Negative values of quantity TOTAL MASS FLUX WALL and peak of HRR

[mst-parkhachev]

Dear colleagues!

There is a test project in which the combustion power increases linearly over 5 seconds.
linear.txt
We noticed that during the combustion process there is a high power peak (column HRR in the file CHID_hrr.csv).
Such a power peak often appears during simulation in more complex projects, which sometimes leads to premature activation of the fire detectors.

We decided to measure how much fuel actually flows into the computational domain. After reviewing User Guide Chapter 22.10.10 "Mass Flow", section "Mass Flow at a Boundary", we assumed that QUANTITY = ‘TOTAL MASS FLUX WALL’ was appropriate for our purposes.
That's why we used the device
&DEVC ID='DEVC_MASS_FLUX' QUANTITY='TOTAL MASS FLUX WALL' SPEC_ID = 'SPEC_1' SURF_ID = 'SURF_1' XB=1,2,1,2,0.5,0.5 IOR = 3 SPATIAL_STATISTIC='SURFACE INTEGRAL' TEMPORAL_STATISTIC='INSTANT VALUE'/

This DEVC produced the following graph

The huge fluctuations in the mass flux and the negative values ​​are confusing in this graph. I suspect that the high peak in HRR is related to the fluctuations in mass flux.
There are several questions regarding this situation:

1. Does QUANTITY = ‘TOTAL MASS FLUX WALL’ correspond to the amount of fuel actually entering the calculation domain? Or should I use a different QUANTITY?
2. What is the physical meaning of the negative mass flux values?
3. Is it true that the HRR peak is related to fluctuations in the fuel flux rate?
4. How can I reduce the magnitude of the HRR peak? Ideally, I would like to completely remove the peak. Can anything affect the magnitude of the peak other than increasing the power rise time?

Thank you for any recommendations on this topic.

[mst-parkhachev]

We use FDS 6.9.1

[rmcdermo]

I will take a look but will have to be tomorrow.

[mcgratta]

Add this line to your input file

&PRES MAX_PRESSURE_ITERATIONS=50, VELOCITY_TOLERANCE=0.005 /

It will damp down some of the initial noise. I don't know what impact the combustion scheme has. You might try using the simple chemistry approach to see if that is less noisy.