Knowledge Graph Construction for Fault Diagnosis of Aircraft Environmental Control System

Abstract

With the continuous improvement of the aircraft environmental control system, the content of maintenance manuals on which the maintenance work is based are constantly enriched, causing inconvenience of quick fault positioning. Maintenance engineers’ experience and knowledge are often required and the labor cost of maintenance work increases. For improving the utilization efficiency of these resources, this paper uses the deep text matching model, BERT, to extract semantic information in the maintenance record provided by China Eastern Airlines. After obtaining the entities of warnings and fault causes and the relationship between them, a knowledge graph for fault diagnosis of civil aircraft environmental control system is constructed. And a fault diagnosis support algorithm is completed, which is conducive to improving fault location and reducing aircraft maintenance costs.