A positivity preserving high‐order finite difference method for compressible two‐fluid flows

Abstract

Any robust computational scheme for compressible flows must retain the hyperbolicity property or the real‐valued sound speed. Failure to maintain hyperbolicity, or the positivity of the square of the speed of sound, causes nonphysical distortions and the blow‐up of numerical simulations. Strong shock waves and interfacial discontinuities are ubiquitous features of the two‐fluid compressible dynamics that can potentially induce positivity‐related failure in a simulation. This article presents a positivity‐preserving algorithm for a high‐order, primitive variable‐based, weighted essentially non‐oscillatory finite difference scheme. The positivity preservation relies on a flux limiting technique that locally adapts high‐order fluxes towards the first order to retain the physical bounds of the solution without loss of high‐order convergence. This positivity preserving scheme has been devised and implemented up to eleventh order in one and two dimensions for a two‐fluid compressible model that consists of a single mass, momentum, and energy equations, as well as an advection of material parameters for capturing the interfaces. Several one‐ and two‐dimensional challenging test problems verify the performance. The scheme effectively retains high order accuracy while allowing for the simulation of several challenging problems that otherwise could not be successfully solved using the base scheme, without any penalty on the CFL condition requirement, and without any significant impact on the CPU times. The scheme represents the first high‐order (up to 11‐order) hyperbolicity‐preserving scheme for the considered two‐fluid compressible flows in the fully Eulerian formulation. The inherent efficiency of finite differences and the new robust positivity preserving quality enable modeling other challenging problems of two‐fluid and two‐phase problems.