Numerical Solution of Multi-Dimensional Compressible Reactive Flow Using a Parallel Wavelet Adaptive Multi-Resolution Method

The aim of this research is to further develop a dynamically adaptive algorithm based on wavelets that is able to solve efficiently multi-dimensional compressible reactive flow problems.

This work demonstrates the great potential for the method to perform direct numerical simulation (DNS) of combustion with detailed chemistry and multi-component diffusion. In particular, it addresses the performance obtained using a massive parallel implementation and demonstrates important savings in memory storage and computational time over conventional methods. In addition, fully-resolved simulations of challenging three dimensional problems involving mixing and combustion processes are performed. These problems are particularly challenging due to their strong multiscale characteristics. For these solutions, it is necessary to combine the advanced numerical techniques applied to modern computational resources.