AIM - model flame spread in a CFD solver

VALIDATION EXPERIMENTS - coupled

OUR RESEARCH GROUP

SOLID PHASE ONLY

GAS PHASE

McFPE group

Coordinated effort

Solid phase x gas phase subgroups

Get involved via discussion group on github

Plan experiments meaningful to contribute

Keep being competitive within the community

apply state of art knowledge

Sensitivity tests

Learn

Pyroboard copy McFPE - main difference is material

Who is actually in it? We need to make clear who is in the team. We need to activelly involve people.

We are two institutional team - we need to decide on the direction of the cooperation. Two institutions BUT one research team?

Make a research road map for future work in the field of fire spread development modelling in TUPO and VSCHT collaboration

Enhance ideas sharing

Assign responsibilities

Check progress - "collective blame"

Better work load division

DOTÁHNOUT VĚCI DO KONCE

Aleš? - CFD aspects? Make it work for him as well for us within the group?

It is not working so far

Do we want this? How tro make it work on everyday basis.

We need each other. Make it a more viable cooperation.

Balance between an in-house level and joint research effort within the group.

No need to search for new topics, but rather specialize on features that we need to move the field forward?

CFD modelling

Experiments - how to master various techniques to measure correctly and be competitive

Programming

Make better use of master students for our interest.

Assign a "senior" research group member to a student who is actively gonna work with the student to make the data work for us.

Don´t use them to explore topics but rather specialize in certain parameter or aspect of problems we are dealing with?

Provide support

Vojta

Jirka

Vašek

BOUNDARY conditions are important - not the idealizied we used in models but the real ones in fire scenarios

we need to do a better job in characterizing thermal feedback in experiments

Repeatability is crucial - Randy says he values repeatability more than more validation scenarios.

Understand what we measure and what is the measurement error

Try to always find a metrics, value measured that is coupled as much as possible to one aspect of physics. Do not mix everything together.

We need to make sure our data from UCEEB are useful for validation of coupled gas and solid phase model

Spatial resolution of the measured data - is it enough to characterize thermal feedback from the growing fire?

Uncombustible wall plus burner

Add heat flux gauges

We need to get into theory of fires adjacent to the wall

We completely ignore soot

Work of Stoliarov on PMMA for McFPE workshop 3 - coupling

Imaging of the flame? I don´t understand what information are they getting out of it. Flame temperature? Study what we can actually measure right and would be valuable via flame imaging.

Validate as much as possible in terms of gas phase with no pyrolysis


Soot yileld - cone calorimeter

Pyrolysate - from coupling STA and GCMS

Total heat flux to a cold walll

How can we asses the convective and radiative contribution?

Continue work of Helena, make it a separate topic that will evolve in a direction we need.

We use prescribed yields

Why is it a problem? Difficult to validate radiation.

Do we want to participate in this modelling effort? Is it meaningful?

Can we help with experiments in the large scale?

Do we have computational power?

Can the heat flux be used as fitness metric in optimization?

I did not get how they exactly used the heat flux data to validate whether the pyrolysis model calibrated in the bench scale gives reasonable predicitons in large scale - they used the heat fluxes as a main validation??? Read the paper!!!

We need to understand contributions of the HRR - chemical, convective, radiative... These are terms that FDS use.

MLR and surface temperature are coupled. Surface temperature is a boundary condition to pyrolysis model

We need to understand flow in the RCT test

We need our RCT boundary conditions before we jump in pyrolysis model validation

Measure flow field

Model flow field

We can´t get a good pyrolysis model prediction when the THERMAL FEEDBACK is not properly resolved

We are decoupling gas phase and solid phase / okay to study but we need to have the coupling IN MIND

In pyrolysis we use simplified and well defined thermal boundary conditions to calibrate and validate our models

Questionable in real fire scenarios.

Fixed thermal condition is good for studying influence of K, rho, cp etc

Thermal feedback is more sensitive to radiative feedback

We have to properly resolve gas phase

apply state of the art techniques on both separately, learn our lesson and then couple

Resolving the boundary layer?

Far field versus near field - different resolution is necessary. D/10 is okay for far field.

Think about how you choose yout grid - think of physics, not geometry.

FDS is a CFD solver. We turned it in the black box.

What would be "D", Q* in a flame by wall? D keeps changing

For a properly spatially reoslved case all various solvers did pretty good job regardless of various models and settings

Wall resolved versus modeled - mesh size?

What are the characteristic spatial scale to keep the necessary physics? - mm

What resolution we need in a boundary layer? How far from the wall we need the fine grid.

Gas temperature in the boundary layer is influenced both by convective and radiative heat transfer - both need to be captured

What is driving the heat transfer? Convection or radiation?

Wall modelled okay for D/15

Where is the line between VALIDATION and CALIBRATION?

Community needs a strong validation set that stays out of the model calibration

Global optimization versus hierarchical approach?

I fix parameters not to make them real values with menanif but to put constrains to optimization not to fit and elephant

Hierarchical aproach seems to fit better the small scale experiments.

Global approach fits better large scale experiments?

Would a global approach lead to problems with extrapolation to different scenarios outside the calibration conditions?

It does not matter. We need to have corresponding boundary conditions in small scale as well as large scale and then we are fine no matter of what model or approach we use.

Do not think of real parameters or real parameter values or real models. Think of real scenarios.

Stop philosophizing, go ahaed and try it!

STA - článek - utilize all we learned apply it on the data, find our way and publish

Úplně jsme se vyprdli na interpretaci DSC dat, to není dobře a měli bychom se na to podívat

Don´t underestimate ourselves.

Don´t be afraid to make mistake.

1D - density, thermal conductivity, thermal capacity of BOTH virgin and char - paper 2

Rozpracovat experimentálně - data kalorimetr v dusíku

Rozpracovat vliv v modelu.

McFPE shows that all codes are still used. It is consistent to try more options. Use both Gpyro and Propti plus FDS.

Use in house code. Use what we actually learned and developed!

Discuss with Jirka and support him in "finishing" the code.

Can somebody actually help him? How?

Automating the data processing as much as possible. Teach us how to work with the tools developed.

Char production and its properties - we dont need to make the model "right" we need to make it work. How we are gonna model char in our approach and its influence on heat transfer?

Char porosity - influence on heat transfer - IAFSS prezentace Swann

He sets criterions for succsesful TGA fit in percents of peak loacation, peak magnitude and residue mass

Char is what make a difference - people stay away from it, because nobody knows what to do with it.

Address char production for pyrolysis modelling - develop our methodology how to work with it.

What experimental work can we do to validate our char thermal parameters work in a pyrolysis model?

Do we make a use of MCC data? Possible cooperation with VŚB.

Cone calorimeter

Boundary conditions!

Vitiated atmosphere attachement

Backing

Sample wrapping

Front surface boundary conditions

Model the cone calorimeter instead of only 1D model to asses convection, flow patterns etc...

What we need? How to transform it to make it for research not standardized test

Measure gas concentrations

Window to see the sample?

We completely ignore emisivity and absorption

Lucka

We assume soot is a gaseous product of combustion at gas temperature, not a solid (hot and radiating) particle

Measure multiple heat fluxes

How is front surface boundary influenced by different heat components? So far we only worked with net heat flux

Try using different fluxes for calibration and validation.

Try new FDS SPyro.

What happens to boundary condition (heat transfer coeff.) once surface charred?

Do we need to resolve it for capturing thermal feedback? Isn't wall function working well if y+ region is ok?

Give them solid background to what we want them to do.

Actively interest in their progress, development, and motivation.

Allow them to participate and gain experience from conferences, international seminars.

Better share things we already know or have. No need to start from scratch.