Simulation Aided Design and Development of an Adaptive Scheme with Optimally Tuned PID Controller for a Large Multistage Flash Seawater Desalination Plant - Part III: Parameter Scheduling Based on Optimal PID Tuning with Unreduced Plant Model in Nonparametric Form

Abstract

Optimal tuning of PID controllers based on integral performance criteria is by now well established, but, the design relies on the standard first order plus dead time (FODT) approximation of the plant model whose adequacy is not always guaranteed and consequently the existing design methods can only be tenned as sub-optimal. In this paper, the need to approximate the plant model is removed and a truly optimal design scheme using the unreduced plant model in its nonparametric form is established. Based on this design, a parameter scheduling scheme for adaptive (optimal) control is proposed. The optimality of perfonnance of the control system is maintained over a range of operating conditions against the variation in the linearized plant behaviour due to its nonlinear character. The scheme is illustrated with reference to the top brine temperature (TBT) control of an actual 18 stage multistage flash (MSF) desalination plant.