Dan Shugar, CEO of NEXTracker, is one of the old-school players in the rise of solar in the United States — he lives and breathes solar trackers.
NEXTracker is the world’s leading tracker vendor, with one-third of the market by shipments in 2018. Other firms claim between 8 percent and 12 percent of the market each, including Array Technologies, PV Hardware, Arctech Solar and Soltec.
According to Wood Mackenzie, the tracker market grew 62 percent in 2019, reaching 23 gigawatts of installations — and the market will grow on average by 11 percent annually from 2019 through 2024.
In the first part of our interview with Shugar, we discussed tracker innovation, bifacial panels and getting acquired by Flex. In this final part of our discussion, we cover the terror of torsional galloping in a solar array.
pv magazine’s Australian staff has taken deep dives into how single-axis trackers behave and fail in wind events, noting that single-axis PV trackers become unstable at relatively modest wind speeds. Third-party engineering firms have studied several different tracker designs that have suffered catastrophic failures in the field as a result of torsional galloping.
A classic (non-PV) example of torsional galloping is the 1940 collapse of the Tacoma Narrows Bridge.
Here’s an example of a torque tube single-axis tracker galloping in the wind.
Some tracker designs have aimed to address this instability with torsional stiffness — but this limits the tracker length, and requires a thick torque tube. Others vendors have used damping, but damping can be of limited benefit in some tracker positions.
The interview has been edited and condensed.
pv magazine: Do you want to talk about torsional galloping?
Dan Shugar: I want to tell you the whole story. First, let me address what happened on that Australian project. [The Oakey 2 site in Queensland.] There are two projects in Australia — projects that were built concurrently. There’s a system with a NEXTracker, and there’s a system with another tracker. That vendor will go unnamed.
There are two systems there, they’re pretty big — a hundred megawatts each. There was a windstorm for two days in the mid-40 miles-an-hour, 45 miles-an-hour plus or minus. When the dust settled on all that, there was zero damage on the NEXTracker system.
The other system was a severe train wreck [see article photo]. The part of their system that didn’t fail was shielded by us because of the way the systems were laid out.
The section where the other vendor was not protected by NEXTracker had damage that went way deep into their system.
We have a long history with really characterizing wind. If you go back to even the PowerLight days, we were doing all these wind tunnel studies. What’s fairly straight forward in understanding wind force on structures is what’s called static evaluation. What’s really hard to do is dynamic evaluation — when things start moving.
You can’t look it up in a book. We pioneered this back in 2015 at NEXTracker — we sponsored the preeminent wind engineering company in the world for structures like this (CPP) to characterize trackers in a dynamic fashion.
We were able to, in a lab environment, replicate and actually create these dynamic instabilities, characterize it, and then re-engineer our product to make sure it was addressed. We published this and we allowed CPP to publish with credit to us in 2015 and 2016. In 2018, we made the decision to publish a lot of stuff which was core IP to us, because some of the other folks either didn’t understand it or were cutting corners — and were irresponsibly damaging the tracker industry.[A NEXTracker blog entry notes the pioneering wind tunnel R&D work done by David Banks and his team at CPP, which investigated torsional instability, vortex lock-in and other factors in single-axis trackers. The CPP work also looked at the potential impact of different stowing approaches, tracker design architectures, and the LCOE implications of proper wind engineering practices.]
Dan Shugar: There were a bunch of catastrophic failures around the world. They’ve happened in Australia, Jordan, Latin America and Europe.
pv magazine: It’s not just the magnitude of the wind, but all these other dynamic variables?
Dan Shugar: Actually, the real high wind speeds are less of concern. It’s the medium wind speeds that are of concern. The reason is that these structures have a natural frequency of oscillation, which is about one Hertz. It can be a little more, a little less depending on the actual details of the structure. Those structures can get excited from a resonance standpoint in winds that are between, let’s say, 35 and 50 miles per hour.
You need to address it. The way we address it at NEXTracker is we put a bunch of dampers on the system, and we actually measure for the wind, and we actually stow into the wind. If you’re walking down the street and the wind’s blowing in your face, you have an umbrella, don’t be horizontal, you want to be into the wind.
That’s what we do with the horizon system. We have a 2P two-portrait product. The largest and most spectacular failures had been 2P, two modules in width.
pv magazine: Why is that?
Dan Shugar: The wind force is a squared function of the cantilever. When you go twice as wide you have 400% more torsional force. What happened was, there were a number of competitors that took what we were doing on one portrait, and then doubled it, scaled it and said, “Hey, we have less foundations,” and got a bunch of orders.
We’re not disparaging a competitor, but we’re saying, “Hey, you can’t do that.” You can’t just take a center gear and rely on the stiffness of the tube to hold this crazy big sail out there.
When we designed our initial product, we wanted to hit the middle part of the market, which is typical reasonable soil, typical reasonable wind, a typical reasonable ground cover ratio.
The way we address it, unlike the other providers, is we have multiple gears all the way down the structure to lock it, to prevent it from going into this torsional instability. From our viewpoint you can’t just rely on the stiffness of a tube in a center mount gear. Some of those guys are doing it now even without dampers, which we think is completely nuts.
Look there’s a lot of ways to solve the engineering problem — we’re just saying, if you go two-portrait in particular, you have to be extraordinarily careful about how you deal with wind force and dynamics.