Model Predictive Control of an RCCI Engine

Abstract

Reactivity controlled compression ignition (RCCI) is a promising combustion strategy that offers high fuel conversion efficiency and ultra low emissions of NOx and soot. One of the main challenges associated with RCCI combustion is the difficulty in simultaneously controlling combustion phasing, engine load, and cyclic variability during transient engine operations. This paper aims at developing a model-based control platform for cycle-to-cycle control of RCCI combustion phasing and load. First, a physics based control oriented model (COM) is developed and experimentally validated for predicting combustion phasing (CA50) and indicated mean effective pressure (IMEP) for transient RCCI operations. Then, a switched model predictive controller (MPC) is designed to track CA50 by adjusting start of injection (SOI) and track IMEP by adjusting total fuel quantity (FQ) injected per cycle, while dual-fuel premixed ratio (PR) is used as a scheduling variable to increase the operating range of the controller. Furthermore, the MPC controller is designed such that it limits engine cyclic variability ( $COV_{IMEP}$ ) for different load and speed conditions. Experimental validation results show that the designed MPC controller is able to track both CA50 and IMEP with average errors of 1.5 crank angle degrees (CAD) and 23.1 kP $a$ respectively, while being able to maintain a $COV_{IMEP}$ of less than 5%.