Research On Predictive Control Strategy Of Online Model Identification

In general, since the actual industrial production process is often nonlinear,time-varying and uncertainty, it is difficult to establish a precise mathematical model of the system. People want to be able to make the object model as simple as possible and hope the control algorithm is simple and easy to implement. Predictive control is a kind of new optimized control algorithm based on the model generated in this case.Since the model predictive control algorithm in the form of requirements are relatively relaxed, in which the dynamic matrix control using step response as an internal model, the algorithm is relatively straightforward and simple.Predictive control of online model identification, is to identification and predict control closely Combination. By getting the nonlinear function relationship between the parameters and the dynamic matrix controller, the function of on-line optimization control of controller is realized. According to the characteristics of the dynamic matrix control to construct the prediction model.Select a appropriate identification algorithm,based on industrial site historical data and use off-line mathod to identify the controlled object.By getting the nonlinear function relationship between the parameters and the load, finally, the function of online model identification is realized. Because of industrial systems are non-linear and complex, the description of the system lacks the characteristics of general model.Therefore, it is of great importance in theory and practice to apply the method to the nonlinear system.In this paper, we used dynamic matrix control algorithm and real data of superheated steam temperature system, finally construct the forecast model. The design of on-line identification of dynamic matrix controller for superheated steam temperature system is realized, design simulation test and make performance analysis.