Model Predictive Functional Control for an Automotive Three-way Catalyst

In this work, a model predictive functional control approach for automotive three-way catalyst oxygen storage state control is demonstrated on a Ford 2.0 liter I4 Duratec SI engine. The control system uses a UEGO sensor for the pre-catalyst air fuel ratio (AFR) measurement and a switching-type HEGO sensor for the post-catalyst measurement. The model predictive controller is the primary control loop within a multi-rate cascade control configuration that adapts the parameters of a post-catalyst HEGO relay controller in an optimal manner using a predictive functional control approach. This relay controller adjusts the target of a delay-compensated feedback controller for the pre-catalyst AFR in order to maintain the post-catalyst HEGO sensor signal within a specified range of the desired target voltage. The use of a multi-rate cascade control structure allows the relay controller to execute within a short control interval in order to ensure that the post-catalyst HEGO sensor does not saturate and allows the model predictive controller to execute within a longer control interval in order to provide sufficient computation time to perform the on-line optimization necessary for predictive control. This cascade structure can also tolerate failure of the catalyst controller to solve the online optimization problem because the post-catalyst relay controller will continue to stabilize the catalyst system without the predictive controller operating.