Tuning of PID Controllers for Networked Control Systems

Abstract

As the use of computer networks grows rapidly, applications, such as networked control systems (NCS), have emerged. NCS suffer from varying time-delays that bring new problems to the control loop. This paper discusses tuning of a discrete-time PID controller for NCS. A discrete-time PID controller is selected as a controller, because it is widely used in industry. Minimising of a cost function is a powerful tool in control design. Here the optimisation technique is used to find the best parameters of a PID controller, when the system has time-varying or stochastic delays. The cost variance is used as a robustness measure of the tuning. Comparisons to the Ziegler-Nichols and IMC tuning methods are given in a case where the process and the controller are distributed over the Internet. The tunings done in simulations are further verified on a real process. Despite of the simplicity of the PID controller adequate performance is still achieved and it is shown that the optimisation tuning is robust in the presence of a varying time-delay