A Diesel Engine Combustion Phasing Optimization Using A Model Guided Extremum Seeking Approach

Optimization of control inputs is of importance for the deployment of powertrain calibrations. In this work, a model guided extremum seeking optimization approach is proposed for the combustion phasing optimization on a diesel engine. The optimization target is to achieve the best compromise between NOx emissions and engine thermal efficiency under certain operating conditions. This work first proposed a 'virtual calibration' process that is applied on a simplified first-principle engine model. Initialized by the obtained 'virtual optimal' parameters, extremum seeking approach is implemented for system online optimization. By adopting this approach, it is likely that the development time and effort associated with engine control calibration can be reduced. Simulations were performed to validate the proposed concept.