Authors: Gabriele Lorini, Alessandro Miotti, Riccardo Scattolini
Addresses: Dipartimento di Elettronica e Informazione, Politecnico di Milano, via Ponzio 34/5, Milano 20133, Italy. ' Dipartimento di Elettronica e Informazione, Politecnico di Milano, via Ponzio 34/5, Milano 20133, Italy. ' Dipartimento di Elettronica e Informazione, Politecnico di Milano, via Ponzio 34/5, Milano 20133, Italy
Abstract: Present-day spark ignition engines for the automotive industry are equipped with a number of actuators, such as the Electronic Throttle (ET), the Exhaust Gas Recirculation (EGR) and the Variable Valve Timing (VVT), which must be properly used to guarantee driveability and reduce fuel consumption and pollutant emissions. However, the underlying control problem is particularly difficult due to the strong interactions between these control variables and the main engine performance variables. In the first part of this paper, a control-oriented mean value model of a gasoline engine is presented and validated with real data. This model is then used in the synthesis of two control schemes with predictive capabilities. The first one is composed by a traditional static controller derived through a suitable optimisation problem and by an innovative dynamic term designed with Model Predictive Control (MPC). In the second scheme, the MPC controller is directly plugged into the loop for control of the torque request. Simulation results are reported to illustrate and compare the potentialities of the proposed approaches.
Keywords: automotive control; gasoline engines; model predictive control; MPC; static map; optimisation; spark ignition engines; modelling; simulation; control synthesis.
International Journal of Modelling, Identification and Control, 2008 Vol.3 No.3, pp.258 - 269
Published online: 28 Aug 2008 *Full-text access for editors Access for subscribers Purchase this article Comment on this article