Research On The Construction And Application Of Knowledge Graph For Fault Diagnosis Of Electric Drone

In recent years,the safety and security of electric UAV operations have been facing serious challenges.Although the traditional fault diagnosis methods can achieve good results with the support of sufficient data,they also have the shortcomings of wasted domain knowledge,lack of guaranteed explanation and difficulty in grasping the fault association among UAV components,and are gradually unable to meet the increasingly complex work requirements.Therefore,this paper integrates knowledge graphs into the field of electric UAV fault diagnosis based on the emerging research hotspot,and realizes the rational and evidence-based electric UAV component association fault diagnosis based on knowledge drive.Therefore,this paper implements electric UAV fault diagnosis based on the emerging research hotspot of the knowledge graph.At present,there are few studies on fault diagnosis knowledge graphs,and the problem of insufficient training data for deep learning models is usually solved by using "pre-training" models,but this method is more restricted in application scenarios and cannot provide valuable training samples for subsequent researchers.Therefore,the construction and application of the knowledge graph for electric UAV fault diagnosis proposed in this paper are as follows:1.To address the problem of insufficient training data for deep learning models,a remotely-supervised machine-labeled human school-based fault diagnosis knowledge extraction data annotation method is proposed.Based on the external a priori knowledge base,the entity relationships between electric UAV fault domains are predefined,heuristic rules entity relationship alignment is performed with the support of the knowledge base,and secondary manual proofreading is performed for the noise data generated by remote supervision,which provides a large amount of trainable corpus for the construction of UAV fault diagnosis knowledge graphs.2.Aiming at the problem that the content knowledge granularity of entity features and relevant information rules of electric UAV fault knowledge is too large to be linked,it is proposed to construct and study the knowledge graph of electric UAV fault information,and adopt the rule-based method and Bi LSTM-CRF model respectively according to the structural characteristics of electric UAV fault data to ensure the accuracy and precision of knowledge extraction.This is also used to achieve autonomy in knowledge extraction and reduce the degree of manual dependency.The data are filled into the fault ontology model and transformed into RDF language to deposit into the graph database Neo4 j to form the final knowledge graph for electric UAV fault diagnosis.3.To verify the effectiveness of knowledge graph in the field of fault diagnosis,design and implement an intelligent question and answer system for fault diagnosis of electric UAV systems,provide reasoned and accurate diagnosis of electric UAV faults,and provide a scientific basis for the construction of a fault diagnosis system based on knowledge graph.