Parameter Optimization of Steam Generator Water Level Control System Based on Piecewise ARX Modeling

In this article, an offline sequential parameter optimization method for cascade proportional-integral-derivative (PID) controllers for the steam generator (SG) water level control system in pressurized water reactor (PWR)-type nuclear power plants is proposed to improve the control performance when the unit output power (UOP, electrical power) changes greatly. We establish the simulators in sequence for the SG feedwater flow and water level processes using a set of piecewise linear autoregressive with exogenous inputs (ARX) models obtained under dynamic mode constraints to optimize the controller parameters constrained by system stability, and conduct simulations in the power ranges of 1000-1010, 575-700, 240-260 and 210-220 MW. The results show that the simulators have good long-term prediction performance, and the PID controllers with optimized parameters have good control performance for the SG feedwater flow and water level control.