Chen-Fang Chang

Total engine optimization and control for SI engines using linear parameter-varying models

The role that supervisory control or upper-level multi-objective optimization can play in total engine control is considered. This involved the control of an SI engine for a combination of air-fuel ratio, torque, fuel consumption and residual gas fraction. The upper level optimization algorithm, based on Model Predictive Control principles, makes use of a ‘reference model.’ This is constructed as a closed-loop interconnection of the engine model and the lower-level classical torque/air charge and lambda controller. This model is derived in a linear parameter varying form, which facilitates the application of the predictive control law. The main contribution is the proposed hierarchical control structure, which has been assessed in simulation and on a test track, and provides a useful separation in tasks and functionality. The solution is also compared briefly with MPC at both upper and lower levels, using different sample rates.

Adaptive intake oxygen estimation in lean-burn engines

The oxygen concentration in the intake manifold of a lean-burn engine is a key variable that is directly related to the quality of combustion. In this paper, an adaptive observer is developed to estimate the intake oxygen concentration in the intake manifold of a lean-burn engine using measurements from existing sensors with minimum computational load. Performance of the observer is evaluated using transient experimental data obtained from a diesel engine.