Development of an Adaptive Mesh CFD Code for High Explosive Blast Simulation

Timothy A. Rose, Peter D. Smith & Shaun A. Forth

In  
Proceedings of the 12th International Symposium on Interaction of the Effects of Munitions with Structures, New Orleans, Louisiana, USA, 2005.

Abstract 
Three-dimensional  Computational Fluid Dynamics (CFD) calculations are extremely computationally expensive, and, until recently, have been the sole preserve of large computing laboratories.  With increasing computational resource available to the engineer on his desktop, however, the possibility of performing well-resolved numerical simulations, routinely as part of building design or hazard assessment, is increasing.

Because blast waves propagate from one place to another, numerical simulations usually require large computational domains. Similarly, the necessity of siting non-reflecting boundaries sufficiently far from the region of interest - so they do not influence the more important regions - also demands a large domain.  This, combined with the necessity of using small cells to obtain well-resolved solutions, implies that significant computational resources are required to treat such problems adequately.

This requirement can be alleviated, to some extent, by the use of computational domains which refine automatically as the blast waves pass through.  Theoretically mesh refinement allows good resolution of pressure and density discontinuities while allowing large regions of smooth flow to be adequately resolved on a much coarser mesh.

Unfortunately, considerable programming effort is required to identify, refine and subsequently unrefine regions in which important flow features reside, and there is also a significant computational overhead - in terms of memory and processing time - associated with these activities.

This paper describes the development of one adaptive mesh CFD code, ftt_air3d, and gives examples of its use.

Download
PDF: tar_isiems05.pdf

AMOR home