You just bought a large-scale solar array. The performance expectations you have for your new array, like a new car, are high. Just like the information on the window sticker that claims great gas mileage, your solar array comes with a forecast of how many MWh it should produce over its lifespan. This information is critical to your financing strategy, since this energy forecast and the cash flow it generates, are the basis for justifying the investment and the promises to your financial backers. However, how do you know that you got what you paid for? How will you know if it starts underperforming and what will you do to restore and maintain peak performance?
As all experienced car owners know, performance varies. You have to maintain the health of your car to get even close to the gas mileage number it claimed upon purchase and more importantly to increase the operating life and resale value of the car. Thirty years ago, it took vigilance and often a trained ear to detect issues developing in your vehicle, but now, cars come with sophisticated computers to tell you if something is wrong and more importantly what to do about it.
Solar arrays can also have a higher level of sophistication, in terms of optimizing their performance, extending their active life and increasing their residual value. Unfortunately, the technology to do this doesn’t come with the basic package – but it’s an option. Until recently, most new array owners have not been offered, or taken advantage of, these new monitoring and optimization options. However, the effectiveness, affordability and availability of these technologies are becoming much more attractive. We are approaching the point where large-scale solar assets will not be considered without it, just like you wouldn’t be offered a car without engine management or anti-lock brakes, or dashboard warning lights.
A commonly used measure of solar array health is the Performance Ratio (PR) – the ratio of the actual energy produced from a solar array compared to the theoretical capability of the solar assets installed. The closer the PR of an array gets to 1.0, the better it is performing. In practice, a PR of 1.0 is not achievable, due to unavoidable issues like DC wiring losses, temperature effects and DC-AC conversion losses. But new and well-maintained large-scale PV arrays can achieve levels of around 0.80. Therefore, you often see the DC-side overbuilt by 20-25 percent to provide a defined AC power rating.
Unfortunately, during use, there are many real world factors that will pull down performance. Panel damage, degradation, soiling, shading, tracker misalignment, and equipment outages all conspire to bring down the PR of a site. A study of large-scale sites of varying ages measured annual Performance Ratios between 0.38 and 0.81, with an average of 0.66, suggesting significant variability in performance and a sizable loss when compared to performance expectations.
Weighing the Factors
Let’s consider a theoretical 10 MW (ac) array. The plan calls for the purchase and installation of 12.5 MW of assets (panels, wiring, racking, trackers, inverters) to offset the anticipated 0.80 PR. At commissioning, the initial testing of the perfectly built array with nice clean panels on a fine spring day confirms a 10 MW (ac) capability, just as specified. However, once in operation, varying environmental factors, zonal soiling and a small amount of undetected degradation in some of the panels, cause the array to perform at the 0.66 level, only providing an output consistent with an 8.25 MW (ac) site. “Whoa! You mean I paid for 12.5 MW of assets and I’m getting 8.25 MW of output?” Put simply, 34 percent of the capital put into the project is non-productive.
In a world where large solar assets are built with 80 percent debt leverage or more, a one percent change in output can equate to a 10 percent change in ROI for the investors. The impact of an unanticipated drop in the performance ratio from 0.80 to 0.66 would probably wipe out any anticipated return from the project. This potential future variability has a major impact on site financial viability, but more importantly on the attractiveness of solar as an investable asset class. A key objective of the industry should be to increase the entitlement level for Performance Ratio beyond the 0.80 level and reduce the long-term risk of assets drifting off that entitlement level. This would reduce the overbuild and hence initial capital outlay, reduce the levelized cost of electricity for the site, increase the ROI for the investors and reduce the long-term financial risk, thus attracting financial backing and possibly reducing insurance premiums.
Most of the issues that drag an array down from peak performance are addressable. They can be classified into three categories:
- Defects and damage including damaged panels, wiring faults, blown fuses and inverter outages;
- Environmental impact including soiling, shade and temperature variations. These are especially impactful if the impairments are uneven across the array, creating a multiplying effect due to panel and string mismatch;
- Degradation of panels over time. Again, this has a larger impact if the degradation rates are not consistent across the panels within an array and large mismatches begin to appear.
Managing Better Output
To deal with these issues, the most productive sites implement an aggressive maintenance scheme based on constant vigilance, rapid fault diagnosis and a predetermined plan of response. In addition, a calendar of predefined preventative maintenance is implemented to try to avoid catastrophic issues such as inverter outages, or to get insight into panel performance compared to warranted levels. It would seem that this should be sufficient to maintain a large-scale array at its optimal level, close to the PR of 0.80. However, these operations and maintenance (O&M) procedures are implemented for most of the large PV systems around the world and measured results don’t support expectations.
The issue is lack of insight. The granularity and precision of monitoring systems currently deployed in large arrays are insufficient to provide an accurate diagnosis of array impairments. Most large sites historically have been built with monitoring confined to the AC meter and inverter. There is no insight into the DC side of the array and yet that’s where the power is generated and where all the potentially debilitating impairments reside. It’s no wonder that most array data points to inverter outages as the primary source of power loss in large-scale systems. It’s the only part of the system that’s been historically monitored. O&M teams were blind as to what was happening on the DC side of the array.
There is a growing trend to deploy sub-array or string-level monitoring, which improves granularity. At the string level, the system is aggregating the output of 15 panels or so, and providing measurements to an accuracy of plus or minus 5 percent. This is a significant improvement over inverter or sub-array monitoring, but at this coarse level of granularity and precision, it is still not possible to detect and pinpoint issues down to the panel level. Thus, even when assisted by string-level monitoring, large site O&M teams can only economically deal with issues such as inverter outages or blown combiner fuses. Issues such as panel damage, degradation, excessive soiling and encroaching shade can go unnoticed.
The next step in this evolution is to look to monitoring at the panel level. Technologies are now available that provide this level of insight in a cost effective way. DC optimizers, microinverters and panel monitors all provide this capability, often with greater accuracy (<0.5 percent) than that provided by string monitoring. With every panel in an array being monitored to this level of accuracy, the O&M organizations finally have the insight to aggressively manage PV arrays. Defective panels and wiring faults are immediately evident; panels can be tracked against warranty levels; threshold cleaning strategies can be deployed and targeted to areas most impacted.
Panel monitoring provides the basis for the most efficient manual optimization. However, large sites also have the option of adding automatic optimization technology, such as DC optimizers. These can mitigate dynamic impairments such as cloud effects, temperature gradients, soiling and many of the effects of mismatch and damage, until such time O&M teams can intervene. The ideal solution is a blend of both aggressive O&M and selective optimization, applied in areas of the array that are most likely to suffer from dynamic environmental factors such as occasional daily or seasonal, shade or zonal soiling (e.g. from an adjacent farm or road).
Panel monitoring provides unprecedented insight, but it also creates a fear of data overload and the ability for O&M teams to distill the mass of data, analyze performance and prioritize an action plan that can guarantee improved financial return. The next logical step in the evolution of large-scale array performance optimization is to couple panel-level data insight with advanced analysis and diagnostics tools that remove the need for human vigilance and interpretation. Arrays become self-analyzing and present actionable information to the O&M team driven by site-specific financial business rules. In effect, arrays become intelligent – self-adapting and optimizing where possible and informing the owners with precise details of when, where and how human intervention is required to maintain the asset at optimal performance over its full 25-year life.
Making large-scale arrays ‘intelligent’ is the path to raising the entitlement for Performance Ratio and to maintain arrays at an optimal level. This requires a combination of panel-level monitoring and selective optimization, coupled with a suite of cloud-based intelligent tools. Back to our car analogy, this is the equivalent of remote sensors, engine management and the on-board computer. You wouldn’t buy a car without these, so why do we still settle for a large-scale solar array without them?
Ray Burgess is the President and CEO of Solar Power Technologies.