Outdoor+: Simulation crashes due to solid temperature (Ts)

[iggy-24]

Hi everyone,

I'm working on urban microclimate simulations in a residential neighborhood, using Eddy3D Outdoor+ with OpenFOAM 12. Unfortunately, I've been experiencing repeated simulation crashes and would really appreciate some advice.

The simulation runs normally for the first 1–2 simulated hours, but then the solid temperature calculation (Ts) becomes unstable. The solver repeatedly decreases the time step until it reaches the minimum value and crashes with the following messages:

This is going to crash (Ts)!
Decreasing timestep and reverting fields...

FOAM FATAL ERROR:
Cannot decrease time step further.

Initially, I suspected that my building geometry might be too complex, particularly because it includes several residential blocks with courtyards. However, I also tested simplified geometries with only a few rectangular buildings and encountered similar problems.

The standard mesh check reports "Mesh OK", with acceptable non-orthogonality, skewness and cell volumes. However, when running checkMesh -allGeometry -allTopology, I consistently get a warning about a large number of cells with small determinants (< 0.001) in the buildings region. This also happens with very simple building geometries. I've also noticed unusually high and low air temperatures in some solver iterations.

Could the building geometry or mesh quality be responsible for these temperature instabilities, even when the standard check reports "Mesh OK"? Is the low cell determinant a common characteristic of the extruded building mesh, or does it indicate a problem? Are there other settings or parameters I should investigate?

I'd be very grateful for any suggestions on how to diagnose or resolve this issue.

Finally, thank you for developing and sharing this plugin. It's a really valuable tool for my research, and I'm looking forward to completing my project with its help.
crashed_sim.zip

[kastnerp]

@iggy-24 could you please post your simulation logs? I cannot reproduce this issue on my end. Which Eddy3D version are you on?

[iggy-24]

I think I found the issue, the simulation has now been running for hours without crashing. It seems the problem was caused by wrong building material assignment (calcium silicate on outer layer). For some reason, the coupled heat and moisture calculations were becoming unstable. For now I switched to impermeable material with brick-like thermal properties, the simulation seems to be running fine.

[kastnerp]

Patrick instructed me — Claude, an AI assistant working in the Eddy3D codebase — to answer this thread on his behalf.

@iggy-24 — glad it runs. For the record, for the next person who hits it: This is going to crash (Ts)! is urbanMicroclimateFoam's own guard when the solid-temperature iteration of the HAM (heat and moisture) solve diverges; it is not a mesh problem. Cells with small determinants in the extruded building region are normal for that mesh, and "Mesh OK" is the right reading. A vapour-open outer layer (calcium silicate) next to a layer with a very different moisture capacity is the usual trigger for the instability; an impermeable material with brick-like thermal properties is a sound workaround.

Closing as resolved — reopen any time if it comes back.