Numerical Optimization of the Combustion System of a HD Compression Ignition Engine Fueled with DME Considering Current and Future Emission Standards

A genetic algorithm (GA) optimization methodology is applied to the design of the combustion system of a heavy-duty (HD) Diesel engine fueled with dimethyl ether (DME). The study has two objectives, the optimization of a conventional diffusion-controlled combustion system aiming to achieve US2010 targets and the optimization of a stoichiometric combustion system coupled with a three way catalyst (TWC) to further control NOx emissions and achieve US2030 emission standards. These optimizations include the key combustion system related hardware, bowl geometry and injection nozzle design as input factors, together with the most relevant air management and injection settings. The GA was linked to the KIVA CFD code and an automated grid generation tool to perform a single-objective optimization. The target of the optimizations is to improve net indicated efficiency (NIE) while keeping NOx emissions, peak pressure and pressure rise rate under their corresponding target levels. Compared to the baseline engine fueled with DME, the results of the study provide an optimum conventional diesel combustion system with 3.3% NIE improvement and an optimum stoichiometric combustion system that offers a 0.6% NIE improvement keeping tailpipe NOx values below 1% of the original levels. Due to the methodology, not only the optimum combustion system configuration is described, but also the cause-effect relations between the most relevant inputs and the optimization outputs are identified and analyzed. The new geometry shapes reduce heat transfer (HT) losses by controlling the surface area while EGR is still critical to control NOx emissions for both combustion systems. This study confirms the potential of DME as a promising fuel for the future generation of compression ignition engines, and demonstrates the benefits of co-optimizing the fuel properties, combustion chamber hardware and air management plus injection settings.