Improving the Accuracy of RCCS Simulator Based on Ensemble Kalman Filter Algorithm

Abstract Data assimilation, an effective approach that utilizes measured data and suitable assimilation algorithms to reduce existing model calculation errors, has its origins in weather forecasting research. In this study, a framework of data assimilation based on Kalman filter (EnKF) algorithm is established for the computational program of the reactor cavity cooling system (RCCS), an important component of high temperature gas-cooled reactor-pebble-bed module (HTR-PM). Original model with uncertain parameters and conservative assumptions gives a result deviating from the measured data. After assimilation, parameters including the thermal conductivity of the water cooling wall, blackness coefficient, material reflectivity, resistance coefficients of various components, and the frontal area of the air cooler, are recalibrated. The modified model significantly improves the accuracy of predicting the residual heat removal power, reducing the calculation error from 16.89% to 1.22% compared to the original model. By applying the EnKF algorithm, this paper expands its application scope within nuclear engineering, providing inspiration for future relative studies.