Research And Application Of Power Communication Equipment Fault Diagnosis Method Based On Knowledge Graph

With the rapid development of the information construction of the national power grid company,the number of communication equipment of the company is increasing,and the difficulty of the disposal of power communication equipment faults is also increasing.If the accumulated historical fault text data can be organized and utilized,the fault information of power communication equipment can be platform integrated and the fault knowledge map of power communication equipment can be established,which can better solve the problem of power communication equipment fault disposal.Therefore,this paper researches the technology related to the construction of fault knowledge map of electric power communication equipment and the intelligent diagnosis method of electric power communication equipment fault,and the main research contents are as follows:(1)According to the characteristics of power communication equipment fault data,a method for constructing power communication equipment fault knowledge graph is proposed.Firstly,a fault entity recognition model of power communication equipment based on BERT-BiGRU-CRF is designed,and BERT(Bidirectional Encoder Representation from Transformers)is used as the vector embedding layer to obtain the vector sequence of fault text,and the obtained vector sequence is input to the Bidirectional Gated Recurrent Unit for semantic coding.Label constraint is carried out by combining conditional random field,so as to complete the construction of entity recognition model.Secondly,based on the acquired entity information and original data characteristics,through in-depth discussion and research with experts in related research fields,a fault relationship extraction method for power communication equipment based on artificial rules is designed to complete the construction of fault text relationship set.Finally,the "entityrelationship-entity" triplet is constructed through the obtained fault entity and relationship information of power communication equipment,and the Neo4j database is used to store the triplet knowledge to complete the construction of the power communication equipment fault knowledge graph.(2)An intelligent diagnosis model of power communication equipment fault based on knowledge graph is proposed by combining different severity fault information generated by power communication equipment in the operation process.Firstly,a WBLA(Word2vec-BiLSTMAttention)based fault severity level identification algorithm is designed to input the obtained fault warning information into the Word2vec(word to vector)model for word embedding vector characterization,and then the obtained vector information is input to the Bidirectional Long ShortTerm Memory neural network.The relevant features in the fault alarms are enhanced by Attention,and the Softmax function is used to achieve accurate recognition of the fault level,and then determine the severity of the fault and plan the order of fault management.A TFIDF-COS based intelligent diagnosis algorithm for power communication equipment fault,on the basis of planning the fault management order,vectorizing the obtained fault cause features,weighting the features using TFIDF model,using cosine similarity function to calculate the similarity,and obtaining the results with high reliability for fault diagnosis to realize the intelligent fault diagnosis of power communication equipment.Diagnosis.(3)Combining the constructed knowledge map of electric power communication equipment faults and the intelligent diagnosis method of electric power communication equipment faults,we design and develop the intelligent diagnosis system of electric power communication equipment faults based on the knowledge map.It realizes the construction of the knowledge map of power communication equipment faults,the recognition of the severity level of power communication equipment faults,the diagnosis of power communication equipment fault-related information and the formulation of disposal measures,so as to realize the intelligence of fault diagnosis and the efficiency of fault processing.