Dear community
When I set the high denisy to true in LB Cumulative Sky Matrix, the mid-hour values do not make sense to me (see images below). Do you agree with me? I wonder if there is a bug in interpolating radiation values.
All the best,
Farhang
1 Like
Hi Farhang,
I am not quite sure if your question is related to what I have in mind or not, but apparently, LBT does not calculate the sub-hourly radiation values from .wea files like DAYSIM does. I have asked this in forum while ago here, might worth checking.
Regards
Amir
Good question, @farhang.tahmasebi .
You are correct that we don’t use sub-hourly values by default, @AMIRTABADKANI . However, there’s a way to create the sky matrix using sub-hourly radiation values now in the LBT Ladybug plugin. You just need to pass the Radiation Data Collections through the LB Convert to Timestep component and that will interpolate the data such that you’ll then you’ll be generating the sky matrix using a higher-timestep Wea. Note that you’ll also need to pass the Radiation data (kWh/m2) through a LB Time Rate of Change component in order to convert it Irradiance (W/m2) before you interpolate. Otherwise, your interpolated values will just be the cumulative radiation over each sub-hourly timestep instead of the average irradiance values that gendaymtx expects.
Here’s what the workflow looks like:
finer_timestep_radiation.gh (25.7 KB)
2 Likes
That is great @chris, and thanks for your great team effort,
The point is that these sub-hourly values are interpolated linearly, while if we generate the same .wea file with the same timesteps with DAYSIM directly, results are different, and I still do not have any clue which approach is more accurate to go for, although I decided to use the generated sub-hourly DAYSIM .wea file instead of LBT component in my workflows.
Hi @AMIRTABADKANI ,
If I am correctly remembering the sub-hourly methods that DAYSIM uses, it has some way that it tries to account for scattered clouds and the way that this might give high irradiance in one timestep and low irradiance in the next timestep. I don’t see anything fundamentally wrong with this approach but I also don’t know of any solid evidence that this scattered clouds model is consistently better than just linear interpolation. I think the authors of the model did a validation with “an office in Stuttgart” but, as far as I know, this validation hasn’t been replicated in other climates.
To lend a little credence to the linear interpolation method, I can confirm that this is what EnergyPlus does in order to convert the hourly EPW values to the simulation timestep (usually every 10 minutes). We know that E+ has been validated a lot and so I would infer that linear interpolation is “good enough” for cases when you’re using the solar data to inform building load balances. Maybe the scattered clouds method might improve some daylight and thermal comfort analyses but my understanding is that the jury is still out on this and we’ll probably need more validation to be confident that this model is better than linear interpolation.
2 Likes