Aviation maintenance is one of those domains where the same engine has a different ID in the maintenance system, the parts ERP, the OEM telemetry feed, and the
regulator's AD database. A question like "which of our tails will breach an open Airworthiness Directive in the next 30 cycles" turns into a multi-day
spreadsheet exercise, and it's the kind of question you can't afford to get wrong. That mismatch is what an ontology should fix.
Airworthy has 18 entity types covering aircraft, engines, components, parts, flight cycles, sensor readings, work orders, task cards, defects, mechanics,
certifications, MRO facilities, ADs, service bulletins, manufacturers, operators, lessors, and lease contracts. Around 30 typed relationships connect them, with cardinalities and identifier bindings on each entity.
I picked this domain because aviation is a setting where ontology grounding is genuinely safety-critical. The
goal is to use it as the context layer for a Fabric Data Agent that can answer fleet compliance questions in plain English instead of joining CSVs by hand.
https%3A%2F%2Fmicrosoft.github.io%2FOntology-Playground%2F%23%2F
