# Plan the maintenance window

An illustrative explanation of the selected contract, not an execution result.

![North Basin’s synthetic failure history and inspection evidence inform Run, Service and Replace options under stated action assumptions. The same actions feed crew scheduling and usable-spares planning, joined as a maintenance proposal. Evidence gaps and acceptance remain visible.](https://openprose.ai/contract-views/v1/equipment-maintenance/49fdd1a49ae4eceb98c71af951c9b4c652296d2c62e9560ac7624ee505f4d263.png)

North Basin's fictional commission combines failure/exposure records with current condition evidence. Those are distinct evidence streams. An estimate from one historical cohort does not automatically describe another asset's present condition or the risk of the whole fleet.

Run, Service and Replace have expressly supplied restoration assumptions. Service need not remove a defect or reset age; replacing one component need not rejuvenate the whole machine. These assumptions and the selected decision policy govern the prospective risk/cost comparison without claiming that a benefit has been measured.

The proposed actions must reconcile with crew intervals, mandatory tasks and the same usable spare inventory. Quarantined stock contributes no usable capacity without release evidence. Planned consumption and reserve stock must remain distinct so part quantities and costs are counted consistently.

Source correspondence:

- [Fleet decision](https://pkg.prose.md/releases/0.4.0-rc.1/reliability/contracts/fleet-decision.md) connects inference, inspection, action choice, scheduling and spares as one proposal.
- [North Basin's commission](https://pkg.prose.md/releases/0.4.0-rc.1/reliability/tests/fleet/program.md) requires the actual calculations, mandatory tasks and proposed work/stock lines.
- [Synthetic records](https://pkg.prose.md/releases/0.4.0-rc.1/reliability/tests/fleet/source.md) separate historical cohorts H, condition evidence C, action assumptions A, calendars T, stock P and future evidence E.

The board omits probabilities, expected costs, selected actions and the detailed reserve calculation. It does not claim a feasible schedule or reliable physical equipment. Those findings need the commissioned calculations and evidence, including adverse or incomplete results. A supported rejection can complete an analytical decision while leaving a feasible-action goal unmet. Recommendations do not perform maintenance, purchase spares, establish accepted operating authority or schedule future monitoring; actual intervention and effectiveness require their own evidence.

## Source basis

Source contract: [Produce a coordinated fleet decision](https://pkg.prose.md/releases/0.4.0-rc.1/reliability/contracts/fleet-decision.md) from `0.4.0-rc.1`, commit `a21646f67fa22e171cc70600437317e274ab4628`. This later view is not part of that release.

Contract SHA-256: `1ba80756d0da7469fdf54cc07a6194e55b6c4eee161296e459aaee799a1f5d7a`.

Image SHA-256: `49fdd1a49ae4eceb98c71af951c9b4c652296d2c62e9560ac7624ee505f4d263`.

If packaged, the text is included; the PNG stays separately hosted. This view adds no requirements.