Modeling and Model-Based Control of Multi-Mode Combustion Engines for Closed-Loop SI/HCCI Mode Transitions with Cam Switching Strategies

Homogeneous charge compression ignition (HCCI) combustion has been investigated by many researchers as a way to improve gasoline engine fuel economy through highly dilute unthrottled operation while maintaining acceptable tailpipe emissions. A major concern for successful implementation of HCCI is that it's feasible operating region is limited to a subset of the full engine regime, which necessitates mode transitions between HCCI and traditional spark ignition (SI) combustion when the HCCI region is entered/exited. The goal of this dissertation is to develop a methodology for control-oriented modeling and model-based feedback control during such SI/HCCI mode transitions. The model-based feedback control approach is sought as an alternative to those in the SI/HCCI transition literature, which predominantly employ open-loop experimentally derived actuator sequences for generation of control input trajectories. A model-based feedback approach has advantages both for calibration simplicity and controller generality, in that open-loop sequences do not have to be tuned, and that use of nonlinear model-based calculations and online measurements allows the controller to inherently generalize across multiple operating points and compensate for case-by-case disturbances.;In the dissertation, a low-order mean value modeling approach for multi-mode SI/HCCI combustion that is tractable for control design is described, and controllers for both the SI to HCCI (SI-HCCI) and HCCI to SI (HCCI-SI) transition are developed based on the modeling approach. The model is shown to fit a wide range of steady-state actuator sweep data containing conditions pertinent to SI/HCCI mode transitions, and is extended to capture transient SI-HCCI transition data through using an augmented residual gas temperature parameter. The mode transition controllers are experimentally shown to carry out SI-HCCI and HCCI-SI transitions in several operating conditions with minimal tuning, though the validation in the SI-HCCI direction is more extensive. The model-based control architecture is also equipped with an online parameter updating routine, to attenuate error in model-based calculations and improve robustness to engine aging and cylinder to cylinder variability. Experimental examples at multiple operating conditions illustrate the ability of the parameter update routine to improve controller performance by using transient data to tune the model parameters for enhanced accuracy during SI-HCCI mode transitions.