Efficient Computational Tool for Comprehensive Thermal Analysis of Military Ground Vehicles

Including radiative heat transfer in full-vehicle thermal analysis simulations results in nonlinear equations with coefficient matrices that are dense, poorly conditioned, and difficult to solve. Further, calculating and storing view factors¯needed for surface-to-surface radiative fluxes¯is computationally prohibitive on fine grids. In the proposed SBIR project, CFDRC will develop, validate and demonstrate an efficient, high-fidelity computational tool for full-vehicle thermal analysis. The overall innovation (Phase I and II) consists of four main aspects: (1) Develop a robust, efficient and highly linear system solver by coupling a Krylov sub-space solver and algebraic multigrid (AMG) preconditioner (2) Investigate influence of linear system implicitness level on the CPU requirements, (3) Develop computationally efficient view-factor and ray tracing calculation algorithms, and (4) Implement a comprehensive thermal model including the radiation solver coupled with a thin shell conduction model. In Phase I, feasibility will be demonstrated by: (1) Determining the optimal combination of solver and preconditioner parameters, (2) Incorporating a thin shell conduction model in an existing radiation code, and (3) Demonstrating the thermal analysis model on a realistic vehicle geometry. In Phase II, we will develop efficient algorithms for view factor and infrared ray tracing calculations. The comprehensive thermal model will be incorporated into a code of interest to the Army, and applied to 3-D thermal analysis of a military ground vehicle.