I’ve worked in the solar sector for over a decade, and I frequently get proposals from distributors and manufacturers asking if I’d like to purchase modules—both in bulk and small volumes.
When I started in the industry back in ‘05, the rate was about $4.10/Watt—just for the solar modules. The lowest price on the most recent solicitation I received was $0.28/Watt. I have peers who have quoted me lower prices.
This means that an energy-generating 3’x5’ 300W solar module now costs less than a 3’x5’ sheet of clear polycarbonate from Home Depot, despite recent tariffs. That is almost the very definition of commoditization.
What all this means is that it’s definitely a buyer’s market.
Right now a utility scale solar power plant in the US can be built and sold for under $2/watt. So the entire power plant, with margin, is being sold for under ½ the cost of just a solar module in 2005.
Of course, low construction costs also mean the costs of energy delivered by the power plant is low. In fact, for the American Southwest, energy delivered by solar power plants are the cheapest in the market, cheaper even than a coal plant.
Simply put, solar is the cheapest power you can get.
Now, most O&M providers don’t concern themselves with power plant pricing…but they should. Pricing can make the difference between keeping your contract or hunting for your next gig.
Which brings us to thermal scans and the folks who are doing that work.
Many times, these thermal scans are mandated—you have to do them as part of the annual contract conditions. That means O&M shops either have their own in-house drone operations team, or they’ve sourced that work to outside operators.
In either case, the data they collect is sent to a Drone Data Services firm that converts that thermal data into a set of defects identifying hot spots or failures located across the array—all of which is very good. The problem is the accuracy of the data.
Most of these Data Service shops use Machine Learning (ML) and Artificial Intelligence (AI) systems to interpret the visual imagery and assign defects to the array. Precision XYZ uses these same techniques, so we understand this well. The key is how the Data Services firm treats the data after the scan, because none of them are 100% accurate in identifying defects.
More frequently than not, these systems make mistakes—identifying False Positives (“defects” that aren’t really defects) and False Negatives (“working” parts that don’t actually work). While both errors create challenges for O&M shops, one of these is definitely worse than the other.
To understand why one of these inaccuracies is worse than the other, you need to go back to those low, low prices.
Today’s solar power plant, sold at under $2/watt, generates energy sold as low as $0.03/kWh. Now, if that plant is in a sunny state, that means your Annual Panel Revenue (or APR) is about $14 year, or roughly $0.038/day.
Four pennies per day per panel. And remember because your panel is slowly degrading that APR will go down every year.
Compare your client’s APR to their Annual Panel Cost (or APC). The APC is the depreciated annual cost of the panel plus its annual O&M cost. Anytime an asset owner’s APC is increasing at a rate faster than their APR is going to give you a bad outcome—which brings us back to those false positives and false negatives.
The last thing you want to do is roll a truck for a non-existent problem—a ghost—particularly if you’re rolling a truck for a panel(s) that only generates $0.04/day. It’s far better to have a False Negative in this case, because even if the panel is actually broken, the lost revenue number will not materially change the Rate of Return for that asset.
The key is sorting and prioritizing your defects—real or imagined—into High Impact and Low Impact defects. Unfortunately, unless your Data Services firm understands APR vs APC, they won’t have a good enough system to sort and prioritize those defects. In these situations, every defect becomes a priority, and god forbid the asset owner sees that report.
Responding to everything can not only cause a loss of faith in the efficacy of thermal scans, but also increase the APC without impacting the APR. While there may be some hidden minor gremlins in the works—False Negatives—they generally don’t move the needle on APR because these are almost always at the panel level.
But for the O&M Provider chasing False Positives, those could end up costing them their contract as they drive up the APC.
So what’s an O&M shop to do?
Precision XYZ is the only Drone Data Services firm in the market that stress tests our processed thermal data by calculating an APR and an APC for each power plant. This allows us to sort all defects, real or false, and then determine which are actually monetarily relevant and need to be investigated for correction.
O&M providers can then easily prioritize their work orders, identifying High Impact vs Low impact problems, efficiently driving up their client’s APR while keeping their client’s APC low.
What’s more, it enables the O&M shop to talk to their clients in a language they understand— dollars and cents. Rather than presenting a report to them with a thousand errors, ghosts and gremlins, you can simply tell them how you intend to increase their revenue while reducing their costs.
And that’s a language that anyone can appreciate.
Get fresh insights from Precision XYZ. No spam, ever.