Adaptive controller independent to model and control method thereof

Abstract

The invention relates to a non-model self-adaptive controller. The controller consists of two forward neural networks which are connected in series, wherein the front neural network is a network controller MLPc; and the back neural controller is a network simulator MLPo. The output of the network controller is taken as the input of the network simulator and a controlled object; a closed loop reverse feedback is arranged between the output of the controlled object and the input of the network controller; and the relation between the output of the network simulator and the output of the controlled object is set as the subtraction relation. The network simulator consists of a time lag multilayer perceptron (TDMLP) and is used for simulating the industrial process; and the network controller which is also a TDMLP is used for exerting the optimal control on the simulator in the precondition of considering model estimation errors. The non-model self-adaptive controller has the advantages of high control accuracy, high system response speed, and simplified network internal mechanism.

Description

technical field [0001] The invention relates to an adaptive controller and a control method thereof, in particular to a model-independent adaptive controller and a control method thereof. Background technique [0002] Industrial processes are characterized by complex physicochemical or thermodynamic processes, generally nonlinear and time-varying, that are difficult or difficult to model accurately. Existing controllers used in industrial processes are generally object-based or experience-based. Among these two types of controllers, object-based controllers have great difficulties in designing controllers due to the characteristics of the above-mentioned industrial processes. Whereas the performance of experience-based controllers largely depends on the experience of experts. Both controllers are difficult to achieve optimal control of industrial processes. In the relevant literature, there have been examples of using neural networks to achieve adaptive control, but the ne...

AI Technical Summary

Problems solved by technology

Both controllers are difficult to achieve optimal control of industrial processes

In the relevant literature, there have been examples of using neural networks to achieve adaptive control, but the neural network structure used is often single, and this structure has two defects: 1. It cannot accurately handle industrial processes with time delays , the control accuracy is limited; 2. The internal mechanism of a single network is complicated, the training process is long, and the system response speed is slow

Images