A multi-level sensitivity analysis for the model parameters of the main steam system in nuclear power plants

The main steam system of a nuclear power plant is a core component of its thermal system, and its operation is typically monitored using simulation models to ensure both efficiency and safety. However, the accuracy of the system model is influenced by the uncertainty of multiple parameters. In this context, sensitivity analysis is essential, as it identifies the most key model parameters, thereby reducing the parameter space and enhancing the efficiency and effectiveness of model calibration. This paper presents a multi-level sensitivity analysis framework that combines the Morris method and the Generalized Likelihood Uncertainty Estimation (GLUE) method. The Morris method is employed as an efficient preliminary screening technique to identify parameters that potentially exert significant influence on model outputs, thereby effectively reducing the dimensionality of the parameter space. Subsequently, the GLUE method is applied to assess the model’s goodness-of-fit to benchmark data using the Nash–Sutcliffe efficiency coefficient, and the posterior distribution of model parameters is obtained to quantify the importance of the parameters. The results show that eight model parameters significantly affect the model output, providing theoretical guidance for model optimization and parameter adjustment of the nuclear power plant’s main steam system. The proposed framework reduces computational time from 26.89 h to 10.73 h, improving efficiency by 60.1% while maintaining high accuracy in key model parameter identification compared to traditional approaches.