An Application of Model Predictive Control to a Gasoline Engine

The number of the actuators in modern gasoline engines increases due to rising requirements regarding emissions and fuel consumption. The coordination of the available actuators (e.g. throttle, ignition plug, exhaust gas recirculation and turbo charging) is a challenging task. An available scheme for efficient and model-based coordination that respects the actuator constraints is model predictive control, but it is commonly perceived to be limited to processes with slow dynamics due to the necessary solution of an incorporated optimisation problem in each sampling instance. In this paper, it is presented that model predictive control on a production electronic control unit (ECU) for gasoline engine control is feasible. The numerical properties of the applied optimisation algorithm are obtained by Hardware-In-The-Loop simulation and the practical applicability is shown by controlling the engine of a Volkswagen T5 transporter in idle running and on a test track.