What Your Maintenance Crew Isn’t Telling You

And every year, the same question is somewhere in the background: Why didn’t we do this sooner?

The answer is always the same, too. Because winter felt manageable. Until it wasn’t.

A familiar pattern across many cold-climate sites

Here’s what a typical cold-climate icing season looks like from the inside.

October arrives. Temperatures drop. The first icing events hit – minor ones, maybe a few hours of reduced output. The ops team notes it. Nobody panics. This is normal. This is winter.

November. As winter progresses, more significant icing events begin to appear. Turbine trips, repeated reset attempts, and occasional site visits become more common. The SCADA alarm fires. An operator attempts a remote reset, but conditions keep retriggering the fault. Eventually a crew is dispatched – not always necessary, but this time the turbine won’t clear remotely. Half a day lost. It goes in the log as “weather-related downtime.” The number looks manageable on a report.

December through February. This is typically where the largest losses occur – not necessarily through dramatic shutdowns, but through sustained underperformance that can be difficult to quantify in real time. A turbine doesn’t need to shut down to be losing you money. Blades with even a thin layer of ice lose aerodynamic efficiency. They spin, they generate, power curve data may show underperformance vs. wind speed – though attribution to icing vs. other causes is not always automatic – but they are producing 15%, 20%, sometimes 30% less than they should be.

That gap compounds across every turbine, every day, across the three months when wind energy potential is at its peak and energy prices are at their highest.
By the time February ends, the cumulative loss from these “non-events” typically dwarfs anything that was formally logged as downtime.

March. The season winds down, but the consequences don’t. Maintenance that was deferred through the winter – because conditions were too marginal, access too risky, the crew stretched too thin – now lands all at once. Equipment that has been working harder to compensate for icing-related inefficiencies shows the wear.

April. The ice clears. The season ends. Someone compares this winter’s output against budget, against last year, against the fleet average. The gap is uncomfortable. The math on what a protection system would have cost starts to look different.
May. They call us.

The cost of the pattern

The frustrating thing is that the decision to wait isn’t irrational. It just looks that way in retrospect.

Many wind farm owners have limited automated attribution of icing losses in real time. While SCADA data exists, distinguishing a production dip caused by low wind, blade ice, or a sensor fault often requires manual analysis or dedicated icing analytics software. Losses get absorbed into “winter variability.” The true number never makes it cleanly onto a dashboard.

What we know from the data: A severe icing event can cut output by 50-80% or force a complete shutdown One day of downtime for a modern turbine (2-3 MW) in good wind conditions could represent 20-70+ MWh in lost generation depending on capacity factor. Multiply that across a fleet, across a season, and you are looking at losses that would have fully justified an ice protection system – often within the first year alone.

But those numbers only become visible after the season ends. By then, the winter is over, the urgency has faded, and the procurement decision drifts to next quarter’s agenda.

Until the next October arrives.

What your crew already knows

There is a gap between what happens on the ground at a cold-climate wind farm in winter and what reaches the owner’s desk.

Maintenance technicians who work these sites develop an intimate understanding of how icing affects operations. They know which turbines are most exposed. They know the difference between a turbine that tripped cleanly and one that has been struggling for days. They know what it feels like to make a go/no-go access decision when conditions are borderline.

They also know something else: the losses your reports show are not the whole story. Deferred maintenance, conservative operating decisions, the cumulative drag of blades that aren’t quite clean – these don’t show up as line items. They show up as a production number that’s slightly lower than it should be, month after month, in the months that should be your most profitable.

Winter is when energy prices are highest. Wind is strongest and most frequent. Cold air is denser, which means more energy per rotation.. Your turbines have the most to give – and icing is quietly taking a significant portion of it, in ways that are easy to underestimate until you look back at a full season’s data.

Why planning starts months before winter

Spring and summer are the right time to install ice protection. Not because the technology requires it – but because the installation process takes time. Designing, manufacturing, shipping, and installing a system typically takes 3-6 months depending on fleet size, turbine model, and lead times.

If you start the conversation in May, you can realistically have protection in place before the next icing season begins. If you wait until October, you are already too late for this winter.

This is the window. It is open right now.

The owners who act on this in spring are the ones who don’t make the same call in April. They’re the ones whose production data next March looks different – and whose maintenance crews stop having to make the kinds of calls that never make it into the reports.

A useful question to ask

If you pulled your full production data from this past winter – not the summarized version, but the raw numbers, hour by hour, turbine by turbine – how confident are you in what you’d find?

If the answer is “very confident, we have clear visibility into our icing losses” then this article wasn’t for you.

If there’s any uncertainty in that answer, it might be worth a conversation.

FabricAir offers a complimentary icing loss assessment that helps wind farm owners estimate the production impact of icing events and evaluate whether ice protection could be economically justified for their site.