A Novel Singular Perturbation Technique for Model-Based Control of Cold Start Hydrocarbon Emission

High hydrocarbon (HC) emission during a cold start still remains one of the major emission control challenges for spark ignition (SI) engines in spite of about three decades of research in this area. This paper proposes a cold start HC emission control strategy based on a reduced order modeling technique. A novel singular perturbation approximation (SPA) technique, based on the balanced realization principle, is developed for a nonlinear experimentally validated cold start emission model. The SPA reduced model is then utilized in the design of a model-based sliding mode controller (SMC). The controller targets to reduce cumulative tailpipe HC emission using a combination of fuel injection, spark timing, and air throttle / idle speed controls. The results from the designed multi-input multi-output (MIMO) reduced order SMC are compared with those from a full order SMC. The results show the reduced SMC outperforms the full order SMC by reducing both engine-out and tailpipe HC emission. This improved performance is caused by setting an optimum priority among the control inputs through the incorporation of SPA that evaluates the significance of the control variables/states on HC emission. The presented SPA approach can be applied for different automotive control applications that face the challenge of how to set the priority among different control actuators which often have conflicting impacts on desired control targets.