Implementation of Low-Level Control Systems for Power Converters Based on Adaptive Artificial Neural Networks

The article discusses an approach to the construction of a highly adaptive artificial neural network, designed to control electrical energy converters under changing loads and operating modes of electrical networks. The authors describe key aspects in artificial neural network design, including the choice of architecture, activation functions, training methods, and determination of training criteria. Particular attention is paid to working under conditions of limited resources of the hardware and software complex of modern control systems. The question of defining the quality criteria of transient processes to evaluate the efficiency of the control system based on artificial neural network is raised. An example of implementing a neural network control system in the Matlab/Simulink computer modeling environment is given. The results confirm the potential of using artificial neural network as an alternative or supplement to classical regulators, which opens new prospects for the development of intelligent control systems in the electrical engineering industry.