Rolling stock maintenance is essential to ensure train reliability. Adapting automation to optimise the efficiency of train maintenance procedures is crucial to maintain a smooth operation of rail transport. Information related to the scheduling and planning requirements for train maintenance procedures, however, adhere to task-specific schemas or worse, exists in unstructured formats, which hinders computerised analysis and impedes interoperability. To fill this gap, a knowledge acquisition pipeline is introduced to semantically organise textual information found in semi-structured maintenance instruction manuals. XML transformations, pattern matching and elastic search over industrial vocabularies are utilised to automatically populate a domain-specific ontology and construct a knowledge graph that describes maintenance tasks. Through demonstration, we exemplify how the resulting knowledge source can be accessed and used in Short-Term Scheduling operations.
|Journal||CEUR Workshop Proceedings|
|Publication status||Accepted/In press - 1 May 2023|
|Event||PLanning And onTology wOrkshop: In conjunction with 33rd International Conference on Automated Planning and Scheduling - Prague, Czech Republic|
Duration: 9 Jul 2023 → 10 Jul 2023
Conference number: 33