Handling Particle Temperature Errors in FDS: A Troubleshooting Guide
Handling Particle Temperature Errors in FDS: A Troubleshooting Guide
If you are running a fire simulation involving sprinklers, water mist, or fuel sprays, you might eventually run into a specific pair of warnings in your .out file:
WARNING Delta TMP_G.Mesh: YY Particle: 1 Tag: XXXXX
WARNING TMP_G_N < TMP_D_N.Mesh: YY Particle: 1 Tag: XXXXX
These warnings are more than just “noise”—they indicate that the interaction between your liquid droplets and the surrounding air has become numerically unstable. Here is what is happening and how to fix it.
What Do These Warnings Actually Mean?
FDS tracks heat transfer between the Gas Phase ($TMP\_G$) and the Droplet/Particle ($TMP\_D$).
Delta TMP_G: This means the gas temperature in a specific cell changed too drastically in a single time step because of a particle. The math is “jumping” too far for the solver to feel confident.TMP_G_N < TMP_D_N: This is a physical impossibility error. It means the numerical solver calculated a scenario where the air became colder than the water droplet that was supposed to be cooling it.
This usually happens when a massive amount of evaporation or heat exchange is forced into a single, small grid cell.
Extract from the FDS USER GUIDE
17.3.6 Warning Messages Related to Droplets An important parameter for any simulation involving liquid droplets is PARTICLES_PER_SECOND on the PROP line, which is specified by the DEVC line that includes the sprinkler or nozzle coordinates. The default value is 5000 particles per second. In some cases, this value might be too low, and you may want to raise it. The reason is that each liquid droplet that FDS explicitly tracks represents many, many more actual droplets, and the effect on the simulation of this one “super drop” may be relatively large. For example, a single droplet within a grid cell might decrease the gas temperature rapidly over the duration of just one time step—something that would not happen if there were many more droplets within the cell. You might even see within the output file named case_name.err warning messages like: WARNING Delta TMP_G. Mesh: 5 Particle: 3245 This means that during the temperature update of a particular grid cell in Mesh 5, the droplet whose index is 3245 caused the cell gas temperature (TMP_G) to change too rapidly. If you only see a few of these messages, it is not a problem given how many droplets and time steps there are. However, if you see these messages persistently, you might want to increase the PARTICLES_PER_SECOND to reduce the likelihood that a single droplet will make such a dramatic change in a single time step.
Common Causes and Solutions
1. The “Fire Hose” Effect (Mass Flow vs. Cell Volume)
If you have a high MASS_FLOW_RATE on a &PROP line and your grid cells are very small, you are essentially dumping a huge amount of liquid into a tiny volume of air. The air “over-cools” instantly.
- The Fix: Check your droplet distribution. If you are using a very fine mist, consider if your mesh is too fine for that specific spray density, or if you can spread the spray across multiple particles using
NPPC(Number of Particles Per Second) on the&PROPline.
2. Nozzle Positioning and “Corner” Issues
If a sprinkler or nozzle is placed exactly on the boundary of a mesh or tucked into a tight corner of an obstruction, the particles might be getting “trapped” in a cell with limited air volume.
- The Fix: Use the
OFFSETparameter on the&PROPor&DEVCline to move the injection point slightly away from the surface. Even a shift of 1–2 cm can move the particle into a “cleaner” part of the grid.
3. Droplet Size (The Surface Area Problem)
Very small droplets have a massive surface-area-to-volume ratio, leading to near-instantaneous evaporation. This can “shock” the temperature solver.
- The Fix: Review your
DIAMETERdistribution on the&PARTline. If yourSIGMAvalue orD_32(Sauter Mean Diameter) is creating an excessive number of tiny droplets, try increasing the minimum diameter.
4. High Velocity at the Mesh Interface
If air is moving at high speed across a mesh boundary (where cell sizes change), and a particle is moving through that same boundary, the interpolation of gas temperature can fail.
- The Fix: Try to keep your spray zones within a single, uniform mesh. Avoid having a sprinkler spray directly across a boundary where a $5\text{ cm}$ mesh meets a $10\text{ cm}$ mesh.
Step-by-Step Triage
- Find the Particle: Use the “Tag” number (e.g.,
1004122441) in Smokeview to locate exactly which particle is causing the trouble. - Check the Mesh: Look at the coordinates $(X, Y, Z)$ of the crash. Is it near a wall? Is it at a mesh interface?
- Adjust the Physics: If the simulation crashes immediately after these warnings, try reducing the flow rate or increasing the droplet size as a test.
- Use Sub-stepping: FDS often tries to resolve this by reducing the time step automatically. If the simulation doesn’t crash, it might be safe to ignore a few of these warnings—but if they appear by the hundreds, your results may be physically inaccurate.
| Issue | Likely Fix |
| Nozzle near ceiling/wall | Increase OFFSET |
| Too much water in small cells | Increase NPPC or coarsen mesh slightly |
| Tiny droplets evaporating too fast | Adjust &PART diameter distribution |
| Mesh Boundary Conflict | Ensure uniform mesh around the spray pattern |
Would you like me to help you check your &PART and &PROP lines for any parameters that might be causing this instability?
CloudHPC is a HPC provider to run engineering simulations on the cloud. CloudHPC provides from 1 to 224 vCPUs for each process in several configuration of HPC infrastructure - both multi-thread and multi-core. Current software ranges includes several CAE, CFD, FEA, FEM software among which OpenFOAM, FDS, Blender and several others.
New users benefit of a FREE trial of 300 vCPU/Hours to be used on the platform in order to test the platform, all each features and verify if it is suitable for their needs
