액체-증기 상 변화 모델은 밀폐된 용기의 자체 가압 프로세스 시뮬레이션에 매우 큰 영향을 미칩니다. Hertz-Knudsen 관계, 에너지 점프 모델 및 그 파생물과 같은 널리 사용되는 액체-증기 상 변화 모델은 실온 유체를 기반으로 개발되었습니다. 액체-증기 전이를 통한 극저온 시뮬레이션에 널리 적용되었지만 각 모델의 성능은 극저온 조건에서 명시적으로 조사 및 비교되지 않았습니다. 본 연구에서는 171가지 일반적인 액체-증기 상 변화 모델을 통합한 통합 다상 솔버가 제안되었으며, 이를 통해 이러한 모델을 실험 데이터와 직접 비교할 수 있습니다. 증발 및 응축 모델의 예측 정확도와 계산 속도를 평가하기 위해 총 <>개의 자체 가압 시뮬레이션이 수행되었습니다. 압력 예측은 최적화 전략이 서로 다른 모델 계수에 크게 의존하는 것으로 나타났습니다. 에너지 점프 모델은 극저온 자체 가압 시뮬레이션에 적합하지 않은 것으로 나타났습니다. 평균 편차와 CPU 소비량에 따르면 Lee 모델과 Tanasawa 모델은 다른 모델보다 안정적이고 효율적인 것으로 입증되었습니다.
Volume 181, December 2021, 121879
Validity evaluation of popular liquid-vapor phase change models for cryogenic self-pressurization process
Author links open overlay panelZhongqi Zuo, Jingyi Wu, Yonghua HuangShow moreAdd to MendeleyShareCite
Liquid-vapor phase change models vitally influence the simulation of self-pressurization processes in closed containers. Popular liquid-vapor phase change models, such as the Hertz-Knudsen relation, energy jump model, and their derivations were developed based on room-temperature fluids. Although they had widely been applied in cryogenic simulations with liquid-vapor transitions, the performance of each model was not explicitly investigated and compared yet under cryogenic conditions. A unified multi-phase solver incorporating four typical liquid-vapor phase change models has been proposed in the present study, which enables direct comparison among those models against experimental data. A total number of 171 self-pressurization simulations were conducted to evaluate the evaporation and condensation models’ prediction accuracy and calculation speed. It was found that the pressure prediction highly depended on the model coefficients, whose optimization strategies differed from each other. The energy jump model was found inadequate for cryogenic self-pressurization simulations. According to the average deviation and CPU consumption, the Lee model and the Tanasawa model were proven to be more stable and more efficient than the others.
The liquid-vapor phase change of cryogenic fluids is widely involved in industrial applications, such as the hydrogen transport vehicles , shipborne liquid natural gas (LNG) containers  and on-orbit cryogenic propellant tanks . These applications require cryogenic fluids to be stored for weeks to months. Although high-performance insulation measures are adopted, heat inevitably enters the tank via radiation and conduction. The self-pressurization in the tank induced by the heat leakage eventually causes the venting loss of the cryogenic fluids and threatens the safety of the craft in long-term missions. To reduce the boil-off loss and extend the cryogenic storage duration, a more comprehensive understanding of the self-pressurization mechanism is needed.
Due to the difficulties and limitations in implementing cryogenic experiments, numerical modeling is a convenient and powerful way to study the self-pressurization process of cryogenic fluids. However, how the phase change models influence the mass and heat transfer under cryogenic conditions is still unsettled . As concluded by Persad and Ward , a seemingly slight variation in the liquid-vapor phase change models can lead to erroneous predictions.
Among the liquid-vapor phase change models, the kinetic theory gas (KTG) based models and the energy jump model are the most popular ones used in recent self-pressurization simulations . The KTG based models, also known as the Hertz-Knudsen relation models, were developed on the concept of the Maxwell-Boltzmann distribution of the gas molecular . The Hertz-Knudsen relation has evolved to several models, including the Schrage model , the Tanasawa model , the Lee model  and the statistical rate theory (SRT) , which will be described in Section 2.2. Since the Schrage model and the Lee model are embedded and configured as the default ones in the commercial CFD solvers Flow-3D® and Ansys Fluent® respectively, they have been widely used in self-pressurization simulations for liquid nitrogen ,  and liquid hydrogen , . The major drawback of the KTG models lies in the difficulty of selecting model coefficients, which were reported in a considerably wide range spanning three magnitudes even for the same working fluid , , , , , . Studies showed that the liquid level, pressure and mass transfer rate are directly influenced by the model coefficients , , , , . Wrong coefficients will lead to deviation or even divergence of the results. The energy jump model is also known as the thermal limitation model. It assumes that the evaporation and condensation at the liquid-vapor interface are induced only by heat conduction. The model is widely adopted in lumped node simulations due to its simplicity , , . To improve the accuracy of mass flux prediction, the energy jump model was modified by including the convection heat transfer , . However, the convection correlations are empirical and developed mainly for room-temperature fluids. Whether the correlation itself can be precisely applied in cryogenic simulations still needs further investigation.
Fig. 1 summarizes the cryogenic simulations involving the modeling of evaporation and condensation processes in recent years. The publication has been increasing rapidly. However, the characteristics of each evaporation and condensation model are not explicitly revealed when simulating self-pressurization. A comparative study of the phase change models is highly needed for cryogenic fluids for a better simulation of the self-pressurization processes.
In the present paper, a unified multi-phase solver incorporating four typical liquid-vapor phase change models, namely the Tanasawa model, the Lee model, the energy jump model, and the modified energy jump model has been proposed, which enables direct comparison among different models. The models are used to simulate the pressure and temperature evolutions in an experimental liquid nitrogen tank in normal gravity, which helps to evaluate themselves in the aspects of accuracy, calculation speed and robustness.
Governing equations for the self-pressurization tank
In the present study, both the fluid domain and the solid wall of the tank are modeled and discretized. The heat transportation at the solid boundaries is considered to be irrelevant with the nearby fluid velocity. Consequently, two sets of the solid and the fluid governing equations can be decoupled and solved separately. The pressures in the cryogenic container are usually from 100 kPa to 300 kPa. Under these conditions, the Knudsen number is far smaller than 0.01, and the fluids are
Self-pressurization results and phase change model comparison
This section compares the simulation results by different phase change models. Section 3.1 compares the pressure and temperature outputs from two KTG based models, namely the Lee model and the Tanasawa model. Section 3.2 presents the pressure predictions from the energy transport models, namely the energy jump model and the modified energy jump model, and compares pressure prediction performances between the KTG based models and the energy transport models. Section 3.3 evaluates the four models
A unified vapor-liquid-solid multi-phase numerical solver has been accomplished for the self pressurization simulation in cryogenic containers. Compared to the early fluid-only solver, the temperature prediction in the vicinity of the tank wall improves significantly. Four liquid-vapor phase change models were integrated into the solver, which enables fair and effective comparison for performances between each other. The pressure and temperature prediction accuracies, and the calculation speed
CRediT authorship contribution statement
Zhongqi Zuo: Data curation, Formal analysis, Writing – original draft, Validation. Jingyi Wu: Conceptualization, Writing – review & editing, Validation. Yonghua Huang: Conceptualization, Formal analysis, Writing – review & editing, Validation.
Declaration of Competing Interest
Authors declare that they have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled, “Validity evaluation of popular liquid-vapor phase change models for cryogenic self-pressurization process”.
This project is supported by the National Natural Science Foundation of China (No. 51936006).
- D. Mori et al.Recent challenges of hydrogen storage technologies for fuel cell vehiclesInt. J. Hydrogen Energy(2009)
- W. Zheng et al.China’s first-phase mars exploration program: Yinghuo-1 orbiterPlanet Space Sci.(2013)
- Y. Tang et al.Experimental study on film condensation characteristics at liquid nitrogen temperaturesAppl. Therm. Eng.(2017)
- C.R. Kharangate et al.Review of computational studies on boiling and condensationInt. J. Heat Mass Transf.(2017)
- I. TanasawaAdvances in condensation heat transferAdvances in Heat Transfer(1991)
- S. Zhu et al.Numerical analysis of nitrogen condensation heat transfer enhancement with liquid film fluctuation at cryogenic temperatureInt. J. Heat Mass Transf.(2020)
- Z.Q. Zuo et al.Effect of baffles on pressurization and thermal stratification in cryogenic tanks under micro-gravityCryogenics(2018)
- Z. Zuo et al.Numerical investigation on full thermodynamic venting process of liquid hydrogen in an on-orbit storage tankInt. J. Hydrogen Energy(2020)
- M. Kassemi et al.Validation of two-phase CFD models for propellant tank self-pressurization: crossing fluid types, scales, and gravity levelsCryogenics(2018)
- R. Zhou et al.Simulations on effects of rated ullage pressure on the evaporation rate of liquid hydrogen tankInt. J. Heat Mass Transf.(2019)
- M. Kassemi et al.Effect of interfacial turbulence and accommodation coefficient on CFD predictions of pressurization and pressure control in cryogenic storage tankCryogenics(2016)
- J. Li et al.Simulation of mass and heat transfer in liquid hydrogen tanks during pressurizingChin. J. Aeronaut.(2019)
- A.J. Savas et al.Thermal analysis of a cryogenic liquid acquisition device under autogenous and non-condensable pressurization schemesInt. J. Heat Mass Transf.(2014)
- C. Ludwig et al.Pressure variations in a cryogenic liquid storage tank subjected to periodic excitationsInt. J. Heat Mass Transf.(2013)
- J.J. Shuang et al.Efficiency analysis of depressurization process and pressure control strategies for liquid hydrogen storage system in microgravityInt. J. Hydrogen Energy(2019)
- K. Bellur et al.A new experiment for investigating evaporation and condensation of cryogenic propellantsCryogenics(2016)
- Z. Liu et al.Thermal physical process in a liquid oxygen tank under different sloshing excitationsInt. Commun. Heat Mass Transf.(2020)
- S. Wu et al.Numerical simulation and experiment verification of the static boil-off rate and temperature field for a new independent type B liquefied natural gas ship mock up tankAppl. Therm. Eng.(2020)
- J. Zheng et al.Thermodynamic modelling and optimization of self-evaporation vapor cooled shield for liquid hydrogen storage tankEnergy Convers. Manage.(2019)
- L. Wang et al.Comparison of three computational models for predicting pressurization characteristics of cryogenic tank during dischargeCryogenics(2015)
There are more references available in the full text version of this article.
Cited by (7)
- Theoretical investigation on heat leakage distribution between vapor and liquid in liquid hydrogen tanks2023, International Journal of Hydrogen EnergyShow abstract
- A review of cryogenic quasi-steady liquid-vapor phase change: Theories, models, and state-of-the-art applications2023, International Journal of Heat and Mass TransferShow abstract
- Progress in physical modelling and numerical simulation of phase transitions in cryogenic pool boiling and cavitation2023, Applied Mathematical ModellingCitation Excerpt :We will not delve into cryogenic evaporation phenomena, that predominantly drive the phase-change in well-insulated storage facilities, and thus are less relevant to spill scenarios. We instead refer the reader to the works of Zuo et al. [31–33]. If the static pressure at any location in a turbomachine drops below a fluid’s saturation pressure, localized evaporation events may occur, followed by rapid collapse of the vapour cavities in a process termed “cavitation” .Show abstract
- Thermodynamic performance in a liquid oxygen tank during active-pressurization under different gas injection temperatures2023, International Communications in Heat and Mass TransferCitation Excerpt :The volume of fluid method is adopted to predict the tank pressurization performance. The associated governing equations could refer to previous published investigations [33–39,41,45,46]. Subjected to external heat input and gas injection, the phase change occurs at the interface and within the tank.Show abstract
- Interfacial mass and energy transport during steady-state evaporation in liquid oxygen storage tanks2022, Applied EnergyCitation Excerpt :However, most of them simply used the Lee model for mass transport as did for regular fluids, and seldom focus on the evaporation itself related to the interfacial temperature distribution or were unable to validate their results against credible experimental data. A recent study proposed an optimized evaporation model for the cryogenic self-pressurization with a thorough comparison between popular phase change models , but still lacked of experimental data to validate the results. A series of experiments have been conducted on the heat and mass transport in a thin liquid layer in the vicinity of the liquid–vapor interface of room-temperature fluids [9–14].Show abstract
- Thermal destratification of cryogenic liquid storage tanks by continuous bubbling of gases2022, International Journal of Hydrogen EnergyCitation Excerpt :It was concluded that a single injector with a larger diameter configuration showed a higher chance of developing a vertical temperature gradient. Zuo et al.  carried out a numerical analysis to investigate the temperature distribution within the LH2 storage tank with a self-pinning spraying bar. They used the SST turbulence model coupled with the 6-DOF model.Show abstract