foot of the seawall result in multiple wave interactions and interferences. Consequently, large wave amplitudes, up to 4.6 m higher than the height of the seawall, were generated and overtopped the wall. Extreme impulsive wave impact forces of up to 286 KN were generated by the waves interacting with the seawall.
  • We measured maximum wave overtopping rates of 0.5–16.1 m3/s/m for our scenarios. The cumulative overtopping water volumes per metre length of the wall were 0.7–65.0 m3/m.
  • Analysis of all the evidence combined with our engineering judgement suggests that the most likely initiating cause of the failure was impulsive wave impact forces destabilising one or more grouted joints between adjacent masonry blocks in the wall. Maximum observed pressures of 286 KN in our simulations are four times greater in magnitude than background pressures leading to block removal and initiating failure. Therefore, the sequence of cascading events was :1) impulsive wave impact force causing damage to masonry, 2) failure of the upper sections of the seawall, 3) loss of infill resulting in a reduction of structural strength in the landward direction, 4) ballast washout as wave overtopping and inbound wave activity increased and 5) progressive structural failure following successive tides.

From a risk mitigation point of view, the stability of the seawall in the face of future energetic cyclonic storm events and sea level rise will become a critical factor in protecting the rail network. Mitigation efforts will involve significant infrastructure investment to strengthen the civil engineering assets combined with improved hazard warning systems consisting of meteorological forecasting and real-time wave observations and instrumentation. These efforts must take into account the amenity value of coastal railway infrastructure to local communities and the significant number of tourists who visit every year. In this regard, public awareness and active engagement in the planning and execution of the project will be crucial in order to secure local stakeholder support for the significant infrastructure project that will be required for future resilience.

Notes

  1. https://www.networkrail.co.uk/..
  2. https://www.flow3d.com/products/flow-3d-hydro/.
  3. https://www.devonmuseums.net/Dawlish-Museum/Devon-Museums/.
  4. https://ntslf.org/.
  5. https://www.datawell.nl/Products/Buoys/DirectionalWaveriderMkIII.aspx.
  6. https://www.bodc.ac.uk/.
  7. https://coastalmonitoring.org/cco/.
  8. https://webapp.navionics.com/#boating@8&key=iactHlwfP.

References

Download references

Acknowledgements

We are grateful to Brunel University London for administering the scholarship awarded to KA. The Flow3D-Hydro used in this research for numerical modelling is licenced to Brunel University London through an academic programme contract. We sincerely thank Prof Harsh Gupta (Editor-in-Chief) and two anonymous reviewers for their constructive review comments.

Funding

This project was funded by the UK Engineering and Physical Sciences Research Council (EPSRC) through a PhD scholarship to Keith Adams.

Author information

Authors and Affiliations

  1. Department of Civil and Environmental Engineering, Brunel University London, Uxbridge, UB8 3PH, UKKeith Adams
  2. Department of Architecture and Civil Engineering, University of Bath, Bath, BA2 7AY, UKMohammad Heidarzadeh

Corresponding author

Correspondence to Keith Adams.

Ethics declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.

Availability of data

All data used in this study are provided in the body of the article.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Cite this article

Adams, K., Heidarzadeh, M. Extratropical cyclone damage to the seawall in Dawlish, UK: eyewitness accounts, sea level analysis and numerical modelling. Nat Hazards (2022). https://doi.org/10.1007/s11069-022-05692-2

Download citation

  • Received17 May 2022
  • Accepted17 October 2022
  • Published14 November 2022
  • DOIhttps://doi.org/10.1007/s11069-022-05692-2

Share this article

Anyone you share the following link with will be able to read this content:Get shareable link

Provided by the Springer Nature SharedIt content-sharing initiative

Keywords

  • Storm surge
  • Cyclone
  • Railway
  • Climate change
  • Infrastructure
  • Resilience
Fig. 1. Schematic of lap welding for 6061/5182 aluminum alloys.

알루미늄 합금 겹침 용접 중 용접 형성, 용융 흐름 및 입자 구조에 대한 사인파 발진 레이저 빔의 영향

린 첸 가오 양 미시 옹 장 춘밍 왕
Lin Chen , Gaoyang Mi , Xiong Zhang , Chunming Wang *
중국 우한시 화중과학기술대학 재료공학부, 430074

Effects of sinusoidal oscillating laser beam on weld formation, melt flow and grain structure during aluminum alloys lap welding

Abstract

A numerical model of 1.5 mm 6061/5182 aluminum alloys thin sheets lap joints under laser sinusoidal oscillation (sine) welding and laser welding (SLW) weld was developed to simulate temperature distribution and melt flow. Unlike the common energy distribution of SLW, the sinusoidal oscillation of laser beam greatly homogenized the energy distribution and reduced the energy peak. The energy peaks were located at both sides of the sine weld, resulting in the tooth-shaped sectional formation. This paper illustrated the effect of the temperature gradient (G) and solidification rate (R) on the solidification microstructure by simulation. Results indicated that the center of the sine weld had a wider area with low G/R, promoting the formation of a wider equiaxed grain zone, and the columnar grains were slenderer because of greater GR. The porosity-free and non-penetration welds were obtained by the laser sinusoidal oscillation. The reasons were that the molten pool volume was enlarged, the volume proportion of keyhole was reduced and the turbulence in the molten pool was gentled, which was observed by the high-speed imaging and simulation results of melt flow. The tensile test of both welds showed a tensile fracture form along the fusion line, and the tensile strength of sine weld was significantly better than that of the SLW weld. This was because that the wider equiaxed grain area reduced the tendency of cracks and the finer grain size close to the fracture location. Defect-free and excellent welds are of great significance to the new energy vehicles industry.

온도 분포 및 용융 흐름을 시뮬레이션하기 위해 레이저 사인파 진동 (사인) 용접 및 레이저 용접 (SLW) 용접에서 1.5mm 6061/5182 알루미늄 합금 박판 랩 조인트 의 수치 모델이 개발되었습니다. SLW의 일반적인 에너지 분포와 달리 레이저 빔의 사인파 진동은 에너지 분포를 크게 균질화하고 에너지 피크를 줄였습니다. 에너지 피크는 사인 용접의 양쪽에 위치하여 톱니 모양의 단면이 형성되었습니다. 이 논문은 온도 구배(G)와 응고 속도 의 영향을 설명했습니다.(R) 시뮬레이션에 의한 응고 미세 구조. 결과는 사인 용접의 중심이 낮은 G/R로 더 넓은 영역을 가짐으로써 더 넓은 등축 결정립 영역의 형성을 촉진하고 더 큰 GR로 인해 주상 결정립 이 더 가늘다는 것을 나타냅니다. 다공성 및 비관통 용접은 레이저 사인파 진동에 의해 얻어졌습니다. 그 이유는 용융 풀의 부피가 확대되고 열쇠 구멍의 부피 비율이 감소하며 용융 풀의 난류가 완만해졌기 때문이며, 이는 용융 흐름의 고속 이미징 및 시뮬레이션 결과에서 관찰되었습니다. 두 용접부 의 인장시험 은 융착선을 따라 인장파괴형태를인장강도사인 용접의 경우 SLW 용접보다 훨씬 우수했습니다. 이는 등축 결정립 영역이 넓을수록 균열 경향이 감소하고 파단 위치에 근접한 입자 크기가 미세 하기 때문입니다. 결함이 없고 우수한 용접은 신에너지 자동차 산업에 매우 중요합니다.

Fig. 1. Schematic of lap welding for 6061/5182 aluminum alloys.
Fig. 1. Schematic of lap welding for 6061/5182 aluminum alloys.
Fig. 2. Finite element mesh.
Fig. 2. Finite element mesh.
Fig. 3. Weld morphologies of cross-section and upper surface for the two welds: (a) sine pattern weld; (b) SLW weld.
Fig. 3. Weld morphologies of cross-section and upper surface for the two welds: (a) sine pattern weld; (b) SLW weld.
Fig. 4. Calculation of laser energy distribution: (a)-(c) sine pattern weld; (d)-(f) SLW weld.
Fig. 4. Calculation of laser energy distribution: (a)-(c) sine pattern weld; (d)-(f) SLW weld.
Fig. 5. The partially melted region of zone A.
Fig. 5. The partially melted region of zone A.
Fig. 6. The simulated profiles of melted region for the two welds: (a) SLW weld; (b) sine pattern weld.
Fig. 6. The simulated profiles of melted region for the two welds: (a) SLW weld; (b) sine pattern weld.
Fig. 7. The temperature field simulation results of cross section for sine pattern weld.
Fig. 7. The temperature field simulation results of cross section for sine pattern weld.
Fig. 8. Dynamic behavior of the molten pool at the same time interval of 0.004 s within one oscillating period: (a) SLW weld; (b) sine pattern weld.
Fig. 8. Dynamic behavior of the molten pool at the same time interval of 0.004 s within one oscillating period: (a) SLW weld; (b) sine pattern weld.
Fig. 9. The temperature field and flow field of the molten pool for the SLW weld: (a)~(f) t = 80 ms~100 ms.
Fig. 9. The temperature field and flow field of the molten pool for the SLW weld: (a)~(f) t = 80 ms~100 ms.
Fig. 10. The temperature field and flow field of the molten pool for the sine pattern weld: (a)~(f) t = 151 ms~171 ms.
Fig. 10. The temperature field and flow field of the molten pool for the sine pattern weld: (a)~(f) t = 151 ms~171 ms.
Fig. 11. The evolution of the molten pool volume and keyhole depth within one period.
Fig. 11. The evolution of the molten pool volume and keyhole depth within one period.
Fig. 12. The X-ray inspection results for the two welds: (a) SLW weld, (b) sine pattern weld.
Fig. 12. The X-ray inspection results for the two welds: (a) SLW weld, (b) sine pattern weld.
Fig. 13. Comparison of the solidification parameters for sine and SLW patterns: (a) the temperature field simulated results of the molten pool upper surfaces; (b) temperature gradient G and solidification rate R along the molten pool boundary isotherm from weld centerline to the fusion boundary; (c) G/R; (d) GR.
Fig. 13. Comparison of the solidification parameters for sine and SLW patterns: (a) the temperature field simulated results of the molten pool upper surfaces; (b) temperature gradient G and solidification rate R along the molten pool boundary isotherm from weld centerline to the fusion boundary; (c) G/R; (d) GR.
Fig. 14. The EBSD results of equiaxed grain zone in the weld center of: (a) sine pattern weld; (b) SLW weld; (c) grain size.
Fig. 14. The EBSD results of equiaxed grain zone in the weld center of: (a) sine pattern weld; (b) SLW weld; (c) grain size.
Fig. 15. (a) EBSD results of horizontal sections of SLW weld and sine pattern weld; (b) The columnar crystal widths of SLW weld and sine pattern weld.
Fig. 15. (a) EBSD results of horizontal sections of SLW weld and sine pattern weld; (b) The columnar crystal widths of SLW weld and sine pattern weld.
Fig. 16. (a) The tensile test results of the two welds; (b) Fracture location of SLW weld; (b) Fracture location of sine pattern weld.
Fig. 16. (a) The tensile test results of the two welds; (b) Fracture location of SLW weld; (b) Fracture location of sine pattern weld.

Keywords

Laser welding, Sinusoidal oscillating, Energy distribution, Numerical simulation, Molten pool flow, Grain structure

References

Chen, X., 2014. Study on laser-MAG Hybrid Weaving Welding Charateristics. Master
thesis. Harbin Institute of Technology, China.
Chen, G., Wang, B., Mao, S., Zhong, P., He, J., 2019. Research on the “∞”-shaped laser
scanning welding process for aluminum alloy. Opt. Laser Technol. 115, 32–41.
Cho, W.-I., Na, S.-J., Cho, M.-H., Lee, J.-S., 2010. Numerical study of alloying element
distribution in CO2 laser–GMA hybrid welding. Comput. Mater. Sci. 49, 792–800.
Cho, W.-I., Na, S.-J., Thomy, C., Vollertsen, F., 2012. Numerical simulation of molten
pool dynamics in high power disk laser welding. J. Mater. Process. Technol. 212,
262–275.
Das, A., Butterworth, I., Masters, I., Williams, D., 2018. Microstructure and mechanical
properties of gap-bridged remote laser welded (RLW) automotive grade AA 5182
joints. Mater. Charact. 145, 697–712.
Fetzer, F., Sommer, M., Weber, R., Weberpals, J.-P., Graf, T., 2018. Reduction of pores by
means of laser beam oscillation during remote welding of AlMgSi. Opt. Lasers Eng.
108, 68–77.
Geng, S., Jiang, P., Shao, X., Guo, L., Gao, X., 2020. Heat transfer and fluid flow and their
effects on the solidification microstructure in full-penetration laser welding of
aluminum sheet. J. Mater. Sci. Technol. 46, 50–63.
Hagenlocher, C., Sommer, M., Fetzer, F., Weber, R., Graf, T., 2018a. Optimization of the
solidification conditions by means of beam oscillation during laser beam welding of
aluminum. Mater. Des. 160, 1178–1185.
Hagenlocher, C., Weller, D., Weber, R., Graf, T., 2018b. Reduction of the hot cracking
susceptibility of laser beam welds in AlMgSi alloys by increasing the number of grain
boundaries. Sci. Technol. Weld. Join. 24, 313–319.
Hagenlocher, C., Fetzer, F., Weller, D., Weber, R., Graf, T., 2019. Explicit analytical
expressions for the influence of welding parameters on the grain structure of laser
beam welds in aluminium alloys. Mater. Des. 174, 107791.
Han, X., Tang, X., Wang, T., Shao, C., Lu, F., Cui, H., 2018. Role of ambient pressure in
keyhole dynamics based on beam transmission path method for laser welding on Al
alloy. Int. J. Adv. Manuf. Technol. 99, 1639–1651.
Hao, K., Li, G., Gao, M., Zeng, X., 2015. Weld formation mechanism of fiber laser
oscillating welding of austenitic stainless steel. J. Mater. Process. Technol. 225,
77–83.
Hirt, C.W., Nichols, B.D., 1981. Volume of fluid (VOF) method for the dynamics of free
boundaries. J. Comput. Phys. 39, 201–225.
Jiang, Z., Chen, X., Li, H., Lei, Z., Chen, Y., Wu, S., Wang, Y., 2020. Grain refinement and
laser energy distribution during laser oscillating welding of Invar alloy. Mater. Des.
186, 108195.
Kaplan, A., 1994. A model of deep penetration laser welding based on calculation of the
keyhole profile. J. Phys. D Appl. Phys. 27, 1805–1814.
Kou, S., 2002. Welding Metallurgy, 2nd ed. Wiley-Interscience, New Jersey, USA.
Kuryntsev, S.V., Gilmutdinov, A.K., 2015. The effect of laser beam wobbling mode in
welding process for structural steels. Int. J. Adv. Manuf. Technol. 81, 1683–1691.
Li, P., Nie, F., Dong, H., Li, S., Yang, G., Zhang, H., 2018. Pulse MIG welding of 6061-T6/
A356-T6 aluminum alloy dissimilar T-joint. J. Mater. Eng. Perform. 27, 4760–4769.
Liu, T., Mu, Z., Hu, R., Pang, S., 2019. Sinusoidal oscillating laser welding of 7075
aluminum alloy: hydrodynamics, porosity formation and optimization. Int. J. Heat
Mass Transf. 140, 346–358.
Seto, N., Katayama, S., Matsunawa, A., 2000. High-speed simultaneous observation of
plasma and keyhole behavior during high power CO2 laser welding: effect of
shielding gas on porosity formation. J. Laser Appl. 12, 245–250.
Tang, Z., Vollertsen, F., 2014. Influence of grain refinement on hot cracking in laser
welding of aluminum. Weld. World 58, 355–366.
Wang, L., Gao, M., Zhang, C., Zeng, X., 2016. Effect of beam oscillating pattern on weld
characterization of laser welding of AA6061-T6 aluminum alloy. Mater. Des. 108,
707–717.
Wang, L., Gao, M., Zeng, X., 2018. Experiment and prediction of weld morphology for
laser oscillating welding of AA6061 aluminium alloy. Sci. Technol. Weld. Join. 24,
334–341.
Yamazaki, Y., Abe, Y., Hioki, Y., Nakatani, M., Kitagawa, A., Nakata, K., 2016.
Fundamental study of narrow-gap welding with oscillation laser beam. Weld. Int. 30,
699–707.
Yuan, Z., Tu, Y., Yuan, T., Zhang, Y., Huang, Y., 2021. Size effects on mechanical
properties of pure industrial aluminum sheet for micro/meso scale plastic
deformation: experiment and modeling. J. Alloys. Compd. 859, 157752.
Zou, J., 2016. Characteristics of laser scanning welding process for 5A06 aluminum alloy
thick plate with narrow gap. Materials Processing Engineering. Harbin Welding
Institute, China. Master thesis.

Flow Field in a Sloped Channel with Damaged and Undamaged Piers: Numerical and Experimental Studies

Flow Field in a Sloped Channel with Damaged and Undamaged Piers: Numerical and Experimental Studies

Ehsan OveiciOmid Tayari & Navid Jalalkamali
KSCE Journal of Civil Engineering volume 25, pages4240–4251 (2021)Cite this article

Abstract

본 논문은 경사가 완만한 수로에서 손상되거나 손상되지 않은 교각 주변의 유동 패턴을 분석했습니다. 실험은 길이가 12m이고 기울기가 0.008인 직선 수로에서 수행되었습니다. Acoustic Doppler Velocimeter(ADV)를 이용하여 3차원 유속 데이터를 수집하였고, 그 결과를 PIV(Particle Image Velocimetry) 데이터와 분석하여 비교하였습니다.

다중 블록 옵션이 있는 취수구의 퇴적물 시뮬레이션(SSIIM)은 이 연구에서 흐름의 수치 시뮬레이션을 위해 통합되었습니다. 일반적으로 비교에서 얻은 결과는 수치 데이터와 실험 데이터 간의 적절한 일치를 나타냅니다. 결과는 모든 경우에 수로 입구에서 2m 거리에서 기복적 수압 점프가 발생했음을 보여주었습니다.

경사진 수로의 최대 베드 전단응력은 2개의 손상 및 손상되지 않은 교각을 설치하기 위한 수평 수로의 12배였습니다. 이와 같은 경사수로 교각의 위치에 따라 상류측 수위는 수평수로의 유사한 조건에 비해 72.5% 감소한 반면, 이 감소량은 경사면에서 다른 경우에 비해 8.3% 감소하였다. 채널 또한 두 교각이 있는 경우 최대 Froude 수는 수평 수로의 5.7배였습니다.

This paper analyzed the flow pattern around damaged and undamaged bridge piers in a channel with a mild slope. The experiments were carried out on a straight channel with a length of 12 meters and a slope of 0.008. Acoustic Doppler velocimeter (ADV) was employed to collect three-dimensional flow velocity data, and the results were analyzed and compared with particle image velocimetry (PIV) data. Sediment Simulation in Intakes with Multiblock option (SSIIM) was incorporated for the numerical simulation of the flow in this study. Generally, the results obtained from the comparisons referred to the appropriate agreement between the numerical and the experimental data. The results showed that an undular hydraulic jump occurred at a distance of two meters from the channel entrance in every case; the maximum bed shear stress in the sloped channel was 12 times that in a horizontal channel for installing two damaged and undamaged piers. With this position of the piers in the sloped channel, the upstream water level underwent a 72.5% reduction compared to similar conditions in a horizontal channel, while the amount of this water level decrease was equal to 8.3% compared to the other cases in a sloped channel. In addition, with the presence of both piers, the maximum Froude number was 5.7 times that in a horizontal channel.

This is a preview of subscription content, access via your institution.

References

Download references

Fig. 4. Numerical modeling of dual spillways: (a) Andong-1; (b) Andong-2; (c) Imha-1; (d) Juam-1; (e) Andong-3; (f) Imha-2; (g) Imha-3; and (h) Juam-3.

Interference of Dual Spillways Operations

Jai Hong Lee, Ph.D., P.E., M.ASCE; Pierre Y. Julien, Ph.D., M.ASCE; and Christopher I. Thornton, Ph.D., P.E., M.ASCE

Abstract

이중 여수로 간섭은 여수로가 서로 가깝게 배치될 때 수압 성능의 손실을 나타냅니다. 배수로 간섭은 물리적 실험과 수치 시뮬레이션을 모두 사용하여 조사됩니다.

이중 여수로 구성의 4개 물리적 모델의 단계 및 배출 측정값을 한국의 4개 댐 부지에서 Flow-3D 계산 결과와 비교합니다.

두 개의 배수로를 함께 사용하는 것을 각 배수로의 단일 작동과 비교합니다. 두 여수로를 동시에 운영할 경우 두 여수로를 통한 총 유량은 최대 7.6%까지 감소합니다.

간섭 계수는 단계 He가 설계 단계 Hd를 초과하고 두 배수로를 분리하는 거리 D가 배수로 너비 W에 비해 짧을 때 가장 중요합니다. 매개변수 DHd/WHe는 계산 및 측정된 간섭 계수와 매우 잘 관련됩니다.

안동댐 설계방류에 대한 홍수경로 예시는 간섭계수를 적용한 경우와 적용하지 않은 경우 저수지 수위의 차이가 42cm임을 보여줍니다. 결과적으로 댐 안전을 위해 추가 여수로의 너비(간섭 계수 포함)를 늘려야 합니다.

Dual spillway interference refers to the loss of hydraulic performance of spillways when they are placed close together. Spillway interference is examined using both physical experiments and numerical simulations. Stage and discharge measurements from four physical models with dual spillways configurations are compared to the Flow-3D computational results at four dam sites in South Korea. The conjunctive use of two spillways is compared with the singular operation of each spillway. When both spillways are operated at the same time, the total flow rate through the two spillways is reduced by up to 7.6%. Interference coefficients are most significant when the stage He exceeds the design stage Hd and when the distance D separating two spillways is short compared to the spillway width W. The parameter DHd/WHecorrelates very well with the calculated and measured interference coefficients. A flood routing example for the design discharge at Andong dam shows a 42 cm difference in reservoir water level with and without application of the interference coefficient. Consequently, the width of additional spillways (including the interference coefficient) should be increased for dam safety.

Fig. 1. Definition sketch for dual spillways
Fig. 1. Definition sketch for dual spillways
Fig. 2. Stage-discharge rating curves for dual spillway operations.
Fig. 2. Stage-discharge rating curves for dual spillway operations.
Fig. 3. Physical modeling of dual spillways: (a) Andong-1; (b) Andong-2; (c) Imha-1; and (d) Juam-1
Fig. 3. Physical modeling of dual spillways: (a) Andong-1; (b) Andong-2; (c) Imha-1; and (d) Juam-1
Fig. 4. Numerical modeling of dual spillways: (a) Andong-1; (b) Andong-2; (c) Imha-1; (d) Juam-1; (e) Andong-3; (f) Imha-2; (g) Imha-3; and (h) Juam-3.
Fig. 4. Numerical modeling of dual spillways: (a) Andong-1; (b) Andong-2; (c) Imha-1; (d) Juam-1; (e) Andong-3; (f) Imha-2; (g) Imha-3; and (h) Juam-3.
Fig. 4. (Continued.)
Fig. 4. (Continued.)
Fig. 5. Meshes and calculation domain for numerical modeling of Andong dam.
Fig. 5. Meshes and calculation domain for numerical modeling of Andong dam.
Fig. 6. Stage-discharge rating curve for existing and additional spillways (Andong-1): (a) existing spillway; (b) additional spillway; and (c) dual spillway simulations.
Fig. 6. Stage-discharge rating curve for existing and additional spillways (Andong-1): (a) existing spillway; (b) additional spillway; and (c) dual spillway simulations.
Fig. 7. Discharge comparison of physical experiments and numerical simulations. The upper panel is the comparative result for the existing spillway (ES) and the lower panel is for the additional spillway (AS) at four dams.
Fig. 7. Discharge comparison of physical experiments and numerical simulations. The upper panel is the comparative result for the existing spillway (ES) and the lower panel is for the additional spillway (AS) at four dams.
Fig. 8. Interference coefficients for dual spillways simulations with various scenarios.
Fig. 8. Interference coefficients for dual spillways simulations with various scenarios.
Fig. 9. Regression model for the distance-width ratio (D=W) and head ratio (Hd=He) by dual spillway simulations
Fig. 9. Regression model for the distance-width ratio (D=W) and head ratio (Hd=He) by dual spillway simulations
Fig. 10. Physical and numerical model validation: (a) numerical modeling; (b) solids of overflow weir of the spillway; and (c) physical models of reservoir and spillway
Fig. 10. Physical and numerical model validation: (a) numerical modeling; (b) solids of overflow weir of the spillway; and (c) physical models of reservoir and spillway
Fig. 11. Interference coefficients for dual spillways operations with various scenarios. The dashed lines indicate the results of the validation model with dual conditions of 1 þ 2, 1 þ 4, 1 þ 6, 3 þ 4, and 4 þ 5.
Fig. 11. Interference coefficients for dual spillways operations with various scenarios. The dashed lines indicate the results of the validation model with dual conditions of 1 þ 2, 1 þ 4, 1 þ 6, 3 þ 4, and 4 þ 5.
Fig. 12. Results of reservoir operations under the PMF at Andong dam.
Fig. 12. Results of reservoir operations under the PMF at Andong dam.

References

Cassidy, J. J. 1965. “Irrotational flow over spillways of finite height.”
J. Eng. Mech. Div. 91 (6): 155–173.
Chanel, P., and J. Doering. 2008. “Assessment of spillway modeling using
computational fluid dynamics.” Can. J. Civ. Eng. 35 (12): 1481–1485.
https://doi.org/10.1139/L08-094.
Chow, V. T. 1959. Open-channel hydraulics, 365–380. New York:
McGraw-Hill.
Ho, D., B. Cooper, K. Riddette, and S. Donohoo. 2006. “Application of
numerical modelling to spillways in Australia.” In Proc., Int. Symp.
on Dams in the Societies of the 21st Century, 22nd Int. Congress on
Large Dams (ICOLD), edited by L. Berga, et al. London: Taylor &
Francis.
Huff, F. A. 1967. “Time distribution of rainfall in heavy storms.” Water
Resour. Res. 3 (4): 1007–1019. https://doi.org/10.1029/WR003i004
p01007.
Kim, D. G., and J. H. Park. 2005. “Analysis of flow structure over ogeespillway in consideration of scale and roughness effects by using CFD
model.” KSCE J. Civ. Eng. 9 (2): 161–169. https://doi.org/10.1007
/BF02829067.
Koutsunis, N. A. 2015. “Impact of climatic changes on downstream hydraulic geometry and its influence on flood hydrograph
routing—Applied to the bluestone dam watershed.” M.S. degree,
Dept. of Civil and Environmental Engineering, Colorado State Univ.
Lee, J. H., and P. Y. Julien. 2016a. “ENSO impacts on temperature over
South Korea.” Int. J. Climatol. 36 (11): 3651. https://doi.org/10.1002
/joc.4581.
Lee, J. H., and P. Y. Julien. 2016b. “Teleconnections of the ENSO and
South Korean precipitation patterns.” J. Hydrol. 534: 237–250.
https://doi.org/10.1016/j.jhydrol.2016.01.011.
Lee, J. H., and P. Y. Julien. 2017. “Influence of the El Nino/southern ˜
oscillation on South Korean streamflow variability.” Hydrol. Processes
31 (12): 2162–2178. https://doi.org/10.1002/hyp.11168.
Li, S., S. Cain, N. Wosnik, C. Miller, H. Kocahan, and R. Wyckoff. 2011.
“Numerical modeling of probable maximum flood flowing through a
system of spillways.” J. Hydraul. Eng. 137 (1): 66–74. https://doi.org
/10.1061/(ASCE)HY.1943-7900.0000279.
MOCT (Ministry of Construction and Transportation). 2003. Hydraulic
model study of Soyanggang multipurpose dam auxiliary spillway.
[In Korean.] Governing City, South Korea: MOCT.
Olsen, N. R., and H. M. Kjellesvig. 1998. “Three-dimensional numerical
flow modeling for estimation of spillway capacity.” J. Hydraul. Res.
36 (5): 775–784. https://doi.org/10.1080/00221689809498602.
Savage, B. M., and M. C. Johnson. 2001. “Flow over ogee spillway:
Physical and numerical model case study.” J. Hydraul. Eng. 127 (8):
640–649. https://doi.org/10.1061/(ASCE)0733-9429(2001)127:8(640).
USACE (US Army Corps of Engineers). 2008. Hydrologic modeling
system HEC-HMS, user’s manual version 3.2. Davis, CA: USACE.
USBR (US Bureau of Reclamation). 1980. Hydraulic laboratory techniques: A water resources technical publication. Denver: US Dept.
of the Interior, Bureau of Reclamation.
Yakhot, V., and S. A. Orszag. 1986. “Renormalization group analysis of
turbulence. I: Basic theory.” J. Sci. Comput. 1 (1): 3–51. https://doi
.org/10.1007/BF01061452.
Yakhot, V., and L. M. Smith. 1992. “The renormalization group, the
e-expansion and derivation of turbulence models.” J. Sci. Comput.
7 (1): 35–61. https://doi.org/10.1007/BF01060210.
Zeng, J., L. Zhang, M. Ansar, E. Damisse, and J. A. Gonzalez-Castro. 2017.
“Applications of computational fluid dynamics to flow ratings at prototype spillways and weirs. I: Data generation and validation.” J. Irrig.
Drain. Eng. 143 (1): 04016072. https://doi.org/10.1061/(ASCE)IR
.1943-4774.0001112.

Fig. 2 Schematic diagram of the experimental Rijke tube

RIJKE 튜브 내부의 열음향 장에 대한 새로운 조사

A novel investigation of the thermoacoustic field inside a Rijke tube

B. EntezamW. Van Moorhem and J. MajdalaniPublished Online:22 Aug 2012 https://doi.org/10.2514/6.1998-2582

Abstract

이 논문에서는 Rijke 튜브 내부의 시간 종속 유동장의 실험 연구 및 계산 시뮬레이션에서 진행한 결과를 제시하고 해석합니다. 기존의 추측과 스케일링 분석을 기반으로 한 이론적 논의가 진행됩니다. 주요 결과에는 열 구동 진동에서 중요한 역할을 하는 것으로 보이는 유사성 매개변수가 포함됩니다. 이 매개변수는 열 섭동을 속도, 압력 및 특성 길이의 제곱과 관련시킵니다. 열 진동을 압력 및 속도 진동의 결합된 효과에 기인하는 간단한 이론은 계산, 실험 및 스케일링 고려 사항을 통해 논의됩니다. 이전의 분석 이론은 열 진동을 속도 또는 압력 진동에 연결했기 때문에 현재 분석 모델은 기존 추측에 동의하고 조정합니다. Rayleigh 기준에 따라 열원은 Rijke-tube 하단에서 1/4의 임계 거리에 위치해야 공명이 발생합니다. 이 관찰은 결합이 최대화되는 임계점이 음향 속도와 압력의 곱인 음향 강도가 가장 큰 공간 위치에 해당하기 때문에 제안된 해석을 확인합니다. 수치 시뮬레이션은 Rijke 튜브 내부의 압력 진동이 열 입력이 증가함에 따라 기하급수적으로 증가한다는 것을 보여줍니다. 충분히 작은 열 입력으로 음향 싱크가 소스를 초과하고 음향 감쇠가 발생합니다. 열 입력이 임계 임계값 이상으로 증가하면 음향 싱크가 불충분해져서 ​​내부 에너지 축적으로 인해 빠른 음향 증폭이 발생합니다.

In this paper, results proceeding from experimental studies and computational simulations of the time-dependent flowfield inside a Rijke tube are presented and interpreted. A theoretical discussion based on existing speculations and scaling analyses is carried out. The main results include a similarity parameter that appears to play an important role in the heat driven oscillations. This parameter relates heat perturbations to velocity, pressure, and the square of a characteristic length. A simple theory that attributes heat oscillations to the combined effects of pressure and velocity oscillations is discussed via computational, experimental, and scaling considerations. Since previous analytical theories link heat oscillations to either velocity or pressure oscillations, the current analytical model agrees with and reconciles between existing speculations. In compliance with the Rayleigh criterion, it is found that the heat source must be positioned at a critical distance of 1/4 from the Rijke-tube lower end for resonance to occur. This observation confirms our proposed interpretation since the critical point where coupling is maximized corresponds to a spatial location where the acoustic intensity, product of both acoustic velocities and pressures, is largest. Numerical simulations show that pressure oscillations inside the Rijke tube grow exponentially with increasing heat input With a sufficiently small heat input, the acoustic sinks exceed the sources and acoustic damping takes place. When the heat input is augmented beyond a critical threshold, acoustic sinks become insufficient causing rapid acoustic amplification by virtue of internal energy accumulation.

Fig. 2 Schematic diagram of the experimental Rijke tube
Fig. 2 Schematic diagram of the experimental Rijke tube
A novel investigation of the thermoacoustic field inside a Rijke tube
A novel investigation of the thermoacoustic field inside a Rijke tube

References

‘Entezam, B., Majdalani, J., and Van Moorhem, W. K.,
“Modeling of a Rijke-Tube Pulse Combustor Using
Computational Fluid Dynamics,” AIAA Paper 97-2718,
Seattle, WA, July 1997.

2George, W., and Reethof, G., “On the Fragility of
Acoustically Agglomerated Submicron Fly Ash
Particles,” Journal of Vibration, Acoustics, Stress, and
Reliability in Design, Vol. 108, July 1986, pp. 322-329.
3Tiwary R., and Reethof, G., “Hydrodynamic
Interaction of Spherical Aerosol Particles in a High
Intensity Acoustic Field,” Journal of Sound and
Vibration, Vol. 108, 1986, pp. 33-49.
4Reethof, G., “Acoustic Agglomeration of Power Plant
Fly Ash for Environmental and Hot Gas Clean-up,”
Transaction of the ASME, Vol. 110, Oct., 1988, pp.
552-557.
5
Song, L., Reethof, G., and Koopmann, G. H., “An
Improved Simulation Model of Acoustic
Agglomeration,” NCA Vol. 5, 89-WA, American
Society of Mechanical Engineers, Winter Annual
Meeting, San Francisco, CA, Dec., 10-15, 1989.
6Reethof, G., Koopmann, G. H., and Dorchak, T.,
“Acoustic Agglomeration for Paniculate Control at
High Temperature and high Pressure – Some Recent
results,” NCA Vol. 4, 89-WA, American Society of
Mechanical Engineers, Winter Annual Meeting, San
Francisco, CA, Dec., 10-15, 1989.
7Richards , G. A., and Bedick, R. C, “Application of
Acoustics in Advanced Energy Systems,” NCA Vol. 3,
89-WA, American Society of Mechanical Engineers,
Winter Annual Meeting, San Francisco, CA, Dec., 10-
15, 1989.
8Yavuzkurt, S., Ha, M. Y., Reethof, G., and Koopmann,
G., “Effect of Acoustic Field on the Combustion of
Coal Particles in a Rat Flame Burner,” Proceedings of
the Ist
Annual Pittsburgh Coal Conference, Pittsburgh,
PA, Sep., 1984, pp. 53-58.
^rice, E. W., “Review of Combustion Instability
Characteristics of Solid Propellants,” Advances in
Tactical Rocket Propulsion, AGARD Conference
Proceedings, No. 1, Part 2, Chap. 5, Technivision
Services, Maidenhead, England, 1968, pp. 141-194.
10Zinn, B.T., “State of the Art and Research Needs of
Pulsating Combustion,” NCA Vol. 19, 84-WA,
American Society of Mechanical Engineers, 1984.
“Rayleigh, J.W.S., The Theory of Sound, Vol. 1 and 2,
Dover Publications, New York, 1945, pp. 231-235.
12Zinn, B.T., Miller, N., Carvalho, J.A. Jr., and Daniel.
B. R., “Pulsating Combustion of Coal in a Rijke Type
Combustor,” Proceedings of the 19th International
Symposium on Combustion, 1982, pp. 1197-1203.
13Evans, R.E., and Putnam, A.A., “Rijke Tube
Apparatus,” Journal of Applied Physics, Vol. 360,
1966.
14Feldman, K. T., “Review of the Literature on Rijke
Thermoacoustic Phenomena, ” Journal of Sound and
Vibration, Vol. 7, 1968, pp. 83-89.
15Carvalho, J.R., Ferreira, C., Bressan, C., and Ferreira,
G., “Definition of Heater Location to Drive Maximum
Amplitude Acoustic Oscillations hi a Rijke Tube,”
Combustion and Flame, Vol. 76, 1989, pp. 17-27.
16Raun, R.L., Beckstead, M. W., Finlinson, J. C. , and
Brooks, K. P., “A Review of Rijke Tubes, Rijke
Burners and Related Devices,” Progress in Energy and
Combustion Science, Vol. 19, 1993, pp. 313-364.
17Chu, B. T., “Stability of Systems Containing a HeatSource-The Rayleigh Criterion, “NACA Research
Memorandum 56D27, 1956.
18Zinn, B. T., Daniel, B. R., and Shesdari, T.S.,
“Application of Pulsating Combustion in the Burning of
Solid Fuels,” Proceedings of the Symposium on Pulse
Combustion Technology for Heating Applications,
Argonne National Laboratory, 1979, pp. 239-248.
19Feldman, K.T., “Review of the Literature on
Soundhauss Thermoacoustic Phenomena ” Journal of
Sound and Vibration, Vol. 7, 1968, pp. 71-82.
20Flow Science Incorporated, Los Alamos, New
Mexico.

Figures-Effects of sinusoidal oscillating laser beam on weld formation, melt flow and grain structure during aluminum alloys lap welding

알루미늄 합금 겹침 용접 중 용접 형성, 용융 흐름 및 입자 구조에 대한 사인파 발진 레이저 빔의 영향

Effects of sinusoidal oscillating laser beam on weld formation, melt flow and grain structure during aluminum alloys lap welding

Lin Chen, Gaoyang Mi, Xiong Zhang, Chunming Wang
School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan, 430074, China

Abstract

레이저 사인파 진동(사인) 용접 및 레이저 용접(SLW)에서 1.5mm 6061/5182 알루미늄 합금 박판 랩 조인트의 수치 모델이 온도 분포와 용융 흐름을 시뮬레이션하기 위해 개발되었습니다.

SLW의 일반적인 에너지 분포와 달리 레이저 빔의 사인파 진동은 에너지 분포를 크게 균질화하고 에너지 피크를 줄였습니다. 에너지 피크는 사인 용접의 양쪽에 위치하여 톱니 모양의 단면이 형성되었습니다. 이 논문은 시뮬레이션을 통해 응고 미세구조에 대한 온도 구배(G)와 응고 속도(R)의 영향을 설명했습니다.

결과는 사인 용접의 중심이 낮은 G/R로 더 넓은 영역을 가짐으로써 더 넓은 등축 결정립 영역의 형성을 촉진하고 더 큰 GR로 인해 주상 결정립이 더 가늘다는 것을 나타냅니다. 다공성 및 비관통 용접은 레이저 사인파 진동에 의해 얻어졌습니다.

그 이유는 용융 풀의 부피가 확대되고 열쇠 구멍의 부피 비율이 감소하며 용융 풀의 난류가 완만해졌기 때문이며, 이는 용융 흐름의 고속 이미징 및 시뮬레이션 결과에서 관찰되었습니다. 두 용접부의 인장시험에서 융착선을 따라 인장파괴 형태를 보였고 사인 용접부의 인장강도가 SLW 용접부보다 유의하게 우수하였습니다.

이는 등축 결정립 영역이 넓을수록 균열 경향이 감소하고 파단 위치에 근접한 입자 크기가 미세하기 때문입니다. 결함이 없고 우수한 용접은 신에너지 자동차 산업에 매우 중요합니다.

A numerical model of 1.5 mm 6061/5182 aluminum alloys thin sheets lap joints under laser sinusoidal oscillation (sine) welding and laser welding (SLW) weld was developed to simulate temperature distribution and melt flow. Unlike the common energy distribution of SLW, the sinusoidal oscillation of laser beam greatly homogenized the energy distribution and reduced the energy peak. The energy peaks were located at both sides of the sine weld, resulting in the tooth-shaped sectional formation. This paper illustrated the effect of the temperature gradient (G) and solidification rate (R) on the solidification microstructure by simulation. Results indicated that the center of the sine weld had a wider area with low G/R, promoting the formation of a wider equiaxed grain zone, and the columnar grains were slenderer because of greater GR. The porosity-free and non-penetration welds were obtained by the laser sinusoidal oscillation. The reasons were that the molten pool volume was enlarged, the volume proportion of keyhole was reduced and the turbulence in the molten pool was gentled, which was observed by the high-speed imaging and simulation results of melt flow. The tensile test of both welds showed a tensile fracture form along the fusion line, and the tensile strength of sine weld was significantly better than that of the SLW weld. This was because that the wider equiaxed grain area reduced the tendency of cracks and the finer grain size close to the fracture location. Defect-free and excellent welds are of great significance to the new energy vehicles industry.

Keywords

Laser weldingSinusoidal oscillatingEnergy distributionNumerical simulationMolten pool flowGrain structure

Figures-Effects of sinusoidal oscillating laser beam on weld formation, melt flow and grain structure during aluminum alloys lap welding
Figures-Effects of sinusoidal oscillating laser beam on weld formation, melt flow and grain structure during aluminum alloys lap welding
Fig. 1. Schematic description of the laser welding process considered in this study.

Analysis of molten pool dynamics in laser welding with beam oscillation and filler wire feeding

Won-Ik Cho, Peer Woizeschke
Bremer Institut für angewandte Strahltechnik GmbH, Klagenfurter Straße 5, Bremen 28359, Germany

Received 30 July 2020, Revised 3 October 2020, Accepted 18 October 2020, Available online 1 November 2020.

Abstract

Molten pool flow and heat transfer in a laser welding process using beam oscillation and filler wire feeding were calculated using computational fluid dynamics (CFD). There are various indirect methods used to analyze the molten pool dynamics in fusion welding. In this work, based on the simulation results, the surface fluctuation was directly measured to enable a more intuitive analysis, and then the signal was analyzed using the Fourier transform and wavelet transform in terms of the beam oscillation frequency and buttonhole formation. The 1st frequency (2 x beam oscillation frequency, the so-called chopping frequency), 2nd frequency (4 x beam oscillation frequency), and beam oscillation frequency components were the main components found. The 1st and 2nd frequency components were caused by the effect of the chopping process and lumped line energy. The beam oscillation frequency component was related to rapid, unstable molten pool behavior. The wavelet transform effectively analyzed the rapid behaviors based on the change of the frequency components over time.

Korea Abstract

빔 진동 및 필러 와이어 공급을 사용하는 레이저 용접 공정에서 용융 풀 흐름 및 열 전달은 CFD (전산 유체 역학)를 사용하여 계산되었습니다. 용융 용접에서 용융 풀 역학을 분석하는 데 사용되는 다양한 간접 방법이 있습니다.

본 연구에서는 시뮬레이션 결과를 바탕으로 보다 직관적 인 분석이 가능하도록 표면 변동을 직접 측정 한 후 빔 발진 주파수 및 버튼 홀 형성 측면에서 푸리에 변환 및 웨이블릿 변환을 사용하여 신호를 분석했습니다.

1 차 주파수 (2 x 빔 발진 주파수, 이른바 초핑 주파수), 2 차 주파수 (4 x 빔 발진 주파수) 및 빔 발진 주파수 성분이 발견 된 주요 구성 요소였습니다. 1 차 및 2 차 주파수 성분은 쵸핑 공정과 집중 라인 에너지의 영향으로 인해 발생했습니다.

빔 진동 주파수 성분은 빠르고 불안정한 용융 풀 동작과 관련이 있습니다. 웨이블릿 변환은 시간 경과에 따른 주파수 구성 요소의 변화를 기반으로 빠른 동작을 효과적으로 분석했습니다.

1 . 소개

융합 용접에서 용융 풀 역학은 용접 결함과 시각적 이음새 품질에 직접적인 영향을 미칩니다. 이러한 역학을 연구하기 위해 고속 카메라를 사용하는 직접 방법과 광학 또는 음향 신호를 사용하는 간접 방법과 같은 다양한 측정 방법을 사용하여 여러 실험 방법을 고려했습니다. 시간 도메인의 원래 신호는 특별히 주파수 도메인에서 변환 된 신호로 변환되어 용융 풀 동작에 영향을 미치는 주파수 성분을 분석합니다. Kotecki et al. (1972)는 고속 카메라를 사용하여 가스 텅스텐 아크 용접에서 용융 풀을 관찰했습니다. [1]. 그들은 120Hz 리플 DC 출력을 가진 용접 전원을 사용할 때 용융 풀 진동 주파수가 120Hz임을 보여주었습니다. 전원을 끈 후 진동 주파수는 용융 풀의 고유 주파수를 나타내는 용융 풀 크기와 관련이 있습니다. 진동은 응고 중에 용접 표면 스케일링을 생성했습니다. Zacksenhouse and Hardt (1983)는 레이저 섀도 잉 동작 측정 기술을 사용하여 가스 텅스텐 아크 용접에서 완전히 관통 된 용융 풀의 동작을 측정했습니다 [2] . 그들은 2.5mm 두께의 강판에서 6mm 풀 반경 (고정 용접)에 대해 용융 풀의 고유 주파수가 18.9Hz라는 것을 발견했습니다. Semak et al. (1995) 고속 카메라를 사용하여 레이저 스폿 용접에서 용융 풀 및 키홀 역학 조사 [3]. 그들은 깊이가 약 3mm이고 반경이 약 3mm 인 용융 풀에서 200Hz의 낮은 체적 진동 주파수를 관찰했습니다. 0.45mm Aendenroomer와 den Ouden (1998)은 강철의 펄스 가스 텅스텐 아크 용접에서 용융 풀 진동을보고했습니다 [4] . 그들은 침투 깊이에 따라 진동 모드 변화를 보였고 주파수는 50Hz에서 150Hz 사이에서 변화했습니다. 주파수는 완전히 침투 된 용융 풀에서 더 낮았습니다. Hermans와 den Ouden (1999)은 단락 가스 금속 아크 용접에서 용융 풀 진동을 분석했습니다. [5]. 그들은 용융 풀의 단락 주파수와 고유 주파수가 같을 때 부분적으로 침투 된 용융 풀의 경우 공정 안정성이 향상되었음을 보여주었습니다. Yudodibroto et al. (2004)는 가스 텅스텐 아크 용접에서 용융 풀 진동에 대한 필러 와이어의 영향을 조사했습니다 [6] . 그들은 금속 전달이 특히 부분적으로 침투 된 용융 풀에서 진동 거동을 방해한다는 것을 보여주었습니다. Geiger et al. (2009) 레이저 키홀 용접에서 발광 분석 [7]. 신호의 주파수 분석을 사용하여 용융 풀 (1.5kHz 미만)과 키홀 (약 3kHz)에 해당하는 진동 주파수 범위를 찾았습니다. Kägeler와 Schmidt (2010)는 레이저 용접에서 용융 풀 크기의 변화를 관찰하기 위해 고속 카메라를 사용했습니다 [8] . 그들은 용융 풀에서 지배적 인 저주파 진동 성분 (100Hz 미만)을 발견했습니다. Shi et al. (2015) 고속 카메라를 사용하여 펄스 가스 텅스텐 아크 용접에서 용융 풀 진동 주파수 분석 [9]. 그들은 용접 침투 깊이가 작을수록 용융 풀의 진동 빈도가 더 높다는 것을 보여주었습니다. 추출 된 진동 주파수는 완전 용입 용접의 경우 85Hz 미만 이었지만 부분 용입 용접의 경우 110Hz에서 125Hz 사이였습니다. Volpp와 Vollertsen (2016)은 레이저 키홀 역학을 분석하기 위해 광학 신호를 사용했습니다 [10] . 그들은 공간 레이저 강도 분포로 인해 0.8에서 154 kHz 사이의 고주파 범위에서 피크를 발견했습니다. 위에서 언급 한 실험적 접근법은 공정 조건, 측정 방법 및 측정 된 위치에 따라 수십 Hz에서 수십 kHz까지 광범위한 용융 풀 역학에 대한 결과를 보여 주었다는 점에 유의해야합니다.

융합 용접에서 용융 풀 역학을 연구하기 위해 분석 접근 방식도 사용되었습니다. Zacksenhouse와 Hardt (1983)는 2.5mm 두께의 강판에서 대칭형 완전 관통 용융 풀의 고유 진동수를 계산했습니다 [2] . 매스 스프링 해석 모델을 사용하여 용융 풀 반경 6mm (고정 용접)에 대해 20.4Hz (실험에서 18.9Hz)의 고유 진동수와 3mm 풀 반경 (연속 용접)에 대해 40Hz의 고유 진동수를 예측했습니다. ). Postacioglu et al. (1989)는 원통형 용융 풀과 키홀을 가정하여 레이저 용접의 용융 풀에서 키홀 진동의 고유 진동수를 계산했습니다 .. 특정 열쇠 구멍 모양의 경우 약 900Hz의 기본 주파수가 계산되었습니다. Postacioglu et al. (1991)은 또한 레이저 용접에서 용접 속도를 고려하기 위해 타원형 용융 풀의 고유 진동수를 계산했습니다 [12] . 그들은 타원형 용융 풀의 모양이 고유 진동수에 영향을 미친다는 것을 보여주었습니다. 고유 진동수는 축의 길이 비율이 낮았으며, 즉 타원의 반장 축과 반 단축의 비율이 낮았습니다. Kroos et al. (1993)은 축 대칭 용융 풀과 키홀을 가정하여 레이저 키홀 용접의 동적 거동에 대한 이론적 모델을 개발했습니다 .. 키홀 폐쇄 시간은 0.1ms였으며 안정성 분석은 약 500Hz의 주파수에서 공진과 같은 진동을 예측했습니다. Maruo와 Hirata (1993)는 완전 관통 아크 용접에서 용융 풀을 모델링했습니다 [14] . 그들은 녹은 웅덩이가 정적 타원 모양을 가지고 있다고 가정했습니다. 그들은 고유 진동수와 진동 모드 사이의 관계를 조사하고 용융 풀 크기가 감소함에 따라 고유 진동수가 증가한다는 것을 보여주었습니다. Klein et al. (1994)는 원통형 키홀 모양을 사용하여 완전 침투 레이저 용접에서 키홀 진동을 연구했습니다 [15] . 그들은 점성 감쇠로 인해 키홀 진동이 낮은 kHz 범위로 제한된다는 것을 보여주었습니다. Klein et al. (1996)은 또한 레이저 출력의 작은 변동이 강한 키홀 진동으로 이어질 수 있음을 보여주었습니다[16] . 그들은 키홀 진동의 주요 공진 주파수 범위가 500 ~ 3500Hz라는 것을 발견했습니다. Andersen et al. (1997)은 고정 가스 텅스텐 아크 용접 [17] 에서 고정 된 원통형 모양을 가정하여 용융 풀의 고유 진동수를 예측 했으며 완전 용입 용접에서 용융 풀 폭이 증가함에 따라 감소하는 것으로 나타났습니다. 3.175mm 두께의 강판의 경우 주파수는 20Hz ~ 100Hz 범위였습니다. 위에 표시된 분석 방법은 일반적으로 단순한 용융 풀 모양을 가정하고 고유 진동수를 계산했습니다. 이것은 단순한 용융 풀 모양으로 고정 용접 공정을 분석하는 데 충분하지만 대부분의 용접 사례를 설명하는 과도 용접 공정에서 용융 풀 역학 분석에는 적합하지 않습니다.

반면에 수치 접근 방식은 고온 및 강한 빛과 같은 실험적 제한없이 자세한 정보를 제공하기 때문에 용융 풀 역학을 분석하는 이점이 있습니다. 전산 유체 역학 (CFD)의 수치 시뮬레이션 기술이 발전함에 따라 용융 풀 역학 분석에 대한 많은 연구가 수행되었습니다. 실제 용융 표면 변화는 VOF (체적 부피) 방법을 사용하여 계산할 수 있습니다. Cho et al. (2010) CO 2 레이저-아크 하이브리드 용접 공정을 위한 수학적 모델 개발 [18], 구형 방울이 생성 된 금속 와이어의 용융 과정이 와이어 공급 속도와 일치한다고 가정합니다. 그들은 필러 와이어가 희석되는 용융 풀 동작을 보여주었습니다. Cho et al. (2012)는 높은 빔 품질과 높은 금속 흡수율로 인해 업계에서 널리 사용되는 디스크 레이저 키홀 용접으로 수학적 모델을 확장했습니다 [19] . 그들은 열쇠 구멍에서 레이저 광선 번들의 다중 반사를 고려하고 용융 풀에서 keyholing과 같은 빠른 표면 변화를 자세히보고했습니다. 최근 CFD 시뮬레이션은 험핑 (Otto et al., 2016 [20] ) 및 기공 (Lin et al., 2017 [21] )과 같은보다 구체적인 현상을 분석하는데도 사용되었습니다 .) 레이저 용접에서. 그러나 용융 풀 역학과 관련된 연구는 거의 수행되지 않았습니다. Ko et al. (2000)은 수치 시뮬레이션을 사용하여 가스 텅스텐 아크 용접 풀의 동적 거동을 조사했습니다 [22] . 그들은 완전히 침투 된 용융 풀이 부분적으로 침투 된 풀보다 낮은 주파수에서 진동한다는 것을 보여주었습니다. 진동은 수십 분의 1 초 내에 무시할 수있는 크기로 감쇠되었습니다. Geiger et al. (2009)는 또한 수치 시뮬레이션을 사용하여 레이저 용접에서 용융 풀 거동을 보여주었습니다 [7]. 그들은 계산 된 증발 속도를 주파수 분석에 사용하여 공정에서 나오는 빛의 실험 결과와 비교했습니다. 판금 레이저 용접에서 중요한 공간 빔 진동 및 추가 필러 재료가있는 공정에 대한 용융 풀 역학에 대한 연구도 불충분합니다. Hu et al. (2018)은 금속 전달 메커니즘을 밝히기 위해 전자빔 3D 프린팅에서 와이어 공급 모델링을 수행했습니다. 그들은 주로 열 입력에 의해 결정되는 액체 브리지 전이, 액적 전이 및 중간 전이의 세 가지 유형의 금속 전달 모드를 보여주었습니다 .. Meng et al. (2020)은 레이저 빔 용접에서 용융 풀에 필러 와이어에 의해 추가 된 추가 요소의 전자기 교반 효과를 모델링했습니다. 용가재의 연속적인 액체 브릿지 이동이 가정되었고, 그 결과 전자기 교반의 영향이 키홀 깊이에 미미한 반면 필러 와이어 혼합을 향상 시켰습니다 [24] . Cho et al. (2017) 용접 방향에 수직 인 1 차원 빔 진동과 용접 라인을 따라 공급되는 필러 와이어를 사용하여 레이저 용접을위한 시뮬레이션 모델 개발 [25]. 그들은 시뮬레이션을 사용하여 특정 용접 현상, 즉 용융 풀의 단추 구멍 형성을 보여주었습니다. Cho et al. (2018)은 다중 반사 수와 전력 흡수량의 푸리에 변환을 사용하여 주파수 영역에서 소위 쵸핑 주파수 (2 x 빔 발진 주파수) 성분을 발견했습니다 [26] . 그러나 그들은 용융 풀 역학을 분석하기 위해 간접 신호를 사용했습니다. 따라서보다 직관적 인 분석을 위해서는 표면의 변동을 직접 측정해야합니다.

이 연구는 이전 연구에서 개발 된 레이저 용접 모델을 사용하여 3 차원 과도 CFD 시뮬레이션을 수행하여 빔 진동 및 필러 와이어 공급을 포함한 레이저 용접 공정에서 용융 풀 역학을 조사합니다. 용융 된 풀 표면의 시간적 변화는 시뮬레이션 결과에서 추출되었습니다. 추출 된 데이터는 주파수 영역뿐만 아니라 시간-주파수 영역에서도 분석되었습니다. 신호 처리를 통해 도출 된 결과는 특징적인 용융 풀 역학을 나타내며 빔 진동 주파수 및 단추 구멍 형성 측면에서 레이저 용접의 역학을 줄일 수있는 잠재력을 제공합니다.

2 . 방법론

그림 1도 1은 용접 방향에 수직 인 1 차원 빔 진동과 용접 라인을 따라 공급되는 필러 와이어를 사용하는 레이저 용접 프로세스의 개략적 설명을 보여줍니다. 1mm 두께의 알루미늄 합금 (AlSi1MgMn) 시트는 시트 표면에 초점을 맞춘 멀티 kW 파이버 레이저 (YLR-8000S, IPG Photonics, USA)를 사용하여 용접되었습니다. 시트는 에어 갭이있는 맞대기 이음으로 정렬되었습니다. 1 차원 스캐너 (ILV DC-Scanner, Ingenieurbüro für Lasertechnik + Verschleiss-Schutz (ILV), 독일)를 사용하여 레이저 빔의 1 차원 정현파 진동을 실현했습니다. 이 스캔 시스템에서 최대 진동 폭은 250Hz의 진동 주파수에서 1.4mm입니다. 오정렬에 대한 공차를 개선하기 위해 동일한 최대 너비 값이 사용되었습니다. 와이어 공급 시스템은 1을 공급했습니다. 2mm 직경의 알루미늄 합금 (AlSi5) 필러 와이어를 일정한 공급 속도로 에어 갭을 채 웁니다. 1mm 에어 갭의 경우 와이어 이송 속도는 용접 속도의 1.5 배 값으로 설정되었으며 참조 실험 조건은 문헌에서 얻었습니다 (Schultz, 2015 참조).[27] ).

그림 1

CFD 시뮬레이션은 레이저 용접에서 열 전달 및 용융 풀 동작을 계산하기 위해 수행되었습니다. 그림 2 는 CFD 시뮬레이션을위한 계산 영역을 보여줍니다. 실온에서 1.2mm 직경의 필러 와이어가 공급되고 레이저 빔이 진동했습니다. 1mm 두께의 공작물이 용접 속도로 왼쪽에서 오른쪽으로 이동했습니다. 0.1mm의 최소 메쉬 크기가 도메인에서 생성되었습니다. 침투 깊이가 더 깊은 이전 연구의 메쉬 테스트 결과는 0.2mm 이하의 메쉬 크기로 시뮬레이션 정확도가 확보 된 것으로 나타 났으므로 [28] 본 연구에서 사용 된 메쉬 크기가 적절할 수 있습니다. 도메인을 구성하는 세포의 수는 약 120 만 개였습니다. 1 번 테이블사용 된 레이저 용접 매개 변수를 보여줍니다. 용융 풀 역학 측면에서 다양한 진동 주파수와 에어 갭 크기가 고려되었으며 12 개의 용접 사례가 표 2 에 나와 있습니다. 표 3 은 시뮬레이션에 사용 된 알루미늄 합금과 순수 알루미늄 (Cho et al., 2018 [26] )의 표면 장력 계수를 제외하고 온도와 무관 한 열-물리적 재료 특성을 보여줍니다 . 여기서 표면 장력 계수는 액체 온도에서 온도와 표면 장력 계수 사이의 선형 관계를 가진 유일한 온도 의존적 ​​특성이었습니다.

그림 2

표 1 . . 레이저 용접 매개 변수.

레이저 용접 매개 변수
레이저 빔 파워3.0kW
빔 허리 반경50µm *
용접 속도6.0m / 분
와이어 공급 속도9.0m / 분
빔 진동 폭1.4mm
빔 진동 주파수100Hz, 150Hz, 200Hz, 250Hz
에어 갭 크기0.8mm, 0.9mm, 1.0mm, 1.1mm

반경은 1.07μm의 파장, 4.2mm • mrad의 빔 품질, 시준 초점 거리 및 초점 렌즈 200mm, 광섬유 직경 100μm의 원형 빔을 가정하여 계산되었습니다.

표 2 . 이 연구에서 고려한 용접 사례.

에어 갭 크기 [mm]진동 주파수 [Hz]
100150200250
0.9사례 1엑스엑스엑스
1.0사례 2사례 4사례 7사례 10
1.1사례 3사례 5사례 8사례 11
1.2엑스사례 6사례 912면

표 3 . 시뮬레이션에 사용 된 열 물리적 재료 특성 (Cho et al., 2018 [26] ).

특성상징
밀도ρ2700kg / m3
열 전도성케이1.7×102Wm K
점도ν1.15×10−삼kg / ms
표면 장력 계수 티엘*γ엘0.871 J / m2
표면 장력 온도 구배 *−1.55×10−4J / m 2 K
표면 장력 계수γγ엘−ㅏ(티−티엘)
비열8.5×102J / kg K
융합 잠열h에스엘3.36×105J / kg
기화 잠열 *hV1.05×107J / kg
Solidus 온도티에스847K
Liquidus 온도티엘905K
끓는점 *티비2743K

순수한 알루미늄.

시뮬레이션을 위해 단상 뉴턴 유체와 비압축성 층류가 가정되었습니다. 질량, 운동량 및 에너지 보존의 지배 방정식을 해결하여 계산 영역에서 속도, 압력 및 온도 분포를 얻었습니다. VOF 방법은 자유 표면 경계를 찾는 데 사용되었습니다. 스칼라 보존 방정식을 추가로 도입하여 용융 풀에서 충전재의 부피 분율을 계산했습니다. 시뮬레이션에 사용 된 레이저 용접의 수학적 모델은 다음과 같습니다. 레이저 빔은 가우스와 같은 전력 밀도 분포를 기반으로 697 개의 광선 에너지 번들로 나뉩니다. 광선 추적 방법을 사용하여 다중 반사를 고려했습니다. 재료에 대한 레이저 빔의 반사 (또는 흡수) 에너지는 프레 넬 반사 모델을 사용하여 계산되었습니다. 온도에 따른 흡수율의 변화를 고려 하였다. 혼합물의 흡수율은베이스 및 충전제 물질 분획의 가중 평균을 사용하여 계산되었습니다. 반동 압력과 부력도 고려되었습니다. 경계 조건으로 에너지와 압력의 균형은 VOF 방법으로 계산 된 자유 표면에서 고려되었습니다. 레이저 용접 모델과 지배 방정식은 FLOW-3D v.11.2 (2017), Flow Science, Inc.에서 유한 차분 방법과 유한 체적 방법을 사용하여 이산화되고 해결되었습니다. 경계 조건으로 에너지와 압력의 균형은 VOF 방법으로 계산 된 자유 표면에서 고려되었습니다. 레이저 용접 모델과 지배 방정식은 FLOW-3D v.11.2 (2017), Flow Science, Inc.에서 유한 차분 방법과 유한 체적 방법을 사용하여 이산화되고 해결되었습니다. 경계 조건으로 에너지와 압력의 균형은 VOF 방법으로 계산 된 자유 표면에서 고려되었습니다. 레이저 용접 모델과 지배 방정식은 FLOW-3D v.11.2 (2017), Flow Science, Inc.에서 유한 차분 방법과 유한 체적 방법을 사용하여 이산화되고 해결되었습니다.[29] . 계산에는 48GB RAM이 장착 된 Intel® Xeon® 프로세서 E5649로 구성된 워크 스테이션이 사용되었습니다. 계산 시스템을 사용하여 0.2 초 레이저 용접을 시뮬레이션하는 데 약 18 시간이 걸렸습니다. 지배 방정식 (Cho and Woizeschke, 2020 [30] ) 및 레이저 용접 모델 (Cho et al., 2018 [26] )에 대한 자세한 설명은 부록 A 에서 확인할 수 있습니다 .

그림 3 은 용융 풀 변동의 직접 측정에 대한 개략적 설명을 보여줍니다. 용융 풀의 역학을 분석하기 위해 시뮬레이션 중에 용융 풀 표면의 시간적 변동 운동을 측정했습니다. 상단 및 하단 표면 모두에서 10kHz의 샘플링 주파수로 변동을 측정 한 반면, 측정 위치는 X 축의 레이저 빔 위치에서 2mm 떨어진 용접 중심선에있었습니다. 그림 4시간 신호를 분석하는 데 사용되는 푸리에 변환 및 웨이블릿 변환의 개략적 설명을 보여줍니다. 측정 된 시간 신호는 고속 푸리에 변환 (FFT) 방법을 사용하여 주파수 영역으로 변환되었습니다. 결과는 측정 기간 동안 평균화 된 주파수 성분의 크기를 보여줍니다. 웨이블릿 변환 방법은 시간-주파수 영역에서 국부적 인 특성을 찾는 데 사용되었습니다. 결과는 주파수 구성 요소의 크기뿐만 아니라 시간 변화도 보여줍니다.

그림 3
그림 4

3 . 결과

이 연구 에서는 표 2에 표시된 12 가지 용접 사례 를 시뮬레이션했습니다. 그림 5 는 3 차원 시뮬레이션 결과를 평면도 와 바닥면으로 보여줍니다. 결과는 용융 된 풀의 거동에 따라 분류 할 수 있습니다 : 단추 구멍 형성 없음 (녹색), 안정 또는 불안정 단추 구멍 있음 (파란색), 불안정한 단추 구멍으로 인한 구멍 결함 (빨간색). 일반적인 열쇠 구멍보다 훨씬 큰 직경을 가진 단추 구멍은 레이저 용접의 특정 진동 조건에서 나타날 수 있습니다 (Vollertsen, 2016 [31]). 진동 주파수가 증가함에 따라 용접 이음 부 코스 및 스케일링 측면에서 시각적 이음새 품질이 향상되었습니다. 고주파에서 스케일링은 무시할 수있을 정도 였고 코스는 균질했습니다. 언더컷 결함의 발생도 감소했습니다. 그러나 관통 결함 부족 (case 7, case 10)이 나타났다. 에어 갭은 단추 구멍 형성에 중요했습니다. 에어 갭 크기가 증가함에 따라 단추 구멍이 더 쉽게 형성되었지만 구멍 결함으로 더 쉽게 남아 있습니다. 안정적인 단추 구멍 형성은 고려 된 공극 조건의 좁은 영역에서만 나타납니다.

그림 5

그림 6 은 시뮬레이션과 실험에서 융합 영역의 모양을 보여줍니다. 버튼 홀이없는 경우 1, 불안정한 버튼 홀 형성이있는 경우 8, 안정적인 버튼 홀 형성이있는 경우 11의 3 가지 경우에 대해 시뮬레이션 결과와 실험 결과를 비교하여 유사성을 나타냈다. 본 연구에서 고려한 용접 조건의 경우 표면 품질 결과는 Fig. 5 와 같이 큰 차이를 보였으 나 단면 융착 영역 [26] 과 형상은 큰 차이를 보이지 않았다.

그림 6

무화과. 7 과 8 은 각각 100Hz와 250Hz의 진동 주파수에서 시뮬레이션 결과를 기반으로 분석 된 용융 풀 역학과 시뮬레이션 및 실험 결과를 보여줍니다. 이전 연구에서 볼 수 있듯이 레이저 빔의 진동 주파수는 단추 구멍 형성과 밀접한 관련이 있습니다 (Cho et al., 2018 [26] 참조 ). 그림 7 (a) 및 (b)는 각각 시뮬레이션 및 실험을 기반으로 한 진동 주파수 100Hz에서 대표적인 용융 풀 동작을 보여줍니다. 완전히 관통 된 키홀 및 버튼 홀 형성은 관찰되지 않았으며 응고 후 거친 비드 표면이 남았습니다. 그림 7(c)와 (d)는 각각 윗면과 바닥면의 표면 변동에 대한 시뮬레이션 결과를 기반으로 한 용융 풀 역학 분석을 보여줍니다. 샘플링 데이터는 상단 표면이 공작물의 상단 표면 위치에서 평균적으로 변동하는 반면 하단 표면은 공작물의 하단 표면 위치에서 평균적으로 변동하는 것으로 나타났습니다. 표면 변동의 푸리에 변환 및 웨이블릿 변환 결과는 명확한 1  주파수 (2 x 빔 발진 주파수, 이른바 초핑 주파수, Cho et al., 2018 [26] 참조 ) 및 2  주파수 (4 x 빔 발진)를 보여줍니다. 주파수) 두 표면의 구성 요소, 그러나 바닥 표면과 첫 번째에 대한 결과주파수 성분이 더 강합니다. 반면 그림 8 (a)와 (b)에서 보는 바와 같이 250Hz의 진동 주파수에서 시뮬레이션과 실험 결과는 안정된 버튼 홀 형성과 응고 후 매끄러운 비드 표면을 나타냈다. 그림 8 의 샘플링 신호의 진폭은 그림 7 의 진폭 보다 작으며 푸리에 변환 및 웨이블릿 변환의 결과에서 중요한 주파수 성분이 발견되지 않았습니다.

Fi 7
그림 8

Fig. 9 는 진동 주파수 200Hz에서 시뮬레이션 결과를 바탕으로 분석 된 용융 풀 역학과 시뮬레이션 및 실험 결과를 보여준다. 이 주파수에서 Fig. 9 (a)와 (b) 에서 보는 바와 같이 , 시뮬레이션과 실험 모두에서 불안정한 buttonhole 거동이 관찰되었다. 바닥면에서 샘플링 데이터의 푸리에 변환 및 웨이블릿 변환의 결과 빔 발진 주파수 성분이 발견되었습니다.

그림 9

4 . 토론

시뮬레이션 및 실험 결과는 비드 표면 품질이 향상되고 빔 진동 주파수가 증가함에 따라 버튼 홀이 형성되는 것으로 나타났습니다. 표면의 변동 데이터에 대한 푸리에 변환 및 웨이블릿 변환의 결과에 따라 다음과 같은 주요 주파수 구성 요소가 발견되었습니다. 1  및 2 버튼 홀 형성이없는 주파수, 불안정한 용융 풀 거동이있는 빔 진동 주파수, 안정적인 버튼 홀 형성이있는 중요한 주파수 성분이 없습니다. 이들 중 불안정한 용융 풀 동작과 관련된 빔 진동 주파수 성분은 완전히 관통 된 키홀과 반복적으로 생성 및 붕괴되는 불안정한 버튼 홀의 특성으로 인해 웨이블릿 변환 결과에서 명확한 실선 형태로 나타나지 않았습니다. 분석 결과는 윗면보다 바닥면에서 더 분명했습니다. 이는 필러 와이어 공급 및 키홀 링 공정에서 강한 하향 흐름으로 인해 용융 풀 역학이 바닥 표면 영역에서 더 강했기 때문입니다. 진동 주파수가 증가함에 따라 용융 풀 역학과 상단 표면과 하단 표면 간의 차이가 감소했습니다.

첫 번째 주파수 (2 x 빔 진동 주파수)는이 연구에서 관찰 된 가장 분명한 구성 요소였습니다. Schultz et al. (2018)은 또한 실험을 통해 동일한 성분을 발견했습니다 [32] , 용융 풀 표면 운동에 대한 푸리에 분석을 수행했습니다. 첫 번째 주파수 성분은 빔 발진주기 당 두 개의 주요 이벤트가 있음을 의미합니다. 이것은 레이저 빔이 빔 진동주기 당 두 번 와이어를 절단하거나 절단하는 프로세스와 일치합니다. 용융 된 와이어 팁은 낮은 진동 주파수에서 고르지 않고 날카로운 모서리를 갖는 것으로 나타났습니다 (Cho et al., 2018 [26] ). 이것은 첫 번째 원인이 될 수 있습니다.용융 된 풀에서 지배적이되는 주파수 성분. 진동 주파수가 증가하면 용융 된 와이어 팁이 더 균일 해 지므로 효과가 감소합니다. 용접 방향으로의 정현파 횡 방향 빔 진동을 통한 에너지 집중도 빔 진동주기 당 두 번 발생합니다. 그림 10 은 발진 주파수에 따른 레이저 빔의 라인 에너지 (단위 길이 당 에너지)의 변화를 보여줍니다. 그림 10 b) 의 라인 에너지 는 레이저 출력을 공정 속도로 나누어 계산했습니다. 여기서 처리 속도는(w이자형엘디나는엔지에스피이자형이자형디)2+(디(에스나는엔유에스영형나는디ㅏ엘wㅏV이자형나는엔에프나는지.10ㅏ))디티)2. 낮은 발진 주파수에서 라인 에너지는 발진 폭의 양쪽 끝에 과도하게 집중됩니다. 이러한 집중된 에너지는 과도한 키홀 링 프로세스를 초래하므로 언더컷 결함이 나타날 수있는 높은 흐름 역학이 발생합니다. 진동 주파수가 증가함에 따라 집중 에너지는 더 작은 조각으로 나뉩니다. 따라서 높은 진동 주파수에서 과도한 키홀 링 및 수반되는 언더컷 결함의 발생이 감소되었습니다. 위에서 언급 한 두 가지 현상 (불균일 한 와이어 팁과 집중된 라인 에너지)은 빔 발진주기 당 두 번 발생하며 발진 주파수가 증가하면 그 효과가 감소합니다. 따라서 저주파 에서 2  주파수 성분 (4 x 빔 발진 주파수)이 나타나는 것은이 두 현상의 동시 작용입니다.

그림 10

두 가지 현상 중 첫 번째 주파수 에 대한 주된 효과 는 집중된 라인 에너지입니다. Cho et al. (2018)은 전력 흡수 데이터를 푸리에 변환을 사용하여 분석했을 때 1  주파수 성분이 더 우세 해졌고, 2  주파수 성분은 발진 주파수가 증가함에 따라 상대적으로 약화 되었음을 보여주었습니다 [26] . 용융 된 와이어 팁은 또한 빈도가 증가함에 따라 더욱 균일 해졌습니다. 결과는 진동 주파수의 증가가 용융 풀에 대한 와이어의 영향을 제거하는 것으로 나타났습니다. 따라서 발진 주파수가 증가함에 따라 라인 에너지 집중의 영향 만 남을 수 있습니다. 그림 10 과 같이, 집중 선 에너지가 작은 조각으로 분할되기 때문에 효과도 감소하지만 최대 값이 변경되지 않았기 때문에 여전히 효과적입니다.

빔 진동 주파수 성분은 불안정한 단추 구멍 및 열쇠 구멍 붕괴를 수반하는 불안정한 용융 풀 동작과 관련이 있습니다. 언더컷 결함이있는 케이스 8 (발진 주파수 200Hz)에서 발진 주파수 성분이 관찰되었습니다. 이것은 특히 완전히 관통 된 열쇠 구멍과 불안정한 단추 구멍에서 불안정한 용융 풀 동작을 보여주었습니다. 경우 10 (진동 주파수 250Hz)의 경우 상대적으로 건강한 비드가 형성 되었으나, 도 11 (a) 와 같이 웨이블릿 변환 결과에서 t1의 시간 간격으로 진동 주파수 성분이 관찰되었다 . 이 시간 간격 t1의 용융 풀 거동은 그림 11에 나와 있습니다.(비). 완전히 관통 된 열쇠 구멍이 즉시 무너지는 것이 분명하게 관찰되었습니다. 이것은 진동 주파수 성분이 불안정한 용융 풀 거동과 밀접한 관련이 있음을 보여줍니다. 발견 된 주파수 성분으로부터 완전히 관통 된 열쇠 구멍과 같은 불안정한 용융 풀 거동을 예측할 수 있습니다. 완전히 관통 된 키홀이 반복적으로 붕괴되기 때문에 빔 진동 주파수 성분은 그림 9 (d) 와 같이 웨이블릿 변환 결과에서 명확한 실선 형태로 보이지 않습니다 .

그림 11

Cho and Woizeschke (2020)에 따르면 단추 구멍 형성은 자체 지속 가능한 카테 노이드처럼 작용하기 때문에 용융 풀 역학을 감소시킬 수 있습니다 [30] . 그림 12 는 버튼 홀 형성 측면에서 t2의 시간 간격에서 용융 풀 거동의 변화를 보여줍니다. 단추 구멍은 t2의 간헐적 인 부분에만 형성되었습니다. 1st 이후이 시간 동안 웨이블릿 변환의 결과로 주파수 성분이 사라졌고, 버튼 홀 형성은 용융 풀 역학을 줄이는 데 효과적이었습니다. 따라서, 웨이블릿 변환의 결과로 주파수 성분이 지워지는 것을 관찰함으로써 버튼 홀 형성을 예측할 수있다. 이와 관련하여 웨이블릿 변환 기술은 시간에 따른 용융 풀 변화를 나타낼 수 있습니다. 이 기술은 향후 용융 풀 동작을 모니터링하는 데 사용될 수 있습니다.

그림 12

5 . 결론

CFD 시뮬레이션 결과를 사용하여 빔 진동 및 필러 와이어 공급을 통한 레이저 용접에서 용융 풀 역학을 분석 할 수있었습니다. 용융 풀 표면의 변동 데이터의 푸리에 변환 및 웨이블릿 변환은 여기서 용융 풀 역학을 분석하는 데 사용되었습니다. 결과는 다음과 같은 결론으로 ​​이어집니다.1.

 주파수 (2 x 빔 발진 주파수, 이른바 초핑 주파수), 2  주파수 (4 x 빔 발진 주파수) 및 빔 발진 주파수 성분은 푸리에 변환 및 웨이블릿 변환 분석에서 발견 된 주요 성분이었습니다.2.

 주파수와 2  주파수 성분 의 출현은 두 가지 사건, 즉 레이저 빔에 의한 필러 와이어의 절단 공정과 집중된 레이저 라인 에너지의 효과의 결과였습니다. 이는 빔 진동주기 당 두 번 발생했습니다. 따라서 두 번째 주파수 성분은 동시 작용으로 인해 발생했습니다. 빔 진동 주파수 성분은 불안정한 용융 풀 동작과 관련이 있습니다. 구성 요소는 열쇠 구멍과 단추 구멍의 붕괴와 함께 나타났습니다.삼.

낮은 발진 주파수에서는 1  주파수와 2  주파수 성분이 함께 나타 났지만 발진 주파수가 증가함에 따라 그 크기가 함께 감소했습니다. 집중 선 에너지는 주파수가 증가함에 따라 최대 값이 변하지 않는 반면, 잘게 잘린 선단이 평평 해져 그 효과가 사라졌기 때문에 쵸핑 프로세스보다 더 큰 영향을 미쳤습니다.4.

용융 풀 거동의 빠른 시간적 변화는 웨이블릿 변환 방법을 사용하여 분석되었습니다. 따라서이 방법은 열쇠 구멍 및 단추 구멍의 형성 및 붕괴와 같은 일시적인 용융 풀 변화를 해석하는 데 사용할 수 있습니다.

CRediT 저자 기여 성명

조원익 : 개념화, 방법론, 소프트웨어, 검증, 형식 분석, 조사, 데이터 큐 레이션, 글쓰기-원고, 글쓰기-검토 및 편집. Peer Woizeschke : 감독, 프로젝트 관리, 작문-검토 및 편집.

경쟁 관심의 선언

저자는이 논문에보고 된 작업에 영향을 미칠 수있는 경쟁적인 재정적 이해 관계 나 개인적 관계가 없다고 선언합니다.

감사의 말

이 작업은 알루미늄 합금 용접 역량 센터 (Centr-Al)에서 수행되었습니다. Deutsche Forschungsgemeinschaft (DFG, 프로젝트 번호 290705638 , “용접 풀 캐비티를 생성하여 레이저 깊은 용입 용접에서 매끄러운 이음매 표면”) 의 자금은 감사하게도 인정됩니다.

부록 A . 사용 된 지배 방정식 및 레이저 용접 모델

1 . 지배 방정식 (Cho 및 Woizeschke [ 30 ])

-대량 보존 방정식,(A1)∇·V→=미디엄˙에스ρ어디, V→속도 벡터입니다. ρ밀도이고 미디엄˙에스필러 와이어를 공급하여 질량 소스의 비율입니다. 단위미디엄에스단위 부피당 질량입니다. WFS (와이어 공급 속도) 및 필러 와이어의 직경과 같은 매스 소스 및 필러 와이어 조건,디w계산 영역에서 다음과 같은 관계가 있습니다.(A2)미디엄=∫미디엄에스디V=미디엄0+씨×ρ×W에프에스×π디w24×티어디, 미디엄총 질량, 미디엄0초기 총 질량, V볼륨입니다.씨단위 변환 계수입니다. 티시간입니다.

-운동량 보존 방정식,(A3)∂V→∂티+V→·∇V→=−1ρ∇피+ν∇2V→−케이V→+미디엄˙에스ρ(V에스→−V→)+지어디, 피압력입니다. ν동적 점도입니다. 케이뭉툭한 영역의 다공성 매체 모델에 대한 항력 계수, V에스→질량 소스에 대한 속도 벡터입니다. 지신체 힘으로 인한 신체 가속도입니다.

-에너지 절약 방정식,(A4)∂h∂티+V→·∇h=1ρ∇·(케이∇티)+h˙에스어디, h특정 엔탈피입니다. 케이열전도율, 티온도이고 h˙에스특정 엔탈피 소스로, Eq 의 질량 소스와 연관됩니다 (A1) . 계산 영역의 총 에너지,이자형다음과 같이 계산됩니다.(A5)이자형=∫미디엄에스h에스디V=∫미디엄에스씨Vw티w디V어디, 씨Vw질량 원의 비열, 티w질량 소스의 온도입니다.

또한, 엔탈피 기반 연속체 모델을 사용하여 고체-액체 상 전이를 고려했습니다.

-VOF 방정식,(A6)∂에프∂티+∇·(V→에프)=에프˙에스어디, 에프유체가 차지하는 부피 분율이며 0과 1 사이의 값을 가지며 에프˙에스질량의 소스와 연결된 유체의 체적 분율의 비율 식. (A1) . 질량 공급원에 해당하는 부피 분율은 다음에 할당됩니다.에프에스.

-스칼라 보존 방정식,(A7)∂Φ∂티+∇·(V→Φ)=Φ˙에스어디, Φ필러 와이어의 스칼라 값입니다. 셀의 유체가 전적으로 필러 와이어로 구성된 경우Φ1이고 유체에 대한 필러 와이어의 부피 분율에 따라 0과 1 사이에서 변경됩니다. Φ˙에스Eq 에서 질량 소스에 연결된 스칼라 소스의 비율입니다 (A1) . 스칼라 소스는 전적으로 필러 와이어이기 때문에 1에 할당됩니다. 확산 효과는 고려되지 않았습니다.

2 . 레이저 용접 모델 (Cho et al. [26] )

흡수율을 계산하기 위해 프레 넬 반사 모델을 사용했습니다. ㅏ=1−ρ씨재료의 표면 상에 도시 된 바와 같이 수학 식. (A8) 원 편광 빔의 경우.(A8)ㅏ=1−ρ씨=1−12(ρ에스+ρ피)어디,ρ에스=(엔1씨영형에스θ−피)2+큐2(엔1씨영형에스θ+피)2+큐2,ρ에스=(피−엔1에스나는엔θ티ㅏ엔θ)2+큐2(피+엔1에스나는엔θ티ㅏ엔θ)2+큐2,피2=12{[엔22−케이22−(엔1에스나는엔θ)2]2+2엔22케이22+[엔22−케이22−(엔1에스나는엔θ)2]},큐2=12{[엔22−케이22−(엔1에스나는엔θ)2]2+2엔22케이22−[엔22−케이22−(엔1에스나는엔θ)2]}.어디, 복잡한 인덱스 엔1과 케이1반사 지수와 공기의 흡수 지수이며 엔2과 케이2공작물을위한 것입니다. θ입사각입니다. 도시 된 바와 같이 수학 식. (A9)에서 , 혼합물의 흡수율은 식에서 얻은 모재 및 필러 와이어 분획의 가중 평균이됩니다 . (A7) .(A9)ㅏ미디엄나는엑스티유아르 자형이자형=Φㅏw나는아르 자형이자형+(1−Φ)ㅏ비ㅏ에스이자형어디, ㅏ비ㅏ에스이자형과 ㅏw나는아르 자형이자형각각 비금속과 필러 와이어의 흡수율입니다.

자유 표면 경계에서의 반동 압력 에이 싱은 Eq. (A10) .(A10)피아르 자형(티)≅0.54피에스ㅏ티(티)=0.54피0이자형엑스피(엘V티−티비아르 자형¯티티비)어디, 피에스ㅏ티포화 압력, 피0대기압입니다. 엘V기화의 잠열, 티비끓는 온도이고 아르 자형¯보편적 인 기체 상수입니다.

참고 문헌

D.J. Kotecki, D.L. Cheever, D.G. Howden
Mechanism of ripple formation during weld solidification
Weld. J., 51 (8) (1972), pp. 386s-391s
Google Scholar
[2]
M. Zacksenhouse, D.E. Hardt
Weld pool impedance identification for size measurement and control
J. Dyn. Syst. Meas. Control, 105 (3) (1983), pp. 179-184
CrossRefView Record in ScopusGoogle Scholar
[3]
V.V. Semak, J.A. Hopkins, M.H. McCay, T.D. McCay
Melt pool dynamics during laser welding
J. Phys. D, 28 (1995), pp. 2443-2450
CrossRefView Record in ScopusGoogle Scholar
[4]
A.J.R. Aendenroomer, G. den Ouden
Weld pool oscillation as a tool for penetration sensing during pulsed GTA welding
Weld. J., 77 (5) (1998), pp. 181s-187s
Google Scholar
[5]
M.J.M. Hermans, G. den Ouden
Process behavior and stability in short circuit gas metal arc welding
Weld. J., 78 (4) (1999), pp. 137-141
View Record in ScopusGoogle Scholar
[6]
B.Y.B. Yudodibroto, M.J.M. Hermans, Y. Hirata, G. den Ouden
Influence of filler wire addition on weld pool oscillation during gas tungsten arc welding
Sci. Technol. Weld. Join., 9 (2) (2004), pp. 163-168
View Record in ScopusGoogle Scholar
[7]
M. Geiger, K.-H. Leitz, H. Koch, A. Otto
A 3D transient model of keyhole and melt pool dynamics in laser beam welding applied to the joining of zinc coated sheets
Prod. Eng. Res. Dev., 3 (2009), pp. 127-136
CrossRefView Record in ScopusGoogle Scholar
[8]
C. Kägeler, M. Schmidt
Frequency-based analysis of weld pool dynamics and keyhole oscillations at laser beam welding of galvanized steel sheets
Phys. Procedia, 5 (2010), pp. 447-453
ArticleDownload PDFView Record in ScopusGoogle Scholar
[9]
Y. Shi, G. Zhang, X.J. Ma, Y.F. Gu, J.K. Huang, D. Fan
Laser-vision-based measurement and analysis of weld pool oscillation frequency in GTAW-P
Weld. J., 94 (2015), pp. 176s-187s
Google Scholar
[10]
J. Volpp, F. Vollertsen
Keyhole stability during laser welding—part I: modelling and evaluation
Prod. Eng.-Res. Dev., 10 (2016), pp. 443-457
CrossRefView Record in ScopusGoogle Scholar
[11]
N. Postacioglu, P. Kapadia, J. Dowden
Capillary waves on the weld pool in penetration welding with a laser
J. Phys. D, 22 (1989), pp. 1050-1061
CrossRefView Record in ScopusGoogle Scholar
[12]
N. Postacioglu, P. Kapadia, J. Dowden
Theory of the oscillations of an ellipsoidal weld pool in laser welding
J. Phys. D, 24 (1991), pp. 1288-1292
CrossRefView Record in ScopusGoogle Scholar
[13]
J. Kroos, U. Gratzke, M. Vicanek, G. Simon
Dynamic behaviour of the keyhole in laser welding
J. Phys. D, 26 (1993), pp. 481-486
View Record in ScopusGoogle Scholar
[14]
H. Maruo, Y. Hirata
Natural frequency and oscillation modes of weld pools. 1st Report: weld pool oscillation in full penetration welding of thin plate
Weld. Int., 7 (8) (1993), pp. 614-619
CrossRefView Record in ScopusGoogle Scholar
[15]
T. Klein, M. Vicanek, J. Kroos, I. Decker, G. Simon
Oscillations of the keyhole in penetration laser beam welding
J. Phys. D, 27 (1994), pp. 2023-2030
CrossRefView Record in ScopusGoogle Scholar
[16]
T. Klein, M. Vicanek, G. Simon
Forced oscillations of the keyhole in penetration laser beam welding
J. Phys. D, 29 (1996), pp. 322-332
View Record in ScopusGoogle Scholar
[17]
K. Andersen, G.E. Cook, R.J. Barnett, A.M. Strauss
Synchronous weld pool oscillation for monitoring and control
IEEE Trans. Ind. Appl., 33 (2) (1997), pp. 464-471
View Record in ScopusGoogle Scholar
[18]
W.-I. Cho, S.-J. Na, M.-H. Cho, J.-S. Lee
Numerical study of alloying element distribution in CO2 laser-GMA hybrid welding
Comput. Mater. Sci., 49 (2010), pp. 792-800
ArticleDownload PDFView Record in ScopusGoogle Scholar
[19]
W.-I. Cho, S.-J. Na, C. Thomy, F. Vollertsen
Numerical simulation of molten pool dynamics in high power disk laser welding
J. Mater. Process. Technol., 212 (2012), pp. 262-275
ArticleDownload PDFView Record in ScopusGoogle Scholar
[20]
A. Otto, A. Patschger, M. Seiler
Numerical and experimental investigations of humping phenomena in laser micro welding
Phys. Procedia, 83 (2016), pp. 1415-1423
ArticleDownload PDFView Record in ScopusGoogle Scholar
[21]
R. Lin, H.-P. Wang, F. Lu, J. Solomon, B.E. Carlson
Numerical study of keyhole dynamics and keyhole-induced porosity formation in remote laser welding of Al alloys
Int. J. Heat Mass Trans., 108 (2017), pp. 244-256
ArticleDownload PDFView Record in ScopusGoogle Scholar
[22]
S.H. Ko, C.D. Yoo, D.F. Farson, S.K. Choi
Mathematical modeling of the dynamic behavior of gas tungsten arc weld pools
Metall. Mater. Trans. B., 31B (2000), pp. 1465-1473
CrossRefView Record in ScopusGoogle Scholar
[23]
R. Hu, X. Chen, G. Yang, S. Gong, S. Pang
Metal transfer in wire feeding-based electron beam 3D printing: modes, dynamics, and transition criterion
Int.