Synthesis of the Adaptive-fuzzy System Regulating the Temperature of Overheated Steam in Heat-electric Objects
Abstract
The article discusses the results of developing an algorithm for calculating the parameters of an adaptive controller when controlling the temperature of superheated steam. The presented control algorithm uses artificial neural networks. The purpose of this scientific work is to develop adaptive control capable of operating under unknown limited external disturbances with varying parameters of the thermoelectric boiler over time. Research methods are based on the provisions of modern areas of control theory, such as adaptive management and identification. Mathematical models are constructed by the analytical method using equations that describe the physical properties of the object. The methodology of creating a temperature control system for a superheater operating under conditions of a priori uncertainty is presented. It is shown that such control systems belong to adaptive systems capable of controlling an object with significant and previously unknown object parameters. It is proposed to use an artificial neuron as an adaptive part of a control system or their combination - artificial neural networks.