Waste Heat Boiler Control System Based on Neural network Technology and Fuzzy Logic

The experience of developing an intelligent system for controlling the temperature of network water at the outlet of a waste heat boiler is considered. When developing an intelligent control system, a number of sequential control tasks were solved: creating a model of a waste heat boiler based on neural network technologies, studying the frequency characteristics of a neural network model, identification of the structure and parameters of a control object, synthesizing a typical controller, creating an algorithm for adapting parameters of a typical controller based on a fuzzy logic apparatus. The waste heat boiler was chosen as the control object. The control object is represented by a multidimensional thermodynamic system with non-linear parameters. Training and verification of the neural network model took place on the data of the technological process of gas energy utilization. Frequency identification methods were used to identify the structure and parameters of the control object. The algorithm for adapting the parameters of a typical controller contains sets of parameters of a typical controller for different system states with different bandwidths. The article presents a description of the stages of synthesis of an intelligent control system, the results of studying a neural network model in the frequency domain, the result of identifying the structure and parameters of a control object, calculating the parameters of a typical controller for different system states, description a fuzzy adaptation algorithm, and modeling results.