This is the first article of Wattch’s three-part Performance Modeling 101 series. Stay tuned to learn how to evaluate modeling tools, create a digital twin, and use it to diagnose issues in installation and ongoing operations.

Takeaways

• A number of KPIs can describe solar performance using different considerations such as system capacity, modeled or real-time weather conditions, and energy output.
• Among industry standard KPIs, we believe EPI is the most useful for O&M and asset management teams to identify and prioritize issues and track portfolio performance.
• We have created our own KPI, the Wattch Health Score, to synthesize the information from multiple other KPIs to more quickly evaluate site health.

By the end of this article, we hope you will have a thorough understanding of the different solar KPIs and their relative usefulness in monitoring applications.

Common solar performance KPIs

Commercial solar operations and maintenance (O&M) and asset management teams have a tall order: to maintain dozens or hundreds of heterogeneous systems with thousands of devices, often among a small team.

Doing this well requires using KPIs that provide quick, actionable insights, so teams can spend more time fixing issues and less time churning through vast spreadsheets. The trouble is, there are so many KPIs thrown around in the solar industry that it’s difficult to know which are the most useful for ongoing operations.

Which is why we’re here to help. Below, we explain each of the most prevalent metrics in loose ascending order of how much we like them for diagnosing issues and allocating resources in real time.

Capacity Factor

The capacity factor is calculated by dividing the actual energy output of the system by the amount of energy the system would have produced if it ran at full output all the time. The capacity factor is typically expressed as a percentage, and a higher capacity factor indicates a better performing system.

A capacity factor of 100% would mean that the system is producing all of the energy it is capable of producing 24/7, which is not possible in reality due to the system’s location, changing weather conditions, and other factors such as soiling and shading, not to mention nighttime.

Due to these myriad factors,  a “good” capacity factor is around 20-25%, which makes sense when you consider that most of the United States experiences 3-5 peak sun-hours per day. However, knowing the ideal capacity factor for a particular site requires a lot of context on the location, time of year, surrounding environment, and attributes of the system.

Additionally, the capacity factor is not standardizable between systems. It’s difficult to say whether a system with a capacity factor of 25% is truly performing “better” than a system with a capacity factor of 20%. It can thus be difficult to use this metric as a way to prioritize issues and allocate resources within a large solar fleet. This is why we believe this metric by itself has limited usefulness in the context of real-time operations and maintenance decision-making.

Performance Ratio (PR)

The PR is calculated by dividing the actual energy output of the system by the expected energy output, based on the system's nameplate capacity and the amount of sunlight it receives. The expected energy output is based on the system’s output in standard test conditions (STC), defined as 25°C cell temperature and 1,000 W/m² incident irradiance.

PR is typically expressed as a percentage, and a higher PR indicates a better performing system. A PR of 100% would mean that the system is always operating perfectly at STC, which is not feasible given that outdoor temperatures are not constant and irradiance changes with the position of the sun and other factors.

While the PR can provide some insight into a system’s function, we don’t believe it is the best metric for optimizing performance in real time. The PR incorporates more real-world information than the capacity factor, but it does not take into account all of the factors that can affect the performance of a solar system, such as module temperature, which affects the output significantly. Additionally, if a system has multiple array orientations, you must perform a weighted average of the insolation onto each orientation to get the true PR.

And because the theoretical ideal PR of 100% is not achievable, it is not always clear what PR a system should target, or whether the gap between the ideal and measured value is due to actionable conditions (such as soiling or module damage) or factors outside of one’s control (such as age-related performance degradation).

Because of this, it can be difficult to tell whether a PR of 85% is the system’s ideal under the current conditions or if it is cause for alarm without further investigation. If you’ve run a PVsyst report, you’ll know that the PR changes drastically from month-to-month, so you would have to go back to your 8760 output to know if you’re hitting your target.

Furthermore, your system could be underperforming but have a higher PR due to more favorable weather conditions than predicted in your performance model (e.g. lower temperatures). This masks underlying issues that you don’t know about unless you look at trends over a large time period (after which you’ve potentially lost a lot of production!)

Finally, similar to the capacity factor, the PR is not standardizable between systems, making prioritization and resource allocation more difficult for large fleets. For these reasons, we believe the PR should be used in conjunction with other performance metrics and not on its own when evaluating solar system function.

Availability

Availability is the percentage of time that assets are producing when they should be (for solar assets, this is during daylight hours). An availability of 100% means that your system or device is continuously operational from dawn to dusk.

Availability is one component of other metrics such as EPI (below) and can be used to explain why EPI is low. For example, if the system’s availability is equal to EPI and both are less than 100%, then some amount of your system’s inverters are offline. This allows you to differentiate equipment failure from system-wide losses of similar magnitude such as soiling. Availability can also be used in place of other KPIs for systems that solely use a site-wide meter to measure production data.

However, if you have access to inverter data, these values should be used. Access to inverter data also likely means that you can use zero-generation or offline alerts in place of this metric to diagnose equipment failure. 

But the most significant limitation of the availability KPI is that it only surfaces issues that involve the total failure of one or more devices. On the other hand, if all devices are online but underperforming, availability doesn't provide any further insights, and may even mask issues if it isn't used in conjunction with other metrics. For this reason, we think that it is best used in conjunction with other KPIs in O&M scenarios.

Specific Yield

The specific yield is a measure of how much energy a solar panel system produces per square meter of surface area. This is usually expressed as kWh / kWp, or energy per unit STC capacity of the modules, which is equivalent to energy per surface area by a scale factor, assuming that the modules are all of the same model.

It is an important metric in the feasibility assessment phase of commercial solar energy projects, as it can help developers and investors to compare different systems and determine which one will be the most efficient and cost-effective. But this also means that specific yield has as much to do with system design as system function. What’s more, maximizing the specific yield of a project may not even be in your organization’s goals, especially when it conflicts with cost goals.

And, like the other metrics we’ve mentioned thus far, there’s no true “north star” value for specific yield. You can measure whether the specific yield is as designed, but this requires context and more calculations, and is different for every system. This makes it difficult to tell at-a-glance whether a system is performing well without knowing the target number.

At the device level, you can use the specific yield to compare performance from one inverter to another to determine whether one is underperforming. However, this doesn’t allow for easy comparisons between sites, nor does it yield accurate results for sites with more than one inverter model, multiple string lengths, or multiple orientations. This metric works in a pinch if you don’t have a digital twin, but our (biased) opinion is that a digital twin is better for modeling device-level performance.

For these reasons, we think that the specific yield is more useful in the design and project evaluation phases than in operations and maintenance scenarios.

Baseline Energy Performance Index (BEPI)

The BEPI is a metric used to track the performance of a solar system over time as it operates. It is calculated by comparing the actual energy production of the system to its predicted, or ideal, energy production, based on factors such as the system's size, location, and weather conditions. A BEPI of 100% means that the system is performing optimally given how it was predicted to perform under modeled weather conditions.

The primary drawback of this metric is that it does not separate environmental effects from performance effects. Is the BEPI low because there is a malfunction in the system, or is the weather simply worse than predicted based on the TMY dataset? On the other hand, is a BEPI near, at or over 100% the result of better-than-average weather or better-than-average performance?

Additionally, the models used to calculate ideal production generally use TMY weather data, which describes the average temperature and irradiance over the last 30 years, and the average is rarely a reality. Not to mention that weather patterns have changed quite a bit over the last 30 years, and even over the last 5 years, so what was normal yesterday may not be normal tomorrow.

BEPI can be a useful metric when evaluating an asset’s performance against what was modeled, and can be good for financiers. BEPI tends to be more accurate and useful when comparing expected and actual performance over months and years rather than days and hours. For this reason, O&M and asset management teams can evaluate real-time performance more accurately using EPI.

Energy Performance Index (EPI)

Similar to the BEPI, the EPI also tracks a solar system’s performance over time. It is calculated by dividing the total energy produced by the system by the total energy input (i.e. irradiance).

The main difference between EPI and BEPI is that EPI uses real-time, local weather data, while BEPI uses 30-year averages. An EPI of 100% means that a solar system is performing as expected based on current weather conditions and the expected performance of each of its components.

Because EPI takes the unique configuration of each solar system into account, it gives operations and maintenance teams the clear goal of bringing the EPI of each site, as well as that of the whole portfolio, to 100%. O&M teams can then easily monitor a solar fleet for deviations from this value and flag them as cause for concern, prioritizing the largest deviations for corrective action.

While no single metric can give you the entire picture of your system’s performance, we believe EPI is the most useful and expedient when evaluating system performance in real time.

That said, not all EPI calculations are created equal due to differences in modeling expected performance. We strongly recommend using a bottom-up, digital twin based model, especially on smaller sites, sites with multiple orientations, or sites with multiple equipment models. This will make your EPI value much more accurate, allowing you to better surface and resolve operational issues.

Why create a new KPI?

To clear the confusion inherent in tracking multiple KPIs over a large portfolio, we developed the Wattch Health Score, a letter grade from A+ to F of solar system performance. The score uses a blend of the metrics mentioned above and our experience monitoring the performance of hundreds of solar systems.

We developed the Wattch Health Score to make it easier for owners and operators to get a sense of how their systems are performing, taking into account the site’s Digital Twin as well as local, real-time weather conditions. O&M teams can use this metric to sort and easily identify underperforming sites for further investigation. You can click on this score to view the Wattch Report Card, displaying multiple KPIs that provide more information about the issue.

We believe that the Wattch Health Score allows O&M teams to spend less time uncovering issues, prioritizing maintenance work, and calculating KPIs in spreadsheets, and more time actually fixing the issues that arise.

“If you only use one solar performance metric this year, make it the Wattch Health Score.” - Wattch

Using KPIs in O&M and asset management

Now that you know what all of these KPIs measure, the natural next question is what to actually do with solar performance KPIs. These metrics can be used to identify underperforming sites, allocate resources, prioritize corrective action, and report on portfolio performance over time.

The good news is, if your organization uses an inverter-agnostic monitoring platform, you shouldn’t have to do this via a spreadsheet. Most of these platforms incorporate common metrics such as BEPI, EPI and specific yield into their user interfaces, though the methodology used to calculate each of these metrics may take different factors into account.

Prioritizing Problem Sites

To prioritize issues across your fleet, it’s important to have all of your sites in a single solar monitoring platform. Once your portfolio is all in one place, you can sort your sites by EPI (or, if you’re using Wattch, by the Wattch Health Score) to determine where the largest performance deviations are.

Ultimately, the Health Score or EPI is just an indicator. The prioritization could involve the distance/cost of repairs, the size of the project/revenue lost, issue severity, or a number of other factors. It’s all based on your organizational goals.

Assessing Portfolio Health

Different organizations use different numbers for reporting and evaluation. If you don’t have your process set in stone, we recommend using EPI to report on portfolio and site performance. By tracking this number over time, you can directly evaluate how your O&M efforts are improving site health.

With a few more calculations, you can also apply revenue metrics to your EPI to see how much money you could be leaving on the table by not addressing operational issues.

Putting it all together

While it’s important to quantitatively track performance, we don’t believe O&M or asset management teams should be spending hours consolidating their portfolios in spreadsheets or calculating performance metrics manually. Put simply, KPIs are supposed to save time, not create more work. With the right tools, you can easily track, diagnose, resolve and report on issues in your portfolio in minimal effort and in real time.

‍