Adaptation Algorithm for the Parameters of the Heat Carrier Temperature Controller of the Waste Heat Boiler Based on Neural Network

The work is devoted to the development of an algorithm for neural network adaptation of the parameters of the PI-controller for the water temperature at the outlet of the waste heat boiler. The flow of water through the heat exchanger of the waste heat boiler chosen as a control action. The control object is a complex thermodynamic system Automatic control systems based on a controller that implements the PI law without adjusting the coefficients, the controller cannot provide a satisfactory quality of transient processes. There are several methods for automatic tuning of the PI controller coefficients, among which one can distinguish autotuning by tabular values, when the previously found controller parameters for different operating conditions of the system are entering to a table. Each set of coefficients are bound to certain conditions. The work proposes to create an algorithm for adapting a PI controller based on a neural network. The neural network generates the values of the PI controller coefficients for different operating points. The values of the regulator coefficients included in the training sample are calculating from the frequency response of the control object. The frequency response were determined from the analysis of the neural network model of the waste heat boiler.