Maintenance analysis fails when the alarm tag, physical asset, failure event, work order and fitted spare cannot be joined reliably. A maintenance data map defines identities, relationships, time rules and ownership so evidence from operations and maintenance describes the same equipment history.
The short answer
Choose a governed equipment identifier and hierarchy, then map every source identifier to it without discarding the source value. Separate symptoms, alarms, functional failures, failure modes, causes, work requests, work orders, tasks, downtime and replaced parts. Define event start and end rules, clocks, status vocabulary, units, coding authority and correction workflow. Preserve the software and document baseline effective at the event time. Validate joins with real cases and publish ownership for unmatched and duplicated records.
- One physical asset has a governed identity across systems.
- Alarm, symptom, failure and cause remain distinct concepts.
- Time and status semantics are documented before aggregation.
- Corrections are traceable and do not rewrite source history silently.
Decision guide
| Record | Key relationship | Required context |
|---|---|---|
| Alarm/event | Source point to affected asset | State, priority, quality and time |
| Failure | Asset function to observed loss and mode | Detection, effect, cause confidence and downtime |
| Work order | Failure or task to performed maintenance | Status, labor, procedure and completion |
| Spare | Removed and installed item to asset position | Part, lot/serial, quantity and effective time |
A practical workflow
- 01
Inventory systems, owners, identifiers, vocabularies and retention.
- 02
Define canonical asset hierarchy and cross-reference tables.
- 03
Model event, failure, work and spare relationships with time rules.
- 04
Test historical cases, unmatched records, duplicates and corrections.
- 05
Publish data contracts, quality KPIs, exception owners and change governance.
Evidence and acceptance
Retain source-system inventory, asset hierarchy revision, identifier map, vocabulary and code lists, event and time semantics, field lineage, transformation rules, sample joins, reconciliation totals, unmatched and duplicate queues, correction approvals, data-quality trend, access classification, interface versions and consumer acceptance.
Common mistakes
- Matching records only by free-text equipment name.
- Treating every alarm as a failure.
- Using work-order close date as failure end without definition.
- Losing removed-versus-installed spare identity.
- Changing hierarchy without effective dates and downstream review.
Responsibility boundary
Primary sources
Technical claims and product boundaries were checked against these primary sources. Fact-checked
Frequently asked questions
Which identifier should be canonical?
Use the governed asset identifier that remains stable through the intended lifecycle and maintain explicit cross-references to source IDs.
Are alarms failure records?
No. Alarms are observations or calls for action; a failure record describes loss of required function with its own evidence.
How are hierarchy changes handled?
Use controlled revisions and effective dates so historical events retain their original context.
Does AutoNexit replace a CMMS?
No. It preserves engineering context and project evidence alongside links to operational systems.