A control orientated combustion model.*

The control of Internal Combustion (IC) engines is dominated by conventional controllers that regulate the engine operating condition to set-points, which are usually derived from steady-state testing. A significant issue is that IC engines can spend much of their time operating under transient conditions, which take the engine away from their optimized steady-state operating conditions. Currently, a variety of ad-hoc methodologies are used to compensate for transient operation, by modifying set-points, gains and feed-forward values for example. It is proposed that Model Predictive Control be used for the online optimization of the combustion processes as a more structured way of controlling transient engine operation. This type of optimization requires a combustion model that captures the response of the combustion system to changes in the control inputs. This paper describes a model that is detailed enough to predict emissions trends, yet is suitable for implementation in a future real-time automotive control unit.