Turbulence modeling is an essential aspect of computational fluid dynamics (CFD) for applications of interest in the automotive industry. In this paper, turbulence models are reviewed in the context of in-cylinder flows in reciprocating engines. While the ubiquitous k - e remains the model of choice in most applications, higherorderstatistical closures and large-eddy simulation have been brought to bear in an effort to increase the predictive capability of multidimensional simulations for these complex flows. The performance of several models in simple engine-like configurations is reviewed and applications to flows of more practical interest are summarized. An assessment of the demonstrated and/or anticipated performance of different modeling approaches is given and suggestions of fruitful paths for further improvement are offered. The nature of turbulence in engines and numerical accuracy are two issues that are identified as being critical in multidimensional modeling of turbulent in-cylinder flows. Probability density function methods and large-eddy simulation are proposed as candidates for next-generation "turbulence models" that have the potential for significant improvements over today's k - £ based approaches.
All Science Journal Classification (ASJC) codes
- Aerospace Engineering
- Fuel Technology
- Mechanical Engineering
- Space and Planetary Science