An RBF neural network predictive control-based control system and control method for a double-input double-output ball mill, the control system comprising an RBF neural network model-based predictive controller, a control quantity initialization module and a controlled object, wherein the controlled object is a double-input double-output ball mill model which outputs a discrete controlled quantity generated after a continuous controlled quantity is discretized and a controlled quantity current set value, and same are inputted to the control quantity initialization module and the predictive controller; the control quantity initialization module outputs a control quantity initial value that is inputted to the predictive controller; and the predictive controller outputs a discrete control vector that is converted by a zero-order holder into a continuous control quantity and then is outputted to the double-input double-output ball mill model. The present control method uses an RBF neural network forward model and an RBF neural network inverse model to achieve predictive control of the controlled object. The described method may control and adjust the system in advance, and is suitable for the control of a large lag system. The controlled quantity has a fast response, and a small overshoot amount, and is simultaneously very robust.
展开▼
