[FDS-SMV Developer Blog] Pyrolysis Model in FDS

[Kevin McGrattan]

In the next few weeks we plan to write a few articles to alert you to some changes made in FDS 5.4. This article focuses on the solid phase pyrolysis model; that is, where you specify a reaction or reactions that the solid undergoes in releasing fuel gases. There is still confusion among users as to the various ways that you can describe a fire. Most simply specify a heat release rate per unit area (HRRPUA). In that case, there is no need to provide more details about the solid phase reaction. HRRPUA is equivalent to a gas burner that you control via RAMPs or similar parameters.

However, if you specify material (MATL) properties, including kinetic parameters, to describe one or more reactions that occur as your solid heats up, be aware that the definition of the parameter REFERENCE_TEMPERATURE has changed in 5.4. In previous versions of FDS, this parameter, along with REFERENCE_RATE, was used to calculate the Arrhenius parameters A and E in the reaction rate expression. These parameters are typically found via thermo-gravimetric analysis (TGA) or similar small-scale measurement techniques. Since these measurements are most often unavailable for a particular material of interest, FDS provides a way, via these "reference" values, of estimating the kinetic parameters. You have also probably seen discussion of genetic algorithms that also are intended to estimate various kinetic parameters when only a partial number are measured directly.

The change made in FDS 5.4 to the definition of REFERENCE_TEMPERATURE was intended to make the FDS pyrolysis model more consistent with current trends in materials testing and analysis. For a good introduction, read Jose Torero's chapter, "Flaming Ignition of Solid Fuels," in the Fourth Edition of the SFPE Handbook. Then read the FDS User's Guide section entitled, "Solid Fuels that do NOT Burn at a Specified Rate." All of this development is focused on the long-term goal of standardizing the process of obtaining material property data. A necessary first step is to understand the meaning of typical TGA results (a good example can be found in Torero's chapter). Then we need to translate this information into FDS inputs.

We'd like to continue the dialog on pyrolysis modeling that started last year, but has recently stalled a bit. We are starting to notice yet again that many users are simply cutting and pasting lines of input from the User's Guide and sample cases without really understanding their meaning. We are developing a suite of verification cases that can be used to check the basic kinetic parameters (a simulated TGA experiment), as well as cases to assess the overall solid phase model (a simulated cone calorimeter measurement). Nick Dembsey, Marc Janssens and Morgan Hurley are continuing their multi-year effort to develop a standard guide for obtaining material properties. Our work in FDS will hopefully move us closer to the goal.

So for those of you with an interest in this area, it would be very useful to get your feedback. I will post this blog to the Discussion Group so that it will appear near the top of the list of issues.

--
Posted By Kevin McGrattan to FDS-SMV Developer Blog at 9/08/2009 11:28:00 AM

[Münch]

Hi Kevin,

I just have some problems to understand the interaction between the
mixture fraction model and the pyrolysis model. As described in the
documentation their is only one fuel in the gas phase, therefore only
one HEAT_OF_COMBUSTION value can be used to heat up the flow through
the combustion process in the gas phase.

In the MATL lines of the pyrolysis model many different materials can
be defined, each with his own HEAT_OF_COMBUSTION. For the following
example I use some short forms:

HEAT_OF_COMBUSTION in the MATL line : HC_p1 and HC_p2 for material 1
and 2
HEAT_OF_COMBUSTION in the REAC line : HC_g

Now, assumed I'm able to adjust the pyrolysis model for two solid fuel
materials to produce a measured mass loss rate. What happens inside
the computation if HC_g = HC_p1 = 0.5 HC_p2?

Because of a radiator both materials start the pyrolysis model and
compute the mass loss rates m_p1 and m_p2. The mixture fraction model
compute a change in the mixture fraction variable Z near the material
surface and therefore the heat release in the gas phase should be m_p1
* HC_p1 and m_p2 * HC_p2. But beause of the one-fuel-fact in the gas
phase and the correction algorithm in FDS5 the computation is

m_p1 * HC_p1/HC_g * HC_g = m_p1 * HC_p1

m_p2 * HC_p2/HC_g * HC_g = m_p1 *2 *HC_p2


Therefore the object with the material 2 lost m_p1 of mass but burned
2*m_p1 in the gas phase.

Is that correct?

Have nice weekend
Matthias

[drjfloyd]

If the HOC of the MATL is different from the HOC of the gas then
pyrolysis rate computed by the pyrolysis model is adjusted. For
example if HOC on REAC is 50,000 and HOC on MATL is 10,000, then if
the pyrolysis rate were computed as 1 kg/m^2/s, it would be adjusted
to 0.2 kg/m^2/s IN THE GAS PHASE. In this way the combustion model
will result in the correct heat release rate. The unadjusted
pyrolysis rate is for updating the solid phase fuel mass.

[Kevin]

Simo should check this, but I would say that the mass loss rate of
fuel gas (mdot) is

mdot = (mdot_1*HoC_1 + mdot_2*HoC_2) / HoC_gas

On Sep 18, 10:41 am, Münch <mue...@inf.fu-berlin.de> wrote:

> > 11:28:00 AM- Hide quoted text -
>
> - Show quoted text -

[shostikk]

I think the gentlemen above already answered. Mathias, I would assume
your sub-indexes 1 and 2 had jumped a little bit
in the last formula. MATL 2 would produce 2 * m_p2 to the gas phase,
but loose only m_p2 in the solid phase.

[Münch]

Simo that right. Thank for all answers.

Matthias

[Kuldeep Prasad]

Jason,

As you (and others) have noted, that the pyrolysis rate computed by
the
pyrolysis model is adjusted if the HOC of the material is different
from the
HOC of the gas. The unadjusted pyrolysis rate is still used for
updating the
solid phase fuel mass. This stems from the fact that FDS has a single
gas phase
heat of combustion, but allows for different materials to burn and the
adjustment
(noted above) is needed to ensure that the energy is conserved.
Unfortunately, this
results in the problem that the solid is being consumed at a rate that
could be
different from the rate at which fuel gases are entering the gaseous
domain.

One solution could be to change the gas phase heat of combustion (as a
function of time) as follows
HoC_gas (t) = mdot_1(t)*HoC_1+ mdot_2(t)*HoC_2. This would ensure that
both mass and energy
are conserved at the gas/solid interface. No adjustments will then be
needed for the
pyrolysis rate.

Do you think that this would improve the formulation or would it
potentially result in other problems.

Kuldeep

[Kevin]

In a large calculation, there might be many things burning in many
different places. Literally dozens of different materials. Currently,
we do the pyrolysis modeling without any mass adjustment. The
adjustment is made when mdot is applied as a boundary condition. Then
we adjust mdot so that the appropriate amount of surrogate fuel gas is
injected into the gas phase. The difference in masses between solid
and gas are insignificant compared to the overall mass budget.

> > - Show quoted text -- Hide quoted text -

[drjfloyd]

That introduces problems. Fuel does not burn the instant it reaches
the gas phase but rather when it reaches O2 under conditions where
combustion can occur. So fuel released into the domain at time t
doesn't burn at time t but at a wide range of times larger than t
depending on how long it takes the fuel to encounter oxygen. At that
future time we have no knowledge of what fraction of the fuel in that
gas cell comes from which fuel stream. Using a hoc based on the
current production rates would not be correct.

On Sep 21, 11:03 am, Kuldeep Prasad <prasad.kuld...@gmail.com> wrote:

[Kuldeep Prasad]

Jason,

Currently in FDS, we are modifying the instantaneous pyrolysis rate to
match the
gas phase HoC. The pyrolysis rate is modified even though we know that
the fuel
gases that are entering the computational domain do not burn at that
time, but
actually burn sometime later when they encounter oxygen. So the
current implementation
has the same problem that you have described.

The approach that I have suggested has the added benefit in that in
balances
the mass at the gas - solid interface.

Kuldeep

> > > - Show quoted text -- Hide quoted text -

[drjfloyd]

Getting the right total heat release is likely to be far more
important to the user than the small amount of mass error associated
with adjusting the pyrolysis rate. With your proposal, all mass
everywhere will burn with the new instantaneous HOC. Unburned fuel
can accumulate and burn at a later time, you cannot base the HOC on
the current mdots of various fuels. With your proposal we conserve
mass at the cost of not adding the correct amount of energy to the
gas.