Input vs output HRR

[Timea Marton]

Hi. 

I am simulating a propane fire in a large room and i want to compare the input and output HRR. I am using a time varying HRR with a ramp. The results show that the input matches perfectly the output in case of the HRR. I was expecting a more fluctuating HRR as an output with a bit more scattered points, and for this reason i was wondering if i did anything wrong. When doing the same simulation but with a constant HRR the output HRR results in a more scattered and fluctuating curve. Why is this difference between the two?

I am have attached the FDS File and the output results gathered in an Excel sheet.

I was also wondering about the soot formation. If i do not define the soot_yield then the smokeview does not generate any smoke? However the user guide says that a default value of 0.01 is used, but still there is no smoke, unless i specify myself the soot_yield? Is there an explanation to this? Also the room fills up with smoke very very quickly, is this just because of the size of the fire?

Thank you in advance for a response.

[Randy McDermott]

Which user guide and code version are you using?  In the latest code, you need to specify your soot yield.  This is stated on p. 126 of current guide.  If 0.01 is stated elsewhere, please let me know.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/ffa94a3a-0b88-4ab0-943a-8c9b75e4589b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Kevin]

Your graph shows the burning rate, not the heat release rate. In theory, the two are directly proportional, but the HRR is the result of the turbulent combustion process. The burning rate is dialed into FDS exactly.

[dr_jfloyd]

Per the User's Guide the default SOOT_YIELD is zero.  You are responsible for setting the soot yield.  We no longer presume a default fuel or a default soot yield in FDS 6 like was done in FDS 5.

Page 122 of the FDS 6.0.1 User's Guide:

SOOT_YIELD The fraction of fuel mass converted into smoke particulate, ys. Note that this parameter is only appropriate when the simple chemistry model is applied. (Default 0.) 

Table 17.23 of the FDS 6.0.1 User's Guide:

SOOT_YIELD Real Section 12.1.1 kg/kg 0.0

Since you are using a fuel in the database of species you should just be specifying your reaction as:

&REAC FUEL='PROPANE',SOOT_YIELD=0.01,CO_YIELD=0.011/  

However if you are simulating a car park, one would presume your fire is a burning vehicle in which case does propane with a soot yield of 0.01 really represent a burning vehicle?

It isn't clear what you are plotting.  The data in the HRR comparison sheet does not seem to match any of the data in the Mesh 1, 3, or 4 sheets (which look like hrr.csv files).   If indeed you are seeing almost no fluctuation in HRR, then that could be an indication your grid resolution is inadequate.

On Wednesday, May 7, 2014 10:57:09 AM UTC-4, Timea Marton wrote:

[Timea Marton]

I might have looked at an older version, because i checked in the new one and indeed it said the soot_yield is zero by default. This answers one of my questions then. Thank you.

However exactly because the HRR is the result of a turbulent combustion process i would expect a more fluctuating graphic. The heat release rate is in the same excel file but in a different sheet. I do know the HRR and burning rate are directly related, but i still wonder why the graphic is this linear.

[Randy McDermott]

Do you expect all fuel to burn before exiting the domain?  HRR is an integrated quantity (spatially integrated over the domain).  If all the fuel burns, there will not be fluctuations.  Put a SCLF or DEVC of QUANTITY='HRRPUV' if you want to see the local fluctuations.

On Wed, May 7, 2014 at 11:21 AM, Timea Marton <marton...@gmail.com> wrote:

I might have looked at an older version, because i checked in the new one and indeed it said the soot_yield is zero by default. This answers one of my questions then. Thank you.

However exactly because the HRR is the result of a turbulent combustion process i would expect a more fluctuating graphic. The heat release rate is in the same excel file but in a different sheet. I do know the HRR and burning rate are directly related, but i still wonder why the graphic is this linear.

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+unsubscribe@googlegroups.com.

To post to this group, send email to fds...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/05f1e5a2-cab7-459d-80a0-38a9ef1803d2%40googlegroups.com.

[Timea Marton]

I have plotted the Output HRR from Mesh 3 against time and compared it with the Input HRR as defined in sheet 'HRR Comparison'. The two match perfectly, and if you plot the HRR from the other meshes those will fit perfectly as well. 

What do you mean by the grid resolution is inadequate? Isn't this in case you have a mesh that is too course? I have optimized the mesh with the characteristic fire diameter, as defined on pg 37. in the user guide. 

[dr_jfloyd]

For the first 960 seconds you have an HRRPUA of ~180 kW/m^2.  Load the HRRPUV smoke3d file in smokeview.  The fire is occupying only the first grid cell above the fire.  At this point in time you are likely underresolved and all the fuel is just burning in that first grid cell.   If you look at the actual numbers you see some oscillation but it is very small.

If you zoom the curve around 1500 s you see there is some variance there in the heat release rate.  However, you have one output every 4 s and FDS outputs the average value over that time period.  This averaging may be smoothing your curve more than you realize.

[Timea Marton]

Yes. All the fuel will burn before exiting the domain. I have done simulations for both well-ventilated fires with large openings and under-ventilated fires with small leakages. Indeed in case of a small opening the fire showed more fluctuation when the fire started to be under ventilated, and almost none before that. This answer the question just 50 % as i am still wondering why the fluctuation is still that much higher in case of a constant HRR with a well ventilated fire?

Also i am using propane just to establish some bench-marks and i will build up the model with time. But off course propane is not the correct fuel to use for a a burning car.

[dr_jfloyd]

This guidance is for buoyant plumes from fires.  You have a fire impinging on a ceiling.  No buoyant plume so maybe this rule of thumb doesn't apply.  Also the guidance indicates the validation studies have used a range from 4 to 16.   You are sitting around 7.  Maybe that is too low.

[Timea Marton]

Yes indeed i am in the range of 7 for the mesh along z axis. Even for a value around 9 (in case of Mesh 4 - Excel) the fluctuation of HRR is low. So in case i want to have a more visible fluctuation i should have an even finer grid defined along the z axis ?

Is it possible in any way to adjust the time-step so the simulation does not last as long , but keep the quality of the grid size? 

And a final question? Does it mean that the grid is not defined correctly if i do not have the energy fluctuating? 

[dr_jfloyd]

If you are under resolved, then decreasing the cell size would help.  However, as I previously noted, right now you have outputs every 4 s.  FDS averages the value over that time period when it writes to the csv file.  Some of your perceived low fluctuation may be due to this (you could try setting NFRAMES to a larger number  or DT_HRR to a smaller number than the default).   

If you force the time step to be larger, you will increase your error and either partially or completely negate any benefit of a finer grid.  Also finer grids require smaller time steps for stability.  These are fairly basic CFD principles.  If we could reliably take larger time steps, we would already do so.

If the results of a simulation indicate behavior that does not seem physical, then that is indicative of either an issue with the grid or some other input error.

[David Sheppard]

FDS developers please correct me if I am wrong on this.

 

The SOOT_YIELD values that are mentioned here and in the FDS User’s Guide are only examples.  Different fuels and burning conditions will create different SOOT_YIELDS.  There is currently no recommended test method for measuring the SOOT_YIELD values to be used as input into FDS.

[dr_jfloyd]

The fact that soot yield is dependent on the specific fuel and burning conditions is a primary reason in FDS 6 that we got rid of the default soot yield that was in earlier versions.  It is now up to the model user to determine an appropriate soot yield for their specific application either by testing on their own and/or developing a yield based on existing published data.