IMC-based diesel oxidation catalyst outlet temperature control with extended state predictor observer

Abstract

It is of significant importance in the thermal regeneration process of diesel particulate filter that the diesel oxidation catalyst (DOC)-outlet gas temperature is accurately controlled to ensure efficient and safe regenerations. However, this is challenging because the DOC is a nonlinear and time-varying system which has complex oxidation reactions and large time delay, together with uncertainties and high frequency disturbances in real applications. To this end, an intuitive and effective control strategy which integrates the internal model control (IMC)-based design with the extended state predictor observer (ESPO) is proposed in this work. Firstly, a reduced first-order plus delay model, which is derived from a control-oriented one-dimensional DOC model, is deployed and experimentally validated to facilitate the model-based control design. Then, an IMC-based proportional–integral–derivative feedback control strategy, and an active disturbance rejection control (ADRC)-based ESPO are explored and designed. The IMC-based controller is designed for required set-point tracking and the ESPO is introduced for disturbance estimation and mitigation. In addition, as the regeneration is usually performed in real driving cycles, a feedforward control law based on the principle of energy balance is also well developed and integrated for compensation for the measurable transient disturbances. Simulation results show that a reduction of about 15% in recovery time can be achieved by introducing the ESPO into the IMC-based design. Performance of the proposed control strategy is finally validated on the engine bench and in on-vehicle real driving conditions under the worldwide harmonized light vehicles test cycle with new and aged DOCs. Experimental results show that, when the ESPO is introduced, it is more accurate to provide adequate set-point tracking in real applications where both the external disturbances and internal parameter uncertainties are presented.