So This Happened …

Picture of Adam Bahret
Adam Bahret
Screen Shot 2022 08 31 at 11.22.07 AM

I remember witnessing these windmills being installed back in 2013, they’re only 10 miles from my house. The other day something, uh, strange happened. One of the blades fell off! We’re talking about a 170-foot blade 500 feet in the air just popping off, how can that happen? And on only a 9-year-old windmill nevertheless. So I was walking my dog, he took off, we got lost in the woods, and then happened to come across the location of the windmill base. To my surprise, the blade was at the base, it fell straight down and was nicely folded over on itself. In most scenarios I imagine where a windmill blade comes flying off, this would look more akin to Godzilla deciding to vacation in your hometown. So, how did we get so lucky? What are the odds? 1%, 0.1%, or even a 0.0001% chance of having the blade drop straight down? My guess is they were much better than that. I bet we had a 75% chance of it landing just as it did, right at the base, nicely folded like a t-shirt on a store shelf.

I reached out to a past customer. He is a lead engineer for a company that makes the big electronics that go into the head units of windmills. We did a great program a few years back using HALT to access what failures looked like. Yes, we blew one up with voltage. It was freakin awesome, but alas, that is a story for another time.

He shared some interesting information; It is apparently a standard function in a windmill control system to monitor vibration and react to any anomalies. In simple terms, “If something feels wrong, stop.” This accomplishes two things, the first being that removing centrifugal forces can stop a small issue from becoming a bigger issue. It also can take a big issue, like blades falling off, and reduce the effect from “horrific destruction” to “clean up on aisle five.”

I smell a well-done Failure Mode Effects Analysis (FMEA).

I did look at how this blade mounting system worked and it is very similar to many aircraft propeller blade mounts. This is because it shares the feature of being a variable pitch blade. This means that the blade can rotate around its own axis to change how its leading edge cuts into the air. This is done so maximum efficiency can be achieved, similar to changing gears on a bicycle. Because of this feature, the blade can not simply be bolted to the hub. Instead,

 

it is held in place by the same bearings that allow it to rotate. You can see the bearing race and the large ring gear that is driven to position the blade.

So why do I suspect a well-done FMEA occured? FMEAs assess the risk of failure modes quantitatively by scoring them in three categories: severity, likelihood of occurrence, and detectability. By scoring these individually, the team can identify means of reducing risk in a more targeted manner. “Severity” and ”Likelihood of Occurrence” are often approached first when reducing risk. “Detectability” is often ignored, in fact, so much so that I have seen teams actually remove detectability scoring entirely.

However, I suspect that working with the detectability aspect is what saved the day here. It is important to point out that the detectability score does not reflect the ability to detect a failure. All the failures listed are detectable because we are listing effects as well. Basically, we are scoring how far in advance we can detect the failure. Something like “vibration may occur in advance of bearing failure.”

So how does this help when assessing risk? Well, identifying this relationship opens the opportunity for investigating whether we can change the effect or mitigate the failure entirely by leveraging or changing the detectability score. Including a prognostic feature in a control system that monitors vibration and then takes precautions is exactly the kind of result that would come from this analysis.

Many locals see the windmill with the sad two blade frown on the horizon as evidence of failure. As a reliability engineer, however, what I see is a nicely folded blade at the base, the best possible outcome. Likely the result of a well-done FMEA that used risk scoring in “Detectability” to drive important design changes, and one that turned out to be very important to me and others on the north shore. Well done windmill FMEA team!

-Adam

Share this post