Decoupling state spatial prediction control method of chemical multivariate processes

The invention relates to a decoupling state spatial prediction control method of chemical multivariate processes. In the conventional control means, control parameters are completely dependent on the experience of technicians, and the control effect is not excellent. The method comprises the following steps of: firstly, building a decoupling state spatial model based on the chemical process, and mining the basic process characteristics; then, building a prediction control loop based on the decoupling state spatial model; and finally, calculating parameters of a prediction controller to carry out prediction control on a process object. In the method provided by the invention, by means of data acquisition, process processing, prediction mechanism, data driving, optimization and the like, a decoupling state spatial prediction control method of chemical multivariate processes is established, with the adoption of the method, the control precision can be improved effectively, and the control smoothness also can be improved. The method provided by the invention can effectively reduce the errors between the ideal technological parameters and the actual technological parameters, ensures that a control device operates in optimal state, and realizes strict control on the technological parameters of the production process.