Summary
In short
- Record what failed and what was done as separate fields. Merging them produces a note that reads as an action and cannot be analysed as a failure.
- Failure coding is what turns corrective history into reliability data. Without a controlled vocabulary the same failure appears under a dozen descriptions and never trends.
- Corrective work is the feedback signal for the preventive regime. Failures on assets under PM indicate the PM is wrong, mistimed or not being done.
- Safety and food safety controls applied during the work belong on the record, because the restart decision depends on them.
- Verification that the asset works is not verification that the cause was addressed. A temporary repair recorded as complete disappears from view.
- Time to repair splits into waiting and working. Where waiting dominates, the problem is spares or access rather than technical capability.
What it is
What it is
What is a corrective work order?
A record of unplanned work to restore an asset to working condition following a failure or fault. It captures what was found, what was done, what parts were used, how long it took, and the controls applied while the work was performed.
Why does the failure description matter so much?
Because the corrective history is the only systematic record of how assets actually fail. Analysis of recurring failures, of which components dominate downtime, and of whether a preventive regime is working all depend on the fault being described consistently. Free text saying broken supports none of it.
When to use it
When to use it, and when not to
This covers unplanned corrective work. Planned maintenance and improvement work have their own records.
Use it for
- Unplanned work to restore function after a failure, fault or performance loss
- Faults found during inspection or operator checks requiring corrective action
- Temporary repairs, recorded as temporary with the permanent fix scheduled
- Work arising from an incident, near miss or audit finding against an asset
- Any intervention where the asset was not performing as intended
Not for
- Preventive maintenance work orders, which are scheduled and have different analysis value
- Modification and improvement work, which changes the asset rather than restoring it
- The permit to work, which authorises the activity where required
- Root cause analysis, which belongs separately for significant or recurring failures
- Asset criticality assessment, which determines how the asset is maintained overall
Standards
What it is built against
Corrective maintenance sits under asset management requirements and, in regulated production, under equipment and hygiene obligations.
| Clause | Requirement | Where it lands |
|---|---|---|
| ISO 55001 cl.8.1 | Operational planning and control of activities to achieve asset management objectives | Header |
| ISO 55001 cl.9.1 | Evaluation of asset performance and the effectiveness of asset management | Verification |
| ISO 9001 cl.7.1.3 | Infrastructure maintained to ensure conformity of products and services | Repair |
| 21 CFR 117.40 | Plant and equipment maintained in a condition preventing contamination of food | Safety and food controls |
| 29 CFR 1910.147 | Energy control applied during servicing and maintenance, including corrective work | Safety and food controls |
| BRCGS Issue 9 cl.4.7 | Maintenance with hygiene clearance and sign-off before equipment returns to production | Verification |
| NFPA 70B 2023 | Documented electrical maintenance programme, with corrective work forming part of the record | Repair |
| ISO 14224 | Collection and exchange of reliability and maintenance data, including failure and cause coding | Fault detail |
What it does not cover
- Preventive maintenance records, which are scheduled work with different analytical value.
- Modification and improvement work, which changes the asset and may trigger conformity questions.
- Permit to work, authorising the activity where the work is high risk.
- Root cause analysis, for significant or recurring failures.
- Hygiene clearance in food production, which is a specific release step before restart.
Filling it in
Filling it in well
Four fields carry the analytical value, and one of them is usually a single line of prose.
What was found: which component, what condition, what symptom. What was done: what was replaced, adjusted or repaired. These are different pieces of information and merging them makes both unusable. The fault supports reliability analysis; the action supports cost and spares planning.
Component, failure mode and, where known, cause. ISO 14224 provides a structure and any consistent internal list works. Without coding, the same bearing failure appears as bearing gone, noisy bearing, drive end failure and seized, and none of them aggregate into a number anyone can act on.
Isolation, permits, hygiene measures, temporary guarding removal and its reinstatement. In food production the hygiene clearance before restart is a specific release step, and in any setting the record of what was removed determines what must be restored before the asset runs again.
A repair intended to hold until the next shutdown should be flagged as temporary with the permanent work raised at the same moment. Temporary repairs recorded as complete become invisible, and they are found later either by the next failure or by an auditor asking why a hose clip is holding a guard on.
Audit findings
Common audit findings
Corrective work order findings mostly concern data quality, which is why they persist.
| Finding | Clause | What fixes it |
|---|---|---|
| Fault and action recorded in a single free text field. | ISO 14224 | Separate them; merged text supports neither reliability nor cost analysis. |
| No failure coding, so recurring failures never aggregate. | ISO 55001 cl.9.1 | Code component and failure mode against a controlled list. |
| Temporary repairs closed as complete. | ISO 55001 cl.8.1 | Flag as temporary and raise the permanent work at the same time. |
| Corrective failures on assets under preventive maintenance not fed back. | ISO 55001 cl.9.1 | A failure under PM means the PM is wrong, mistimed or not being performed. |
| Energy control not recorded for corrective work. | 1910.147 | Corrective work is servicing; the isolation applies and belongs on the record. |
| Hygiene clearance not evidenced before restart in food areas. | BRCGS 4.7 | Clearance is a release step, not a formality; record it explicitly. |
| Downtime recorded as a single figure with no split between waiting and working. | ISO 55001 cl.9.1 | Split it; where waiting dominates, the fix is spares or access rather than skills. |
| Guards or interlocks removed during work with no reinstatement verification. | 1910.212 | Record removal and verify reinstatement before restart. |
| Recurring failures on one asset never triggering analysis. | ISO 55001 cl.9.1 | Set a threshold that raises root cause analysis automatically. |
| Parts used not recorded against the asset. | ISO 55001 cl.8.1 | Parts history is the fastest route to identifying a component that keeps failing. |
Worked case
Case in point: eleven ways to say the same thing
A plant reviewed two years of corrective work orders on a filling line to understand why availability was falling. The system held over four hundred corrective jobs against the line, each with a completed notes field.
The most frequent failure, once someone read every entry, was a particular drive coupling. It appeared as coupling worn, coupling replaced, drive noise, drive fault, replaced coupling and bolts, vibration on drive, and five other descriptions, spread across four technicians and two shifts. No report had ever surfaced it because no two entries matched.
The line had been under a preventive regime throughout, and the coupling was not on it.
Definitions
Definitions and key terms
- Corrective maintenance
- Unplanned work to restore an asset to working condition after a failure or fault.
- Failure mode
- The manner in which a component fails, coded consistently so that recurrences aggregate.
- Mean time between failures
- A reliability measure only computable where failures are recorded consistently against the asset.
- Temporary repair
- A fix intended to hold until a scheduled opportunity, requiring the permanent work to be raised at the same time.
- Hygiene clearance
- The release step in food production confirming an asset is clean and safe to return to production after maintenance.
- Waiting time
- Downtime spent waiting for parts, access or authorisation, distinct from time spent working on the asset.
- ISO 14224
- The standard for collecting and exchanging reliability and maintenance data, providing failure coding structures.
- PM effectiveness
- Whether the preventive regime is preventing failures, measured by corrective work on assets that are under it.
FAQ
Frequently asked questions
Why separate the fault from the action?+
Because they answer different questions. The fault, which component and what condition, supports reliability analysis and tells you what to prevent. The action, what was replaced or repaired, supports cost, labour and spares planning. Merged into one note, they support neither, and the asset history becomes prose that must be read rather than data that can be queried.
What does failure coding actually give us?+
The ability to see recurrence. Without a controlled list, the same failure is described a dozen ways by different technicians and never aggregates, so the dominant failure mode on an asset stays invisible in a history that records it repeatedly. Coding takes seconds at the point of closing and is what converts maintenance activity into reliability information.
What should corrective work tell us about preventive maintenance?+
Whether it is working. A failure on an asset under a preventive regime means the PM did not cover that component, was scheduled at the wrong interval, or was not actually performed. That feedback loop is the main reason to analyse corrective work, and in most operations it does not exist.
How should temporary repairs be handled?+
Flagged as temporary on the record, with the permanent work raised at the same moment rather than later. A temporary fix closed as complete becomes invisible immediately, and it is then found either by the failure it causes or by someone asking why a component is held on with a hose clip.
Why split downtime into waiting and working?+
Because they have different fixes. Where the technician was working for twenty minutes and waiting three hours for a part, the constraint is stores rather than skills, and training or headcount will not touch it. Most systems record a single downtime figure and conceal which of the two is the actual problem.
The agents
What the agents do with it
The work order is a record of a repair. What fails is the coding that never happened and the temporary fix that closed as complete.
Separates fault from action, applies failure coding at closure, and flags temporary repairs with the permanent work raised alongside.
Aggregates coded failures by asset and component, surfacing dominant failure modes that free text history conceals.
Links isolation, permits and guard removal to the job, and requires reinstatement verification before restart.
Holds hygiene clearance as a release step in food areas, so restart depends on it rather than on the repair being finished.
This template lives in KnowMaintain — asset maintenance. Work orders, planned maintenance, calibration, reliability and shutdowns.
Meet KnowMaintain→Sources
Sources
- ISO 55001:2014, asset management, clauses 8.1 and 9.1
- ISO 14224:2016, collection and exchange of reliability and maintenance data
- 21 CFR 117.40, plant and equipment requirements, FDA
- 29 CFR 1910.147, control of hazardous energy, OSHA
- BRCGS Food Safety Issue 9 clause 4.7, maintenance