Towards combining nonlinear and predictive control of diesel engines

The paper presents the results of a model predictive controller development for diesel engine air path control. The objective is to regulate the intake manifold pressure (MAP) and exhaust gas recirculation (EGR) rate estimate to the specified set-points by coordinated control of the Variable Geometry Turbine (VGT), EGR valve, and EGR throttle. The approach combines elements of nonlinear control (partial nonlinear inversion) and explicit Model Predictive Control (MPC) based on reduced order linear models. The application of nonlinear control renders the model “more linear”, which facilitates the application of predictive control and is shown to reduce the computational complexity. For example, after the nonlinear inversion, the DC gain reversal that is known to complicate the development of effective feedback controllers in diesel engines disappears. The number of controller zones that cover the engine operating region is reduced. The ability of the controller to handle input constraints and intake manifold pressure constraints while following the specified set-points is demonstrated and validated using nonlinear model simulations. Preliminary experimental results are reported.