THE WHOLE PLANT IMPERATIVE • PART 1
For twenty years, deferring maintenance was a safe bet. On today’s grid, it may be the most expensive one you can make.
Start with the worst possible afternoon. It’s the hottest day of the summer, the grid is leaning on every megawatt it can find, and prices are pinned near the cap — when your unit trips on a tube leak that first showed up as a hairline indication eighteen months ago, on a list nobody got back to.
Ten years ago, that tube leak was a deferral line and a shrug. Today it’s the difference between a record revenue day and a seven-figure hole. Nothing about the tube changed. Everything around it did.
That gap — between what a lapse used to cost and what it costs now — is where this series begins. It quietly reframes every maintenance decision that comes after.
A maintenance backlog used to be a budgeting problem. You deferred a few items, carried a little more risk, and the plant kept running. With comfortable reserve margins and steady baseload demand, the grid could shrug off a unit that wasn’t quite at its best.
That tolerance is gone.
Three forces changed that, all at once. Together they moved maintenance out of the budget meeting and into the risk register. If your strategy still runs on the old stakes, your exposure is far bigger than the line item shows.
The grid needs every megawatt you’ve got
Flat demand is over. Load growth is back, and it’s steep — data centers and AI are pulling power faster than utilities modeled, electrification is stacking load onto the system, and demand forecasts are being revised up for the first time in a generation. At the same time, plants that were supposed to retire are being told to keep running. The result: your fleet is load-bearing again in a way it hasn’t been in decades. A megawatt you can’t deliver because of a preventable failure isn’t lost revenue — it’s capacity the grid was counting on and didn’t get.
For years, the system carried enough slack to cover a weak unit. Reserve margins were comfortable; if you limped, someone else picked up the load. That cushion is thinning — retirements are outrunning new dispatchable capacity, and not all of the replacements show up when demand actually peaks. The margin that used to hide an underperforming plant is disappearing, and it isn’t coming back on the timeline your maintenance plan quietly assumes.
You’re running the machine harder than it was built for
The plants carrying that load were built for steady baseload. They don’t run that way anymore. To chase intermittent renewables, thermal units cycle — start, stop, ramp, repeat — far more than their designers ever intended. Every cycle drives thermal and mechanical fatigue straight into the systems maintenance exists to protect: HRSG tubes, headers, attemperators, casings. Cycling doesn’t just age the plant faster. It accelerates the exact degradation that turns a skipped inspection into a forced outage.
And this isn’t a peaker problem anymore. Units that once ran flat as baseload — designed around a few hundred starts across their whole life — are now piling up starts and load swings at multiples of what their metallurgy was specified for. The damage is cumulative and nearly invisible: creep-fatigue in headers, cracking at tube-to-header welds, distortion working into casings and attemperators. It doesn’t announce itself. It surfaces as the forced outage you didn’t schedule, on the system you weren’t watching.
The math turned brutal
Here’s the shift that matters most — and the one least reflected in maintenance budgets: an unplanned outage no longer just costs the repair. In a capacity market, missing a scarcity event can trigger penalties that dwarf the fix, and in the worst cases exceed a unit’s entire annual capacity payment. In Winter Storm Elliott, PJM generators faced non-performance charges measured in the billions. Add the scarcity revenue you forfeit while you’re down — the hours power is worth the most are exactly the hours you can’t afford to be offline — and the cost of a forced outage comes completely unhooked from the cost of the fix. The repair might be fifty thousand dollars. The event around it can be seven figures.
And the penalty isn’t the whole bill. A trip at the wrong moment can put you crosswise with pay-for-performance capacity obligations, PPA availability guarantees, and the reliability metrics regulators now track by name. The economics have gone asymmetric: the savings from deferring a job are small, known, and booked this year; the cost of the failure it invites is large, uncertain, and arrives exactly when you can least absorb it. That’s a poor trade in any market. In this one, it’s a dangerous one.
Put the three together and the conclusion is hard to dodge. Deferring maintenance used to cost, roughly, the price of the maintenance. Now it costs the outage, plus the penalty, plus the scarcity revenue you never earned — on a fleet the grid can’t spare, run harder than it was ever designed to be.
And it plays out in public now. When a modern grid on the Iberian peninsula went dark in seconds in the spring of 2025, the whole industry felt it. Regulators, NERC, and the public watch plant reliability in a way they didn’t a decade ago. Being the unit that trips during a grid event is no longer just an operating problem — it’s a regulatory and reputational one.
Maintenance didn’t get more important because the work changed. It got more important because everything around the work changed.
So the question isn’t whether maintenance earns its keep. It’s whether your plant is managed to the stakes as they are now — the whole plant, not just the parts that used to matter most. That’s where this series goes next.
For more information on how Groome can help make a difference, visit our Gas Turbine Plant Performance Solutions page.
Read Part 2 → MORE THAN A TURBINE.
Missed the opener? Start here → YOUR TURBINE IS FINE. THE ASSETS AROUND IT MAY NOT BE.
