Maintenance Windows & Availability

Planned maintenance is scheduled into the redundancy margin so it never reduces the power available to the mission.

Planned downtime is still downtime

Availability counts both unplanned and planned outages. A unit taken down for scheduled maintenance is not producing, so maintenance must be scheduled so that it never coincides with reduced margin. The +1 unit exists precisely so that one unit can be worked on while the rest carry the load.

Scheduling against the margin

Maintenance is planned into the redundancy: only as many units are taken offline at once as the +1 (or +2) margin allows, and never during a period of elevated risk. Condition monitoring guides timing so that work is done before a component fails rather than after, converting potential unplanned outages into scheduled ones.

• Schedule maintenance within the N+1 margin
• Never remove more units than the margin covers
• Use condition monitoring to plan before failure
• Coordinate across units to avoid concurrent outages

Closing the gate honestly

The single-unit availability of 0.86–0.995 already includes maintenance time. The way an installation reaches high system availability is by absorbing every planned outage into spare capacity, so that from the mission's perspective the power is continuous even while individual units cycle through service.

Maintenance planning is design-stage and site-specific, and its central rule is that planned outages are absorbed entirely by the redundancy margin, so from the mission's perspective power is continuous even as individual units cycle through service. The practical rule is simple: the number of units allowed offline for planned work at any moment never exceeds the redundancy margin, and high-risk periods are kept clear of scheduled outages entirely.