XGBoost-Based Intelligent Decision Making of HVDC System with Knowledge Graph

This study aims to achieve intelligent decision making in HVDC systems the framework of knowledge graphs (KGs). First, whole life cycle KG an system was established by combining making. Then, fault diagnosis studied as a typical case study, and decision-making method for based on XGBoost that significantly improved speed, accuracy, robustness designed. It is noteworthy dataset used this extracted KGs, accordingly combined. Four kinds data from KGs were firstly preprocessed, their features simultaneously trained. sensitive weights set, pre-computed sample put into model training. Finally, trained test set substituted classification after training obtain results, recognition accuracy calculated means comparison with standard labels. To further verify effectiveness proposed method, back propagation (BP) neural network, probabilistic network (PNN), tree adopted validation same dataset. The experimental results show paper could over 87% multiple groups tests, being higher than those its competitors. Therefore, can effectively identify diagnose faults under different operation conditions.