Design and Implementation of Linear Model Predictive Siso Boost Pressure Controller for a Series Sequential Diesel Engine

This paper focuses on the design and implementation of model predictive controller (MPC) for a boost pressure control of series sequential diesel engine. Boost pressure control is critical to satisfy diesel engine performance and driveability requirements as well as increasing volumetric efficieny. In this study, Control oriented linear models are generated by using system identification methods in order to be used in output prediction models. Prediction models are identified for 5 different engine operating regions to increase the accuracy of linear models. Based on state-space prediction models, Controller design is performed considering Kalman Filter tuning, constraint definitions, controller weights. Engine dynamometer testing have been performed to define input and input rate constraints. MPC design is performed for online optimization method. Nonlinear engine model is modeled in high fidelity simulation environment. Results are shown that MPC is capable of showing better setpoint tracking while satisying contraints explicitly than conventional PID (Proportional-Integral-Derivative) controllers. Relatively easy tuning, ability to handle constraints and incorporation of models makes MPC attractive to Automotive control community.