Physics-Based Modeling and MPC for the Air Path of a Two-Stage Turbocharged SI Engine with Low Pressure EGR

For increasing the efficiency and simultaneously decreasing pollutant emissions of internal combustion engines, innovative air path concepts such as turbocharged gasoline engines with exhaust gas recirculation (EGR) are in the focus of current research. These air and EGR path concepts impose high demands on the process control due to the nonlinearities and cross-couplings. This contribution presents a physics-based modeling approach and a nonlinear model predictive controller (NMPC) for the air path control of a sequentially two-stage turbocharged gasoline engine with low pressure EGR. By using EGR at high loads, the in-cylinder temperature can be lowered, reducing the knock probability, while at the same time preventing the need for enrichment of the air/fuel ratio. As air and EGR path are cross-coupled and show different delay times, a model predictive control (MPC) concept is proposed. Therefore, the air and EGR path are modeled in a physical manner, where possible. The model set up is based on a small amount of dyno-run measurement data acquired with a test vehicle. The model is built up such that it can be used within a MPC algorithm. The proposed control concept consists of an extended Kalman Filter for state and disturbance estimation and a NMPC controller considering the dynamic behavior of the air and EGR path. Subsequently, the nonlinear control concept for the two-stage turbocharged spark-ignited (SI) engine with Low Pressure EGR is implemented on a rapid control prototyping hardware, validated via simulative tests and compared to a linear MPC.