Generalized Predictive Control of Steam Generator Level Considering Operating Constraints

Abstract The steam generator (SG) is a crucial component of the pressurized water reactor (PWR). It exhibits characteristics that are nonlinear, time-varying, nonminimum-phase, and exhibit a small stability margin and large time delay. The traditional PID control system demonstrates suboptimal control performance, particularly at low loads. In order to guarantee the safe and cost-effective operation of PWR, it is imperative to implement an advanced and effective control for SG water level. This paper proposes a generalized predictive control (GPC) scheme considering constraints for SG level. In this scheme, the effects of measurable disturbances are included, and a dual-input, single-output CARIMA model is constructed to describe the effects of steam flow and feedwater flow on the level. A recursive least squares method (RLS) with a variable forgetting factor is employed to perform online discriminative corrections of the model parameters. The SG operating constraints are transformed into limitations on the amount of output variation of the GPC controller by the introduction of a softening factor. Typical operating conditions are simulated and the control performance is compared with PID control. The simulation results demonstrate that the GPC scheme considering constraints exhibits rapid response speed, small settling time, robust disturbance-rejection capability, and the capacity to address the time-varying characteristics at varying power levels. Furthermore, it demonstrates superior control efficacy at low power levels in comparison with PID controller.