Time calculation

[sgonzalez]

Hi,

We have got Pyrosim version 2010.2.1621. We can calculate the simulation of a fire with the program and It cost four hours.

When we incorporte a exhaust fan, it cost twenty days.

The building have two floors and a surface of 15.000 m2 with a lot of rooms.

The ram memory is 2 GB.

Is it normal the time calculation?

Thanks

[Charlie Thornton]

Going from a run time of 4 hours to 480 hours is a bigger increase that I would expect to see. Adding a vent, or a burner, or anything else that complicates the simulation tends to make the simulation run time increase. However, in your case I suspect that you crossed some sort of computing power threshold on your hardware or OS. If I had to guess, I'd say that after adding the vent the simulation no longer fit fully within your RAM and the computer was having to swap memory to the disk for every iteration (i.e. thrashing).

I suggest reducing the mesh resolution or, if that isn't an option, breaking the problem up into multiple meshes of roughly equal size.

[sgonzalez]

Hi

What is the minimum mesh resolution accepted?

I have cells 1x1x1 m. If I put cells 2x2x2 m the simulation isn´t correct. The simulation runnig but it isn´t correct.

Thank you

[Charlie Thornton]

Officially, the only way to determine the correct mesh resolution is by doing a sensitivity study. If you want to estimate a reasonable mesh, consider the general guidelines:

1. The mesh should be fine enough to properly resolve your geometry.

2. The cells should be cubes or very close to cubes.

3. The size of the cells should be 0.1 - 0.2 of the characteristic fire diameter based on heat release rate. (calculator)

That should get you in the right ballpark. Often, 0.2 of D* gives a finer resolution that people were hoping for. If you want to ignore guideline #3, you can also do runs with smaller and smaller cells to find the resolution where the answer converges (a mesh sensitivity study).