Construction of Avionics Display Equipment Fault Knowledge Graph Based on BiLSTM-CRF

With the development of avionics technology, fault diagnosis of avionics display equipment plays an increasingly important role in health management. However, due to the complexity of the software of civil aircraft display systems, their failure modes have become more and more complicated. Traditional fault diagnosis methods rely on logical reasoning and fault tree analysis after the occurrence of faults, which are inefficient and difficult to accurately locate the cause of faults when dealing with complex systems, and do not take into account the fact that there is a lot of valid information hidden in the large amount of testing and logging information generated during the development and testing process of civil aircraft display software. This study proposes a knowledge graph construction method for avionics display equipment health management based on BiLSTM-CRF, aiming to visualise health management related fault information. By using BiLSTM-CRF model for text extraction of fault information, constructing knowledge triples by means of knowledge matching and rules, and visualising health management related fault information in Neo4j, it provides an effective tool for health management of avionics equipment.