How important is yield uncertainty?

Investors do not trust the forecasts they receive from project sponsors

From our previous round table discussion on uncertainty at the PVPMC 2024 we discovered that tax-equity investors do not trust the accuracy of the forecasts they receive.

We learned that there isn’t a standardised approach for determining uncertainty of forecasts and that there is much room for improvement.

Survey results

We surveyed the audience at the ACP Resource and Technology Conference 2024 to find out how much uncertainty matters to utility-scale solar. The survey included questions about current industry practice and how it can be improved.

The key takeaways from the survey are

• uncertainty is very important in the US market,
• both internal and external stakeholders depend upon uncertainty calculations,
• the most common practice is a simple sum-of-squares approach in Excel,
• most respondents include IAV (84%), model (84%), irradiance calibration (79%), degradation (74%) and availability (68%) in their calculations,
• about 75% of respondents are interested in a more advanced approach.

A full summary of the results is available as a pdf download.

New free software for yield uncertainty

At the conference Keith McIntosh described an advanced approach to uncertainty. This approach will help the industry more accurately quantify the uncertainty of a forecast, whether generated by PVSyst, SunSolve, SolarFarmer, PVLib or other tools.

As a result of our investigations, we are building new software that will address many of the shortcomings with the simpler sum-of-squares method by employing a rapid Monte Carlo solver. We plan to eventually support

• codependencies between inputs,
• degradation and curtailment,
• yield uncertainty over multiple years (10-30),
• stress testing extreme scenarios,
• risk assessment / impact matrix,
• asymmetric distributions.

If you would like to be included in a pilot release of the new tool, send us a message.