Comparing SEBE algorithm with real electrical production

[Andrei99537]

Hello,
I have a question regarding a solar analysis I conducted to compare the accuracy of the SEBE algorithm with ArcGIS's Area Solar Radiation and actual PV electrical production. To obtain the results, I used a DSM, meteorological file, and shapefiles representing only the rooftop areas where solar PVs are installed.

Using ArcGIS's solar radiation algorithm, I obtained data that I converted to MWh/year of electrical power production. However, when I used the SEBE algorithm, I got the final results in kWh/year (I believe) during the last step, "Zonal statistics", but the difference between the two results and the actual electrical production is significant. Do you have any ideas on what conversions I should perform to obtain similar values using the SEBE algorithm? If i try to convert from Kwh to Mwh it doesn't help

As example, this is the actual value of total electricity produced for one of the building where PV panels are planted:

This is the estimation of electrical production power using the ArcGis's algorithm (MWh/year):

And this is the result that i obtain by using the SEBE algorithm (Kwh):

Also, the point value of the Energy_mp_slope65_roof raster seem to be a little higher

Thank you so much!

[biglimp]

Interesting comparison. SEBE output is kWh/year if a metfile that is one year long is used. If you have multiple years in your metfile, the model calculates an average still giving you kWh/year. If you are using a shorter time length in your metfile, you get values for that period only. The file is still called Energyyearroof.tif (un-pedagogic from our side).

How do you convert irradiance to energy production?

Maybe you can attach your final result and the metfile for us to have a look at.

[biglimp]

Hm, there is something wrong here. I suspect that your CRS is causing issues. Can you provide me with your raw output (Energyroofyear.tif) and the textfile RunInfoSEBE.txt which should be olcated in your output folder as well.

[Andrei99537]

This should be it

Energyyearroof_runinfo.zip

[biglimp]

Something must have gone wrong when you ran the model. I just tried to run with your data for the DSM that you send. Here is what it should look like. Now I get around 1950 kWh on the roof that you are investigating Try to run the model again. Maybe you switsched aspect and height raster?

[Andrei99537]

I ran the model again and indeed, now i get a mean of 2299.565 Kwh/year. i tried to convert it in energy, by multiplying the area * mean / 1000 (in order to obtain Mwh/year) to extract usable area, then i used the follow statement as electricity production:

"The United States Environmental Protection Agency (EPA) provides a conservative best estimate of 16 percent efficiency and 86 percent performance ratio. These values mean that the solar panels are capable of converting 16 percent of incoming solar energy into electricity, and then 86 percent of that electricity is preserved as it goes through the installation. (ESRI)", which means "Usable_irad_area * 0.16 * 0.86

My final value of electricity production is 196.020 Mwh/year for building on top of the map, which is a value that is more than double of the true quantity.

However, the fact is that the total amount of electricity produced by all the PV panels in my area of study matches the total amount of electricity that is produced for all of the buildings in my area for the whole year >>Refference: 519.44 Mwh/year; My calculation: 524.45 Mwh/year (even tho the quantities of electricity per building do not match at all)

Any sugestions or opinions regarding what could i do different?

Thank you so much for your time!

[biglimp]

No real good ideas here. If the values from SEBE match other models (ArcGIS and GRASS), then it must be the area and electricity production conversion that is the problem. Sorry, I dont know how I can assist you further.