Feature
Knowledge Graph
A semantic OWL ontology encodes decades of industrial maintenance expertise into machine-readable rules — giving DiagAI agents the contextual understanding no pure ML model can match.
50+
Entity types
OWL
Ontology format
SWRL
Rule language
Capabilities
Semantic Reasoning
SWRL rules encode expert knowledge like "if torque > 2σ AND temp_diff > 1.5σ THEN probable bearing failure" — logic no ML model alone can express.
Causal Chain Mapping
The graph stores directed causal edges between entities. When an anomaly occurs, agents traverse causal chains to find the most likely origin.
Cross-Domain Patterns
Failure patterns learned from one machine type are transferable to similar equipment via OWL class hierarchies and property inheritance.
Contextual Alerts
Instead of 'anomaly detected', DiagAI returns 'probable bearing wear caused by thermal overload — see recommended maintenance action #3'.
Entity types in the ontology
EquipmentSensorFailure ModeSymptomRoot CauseMaintenance ActionComponentEnvironmentOperatorProcess+40 more
Sample SWRL rule
// Bearing overload detection rule
Equipment(?e) ∧ hasTorqueZScore(?e, ?t) ∧ greaterThan(?t, 2.0)
hasTempDiffZScore(?e, ?d) ∧ greaterThan(?d, 1.5)
→ probableCause(?e, BearingWear)