Configuring Smith Predictor Parameters for a Variable Line Feature
Abstract
In this article, we consider a modification of the Smith predictor for a linear object with variable parameters. The Smith predictor is an effective tool for controlling a system with delay, but its use requires an accurate estimate of the parameters of the control object. Thus, for objects with variable parameters, it is necessary to introduce an additional adaptation loop into the structure of the Smith predictor. The purpose of the study is to develop a modification of the Smith predictor, which allows you to work under the conditions of a change in the gain of the control object described by the transfer function. Results Various options for organizing the adaptation loop of the Smith predictor were considered: based on the MIT rule, Lyapunov functions, based on a linear PID controller and an adaptive controller with a reference model using the sensitivity functions of the closed-loop system relative to the PID controller parameters. On the basis of computer experiments, the choice of an adaptation loop that ensures the operation of the Smith predictor under the conditions of a change in the transfer function of the control object isjustified.