Prediction and Interpretability Analysis of Key Parameters in Nuclear Power Plant Small-Break LOCA Using LightGBM
Bo Pang, Guoxu Qin, Yuanfeng Lin, Qingyu Huang, Y L Zhang, Siyuan Zhang, Qingzhong Ai, Guanghui Yuan, Jingyi Wan
The full-scope simulator plays a critical role in nuclear power plant emergency drills, personnel training, and accident analysis. Traditional system programs lack sufficient computational performance to meet real-time requirements when simulating complex accident scenarios in reactor systems. This study focuses on the small-break loss-of-coolant accident (SBLOCA) in nuclear power plants, generating large-scale datasets through digital simulations. After data preprocessing and normalization, a light gradient boosting decision tree (LightGBM) regression model was developed using machine learning algorithms. SHAP (SHapley Additive exPlanations) analysis identified the contributing factors, enabling the model to predict key parameters such as peak fuel cladding temperature, primary reactor coolant pressure, and pressurizer water level. The model achieved a mean square error (MSE) below 0.002 and a coefficient of determination (R2) exceeding 0.98, with a prediction speed approximately 32,500 times faster than traditional system programs, requiring less than 4×10−4 seconds per data point. This study provides a novel solution for complex condition simulations in nuclear power plant full-scope simulators.