Design of Hydropower Plant PID Controller Parameters Using Artificial Neural Networks

This paper represents the artificial neural networks indirect implementation in non-linear control systems. Due to the popularity of the proportional, integral, and derivative, PID controller in control systems, and the process industry, this paper aims to represent the controller constant parameters design. Ziegler-Nichols method is used for setting up the lower and higher limit parameters. Optimizing the PID constant parameters, the ones with the lowest possible mean squared error is done with the artificial neural networks. Python is used as software for proportional, integral, and derivative constant obtaining, and later the results are implemented in the MATLAB/Simulink simulation. The working model is a non-linear hydropower plant with two parallel turbines placed on the same supply pipeline. Simulation models represent the hydropower plant black start. The artificial neural network implementation means providing large input and target data sets for network training and validation. Other important parameters in the network setting are the number of neurons, the number of layers, the type of activation function, the learning rate, the optimizer type, and others.