High-order computation of burning propellant surface and simulation of fluid flow in solid rocket chamber

Abstract

In this paper, we present a numerical approach for predicting fluid flows in solid rocket motor (SRM) chambers. We use a novel high-order technique to track the burning grain surface. Spectral convergence toward the exact burning surface is achieved thanks to Fourier differentiation. In addition, we make use of a body-fitted mesh deforming with the burning surface and present a method to avoid manual remeshing. We describe several methods to deform the volume mesh and to keep good mesh element quality during the computation. We then couple the surface and volume approaches. The resulting coupled method is able to handle the formation of geometric singularities on the burning surface while keeping constant surface and volume mesh topology. This geometrical approach is integrated into a complex code for compressible, multi-species, turbulent flow simulations. Applications to the simulation of the internal flow in realistic solid rocket motors with complex grain geometry are then presented.

Simple Surface-Tracking Methods for Grain Burnback Analysis

“…Usually the marker methods are harder to implement, especially for the surface tracking of merging and splitting fronts in complicated three-dimensional domains. A very recent and much more accurate front-tracking method for internal ballistics is presented in [13]. A spectral method is proposed to accurately represent the burning front.…”

“…Based on the previous methodologies, many researchers in academies or rocket companies in the United States and several European countries have developed sophisticated internal ballistics codes to deal with fluid-structure interactions with moving grain boundary [13][14][15][16][17][18]. A substantial use of the level set and its related methods has been observed in these sophisticated internal ballistics codes.…”

Reverse Design of Solid Propellant Grain for a Performance-Matching Goal: Shape Optimization via Evolutionary Neural Network

“…Non-meshing methods include the analytical method [15,16], general coordinate method [17,18], and CAD modeling method [19,20]. Meshing methods include the minimum distance method [21,22], moving grid method [23,24], level-set method [9,25,26], and fast marching method [27]. On the one hand, the non-meshing methods are accurate, computationally efficient, and reliable.…”