Research on fault diagnosis of CTCS on-board equipment based on knowledge graph

On-board equipment is a critical component of the Chinese train control system (CTCS), characterized by a complex organizational structure. Traditional CTCS on-board equipment faces three major challenges: low efficiency in fault diagnosis, difficulties in reusing diagnosis knowledge, and high labor costs. These drawbacks necessitate the adoption of intelligent means to quickly identify fault causes, extract fault diagnosis methods, and assist operation and maintenance personnel through decision support. Based on an analysis of the limitations of existing fault diagnosis techniques and an in-depth study of knowledge graph technology, this paper proposes an architectural design for developing fault diagnosis knowledge graphs for on-board equipment based on a middle platform and microservices. Merits from both the bottom-up and top-down approaches were combined to construct these knowledge graphs. The graphs allowed for the graphic storage and visual display of fault diagnosis data for CTCS on-board equipment across multiple stages, including knowledge extraction, ontology construction, knowledge reasoning, and knowledge storage. A domain model was established by adopting a domain-driven design method in the knowledge extraction stage, which was vital to establishing these graphs. In the crucial knowledge reasoning stage, a deduction lattice algorithm was introduced to create a rule decision tree that supported quick inference of equipment failures. The knowledge graph model layer and data layer are separated by a business middle platform and a data middle platform within the entire system, which is ultimately constructed following a micro-service architecture. Subsequent applications show that the established fault knowledge graphs for CTCS on-board equipment effectively facilitate the retrieval of data related to on-board equipment faults, provide fault analysis and diagnosis methods, and support operation and maintenance personnel in intelligent decisions on fault diagnosis, thereby improving operational and maintenance efficiency and reducing maintenance costs.