Research About The Advanced Model Predictive Control Technology Based On System Identification

With the increasing uncertainty in the modern industry,traditional control methods are not sufficient to deal with complex industrial process.The application of advanced control theory is imperative,and the system identification is the foundation of advanced control theory applications.Based on the research of the two identification methods under the closedloop conditions,this paper improves the model predictive control,combines the fuzzy impulse control with the improved model predictive control,and uses the Managed Pressure Drilling(MPD)process model to perform the control effect test finally.The main research contents of this paper are as follows:(1)The design and implementation of closed-loop identification algorithm.A recognition method based on closed-loop dynamic data is proposed.At first,using closed-loop dynamic data control system to get the opened-loop data,and then get the opened-loop system model via the method of split ring.In this paper,a closed-loop identification method based on frequency analysis is also proposed.Through the conversion of the frequency characteristics of the process objects under open and closed loop conditions,the relationship between the parameters of the opened-loop process object model and the closed-loop frequency characteristic is obtained,and then the opened-loop model of the process object is obtained.(2)The research on model predictive controller technology improvement.According to the characteristics of the model predictive control algorithm,the dynamic matrix simplification factor is proposed to improve the model predictive controller.The dynamic matrix simplification factor is used to simplify the inverse operation in the control process.The calculation of the rolling optimization in the model predictive control is reduced and the real-time performance of the predictive control algorithm is improved.(3)Design and research of advanced predictive control algorithm.The impulse control is realized according to the basic principle of impulse control.Then the fuzzy control is used to control the key parameters in the impulse control,and the control performance is improved.The two control algorithms are combined with the improved model predictive control respectively.The dynamic performance of the impulse control and the steady performance of the model predictive control are fully utilized.The impulse-prediction control and the fuzzy impulse-prediction control are realized.The closed-loop identification algorithm based on dynamic data and the fuzzy impulse-predictive control algorithm are combined to form the model self-tuning fuzzy impulse predictive control algorithm.The mathematical model in the fuzzy impulse-predictive controller is corrected in real time according to the control performance index,so as to achieve a better control effect.(4)Application of advanced prediction control algorithm.Firstly,analyze the process of MPD to build its differential equation model.Then the MPD process model is obtained based on dynamic data closed-loop identification of differential equation models.According to the process model,improved model predictive control,impulse control,fuzzy impulse control,impulse-predictive control was realized respectively,fuzzy impulse-predictive control and model self-tuning fuzzy impulse-predictive control algorithm are applied in practice,and the control performance of these advanced control algorithms are analyzed and compared,which lays a foundation for further improving the control effect of MPD.In this paper,through the study of closed-loop System Identification,model predictive control and impulse control,two closed-loop identification algorithms are designed and implemented,and the model predictive control algorithm is improved.Based on the impulse control algorithm,a fuzzy impulse control algorithm is implemented.Combining with model predictive control respectively,two advanced predictive control algorithms are formed,and a closed-loop identification algorithm based on dynamic data is added to fuzzy implusepredictive control algorithm to form model self-tuning control algorithm.Finally,MPD model is used to test the control effect.The experimental results prove that the proposed advanced predictive control technology has a better control performance and lays a foundation for further improving the control effect of the MPD.