Reducing pollutant emissions from IC engines
Over the years, computational fluid dynamics (CFD) has become an established tool for the design and optimisation of internal combustion engines. Advanced numerical models have proven their value in drawing up a detailed description of ignition timing and heat release required to capture the pathways of nitrogen oxide (NOx) formation. This numerical approach was particularly attractive for the industrial partners of the MINNOX project, the benefit of which would have been a drastically reduced need for time-consuming experimental research. Major names in the European automotive industry joined their expertise to develop an advanced model that accounts for all important physical processes occurring in IC engine cylinders. In commercial software packages for computational fluid dynamics simulations, turbulence models are based on a number of assumptions and therefore are effective for certain types of flows. However, resolving the turbulence effects close to the walls of the IC engine is of crucial importance for accurately predicting the characteristics of a wall-bound turbulent flow. Research work at the laboratories of DaimlerChrysler AG in Germany aimed to improve the KIVA-3 and STAR-CD computer programmes for numerical calculations of three-dimensional chemically reactive fluid flows. More specifically, a unified and generalised treatment of wall boundary conditions was introduced to achieve a higher level of accuracy in wall friction predictions. Without extremely dense computational grids and excessive computational resources usually required, Reynolds-averaged Navier-Stokes (RANS) calculations and Large eddy simulations (LENS) could be extended up to the cylinder wall. Promising high flexibility along with simplified computations, the KIVA-3 and STAR-CD code, implementing the proposed method, have been used for simulating steady and pulsating fluid flows in a one-cylinder engine. Computations under a wide range of operating conditions were all in good agreement with Direct numerical simulations (DNS) as well as with data collected at the experimental setup developed during the course of the project.
