Very high order finite volume solver for multi component two-phase flow with phase change using a posteriori Multi-dimensional Optimal Order Detection

In this work we propose a very high-order compressible finite volume scheme with a posteriori stabilization for the computation of multi-component two-phase flow with phase change. It is based on finite volume approach using moving least squares (MLS) reproducing kernels for high order reconstruction of the Riemann states. Increased robustness is achieved by using the multi-dimensional optimal order detection (MOOD) method to get a high-accurate and low-dissipation scheme while maintaining boundedness and preventing numerical oscillations at interfaces and strong gradient zones. The properties of the proposed framework are demonstrated on classical test problems starting with convergence order verification on simple scalar advection test cases. More complex shock and more stringent tube tests with various water, steam and air concentration are then simulated and compared with available references in the literature. Finally, the ability of the proposed approach to compute multi-component flows with phase change is illustrated with the simulation of a liquid oxygen jet in gaseous hydrogen.