Self Tuning Controller on the PLC

Abstract

Tasks in process control are quite heterogeneous. In the past two types of standard process controllers were developed: the Process Controllers, which originate in standard PID controllers and Programmable Logic Controllers (PLC), which replaced relay logic. Their functionality is at present quite different, but it converges. We expect that in the future there will be only one type of process controllers. The purpose of this paper is to describe a simple identification and PID controller tuning procedure and its implementation on PLC as a selftuning procedure. The essential point of the method is the application of the Pseudo Random Binary Signal and the evaluation of the crosscorrelation function between input and output signals. That results in the estimated impulse response of the process. We improved the method by applying the signal twice in succession, so all signals became periodic. The period of 127 (7 bit PRBS register) is sufficient for practical application. The procedure estimates the controller parameters by Chien, Chrones, Reswick tuning rules. Due to these rules the method is applicable to proportional overdumped processes only. The method has proven itself as a very robust method. In this paper also the implementation of the method on Mitsubishi PLC series AxN and its application to the Elwe hydraulic plant is described.