A maintenance warning must arrive early enough for the actual response to finish before a justified intervention boundary. Count the time to collect and deliver evidence, obtain attention, inspect, diagnose, approve work, obtain parts, secure access and complete the intervention. Fast notification is only one part of that chain.
Compare the available opportunity with the response path, including waiting and uncertainty. If a required duration or boundary is unknown, feasibility remains unresolved. A warning that looks early on a dashboard can still arrive too late for the people, materials and permissions needed to act.
Define the opportunity independently of the response plan
The useful opportunity begins when relevant evidence first becomes detectable by the intended method. It ends at the latest justified intervention point, not necessarily at complete failure. That boundary needs an asset-specific technical basis. A convenient service slot or a supplier's optimistic estimate cannot establish it.
Hypothetical worked scenario
Consider gear-tooth deterioration in G-08, the drive gearbox of a finished-goods roller conveyor. Loss of drive would interrupt material movement. The maintenance manager records the timing exercise as WTB-08 in the maintenance register, linking a separate qualified condition assessment and the proposed response stages.
- Starting point
- T0, earliest detectable evidence in this exercise.
- Supplied boundary
- Intervention completed by T0 plus 30 elapsed hours.
- Boundary authority
- Separate qualified engineering assessment, assumed for the exercise.
- Accountable owner
- Maintenance manager with qualified engineering and operations review.
All later durations are invented planning numbers too. No condition assessment has actually established this horizon for a real asset. It is neither an operating limit nor permission to continue running for 30 hours. Real use requires the qualified owner to substantiate the boundary and its uncertainty separately from this worksheet.
Separate evidence time from attention and completion
A source observation time, notification presentation time, human acknowledgement and verified work completion describe different events. Do not use a message's send time as evidence that a person received it. In an actual review, an unobserved delivery or acknowledgement time stays unknown.
T0 is not automatically the first timestamp in a database. Evidence may have become detectable before the first retained observation. Establish that distinction through a justified technical assessment or a controlled exercise. If it cannot be established, the available opportunity is uncertain rather than equal to the recorded history.
- T0: relevant evidence becomes detectable.
- Hour 2: collection and assessment produce the warning.
- Hour 3: notification is presented to the assigned role.
- Hour 6: that role acknowledges it.
- Hour 27: intervention and required completion checks finish.
- Hour 30: supplied intervention boundary.
The OPC Foundation's public acknowledgement definition concerns an operator recognizing an alarm. Applying that distinction here, an acknowledgement does not establish that inspection, repair or completion checks have happened.
The first six hours have already been consumed before inspection begins. Improving message delivery alone cannot recover time spent waiting for diagnosis or an unavailable crew. Record who is expected to respond, their coverage and the meaning of acknowledgement, rather than treating an accepted message as a completed maintenance outcome.
Build a complete elapsed-time budget
For the first calculation, assume every stage must finish before the next starts. Each duration includes its local waiting time. Keep the durations separate from cumulative finish times so they are not counted twice. All figures below are elapsed hours, including overnight time.
| Stage | Duration, hours | Finish after T0, hours |
|---|---|---|
| Collect and assess evidence | 2 | 2 |
| Deliver and present notification | 1 | 3 |
| Wait for acknowledgement | 3 | 6 |
| Inspect | 2 | 8 |
| Diagnose and specify work | 2 | 10 |
| Obtain work approval | 2 | 12 |
| Obtain required parts | 8 | 20 |
| Arrange authorized access | 4 | 24 |
| Complete work and verification | 3 | 27 |
The total is 2 + 1 + 3 + 2 + 2 + 2 + 8 + 4 + 3 = 27 hours. Subtracting it from the supplied opportunity gives 30 minus 27 = 3 hours of planning margin. That is arithmetic headroom, not evidence that the response is sufficiently reliable.
At notification presentation, hour 3, the boundary is 30 minus 3 = 27 hours away. The remaining response takes 27 minus 3 = 24 hours, leaving the same 3-hour margin. This is the distinction between available warning time at presentation and time still needed to complete the response.
The assumed people and suppliers must be available at the resulting clock times. These durations are not person-hours or shift lengths. Convert a quoted business-hours lead time into elapsed calendar time before comparing it with the opportunity. The listed waits are assumptions, not instructions to delay a ready response.
Completion includes the agreed verification, not merely finishing a physical adjustment. If authorization to resume service requires another review, include that dependency wherever it belongs. Avoid excluding administrative waiting because the hands-on repair appears short.
Follow dependencies instead of adding everything
Some stages can overlap; others cannot begin until a finding or authorization exists. The critical path is the dependent sequence that determines earliest completion under the stated resource and calendar constraints. Draw those relationships before assuming that parallel work will shorten the response.
- Common path: evidence through work approval takes 12 hours.
- Possible branches: parts take 8 hours; access arrangements take 4.
- Join: both branches must be complete before work starts.
- Finish: work and verification take 3 more hours.
Under that changed assumption, completion is 12 + max(8, 4) + 3 = 23 hours, leaving 30 minus 23 = 7 hours. The branch taking longer controls the join; adding both would count overlapping time twice.
Do not claim this improvement merely because two teams exist. Confirm that parts ordering and access planning can proceed together, with the required permissions and resources. A fixed shutdown slot needs calendar scheduling, not just this formula. Missing the slot could move completion later even when a task itself takes only a few hours.
Challenge the margin with delays and uncertainty
Return to the sequential case to test sensitivity without changing several assumptions at once. A parts estimate is not an observed delivery, and an average response time can conceal a shift queue or a weekend. Ask which delays are plausible for the actual staffing and supply arrangements.
| Parts duration, hours | Total response, hours | Margin against 30 hours |
|---|---|---|
| 8 | 27 | 3 hours. |
| 11 | 30 | Zero; no headroom. |
| 12 | 31 | Minus 1 hour; completion is late. |
| Unknown | Unresolved | Cannot establish feasibility. |
If parts take 8 to 12 hours, these assumptions produce 27 to 31 hours of response and margin from 3 to minus 1 hour. This is a sensitivity range, not a statistical confidence interval. Finishing exactly at the boundary provides no allowance for uncertainty.
Delays can also share a cause. One busy shutdown can consume both the inspection crew and the access window. Reviewing each average independently would miss that combined constraint; examine the dependent schedule under the same operating situation.
The opportunity itself may be uncertain, especially if the earliest detectable point is not known. Do not compensate by using a later starting timestamp or extending the boundary to fit the plan. Return an unsupported boundary or duration to the qualified owner as an unresolved feasibility question.
Separate warning quality from response capacity
A warning can identify a useful condition while the organization lacks the capacity to act. Conversely, a fully staffed response cannot rescue a warning first produced after the justified boundary. Diagnose these as different limitations before buying a faster model.
| Finding | Limitation | Next review |
|---|---|---|
| Evidence appears too late | Detection opportunity. | Can a justified earlier indication change the outcome? |
| Early evidence, unavailable crew | Response capacity. | Review staffing and authorized service arrangements. |
| Many unhelpful warnings | Attention and investigation burden. | Review usefulness without weakening required alarms. |
| Boundary or delays unsupported | Unknown feasibility. | Assign the missing evidence; do not claim success. |
HSE's alarm-management principles call for useful alarms with defined responses and sufficient time to act. Applying that principle to this maintenance exercise means considering attention and response effort alongside detection.
Unnecessary investigations can occupy the same people needed for consequential work. Review that burden with operations, while preserving mandatory, original equipment manufacturer and qualified safety requirements. This budget does not authorize alarm changes, maintenance delays or operation until failure.
Test the complete path before relying on the warning
Use a tabletop exercise or a separately authorized, non-disruptive rehearsal. Do not create an equipment fault to test this worksheet. Replace planning assumptions with traceable observations where available, and keep untested periods and missing milestones visible.
| Test question | Evidence to retain |
|---|---|
| What starts and ends the opportunity? | Qualified basis, event definition and uncertainty. |
| When did attention actually occur? | Source, notification presentation and acknowledgement times. |
| What controlled completion? | Dependencies, queues, parts, access and verified finish time. |
| What remains unresolved? | Named reviewer, missing evidence and next review date. |
Conventional software can calculate elapsed time and expose overdue stages. AI can summarize referenced records and inconsistencies. Qualified people retain boundary, work and return-to-service authority. The useful business question is whether this warning can change a maintenance outcome with the response capacity actually available.
WTB-08 remains a planning exercise until the engineer substantiates the opportunity and the maintenance planner checks queues, coverage and overlap assumptions. Assign those unresolved questions to named reviewers with a review date. Existing approved responses continue to govern while feasibility remains unproven.