Performance Assessment and Diagnosis for a SCR Denitration System under Model Predictive Control

Model predictive control has been applied to SCR denitration system because of its superior ability of dealing with nonlinearity, large time delay and process constraints. An issue that can't be ignored during operation is the change of process dynamics due to the consumption of catalyzer or other reasons, which will impair the SCR control performance a lot. In this paper, a method based on nominal sensitivity function is conducted to assess the MPC controller performance of a SCR system. And a performance diagnosis method in view of relative sensitivity function is implemented once degraded performance is detected, which can discriminate between model-plant mismatch and unmeasured disturbances. Simulation scenarios are given to verify the feasibility and accuracy of the proposed method.