(A) study on microcomputer-based adaptive control system of a steam generator

Abstract

A MACS (Microcomputer-based Adaptive Control System of a Steam Generator) is developed to control the steam generator water level from 0\% to 100\% load. The important advantages of the MACS over conventional controllers are its dependence on one measurement, lower cost, and greater flexibility in the implementation of the MACS for a real plant. The conventional P.I.D. controllers are replaced by APPLE micro-computer software with a simple digital circuit in the present design. The stochastic selftuning regulator scheme in the control algorithm is adopted. This scheme can calculate an adaptive, optimal valve position in response to changes in the dynamics of the process and disturbances. The self-tuning regulator takes the form of adaptive control law which combines the recursive least squares estimator with the one-step ahead controller. In computer simulation studies, a simplified discrete model is used to simulate the transient response of the water level of an actual steam generator. It is proved that the recursive least squares estimator provides the rapid parameter estimation and global convergence to minimize the square error between the measured output and a-posteriori predicted output. Also, it is shown that the self-tuning regulator developed here can cnotrol the steam generator water level at low load. Especially, it is found that the control input effort decreases with increasing weighting factor. The experimental system is made to simulate the secondary loop in the nuclear power plant. From this system, the real water level data of the steam generator is directly obtained. A software is implemented in the micro-computer to receive the water level measurements and calculated control signal based on the stochastic self-tuning regulator algorithm. Through the experimental tests using the MACS, the same conclusions are reached as those obtained in the computer simulation studies.