Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question

Issue with Solar Radiation collection

asked 2018-02-07 18:22:55 -0500

carlobianchi89's avatar

updated 2018-02-09 13:51:23 -0500

I am having hard times with the way EP treats incoming solar radiation. I created a simple test case: a cubic building, perfectly aligned on the north direction, only red-brick, no surrounding objects, ideal loads, constant internal room temperature, constant air temperature, constant wind speed, constant humidity. The only thing that varies is the incoming solar radiation. I used a model to determine the clear-sky radiation in my location, the TMY3 may contain some weird effects due to the clouds' presence. I set EP to output "Surface Outside Face Incident Solar Radiation Rate per Area [W/m2]" on each surface. I checked out January 1 (the first 24h of the year) and something weird happens.

All the surfaces except west and east show a symmetrical collection of radiation during the day, as expected. You can see that in the picture attached: C:\fakepath\SolRad.jpg

East and West are not supposed to be symmetric, they receive direct radiation in different parts of the day. However, the peak of incoming radiation for the west side is much higher than the peak for the east side.

How is it possible? What possibly can alter the collection of solar radiation on those 2 surfaces?

Note the material of each surface is the same. Attached you can find the plot for January 1, I can provide the .idf and the .epw file if you want to replicate the analysis.

edit retag flag offensive close merge delete

3 Answers

Sort by ยป oldest newest most voted

answered 2018-02-12 23:39:46 -0500

carlobianchi89's avatar

I think I found the problem. For some reasons there are some differences between controlling the incoming Diffuse and Direct fluxes through actuators and directing modifying the EPW file with the desired Direct/Diffuse fluxes.

I was using the TMY3 file for Salt Lake City, UT and I was inputing a .csv with customized values of Direct/Diffuse fluxes. The file was read through Schedule:file and, through the appropriate actuators, I assigned the imported values to the Direct and Diffuse components. For some unknown reasons this didn't work.

I tried to substitute the 2 customized Direct/Diffuse fluxes into the TMY3 file and everything worked. Now East and West facades are symmetrical!

edit flag offensive delete link more

answered 2018-02-09 23:05:04 -0500

Joe Huang's avatar

I've just noticed that you have also posted this same question on EnergyPlus_Support, to which I had questioned the constant direct and diffuse radiation in your epw file. After you explained you were overwriting the solar in EP using actuators (first I've heard of this feature) as you also explained here, I thought of something else that might explain the dissymmetry between the solar gain on the east and west facades. (following is quoted from my post on EnergyPlus_Support):

"Since you said you used a Monteith model that I've not heard of to generate the hourly profiles of solar radiation, I wonder if it's of European origin, in which case it might be giving you the solar radiation at the hour, while EnergyPlus and other North American simulation programs expect the solar radiation to be for the previous hour or averaged for 30 minutes before the time stamp. If this is what's going on, the solar you've calculated would be 30 minutes later than what EnergyPlus assumes it to be. This difference can have a major effect on the radiation during the sunrise and sunset hours, since in addition to the 30 minute shift, EnergyPlus will be cutting off the radiation when it thinks the sun is below the horizon."

"The easiest way to test if that's the cause of the dissymmetry between the east and west solar loads is to shift your building east by 7.5 degrees longitude. You could of course also calculate the solar at the 30 minute mark instead of on the hour."

Let me if that helps."

As a secondary comment: I don't think it's wise to post the same question simultaneously on multiple bulletin boards, or at least you should point out that fact and keep readers informed of all responses. Otherwise, as in this case, I'm reading stuff now that would have answered my questions on EnergyPlus_Support.


edit flag offensive delete link more


Dear Joe, The EnergyPlus_Support group was very little responsive in the past with some questions I posted, so I do think it is wise to post on multiple forums to increase the possibilities to get some help. You're right, I should let you know if the problem is solved on different forums. And that's what I was going to do IN CASE the problem was solved. If you read the comments above, I thought I solved the problem, but actually I didn't. So what should I update you about.

carlobianchi89's avatar carlobianchi89  ( 2018-02-10 11:27:35 -0500 )edit

I am about to test your option, even though I am a little skeptical about. Right now, including the updates I talked about in the question above, all the fluxes are symmetrical, they are just very high, too high. So I don't think it's a matter of rotating the building. Plus, 7.5 degrees is a relatively small shift, my problem seems much "bigger". I'll give it a try and I will let you know if it solves the issue.

carlobianchi89's avatar carlobianchi89  ( 2018-02-10 11:29:58 -0500 )edit

Oh, and I am sorry you've never heard about the actuators. They are quite useful features, especially if you want to customize surface-specific variables, rather than using the weather variables from the EPW file, averaged for all the domain. In my case I could have customized the EPW file with the radiation values from the Monteith model, true. But this way I can keep the same EPW file for all the simulations and modify only a simple .CSV file to test all the radiation models I want. That's pretty convenient!

carlobianchi89's avatar carlobianchi89  ( 2018-02-10 11:34:10 -0500 )edit

I'm afraid you misunderstood my suggestion. I didn't say rotate the building by 7.5 degrees. I said move the building location 7.5 degree east in longitude.

Joe Huang's avatar Joe Huang  ( 2018-02-11 18:26:53 -0500 )edit

I'm afraid you misunderstood my suggestion. I didn't say rotate the building 7.5 degrees. I said modify the building location by 7.5 degrees in longitude. These two changes are completely different.

Joe Huang's avatar Joe Huang  ( 2018-02-11 18:33:07 -0500 )edit

answered 2018-02-07 21:01:50 -0500

updated 2018-02-08 21:05:30 -0500

What is your weather file location? Western facades with higher solar gains is a pretty common result to see. The solar radiation on a surface is dependent on the direct normal, diffuse horizontal, and global horizontal radiation. As you may know, direct normal radiation is often highest between late morning and late afternoon, such as between noon and 5 pm. If you plot the Direct Normal Radiation from your weather file, I suspect you will see that during the time the peak occurs on the western facade, its higher, and during the time the peak occurs on the eastern facade, it is much lower.

When I plot results for a box model using a Sydney weather file, I get very similar results as you, with the western facade at almost double the peak of the east.

Update: since you overwrote your weather file data, then my above explanation doesn't hold true like it would for typical weather data. Though when I look at your model and plot the results, I don't get the same plot that you attached. Double check that you are outputting Surface Outside Face Incident Solar Radiation Rate per Area [W/m2]. When I plot this, the results are not identical but are very close to being mirror opposites of each other for the east and west facade on Jan 1, with a peak at 2597 W/m2 for Surface 3 and 2384 W/m2 for Surface 4.

You also need to consider Site Ground Reflected Solar Radiationwhich has some slight variation.

edit flag offensive delete link more


That's very interesting... My weather file is located in Salt Lake City, Utah. You can find all the data I used for my simulations here:

I overwrote the first 24h of direct normal and diffuse horizontal with data from the Monteith model (clear sky model), so they are expected to be symmetrical. You can try, they are in the file in the dropbox link.

Also, if the 2 radiation components were not symmetrical the south facade and the roof have a symmetrical collection?

carlobianchi89's avatar carlobianchi89  ( 2018-02-07 21:33:58 -0500 )edit

Actually I think I got what you mean! I am gonna make some more tests, but I think you're right! Thanks!

carlobianchi89's avatar carlobianchi89  ( 2018-02-07 21:50:26 -0500 )edit

Ok, good. Let me know if something still doesn't make sense and I'll look at your files. Its an interesting question!

Anna Osborne Brannon's avatar Anna Osborne Brannon  ( 2018-02-08 00:16:22 -0500 )edit

I did some tests and I am not sure I follow you. I plotted direct and diffuse radiations and they are perfectly symmetrical and perfectly centered during the 24h. I understand when you say: "If you plot the Direct Normal Radiation from your weather file, I suspect you will see that during the time the peak occurs on the western facade, its higher, and during the time the peak occurs on the eastern facade, it is much lower." My question, though, is: why?

carlobianchi89's avatar carlobianchi89  ( 2018-02-08 11:35:42 -0500 )edit

If my building is perfectly cubic, perfectly aligned with the cardinal directions, if direct and diffuse components are perfectly symmetric, why the peaks in the 2 lateral facades are not?

carlobianchi89's avatar carlobianchi89  ( 2018-02-08 11:49:33 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer


Question Tools

1 follower


Asked: 2018-02-07 18:22:55 -0500

Seen: 589 times

Last updated: Feb 12 '18