Modeling and control of a nuclear pressurized water reactor using advanced model predictive control strategies

Abstract

The nonlinear dynamics and multiple interacting loops in pressurized water reactors (PWRs) make control design challenging, particularly when load-following and safety constraints must be handled simultaneously. Building on the detailed 27-state PWR model developed by Vajpayee et al. (2020), this paper presents a MATLAB/Simulink implementation of the complete reactor-steam-generator-pressurizer system, utilizing it as a benchmark to compare several standard control strategies. The model includes point-kinetics, thermal–hydraulic, piping, plenum, steam-generator, and pressurizer dynamics, and is validated against the transient responses reported in Vajpayee et al. (2020) for rod, heater, turbine-valve, and feedwater disturbances. Four controllers are implemented for the same multivariable control problem (reactor power, pressurizer pressure, steam pressure, and pressurizer level): a proportional-integral (PI) controller, a linear model predictive controller (MPC), a gain-scheduled MPC (GSMPC), and a nonlinear MPC (NMPC). These controllers were tested under 2 scenarios: power ramping (up and down) to evaluate their regulatory performance and feed-water disturbances to evaluate disturbance rejection. For each controller, we document the design workflow, tuning parameters, and constraints. Closed-loop performance is evaluated using standard integral error indices and selected load-following and disturbance-rejection scenarios, together with average computation time per sampling instant. The results show that, with fair tuning, MPC-based strategies significantly reduce power-tracking integral time absolute error and settling time compared with PI control, while GSMPC can approximate NMPC performance over a wide operating range at lower computational cost, which proves its feasibility for real-time applications. The implementation details and benchmark scenarios provided here are intended to support reproducible studies and further development of advanced control methods for PWRs.