LAUSR

Abstract Partially Premixed Combustion (PPC) has been proved a high efficiency combustion concept, along with ultra-low soot and NOx emissions. However, the nature of high pressure rise rate prevents PPC from operating range extension and further practical use. This paper aims to optimize mixture preparation to achieve the benefits of PPC with the constraints of engine states through injection strategy during transient operation. A control-oriented model (COM) is developed based on double-Wiebe function to predict combustion process, where a new linear algorithm is proposed to identify the model parameters. The root mean square error (RMSE) of the predicted cylinder pressure is less than 2.58 bar and peak error is less than 5% against experimental measurements of steady states. A constrained model predictive controller (MPC) is designed and implemented in a PPC engine. Simulation and experiment results show that the proposed controller manipulates injection events to optimize premixing period and fuel distribution towards more benefits of PPC concept. In the testing scenario, soot, NOx and pressure rise rate are regulated within 0.15 mg/m3, 400 ppm and 8 bar/deg, respectively. Consequently, cumulative soot and NOx emissions are reduced by 43.2% and 6.8% in the whole transient cycle.