Reduced Order Modeling of Engine Transients for Gasoline Compression Ignition Combustion Control

This work focuses on reducing the computational effort of a 0-dimensional combustion model developed for compression ignition engines operating on gasoline-like fuels. As in-cylinder stratification significantly contributes to the ignition delay, which in turn substantially influences the entire gasoline compression ignition combustion process, previous modeling efforts relied on the results of a 1-dimensional spray model to estimate the in-cylinder fuel stratification. Insights obtained from the detailed spray model are leveraged within this approach and applied to a reduced order model describing the spray propagation. Using this computationally efficient combustion model showed a reduction in simulation time by three orders of magnitude for an entire engine cycle over the combustion model with the 1-dimensional spray model. Capturing only the basic features of the spray propagation did not show a substantial decrease in prediction accuracy compared to the 1-dimensional spray model proposed in a previous publication. The reduced order model is able to predict CA50 within ±1CAD for most conditions. Having an estimate of the heat release within the engine as a function of various parameters, combined with first-order transfer functions reflecting the intake dynamics of the engine, allows using this as a full engine model suitable for transient engine simulations. These intake gas dynamics were obtained by applying step changes to engine parameters of interest. From a controls perspective, this engine model is utilized as a plant in a control system to simulate the closed-loop response based on various changes in input parameters. Engine responses and control actions for reference tracking in combustion phasing were investigated for early as well as late pilot injection strategies.