Construction and application of knowledge graph for intelligent decision-making of power grid fault handling

In view of the increasing pressure of power grid fault handling due to complex meteorological and uncertain factors in regional power grid, a knowledge graph construction method for intelligent decision-making of power grid fault handling is proposed. The key entity information of power grid fault handling plan is identified based on the bidirectional encoder representations from transformers, bidirectional long short-term memory network, and conditional random field (BERT-BiLSTM-CRF). The fault handling knowledge map is established by syntactic relation mapping between entities. In order to solve the emergency control and fault recovery requirements when power grid faults are abnormal, an application scheme of intelligent assisted decision push based on fault handling knowledge graph is proposed. According to the data validation of a control center, compared with other methods, the proposed model has a higher accuracy of entity identification, and the knowledge graph can provide auxiliary decision support for the dispatchers under the abnormal fault scenarios.