Numerical solution procedure for Hall electric field of the hypersonic magnetohydrodynamic heat shield system

Abstract

Magnetohydrodynamic (MHD) heat shield system is a novel-concept thermal protection technique for hypersonic vehicles, which has been proved by lots of researchers with both numerical and experimental methods. Most of researchers neglect the Hall effect in their researches. However, in the hypersonic reentry process, the Hall effect is sometimes so significant that the electric current distribution in the shock layer can be changed by the induced electric field. Consequently, the Lorentz force as well as the Joule heat is varied, and thus the efficiency of the MHD heat shield system is affected.In order to analyze the influence of Hall effect, the induced electric field must be taken into consideration. According to the weakly-ionized characteristics of hypersonic flow post bow shock, the magneto-Reynolds number is assumed to be small. Therefore, the Maxwell equations are simplified with the generalized Ohm's law, and the induced electric field is governed by the potential Possion equation. Numerical methods are hence established to solve the Hall electric field equations in the thermochemical nonequilibrium flow field. The electric potential Poisson equation is of significant rigidity and difficult to solve for two reasons. One is that the coefficient matrix may not be diagonally dominant when the Hall parameter is large in the shock layer, and the other is that this matrix including the electric conductivity is discontinuous across the shock. In this paper, a virtual stepping factor is included to strengthen the diagonal dominance and improve the computational stability. Moreover, approximate factor and alternating direction implicit method are employed for further improving the stability. With these methods, a FORTRAN code is written and validated by comparing the numerical results with the analytical ones as well as results available from previous references. After that, relation between the convergence property and the virtual stepping factor is revealed by theoretical analysis and numerical simulations. Based on these work, a local variable stepping factor method is proposed to accelerate the iterating process. Results show that the convergence property is closely related to the mesh density and Hall parameter, and there exists a best stepping factor for a particular mesh as well as a particular Hall parameter. Since the best stepping factor varies a lot for different meshes and different Hall parameter, its appropriate value is hard to choose. The best value of stepping factor coefficient still exists in the local step factor method, but its value range is relatively smaller. More importantly, the local stepping factor method yields better convergence property than the regular constant one when employing a locally refined mesh.