Getting numerical instability at the start of the simulation

[Navaneeth Ramesh]

Hello everyone,

I am trying to run an FDS simulation of a stadium by incorporating the effect of wind. I have included the wind profile into the model using the configure wind option under the environment tab. I received an ifc model of the stadium which I have imported into pyrosim and have meshed just the region I want to simulate. Upon starting the simulation on FDS it ran uptil a time step of 1.01s before it threw me a numerical instability error. I had tried running the same simulation by considering the entire stadium model at first but it threw me a numerical instability error after a time step of 0.04s. The initial model had a mesh count of over 17 million cells. Fearing that it was the size that could be a problem I considered just a small part of the stadium and brought down the mesh size to a little over 4 million but I am still getting the same numerical instability error. Could anyone please let me know where I am going wrong? I have attached the pyrosim models that I have mentioned about. 

Thank you in advance.

Stadium 17 million mesh: https://we.tl/t-EJexOa6tI8

Stadium 4 million mesh: https://we.tl/t-4jIcg8GJ9T

[fde]

It is not reasonable to debug a complex model. You can simplify it in steps to see where it fails. 

Are you running with the latest version?

Is there any obstruction touching with the wind boundary surface?

Is there any mechanical ventilation in the model (supply, extract)?

[Navaneeth Ramesh]

Yes I understand it is a complex model to debug.

Yes I am running it in the latest version of FDS 6.8. No, there aren't any obstructions touching the wind boundary surface. There aren't any mechanical vents in the model.

[Rahul]

One thing you could try is to check the 3D slice FDS generates to find the region which is causing the numerical instability. 

Another way to find the error would be to remove obstructions that might not have a significant influence on the results. Usually, an IFC model has too many details that might not be important for a fire simulation. There could be several small pressure zones where there can be pressure fluctuations. 

Check the CHID_hrr.csv to see how many pressure zones you might have. Simplifying your input file will help.

You could also improve your mesh definitions, it will be easier to debug. 

[Kevin McGrattan]

Can you run the case without the wind?

[Navaneeth Ramesh]

Yes I did check the   CHID_hrr.csv and noticed too many pressure zones like you mentioned. I am trying to get rid of those obstructions from the model that is not required.

[Navaneeth Ramesh]

Yes I was able to run the simulation without the wind. Although I did not run it for the entire simulation duration. I ran it till a time step of approx.100s until which I did not face any issues.

[o...@aquacoustics.biz]

Try ramping your wind profile from over an appropriate time before t=0 to establish the field slowly.  Introducing a velocity instantly around geometry can result in physically impossible pressure profiles.  These transients are likely the cause of your numerical instability.

[Kevin McGrattan]

If you impose the wind close to your geometry, it can lead to an unnatural flow. Typically, wind is imposed at the boundary of the domain far from the building.