Some Current Challenges in Calculating Engine Flows

Peter J. O'Rourke
Theoretical Fluid Dynamics Division
Los Alamos National Laboratory
Los Alamos, New Mexico

Computer codes being developed to effectively utilize massively parallel architectures were described. Numerical and sub-model improvements are being incorporated that have broad applicability both to reactive flows in engines and other combustion devices and also to non-reactive flows. Examples of some recent numerical turbulence modeling improvements were given. The goal of this work is the development of three-dimensional computer calculations for use as an adjunct to experiments in the design of internal combustion engines and improvement in the precision of their descriptions of in-cylinder flows. The increased computational power of massively parallel computers offers the promise of improved accuracy by permitting both increased numerical resolution and the use of more refined sub-models.