Model Comparison

All around LEONARDO 4 (and previous 2014)
Ila
Posts: 39
Joined: Sun Sep 16, 2018 4:22 pm

Model Comparison

Postby Ila » Wed Jun 24, 2020 7:56 am

Hi Envi_met team,
I've simulated two models: reference urban canyon and green scenario with living walls applied at the ground floor of the south exposed building. The model areas and SIM files are the same, except for the living wall.
I compared the output with Leonardo, but I noted weird patterns, especially in Mean Radiant Temperature and PET.
The air temperature comparison seems consistent, with a very slight effect adjacent to the facade.
The mean radiant temp has unexpected differences around every building and in the shading area. Even if they are very low and probably negligible, I'm not able to explain them.
The PET outputs look the same, but the comparison map shows a weird pattern, with a lot of strange "pixels". (similar smalls squares occurred at different hours)
Have you got ideas about possible explanations? May they are Leonardo bugs, connected to the visualization?
Thanks for your support,
Best regards,
Ilaria
Attachments
Screenshot (904).png
PET comparison
Screenshot (917).png
PET green scenario 16h
Screenshot (915).png
PET reference canyon 16h
Screenshot (893).png
Comparison Mean rad temp 16h
Screenshot (886).png
Comparison Tair 16h

helge
Posts: 114
Joined: Mon May 15, 2017 12:27 pm

Re: Model Comparison

Postby helge » Wed Jun 24, 2020 8:58 am

Dear Ilaria,

Thank you very much to bring that to our attention. Could you please upload screenshots of absolute PET for 16:00h with and without the greening?
Also it would help if you could upload (to wetransfer) the model results of biomet for 16:00 (edt and edx file) for the simulations with and without the greening?

Best
Helge

Ila
Posts: 39
Joined: Sun Sep 16, 2018 4:22 pm

Re: Model Comparison

Postby Ila » Wed Jun 24, 2020 10:17 am

Sure Helge,
I confused the screenshot, sorry.
Here we go the wetrasnfer link with the useful files.
https://we.tl/t-CgLJiCH0ra

Thank you very much.

Regards

Ilaria

Tim
Posts: 1496
Joined: Mon May 15, 2017 3:46 pm

Re: Model Comparison

Postby Tim » Wed Jun 24, 2020 12:26 pm

Hi Ilaria,

We could locate the issue and are trying to fix it. Could you also upload the atmosphere outputs of one of those hours?

Best regards,
Tim

Ila
Posts: 39
Joined: Sun Sep 16, 2018 4:22 pm

Re: Model Comparison

Postby Ila » Wed Jun 24, 2020 2:02 pm

Hi Tim,
thanks for your support, I hope that you could fix it.
https://we.tl/t-Eu3G14ds0v

This weird pattern occurs in different hours and It is present also in the next simulation that I've done.
This sim has the same area too, but the green walls are applied in both buildings of the canyon, not only in south one, as before.

I'm using Envi-met 4.4.4 Winter 19/20 with Leonardo 4.4.4
Currently, I'm not updating the latest version, because I've started my research and the validation step with this one and I'd like to go on with the same software.
Should I upgrade to solve the problems?
Thank you

Ilaria
Attachments
Comparison PET 17 reference- total green.png

Ila
Posts: 39
Joined: Sun Sep 16, 2018 4:22 pm

Re: Model Comparison

Postby Ila » Tue Jun 30, 2020 7:17 am

Hi Tim and Helge,
have you got some news about the issue? It is related to the simulation or only with Leonardo visualisation?
Waiting for your response, thank you for your kind attention.

Ilaria

Tim
Posts: 1496
Joined: Mon May 15, 2017 3:46 pm

Re: Model Comparison

Postby Tim » Tue Jun 30, 2020 7:27 am

Hi Ilaria,

We could not fix the bug to be clear but we are however not sure if it is a bug at all. It is about Biomet not Leonardo. If you check the values of the cells without comparison, you will see that they are moving in 0.2 K steps between each cell. A comparison of two simulations with those large step sizes will then give you these results for your simulation and the PET analysis. We are now checking if the step sizes needs to be smaller. However, also steps of 0.05 K will give you the result pictures that you screenshotted above. However, the Biomet simulation would then take much longer.
We are not in conclusion yet but I think that it will not be changed in near future.

One of the reasons for this behaviour probably is your simulation. Because of the large blocking buildings and only very small open space at the borders, wind cannot flow through the model as it should. It is pressed around the buildings forming those extreme patterns, especially because the wind ist not coming from west or east but from north where the huge wall of a building exists. I would rather rerun the model area twice or thrice as big to get more open space and thus allow a better wind flow. Then in a coarser resolution to save grid cells and simulation time. I imagine I gave you the tip in another post about this model already. However, it will improve the results in general but I cannot tell you if it will improve the issue with the step sizes in PET.

Best regards,
Tim

Ila
Posts: 39
Joined: Sun Sep 16, 2018 4:22 pm

Re: Model Comparison

Postby Ila » Tue Jun 30, 2020 9:41 am

Hi Tim,
thanks for the explanations...
Concerning the increase of the model dimensions, I've some doubts:
I need 1m resolution for the type of analysis, so I worry about a very long simulation time...

- I've got a 3D model with specific facade materials and green walls: do I have to reconvert in 2.5 model to add empty cells at the borders and then model the building again? ( function "add cells" in the "tool" windows). Could I leave them made by default loamy soil?
Or is it better to create a new model area with new established dimensions?

- I know that you usually suggest preferring "normal" cells instead of nesting grids, but with nesting grids, I could save "space" and "time" thanks to their fast and progressive increasing, avoiding to re-model the area. Are they a possible solution to increase open space?

Have you got some advice about how to solve the open space issue? Thank you for your support.

Regards
Ilaria

Tim
Posts: 1496
Joined: Mon May 15, 2017 3:46 pm

Re: Model Comparison

Postby Tim » Tue Jun 30, 2020 1:06 pm

Hi Ilaria,

Well, if you really need it in 1 m resolution, it is probably not praticable.
If you want to Add empty cells in Spaces, you need to reconvert to 2.5D, yes. I'd rather try to create a new model area.
Maybe the nesting grids will help you with that issue. You could try that.

Best regards,
Tim

Ila
Posts: 39
Joined: Sun Sep 16, 2018 4:22 pm

Re: Model Comparison

Postby Ila » Tue Jun 30, 2020 2:22 pm

Thank you very much Tim for your response.
Best regards,

Ilaria


Return to “LEONARDO 4”

Who is online

Users browsing this forum: No registered users and 1 guest