Open biglimp opened 2 years ago
@biglimp are you making a custom building type?
No, I am using the existing ones from the pickle coming with the pypi downlaod.
@biglimp this is something that we've run into before (there's a note in the code about it).
The problem may be due to large timesteps (i.e. 3600s) causing instability in the numerical differentiation. Is your timestep very large, and if so, can you try rerunning with something smaller, like 300s?
And I agree, best to update UWG to its latest version.
Timesteps (dtSim) is already in 300s. I also tried to change the building type to MidRiseApartments but still the same error.
Some additional information. I played around a bit and found that if I increased the verToHor parameter I could run the files in error.zip without error. The error area has a few large footprint buildings which might cause the model to fail...
Any updates on this issue?
I debugged the code a bit more using the input in the attached error.zip and found out that it is the quickservicerestaurant building class that is causing the issue. This class is unstable using low verToHor numbers creating T_wallex and T_wallin to become very large negative numbers, which in turn finally cause the saturation pressure calculation too fail.
Why is building classes with no fractions in the uwg-file included in the calculation in the first place?
Hi, Just to say I ran into this error too. Increasing verToHor to 0.2 was enough to proceed to the calculation, but it’s a poor way to solve the issue.
I am experiencing an error where a log of zero or a negative value probably happens. Input files are attached.
Version of uwg according to
pip list
is 5.8.7 (using uwg.version gives me 5.4.3, maybe needs to be updated?).The region is Gothenburg, Sweden and I have successfully run uwg on similar settings (also attached). I assume there is something strange with the Warehouse building type?
error.zip success.zip