The actual industrial production process often has the characteristics of large time delay,nonlinear,multi variable coupling,and it is difficult to control.The generalized predictive control algorithms for complex control objects mainly has the characteristics of model prediction,rolling optimization and feedback correction and on-line identification etc.This algorithm overcomes the influence of the uncertainty of the object,modeling error and other factors,and has good control effect and strong robustness.This paper introduces the traditional generalized predictive control algorithm and the improved generalized predictive control algorithm.The control increment in the optimal control law is solved directly according to the input and output.This method avoids solving the Diophantine equation online and has good real time performance.And then constrain the incremental control over the next moment,and make it proportional to the increment of control at this time.This algorithm transforms the original matrix operation into vector operation to avoid the online solution of the inverse matrix.Combing these two ideas constitute an implicit generalized predictive control algorithm for introducing staged control strategies,which has advantages of both two algorithms.The process control experiment with MATLAB simulation has some limitations in the laboratory.The static characteristics of the system,the nonlinear of the controlled object and the constraints in the actual process are difficult to be expressed by MATLAB software,also the control algorithm is not allowed experiment in the continuous control process repeatedly.Therefore,the actual physical model of four tank multi variable coupling large delay system in the laboratory in accordance with the characteristics of industrial production process.This physical platform can strictly conduct experiment study according to the constraints in the actual production process.In order to show the process of the whole system,the interface operation is more flexible,and the experiment results are clear,we combine LabVIEW virtual instrument with MATLAB software to set up a set of man-machine interface.In this way,it can not only take full advantages of the LabVIEW virtual instrument to replace text line creation application program and the interface friendly features,but also play the powerful computing function of MATLAB software.This physical model of the four tank and the design of the LabVIEW operation platform constitute a multi variable coupled large delay physical system.The improved generalized predictive control algorithm is applied to the multi-variable coupling large delay physical platform.And the feasibility and effectiveness of the control algorithm are verified by experimental research.The experiment results showed that the system has good control quality and can effectively overcome the variable coupling and delay in the model drift,constant value disturbance and external disturbance. |