A fuzzy controller based on self-tuning rules for the nuclear steam generator water level.pdf

K S M E International Journal, Vol. 11, No. 5, pp. 485~ 493, 1997 A Fuzzy Controller Based on Self-Tuning Rules for the Nuclear Steam Generator Water Level 485 Man Gyun Na* and Jae Hyung Lim** (Received December 5, 199.5) It was demonstrated that the fuzzy control method could be applied successfull~ to control and regulate the nuclear steam generator water level through previous works. However, this method has some limitations from the fact that its performance largely depends on the tuning method and how to settle the rule base. In this work, a fuzzy control algorilhm based on self --tuning rules is investigated to find the optimal tuning point and to generate the control rule automatically. It adopts an on-l ine iterative method and the objective error function which contains the variation errors of all time-steps. The triangular membership functions based on the mass flow rate and water level errors are applied. And also, a torgetting factor is used to account for an exponential decay of the past data in tracking the slow drift in system paramelers. Thus it controls the steam generator water level flom low power to lull power with a little transients. This control algorithm is validated by computer simulations. Key Words: Fuzzy Control, Membership Function, Self-Tuning Rule I. In troduct ion The proper control of the water level of a nuclear steam generator is of much importance in order to secure the sufficient cooling source of the nuclear reactor and the good performance of the steam separator and dryer, and to prevent the damage of turbine blades. The steam generator level control has been a main culprit of un- expecled shutdown of nuclear power plants. The non rainimum phase effects are significantly greater at low power, which makes more perilous the use of a high gain of the control loop at the reduced power level. Main objective of this work is to design a controller that not only makes the feed water mass flow rate tra