Translating SunEye Solar Access percentage to SAM-modeled energy losses

We're using SAM to model solar PV systems in upstate NY. We are using actual shading data from a Solmetric SunEye to feed our SAM models. As a QC mechanism, we're modeling each system with the actual shading data and with no shading data in SAMul, and calculating an annual %kWh loss due to shading. We then compared this % annual kWh loss to the annual percentage from the SunEye Shading Summary file, and we're finding the numbers are significantly different. For example, one system had a 2.75% shading factor from the SunEye shading summary file, but this only translated to a ~1% SAM-modeled annual kWh loss.

This leads to a couple questions:
1. Has anyone performed similar benchmarking tests to translate SunEye solar access percentages to SAM-modeled kWh losses?
2. Does anyone actually know how the annual SunEye solar access percentage is calculated based on the 15-minute shading data (where each 15-minute interval gets a 0, 0.5, or 1)?
3. Does this make sense that there's not a 1-to-1 relationship between the SunEye solar access number, which is defined as "Actual insolation compared to insolation with no shading, in field of view of the panel." It seems like, as a rule-of-thumb, solar installers regularly take the simple approach of multiplying an expected annual kWh by the SunEye shading factor to get a revised estimate. Maybe they're underestimating solar production by using this method?

Hi, I am with Solmetric, and may have some ideas about the SunEye calculations, in response to your questions.

1. We have limited experience in benchmarks of SunEye solar access vs. SAM kwh calculations. We would very much like to get other people's inputs.

2. SunEye shading files are available in different forms. If you just look at one skyline, the shading is 0 (shade) or 1 (no shade). If you take multiple skylines and average them, you typically get fractional values. For example, with two skylines averaged, values are 0, 0.5, and 1. With four skylines averaged, the values can be 0, 0.25, 0.5, 0.75, and 1. For example if one of the four skylines in that 15 minute interval is shaded, you get 0.25. These values come out of the file named AverageShading.csv.

3. I am not sure why the 1-to-1 relationship is not closer, but I have some ideas. The way that solar access is defined is that shading is removing both direct AND diffuse radiation from the panel. Most PV models, and SAM, I think, do model the system with non-zero diffuse radiation when shaded. This could account for the difference, or at least some of it. In my experience many installers are using Solar Access as a derate factor, usually by month. In this case string effects like imbalance are not included. To some degree this works in the opposite direction from the diffuse issue. The PV model will also include temperature which can have significant impact, and the SunEye readings will not include that.

I welcome further discussion on this topic.
Feel free to email me at peter@solmetric.com.

As an update, we further researched #2 from Peter's posting above, specifically focusing on diffuse v/s direct radiation. We performed the following:

First, using SAMul, we modeled each of our ~20 PV systems with no shading present, and we noted the annual estimated kWh.

Next, we modeled each site with 100% (full) shading, and noted annual estimated kWh.

We finally divided the annual kWh with full shading by the kWh with no shading, and found that about ~40% of annual energy produced can be attributed to diffuse radiation.

This seems high, but would explain the differences between the SunEye shading factor and what we're seeing in SAM.

Any insight would be appreciated.

In response to Peter's #3:

When you import shading data from an external program in SAM (PVsyst near shading, SunEye, or Solar Pathfinder, the data only affects direct radiation. SAM applies shading factors in the data to the direct component of the radiation incident on the array.

When you check Enable Sky Diffuse Shading Factor and specify a Sky Diffuse Shading Factor value, SAM applies the value to the diffuse component. There is only one diffuse shading factor value that SAM assumes applies to all hours of the year.

In other words, unless you enable the diffuse shading factor option, SAM only removes direct radiation from the panel.

However, as of SAM 2011.12.2, on the Array page you can specify monthly soiling factors that apply to the total incident radiation (beam and diffuse), that would contribute to the difference.

Best regards,
Paul.