高速跨介质航行器是一种具有前景新型海洋装备。它的工作环境较为复杂，涉及空化相变、空化湍流、多相流以及各相流体可压缩性。这四方面因素中仍然存留着前沿科学问题有待解决，但是作为重要研究手段的数值模拟技术受制于各大商业与开源软件未能提供良好求解器的现状。因此，开发用户自定义的求解器将为高速跨介质航行器相关科学问题的研究提供重要支撑工具。为此本文参考了开源软件OpenFOAM中多个现成求解器的实现方式，在其框架下开发了一款针对非定常空化流动的多相可压缩求解器PBCavitatingFoam。它采用了可压缩形式的压力-速度耦合算法，包含有多种空化模型，能够进行大涡模拟。该求解器首先完成了适用性验证，并在接下来被应用于空泡脱落机制以及自由面-空化相互作用这两个前沿问题的研究中，主要包含以下几方面工作：
（1）通过三个简单的基本算例对求解器各方面的功能进行验证。其中，以一维激波管算例对求解器基于压力的可压缩性算法进行验证，以液体中单气泡的脉动算例对求解器的多相流模拟功能进行验证，以圆柱空化绕流算例对求解器的空化相变模拟功能进行验证，并且以弹体入水算例验证了求解器综合模拟多种效应的能力。在以上的验证算例中，求解器均能够与精确解或文献中的数据较好地吻合，因此证实了求解器具备良好的适用性。
（2）非定常空化流动中是否存在与可压缩性相关的空泡脱落机制是当前的一个前沿问题。为此，本文以锥头回转体空化绕流为例，使用PBCavitatingFoam进行数值模拟，并通过基于霍普金森杆与高速摄影的实验开展对比分析。对比结果表明，数值模拟能够正确预测空泡形态随时间的演化过程，并给出无量纲空泡长度-无量纲时间图线的趋势，同时区分出空泡演化的四个阶段。回射流与压力波导致的空泡脱落机制出现在不同的阶段，两者的主要区别在于空泡后缘与回射流前缘的空间关系。通过对比不可压求解器的模拟结果，这种区别受到了进一步分析，由此证实了PBCavitatingFoam能够模拟空化流动中压力波主导的空泡脱落。此外还讨论了不可凝结气体的作用以及二维轴对称化简的可行性。
（3）针对同时包含多相流与空化相变的NACA 0015水翼近自由面空化绕流问题，首先以量纲分析法推导了包括弗劳德数、攻角、浸深比以及空化数的主控无量纲数。其次，通过深水工况开展了网格无关性验证以及与实验数据的对比，说明了数值计算的合理性。接下来阐述了四种自由面-空泡流态，包括伴随空泡体积增大的自由面上升，脱落空泡上浮溃灭导致的自由面射流，伴随高频涡流扰动的自由面波纹，不伴随空化的液面上扬。最后以弗劳德数与浸深比为主，讨论了主控无量纲数对升力系数、阻力系数、升阻比以及频率的影响，发现这些参数对主控无量纲数变化的敏感程度不同，并且空化数的变化能够使参数的变化规律发生本质的改变。

The high-speed cross-medium marine vehicle is promising as a new kind of marine engineering equipment. Its working environment is complex as involving mainly four factors including cavitation phase change, turbulent cavitating flow and compressibility of each phase, which contain scientific problems need further explaination. However, corresponding numerical simulation, as one of the most important research methods, is limited by lack of available solvers in either commercial or open source softs. Therefore, development of a specific solver can provide an essential tool for investigating scientific problems for the high-speed cross-medium marine vehicle. And this thesis established a multiphase compressible PBCavitatingFoam solver for cavitating flow in the framework of open source OpenFOAM according to implementation of solvers it provides. This solver is pressure-based, possessing frequently-used cavitation models and being able to perform Large Eddy Simulation. Usability of this solver is first verified and then frontier problems are investigated, including shedding mechanism and interaction between free surface and cavitating flow. These works are as follow:
(1) The new solver first goes through three simple tests to demonstrate its usability, including a series of shock tube cases to test its pressure-velocity coupling algorithm performance in compressible flow, a subaqueous bubble oscillation case to test its usability on multiphase flow, a case about cavitating flow around a cylinder to test its usability on cavitation, and a case about water entry to test its usability on multiple effect problems. Results predicted by the solver agree well with analytical solution, literature or experiments, which demonstrates its usability.
(2) It is frontier problem whether compressibility is essential for shedding mechanism in unsteady cavitating flow. Thus, cavitating flow around an axisymmetric projectile is simulated by PBCavitatingFoam, and comparative study is carried out with experiments based on SHPB technology and high-speed photography. Results shows that evolution of cavity shape is correctly predicted. The trend on diagram of dimensionless cavity length vs. dimensionless is also reasonable, and four stages is marked out. Shedding dominated by re-entrant jet and pressure waves are observed in different stages, and these two situations mainly differs in spatial relationship between cavity closure and re-entrant jet front. By comparing results from an incompressible solver, PBCavitatingFoam’s ability to simulate pressure-wave-induced shedding mechanism is verified. In addition, effect of non-condensable gas is investigated and feasibility of using 2D-axisymmetric mesh as simplification is confirmed.
(3) To investigate cavitating flow around 2-D NACA 0015 hydrofoil near free surface, important dimensionless numbers, including Froude number, angles of attack, immersion ratio and cavitation number, are first deduced with dimensional analysis. Comparison with experiment data and grid convergence study is performed with submerging case. Then four types of free surface-cavity flow patterns are illustrated, including free surface bulge caused by increasing cavity volume, jet on free surface caused by collapse of lifting shedding cavity, ripple induced by high frequency vortex-shedding and continuous rising liquid phase without cavitation. Finally, the influence of dimensionless numbers on parameters, including lift coefficient, drag coefficient, lift-drag ratio and frequency ratio, is investigated with focus on Froude number and immersion ratio. It is found that sensitivity to variation of dimensionless numbers is different for each parameter, and distribution of parameters changes substantially with variation of cavitation number.