An Integrated COMUS Framework for Multi-Objective Optimization and Dynamic Uncertainty Analysis of the Pressurized Water Reactor Core Power Control System

To achieve control optimization and a performance evaluation of the pressurized water reactor core power system, a COMUS (control-oriented multi-objective optimization with dynamic uncertainty and sensitivity analysis) framework is proposed. First, the reactor core power temperature dual-channel control system is established. Subsequently, a MOPSO (multi-objective particle swarm optimization) algorithm is employed to perform the offline optimization of the controller parameters, with the aim to improve the system’s power tracking capability and transient response performance.Building upon an optimized control system, this paper combines Latin hypercube sampling (LHS) with the statistical quantitative analysis method (SQAM) to establish the SQAM-LHS method. This method is used to evaluate the propagation characteristics of the multiphysics parameter uncertainties during transient processes and their impact on the system output. Additionally, a global sensitivity analysis based on the PAWN (probabilistic analysis with numerical uncertainties) method is conducted to identify the key input parameters that affect the dynamic performance of the control system.Finally, the COMUS framework is applied to two typical operating conditions. The results show that uncertainties in the input parameters cause varying degrees of fluctuation in the output parameters. Total reactivity is the most sensitive to parameter uncertainties and exhibits more pronounced dynamic fluctuations. Near the 40% full-power level, the fuel temperature feedback coefficient is a key parameter affecting reactor core relative power, coolant temperature deviation, and fuel temperature deviation. Meanwhile, the decay constant has a significant impact on the stabilization time of total reactivity and related dynamic performance indicators.