Research On PID Dynamic Compensation Method For Model Predictive Control Algorithm

Model Predictive Control (MPC) strategy combined PID cascade control structure obtains excellent set-point tracking performance and robustness using conventional PID control loop to reject disturbance. This control strategy synthesizes performance merit of MPC and structure excellence of cascade control, which makes this control technology implement widely in chemical process engineering. But model mismatch appears in actual application, just because of PID parameters tuning but not generalized plants structure changing. To avoid time-consuming and unnecessary step-test and process identification, accurate models of generalized plants are obtained to reduce worse control level by reason of model mismatch, which became a problem in chemical industry. Meanwhile, in respect that the dynamic model of conventional PID control loop is difficult to obtain, it is a handicap to improve MPC control performance using PID parameters tuning, which counterworks advanced research to the influence of conventional PID parameters on the performance of MPC in MPC-PID cascade control. In this thesis, based on model conformation and dynamic matrix operation of MPC, via researching simulation of MPC-PID cascade control, above industrial problems have been researched and solved.The main content of the thesis and its innovations are as following:1. A kind of PID dynamic model compensation method in MPC is proposed, for the mismatch of generalized plant due to PID controller parameter tuning. Dynamic matrices based on MPC algorithm are computed to generate generalized plant model with new PID parameters, which avoids re-testing and identification of whole process. The efficiency of this PID dynamic model compensation method is illustrated by simulation results.2. PID controllers with two different purposes of set-point tracking and disturbance rejection are designed based on Intern Model Control (IMC) theory and Direct Synthesis (DS) method. For the structure characteristic of MPC-PID cascade control, the performances and effect of PID controller for disturbance rejection are focused on. At last, the conclusion is presented that purpose of disturbance rejection for PID parameters should be taken in MPC-PID cascade control.3. The influence of PID parameters and PID system robustness on MPC performances is analyzed. It makes possible using PID dynamic compensation method to obtain the dynamics of conventional PID control loops. And the dynamics of conventional PID control loops with changed structure can be compensated to dynamics of generalized plants. With different PID parameters and PID system robustness, MPC performances are recorded. Based on the above-mentioned, qualitative suggestions to improve performance of MPC-PID cascade control are proposed.