Our understanding of solar asset performance is changing. kWh Analytics recently published the 2020 Solar Generation Index (SGI), an industry-wide validation study, that found operating assets are underperforming by an average of 6.3% as compared to their P50s, on a weather-adjusted basis. Jason Kaminsky, the Chief Operating Officer at kWh Analytics, had the opportunity to discuss these findings and how the industry is adapting with leaders at Arevon, Clean Capital, VDE Americas, and Solargis during this year’s Solar Asset Management North America (SAMNA) virtual conference.
Here are three key takeaways from asset managers, owner’s engineers, and weather satellite companies:
1. The “Swinging Pendulum” of Performance Estimates
While the 6.3% underperformance results from the SGI were startling, no one on the panel was surprised. The panelists described maturation of solar coinciding with a gradual shift toward aggressive production assumptions. It was agreed that if you looked at projects five to ten years ago, it was common for them to outperform their production estimates. In contrast, Anand Narayanan, Vice President of Asset Management at Arevon, noted that in today’s competitive landscape with pressure on margins and new modeling complexities, assets are challenged to perform above their P50. He advocates that additional scrutiny of production assumptions is necessary to truly understand performance limitations and the probability of meeting P50 estimates.
When asked to diagnose the reason for this, Brian Grenko, Vice President at VDE Americas, attributed this swing as a gradual change in the assumptions used in technical and financial modeling that results in “death by a thousand cuts.” Grenko summarized it best when he said that today’s P50s represent expectations only “when everything goes as planned.”
2. It’s All About the Data
Data is key to managing solar P50
estimates. In addition to the macro trends identified in reports like the SGI,
panelists also discussed the value of site and portfolio-specific data to
improve underwriting and diagnose underperformance issues. This starts from
understanding the input data to P50 modeling.
Narayanan emphasized the value of leveraging Arevon’s existing operating fleet to support diligence: “As the largest owner of solar assets in California, Arevon has access to generation and weather data to compare performance numbers and underlying assumptions.” This information ensures asset management is comfortable with the underwriting before they manage the asset. Kaminsky added that the kWh Analytics Solar Technology Asset Risk (STAR) Comps reports provide generation and weather data for projects across the country, and these reports are used primarily for asset due diligence and asset management.
Once acquired and operating, the focus shifts to monitoring plant performance and having the right tools to diagnose drivers of underperformance. An all too familiar goose chase in asset management is verifying the impact of weather.
Zoe Berkery, Head of Asset Management at CleanCapital, shared that she’s seen inconsistencies when comparing on-site pyranometer results to a weather file used by the IE at the inception of the project. “On-site pyranometers can be very expensive to upkeep and are sensitive to soiling,” she added, “Which has led the team to explore satellite-based weather options. The updated approach has led to more consistent results across projects and reduces variability.” Giridaran Srinivasan, Business Consultant at Solargis, concurred that ground-based readings suffer from several data quality challenges including “data logging issues, calibration errors, and lack of sensor cleaning.” Kaminsky added that one way to address these challenges is by using satellite data run at scale through production modeling software.
3. The Evolving Role of Asset Management Teams
With growing scrutiny of underwriting accuracy, it is unsurprising that asset management teams are playing a larger role in the project development lifecycle. Narayanan and Berkery both confirmed that their asset management teams are pulled in earlier and earlier to evaluate production assumptions. Narayanan explained that this is driven by his team’s access and understanding of plant data: “We are looking at plants on a daily basis and can identify the factors that have not been modeled properly and make sure those are taken into account in diligence.”
Solar is a maturing asset class with another trillion dollars to put to work over the next six years. As an industry, we have the tools to course correct the systemic miscalculation of solar generation and guide the evolution of solar. A combination of objective market data, analysis with industry benchmarks, and coordinated effort will be paramount to accurately diligence and manage the industry’s growing solar fleet.