Optimal Cold-Start Control Strategies for Gasoline Engines

Abstract

In the first part of this thesis, generally valid rules for optimal control strategies during the catalyst heating phase at a high value of ignition retardation are identified experimentally. To this end, the effects of variations in control strategies are analyzed. Specifically, variations are analyzed in the center of combustion $\theta_{\mathrm{50}}$, in the air-to-fuel ratio $\lambda$ and in maldistributions in both quantities among individual cylinders on the behavior of the engine in idling conditions after a cold start. This behavior includes the fuel consumption, the heat-up behavior of the three-way catalytic converter, and the cumulative tailpipe emission of HC, CO, and NO$_\mathrm{x}$. A dedicated cylinder-individual, model-based, multi-variable controller is developed and used in experiments in order to isolate the effects of the individual control strategy variations as much as possible. An optimal control problem for a gasoline engine at a cold start is formulated which is used to interpret the experimental data obtained. The corresponding goal is to minimize the fuel consumption during an initial idling phase of a fixed duration while guaranteeing that the three-way catalytic converter reaches a sufficiently high final temperature and at the same time ensuring that the cumulative emissions stay below a given limit. The experimental data indicates that the engine should be operated with a maximum ignition retardation and at an air-to-fuel ratio of 5\%-10\% lean in order to reach any temperature inside the three-way catalytic converter as quickly as possible concurrently with minimum tailpipe emissions and at a minimum possible fuel consumption. In the second part of this thesis, trajectory tracking algorithms for gasoline engines are devised. Specifically, a simultaneous and precise reference tracking in engine speed, air-to-fuel ratio, and center of combustion is enabled. Such a tracking of multiple reference trajectories requires a coordinated control action for the air path, the fuel path, and the ignition timing actuators. Combining a dedicated feedforward and feedback controller structure and multivariable model-based norm-optimal parallel iterative learning control strategies, feedforward control trajectories are generated that enable a precise tracking of desired reference trajectories. Experimental results show the effectiveness of the proposed methodology.