Planned outage management for power plants

01Not the other outage management

In utilities, "outage management system" usually means the software that tracks distribution outages and restores customers after a storm. This guide is about the other kind: the planned maintenance outage on a generating unit, when a boiler, turbine, generator or combined-cycle train is taken off line for inspection, repair and upgrade. The planning problem is the same as a refinery turnaround, with a fixed window and a price on every hour.

02What makes a generation outage different

03Scope and the long-lead calendar

Build the scope from the last outage's findings, condition monitoring, OEM bulletins and the inspection interval. Freeze it 12 months or more ahead; on a major turbine or generator outage the long-lead parts were ordered before the scope was frozen, so the freeze is partly a confirmation of what is already on order. Keep one change process, and make every addition after the freeze visible the day it is added, with its hours and cost.

For each scope item record the work package, the long-lead flag and the lead time. Spare parts policy matters here: a unit with a spare rotor plans a different outage from one without.

04Work packages for a compressed window

A generation outage does most of its work in a short span, so the packages have to be execution-ready before the unit comes off line. Each package needs its steps, crew, duration, isolations, permits and drawings, and the activity date from the schedule. Readiness is calculated from the packages, not asserted: the share of packages complete, the share with material staged, and the list of exceptions.

05Material: parts staged against the day

Follow every item from requisition and purchase order through receipt, staging, issuing and installation, with the package and the activity date on each. Two controls carry the outage:

1. Shortages shown against the activity date, so a part promised for day eight that the schedule needs on day five is visible in planning, not on day five.

2. Material test reports and inspection records kept with the item, so a pressure part can be proven years later.

Material is a large share of outage spend and tracking usually stops at the warehouse dock. On an outage with OEM-supplied parts arriving on their own paperwork, that gap is where the surprises are.

06Contractors, OEMs and daily cost

Contractors and OEM crews estimate and report in one format inside the owner's system, so bids compare line by line and the daily cost report exists at the end of the shift. Track headcount planned against actual by contractor by day. On a compressed window, a crew that runs over on day three cannot be recovered on day twenty; the daily headcount is the earliest signal you will get.

The forecast at completion moves by shift. Scope changes are approved before the work starts and reach the forecast the same day. Leadership should be able to answer "what will this outage cost and when does the unit return" at the start of any shift.

07Execution and return to service

Capture progress, QA and discovery work in the field, by shift, on a phone. Close each package within hours of the work. The return-to-service decision then rests on closed packages and completed inspections, not on a verbal round-up. On one site, field capture with QA notified at once moved close-out from over 72 hours to under 24.

After restart, reconcile cost, close purchase orders, and keep the packages. The next outage on the same unit starts from this one.

08A four-week outage, by the calendar

09Common mistakes on generation outages

1. Treating the OEM's schedule as outside the plan, so its slips surprise everyone.

2. Receiving OEM parts on the OEM's paperwork and never tying them to a package.

3. Discovery scope approved verbally in week one and costed in week four.

4. Daily headcount not compared with the plan until the invoices arrive.

5. The return-to-service decision made from a verbal round-up rather than closed packages.

10The number that matters

On a generation outage the whole plan reduces to one number: the day the unit returns to service, and the confidence in it. Every control above exists to make that number true earlier. Packages complete and material staged before the unit comes off line protect the first week. Daily headcount and a forecast that moves by shift protect the middle weeks, when a slip can still be recovered with crews and overtime. Closed packages and completed inspections protect the last days, when the decision to synchronize rests on evidence rather than a round-up. A site that can state the return date and its confidence at the start of every shift, from the system, is running the outage. A site that finds out on the last day is being run by it.