Make HRRPUL work across meshes

[rmcdermo]

Currently, HRRPUL only integrates across the mesh where the DEVC lives. I think that at this point we need to make this multi-mesh compatible. It is important for diagnostic purposes that we are able to do this easily, and this usually means putting a line of DEVC in the middle of the burner and checking that SUM(HRRPUV*DZ) gives the correct HRR. And we are also wanting more and more to check the grid dependence of HRRPUL(Z) in our attempts to develop grid independent fire growth models.

Here is a simple test case. The burner is 1 m2 and has HRRPUA=100 kW/m2. The _hrr.csv file shows 100 kW. The _line.csv file will integrate to 100 kW if there is 1 mesh. It will integrate to 25 kW if I split the domain into 4 meshes.

test.fds.txt

[mcgratta]

I think this feature can be replicated by the more general approach where we can have a linear array of points, each of which uses SPATIAL_STATs. I'll see.

[mcgratta]

&DEVC XB=0,0,0,0,0.05,1.95, POINTS=20, QUANTITY='HRRPUV', ID='hrrpul', DX=1, DY=1, DZ=0.05, SPATIAL_STATISTIC='VOLUME INTEGRAL' /

This yields the HRR in each horizontal slice. You can get it to print out hrrpul by using a CONVERSION_FACTOR of 0.1 and UNITS='kW/m'

[rmcdermo]

Yep. Works great. Sorry, I did not see that in the user guide.

[mcgratta]

The general methodology was put in after HRRPUL. We should probably remove HRRPUL, but we have to first replace it in all the V&V. I'll do that.

[Ranhill]

Can you please explain a bit about the physical significance of conversion_factor being 0.1 here? I do understand when the conversion_factor is used. But why it is 0.1 here is not clear to me.

[mcgratta]

@Ranhill What are you talking about?

[drjfloyd]

I think he is referring to your post two weeks ago where you wrote:

This yields the HRR in each horizontal slice. You can get it to print out hrrpul by using a CONVERSION_FACTOR of 0.1 and UNITS='kW/m'

[mcgratta]

OK, the CONVERSION_FACTOR is sometimes a convenient way to change units of the output. The 0.1 in this case is related to the particular grid size of this case.

[Ranhill]

Thanks for the answer. Still not clear.

If we are multiplying HRR*grid_length (0.1) then we will get kW.m. Or, we take the conversion factor as 10 (reciprocal of 0.1m).

Otherwise, I think we should get HRRPUA (not HRR) as an output and then multiply it by 0.1(m).

HRRPUA*grid_length = kW/m

[mcgratta]

You may be right. I probably have it backwards.

[mcgratta]

HRRPUL has been replaced by something like this

&DEVC XB=0.0,0.0,0.0,0.0,4.50,337.50, QUANTITY='HRRPUV', DX=72., DY=72., DZ=4.5, SPATIAL_STATISTIC='VOLUME INTEGRAL', POINTS=38, Z_ID='Height', ID='HRR' /

That is, at each POINT, we integrate HRRPUV over a volume that is bounded by x-72<x<x+72, and similar for the other coordinate directions. The result is an array of HRRs, not HRRPUL, but for the purpose of estimating a flame height, we use the PERCENTILE function to figure out that height at which 97% (or whatever) of the total energy lies.