AN ADAPTIVE SCHEME WITH AN OPTIMALLY TUNED PID CONTROLLER FOR A LARGEMSF DESALINATION PLANT

This paper reports on the experience in the course of development of a n adaptive control strategy for an 18-stage multistage flash (MSF) des alination plant presently operating in the Arabian Gulf region. The pa rameter-scheduling strategy maintains optimality of PID controllers ov er the operating region and is derived by simulation using a detailed phenomenological dynamic model. An optimal method of reducing the mode l into first-order-dead-time (FODT) form and a non-parametric model-ba sed simulation facility are presented. Based on a relative gain array analysis, an appropriate control structure has been established. A num ber of integral performance criteria have been used, without resorting to the use of a reduced model, directly with the non-parametric simul ation facility for optimal controller tuning, The controller parameter s are given as a vector function of the top brine temperature (TBT) an d the brine recycle flow rate, which are key variables for the TBT con trol.