Development of calibration tools for model based engine control strategies

In order to meet the legislation requirements on pollutant emissions, spark-ignition engines equipped by a three-way catalyst require a precise control of the Air Fuel Ratio (AFR) fed to the combustion chamber. A stoichiometric mixture is necessary for the proper working of the catalyst. The AFR control is fundamentally composed of two parts: Open Loop (OLC) and Closed Loop (CLC). The AFR OLC has to compensate the wall-wetting phenomenon, where the fuel entering the combustion chamber differs from the fuel injected by the quantity exchanged with a fluid film on the intake manifold wall; this is obtained by varying the amount of fuel with respect to the base one. A critical part of the AFR OLC is the knowledge of the correct instantaneous air mass flow that is necessary to calculate the fuel mass base value during transient. Unfortunately, between the establishment of a certain air mass flow value and the actuation of corresponding injection timing there is an unavoidable delay (in the worst case about twelve Top Dead centers, TDC). In order to compensate this delay, an algorithm to predict the air mass flow or the manifold pressure value, on speed density systems, is necessary. The calibration of the Manifold Pressure Predictor (MPP) parameters is very hard. In this paper a brief description is given of the MPP algorithm and of a calibration procedure, using an ad hoc developed software tool, the Manifold Pressure Predictor Calibration Toolbox (MPPTC). The use of this procedure and tool allows a more rapid and more efficient optimization of the MPP parameters. The tool is based on a Simulink model of the intake manifold and of the MPP algorithm and on a minimization procedure that uses a quadratic index, dependent on the difference between the measured manifold pressure and the predicted manifold pressure calculated by the Simulink model.