Problems with radiation result

Hi Mostapha, and everyone!

I simulated some days and hours using ladybug e honeybee and I set São Paulo epw.

So, the december is the summer and june winter. The corectly result would be december, march, september, june.

I did two kind of simulation. The first one - just one hour 10h and then 15:30. The second, 10:00 to 15:30h. I think that’s something wrong with the results kWh/m² because the biggest values for radiation, are for winter. And the results simulation 10:00 to 15:30h the result are different too, the biggest values for winter (june), then september, march, and them december (summer)

The results are (kWh/m²)

10:00h 15:30h 10 to15:30h

21/03 0,69 1,15 2,61

21/06 1,14 1,13 3,71

23/09 0,96 0,90 2,79

21/12 1,31 1,22 2,45

I will be very gratiful with your answer I’m using this software to a important academic work, and in my Country Its not commom use this software, I don’t know anyone that could help me with this. I’d like to encourage university start to use this kind of software.

Thank you

Camila (50.2 KB)
BRA_Sao.Paulo-Congonhas.837800_SWERA.epw (1.49 MB)

Hi Camila,

I took a look at your file. Seems ok. I suggest looking at annual direct normal radiation data for the Sao Paulo location. That should shed more light I believe.

Hi Devang,

How/where can I take a look annual direct normal radiation data for the Sao Paulo?


Please find the file attached. You may observe that the data for the dry bulb temperature is probably in line with your expectation, but the direct normal radiation data is not.

Sao Paulo Annual (421 KB)

And how could I solve this problem?The problem is with epw?


I believe the problem is neither with Ladybug nor with the epw file. The TMY (Typical Meteorological Year) format used for most weather data files was developed by Sandia national library. The method is based on 12 years of measured data. Starting with January, January which is closest to the mean of twelve Januaries is selected based on a composite weighing of solar radiation, temperature, humidity and wind velocity. The same process is applied for the following months. So as a result, each TMY month most likely correspond to measured data from a different year. That is how the weather data file is composed. Due to this averaging, the data sometimes surprise the residents of that location.

I understand.

I opened your file attached, is exactly this ! The temperature is ok, but radiance shuold be like temperature. What could I do ?

Doesn’t have sense, just some part of the world could use the software. So I don’t have any idea how to use this software and get reliable results to radiance.

Do you know something that I could do to have reliable results for my simulations?

Someone knows if is it possible edit Typical Meteorological Year values to get cerectly results to my location ?

I suggest you to check some days before/after the ones you are trying to check. It may be that for the specific day the atmospheric conditions affected the direct radiation. The assumption that the 21 of the month is the critical day is true but mostly in theory.


How could I know if my algorithm’s results (attached other day) are from direct normal radiation or diffuse horizontal radiation or global horizontal radiation?

Because looking at Devang file, I think the results that I think that’s incorrect its from direct normal radiation. Could I change results to get global horizontal radiation ?

How could I modify the algorithm to get global horizontal radiation?

I’m desperate about this !

I really don’t understand your worries.

First for all the radiationAnalysis component uses the selectSkyMtx which uses the Global Radiation (unless you ask to remove direct or diffuse).

Second, i’ve made some tests, and the results are pretty much consistent to your expectations (December the highest, June the lowest). But i also get a case where the 21.6 at some hours is higher than 21.3, so i changed the day (as i wrote previously) to the 30.3 and then the results are fine (they were fine also before, but just for your peace of mind).

Attached the file i worked on. I plotted a 3dChart also so you can see the differences in the data through the year.

See red arrow to see the changed day.

Hope this makes sense for you now.

-A. (942 KB)

First of all thank you very much to help me !

I’m understanding what you did. It’s different from what I did. I was comparing the values of colors variations (red to blue), in this case june was the highest number. (In my way)

You put maximum and minimun values to legend par. (And the number was satisfying. Do you know if this results, attached, is in kWh/m²)

I need to see the image too, not only the numbers, I understand that in your way I saw only the number results, isn’t it?

Look the image december and june (attached). June is more “red”. Like your way results, the corect is the oposit. How could I get corect images ? (understand my question?)

Sorry for so many questions


You are comparing 5 hours on June 21st with 5 hours on December 21st. That is just not enough data for one to draw any conclusions. Because as per your argument, it is impossible for any five hours in June to be warmer than any five hours in December. Direct-Normal radiation is purely a function of the amount radiation received directly from the sun. So, if the five hours in December 21 in your calculation were cloudy, the amount of direct-radiation for those hours will be zero. Secondly, your statement that… “The temperature is ok, but radiance shuold be like temperature…” is not correct. It can be that you feel warmer on a cloudy day than you do on a sunny day.

The monthly weather data for Sao Paulo indicates that most of the rainfall occurs in the months of January and December. This implies that the direct-normal radiation for those months will be low, but the global-horizontal radiation will be high. And this is consistent with the results you are getting.

Try comparing more number of hours in Winter versus Summer. Something like (10AM to 3PM)* ( June 1 - June 30) versus (10AM to 3PM)*(December 1 - December 30). Finally, for the climate in Sao Paulo, it makes sense to compare direct-normal and global horizontal radiations separately.

Completely agree with Sarith.

Radiation and temperature are not the same.

If the checking is on a specific geometry, there is some logic that in winter time you get MORE radiation in south (north, in Brasil). The reason is simple: the sun is lower and more perpendicular to the wall.

I’ve been doing some tests so you can get this picture better. See radiation roses for your 4 analysis period. You see there that in june (your winter) the radiation is higher than in june for the north facade (you can also see there the amount of diffuse and direct).

I also added some basic geometry to illustrate the orientation influence. You see there the same as in the roses.

I’m sure that after this you’ll be more clear in the topic.

-A. (696 KB)

I understood your point and agree too. But in this argument “sun is lower and more perpendicular to the wall” June has more radiation on the wall, then September/march and then December. In addition, this really make sense. (But didn´t happen on this way with my results)

I did table and graphic (with numbers results for colors variations – red to blue) to expose better what is my point.

Looking radiation roses, the sequence - September, June, December and March (my results to 10:00h) is similar what happened to direct radiation.

The sequence March, December, June and September, is similar what happened with diffuse radiance. (My results to 15:30h)

Then the period 10:00 to 15:30h is similar to the graphics diffuse radiance too.

So I’m thinking what’s happening with my results, has more relation with radiation diffuse or direct (radiation roses), then with sun position. Because on this way March and September would be similar radiation. Do you agree? Is it correctly?

If this argument proceed, I did not found any theory that says that on the morning has more direct radiance or in the evening has more diffuse radiance. Do you know where these climate information are taken? To know if are reliable!

Thank you very much

Results.pdf (279 KB)

The source of the weather files can be found here.