РАЗРАБОТКА ГИБРИДНОГО НЕЙРО-ПИД-РЕГУЛЯТОРА ДЛЯ УПРАВЛЕНИЯ АКТИВНЫМ ПОДШИПНИКОМ ЖИДКОСТНОГО ТРЕНИЯ

Abstract

The article proposes a method for controlling an active fluid-film bearing, based on the use of a classical PID controller in conjunction with an artificial neural network. The regulator coefficients are not constant numbers, but are chosen by the network depending on the state of the controlled system. To implement such a control scheme, the coefficients are selected using a particle swarm optimization algorithm, which constitutes the training dataset, and an ANN is trained using the dataset. The controlled object is represented with a model operating in the Simulink environment.