Model Predictive Control (MPC) of an Advanced Multi-Cylinder Engine for Transient Operations

Abstract: The transportation sector is one of the main contributors to greenhouse gas CO2 and emission of toxic gas. In this sector, advanced technologies and innovations to increase engine efficiency is an important research topic. Low-Temperature Combustion (LTC) concepts are effective approaches for high-efficiency engines. Partially Premixed Combustion (PPC) is one type of low-temperature combustion which is promising with respective to efficiency and emissions. PPC is highly influenced by both the inlet charge condition and the injection process. The optimal EGR condition at low load (6 bar) is around 50 % where the smallest energy loss is found as a trade-off between heat transfer and exhaust-gas enthalpy-waste. Lambda close to 1.4 results from the trade-off of high gas-exchange efficiency with low lambda and high thermodynamic efficiency with high lambda. The multiple-injection contributes to the high efficiency of PPC by controlling the fuel in-cylinder stratification directly. Due to the sensitivity of PPC to both the inlet charge condition and the injection process, the cylinder pressure based feedback is necessary for PPC control. A gradient method was used to estimate the polytropic exponent through compression, combustion and expansion and pressure off-set to calculate a more accurate heat release, which was used for combustion timing calculation and feedback control.The air-system was matched for PPC using high EGR and multiple injections supported by BorgWarner. A novel controller called receding horizon sliding model control (RHSC) was investigated and validated on the air-system control. A combustion model suitable for high EGR and multiple injections was investigated and a control scheme was proposed for PPC multi-cylinder engine control for both air and fuel path. The control scheme was validated in transient control. During the stable operation, peak gross indicated efficiency of 51.5 % and peak net indicated efficiency of 48.7 % were achieved. During the transient operation, net specific fuel consumption was 174 g/kWh and the average net indicated efficiency was 47.5 %.