MaCFP-3, NIST-Gasification-Apparatus

[isaac.l...@gmail.com]

Please use this thread to post questions regarding the NIST-Gasification-Apparatus Target Case: Bench-scale thermal degradation experiments conducted in the NIST gasification apparatus, providing validation data for PMMA pyrolysis models.

Further details regarding the test setup, experimental measurements recorded, and modeling targets of interest are available on the MaCFP Github Repo:

https://github.com/MaCFP/matl-db/tree/master/PMMA/Validation_Data/NIST_Gasification_Apparatus

[Jason Floyd]

https://github.com/MaCFP/matl-db/blob/master/PMMA/Material_Properties/MaCFP_PMMA_BUW-FZJ_B.json

In this file the A and E for the first parallel reaction have the same value 56864.9.  While it is possible these are the same value,  it is certainly not expected that they would be identical.

The density of the second material making up 97.5 % of the initial mass is 842.116 kg.  This is ~2/3 the expected value.

Are the values in this file correct?

[Morgan Bruns]

Yes, this is a typo on the activation energy. My mistake. I'll fix it on the repo this afternoon.

You are also right about the densities. I'll need to sort this out, but their two initial components are both at density of 1207.84 kg/m^3, and both reactions produce a residual of 842.116 kg/m^3.

[Jason Floyd]

Thanks Morgan.

[tristan...@gmail.com]

Dear Jason,

there are two different decomposition reactions of the solid defined. Each of them transforms some of the starting material into gas and residue: 2.5% are associated to the first reaction happening at about 200°C, while the rest is reacting in the primary reaction. The PMMA has a density of 1207 and the residue has a density of 842 kg/m³.

Please refer also to our full data set: https://zenodo.org/record/4704571#.ZCsb0vbP2Uk

If you have further questions please let me know.

Best,

Tristan

[Morgan Bruns]

The kinetic parameters are now correct on the GitHub repo.

I was trying my best to keep the reaction network description minimal, but clearly this case broke that formalism. You can see in the updated *.json file how this works:  https://github.com/MaCFP/matl-db/blob/master/PMMA/Material_Properties/MaCFP_PMMA_BUW-FZJ_B.json

Basically, it now labels each component explicitly so it should be fully generalizable even to the types of complex reaction networks that we'd never see. So the BUW-FZJ_B reaction network is 

    PMMA_1 ---> 0.001 Residue + 0.999 Gas    [equation (1) in the document Tristan linked]

    PMMA_2 ---> 0.001 Residue + 0.999 Gas    [equation (2) in the document Tristan linked]

The whole material and reaction description is now expressed in the *.json file as:

"Composition": {
    "Initial Components": [1, 2],
    "Initial Mass Fraction": [0.025, 0.975],
    "Final Components": [3]
},

"Kinetics": {
    "Reactants": [ [1], [2] ],
    "Products": [ [3], [3] ],
    "Pre-exponential": [56864.9,6.71256e12],
    "Activation Energy": [610756.0,180974.37],
    "Reaction Order": [1, 1],
    "Solid Yield": [0.001, 0.001]
},

The virgin PMMA is broken into two materials (components 1 and 2) with the same thermophysical properties, but which undergo two distinct reactions. But both reactions produce the same residual (component 3). And so, for instance, three densities must be expressed (for components labeled 1, 2, and 3) as

"Density": {
    "Form": "Component Specific",
    "Value": [1207.84,1207.84,842.116]
},

I will still need to apply this to the other materials, but they are mostly less ambiguous.

(Tristan, please correct me if my understanding of your reaction mechanism is not correct!)

[Tristan Hehnen]

Thank you , Morgan!

Yes, this seems to be correct.

Also, I'm slightly afraid to submit the next data set now... :D

Best,

Tristan

[Jason Floyd]

Thanks Tristan and Morgan.

Does any of this apply to the BUW-FZJ_ data? Is the second density in that file also the residue? And looking at that _A json file again, the emissivities are 0.7 and 0.4 which seems low for black PMMA.

[Morgan Bruns]

Those emissivities are what were reported. The values are based on a global optimization to CAPA II data, and optimization can sometimes give non-physical values. This is an issue that the community needs to discuss since global optimization is widely used.

In BUW-FZJ_A, they are using the same reaction mechanism as in BUW-FZJ_B described above.

[Tristan Hehnen]

Hi Jason,

It seems to me there is something missing in the first sentence, I guess it is on the dataset of approach A.

For approach A, the parameters are as follows:

PMMA

density: 1168.49

emissivity: 0.768990

Residue

density: 1085.97

emissivity: 0.407040

These values are generated during the inverse modelling process (IMP) and depend on the experiment, the parameters chosen to be adjusted during the IMP and the model. The parameter set is thus necessarily an effective one.

I believe we used the CAPA II data as target for the thermophysical parameters here, without gas phase simulation, so the emissivity values might be a bit more arbitrary. 

Hope this helps,

Tristan

[Tristan Hehnen]

Thank you, Morgan!

Yes, the pyrolysis reaction scheme is the same, just the way we got to the individual parameters is different.

Approach A: pyrolysis  reaction  kinetics in first IMP step with TGA data as target; thermophysical parameters in second step with CAPA II mass loss data as target and using the reaction kinetics from the first step

Approach B: pyrolysis  reaction  kinetics and thermophysical parameters are determined simultaneously by running TGA and CAPA II simulations in parallel.

[Tristan Hehnen]

Jason, I've added another RAR archive to the data repository (https://zenodo.org/record/7803832#.ZC34DPbP2Uk) that contains the FDS files, sorted by approach A/B as a summary.

You can simply download the 'BestParameterInputFDS.rar' (~2.5 MB) and see exactly how we set it up.

It might be nicer this way, than crawling through the full thing.

[Jason Floyd]

Thanks Tristan. I see my erorr now. I've run FDS for the R3 gassification test with all the json files with full parameter sets (k,rho,c,e, A, and E). The A file was taking twice as long as the gassification test to pyrolyze the sample in the json file.  My error was I had mistakenly applied the 0.4 emissivty as the emissivity for PMMA 2.

[Tristan Hehnen]

Okay

[Tristan Hehnen]

How do we want to simulate the NIST-Gasification-Apparatus in FDS, for the proposed validation step?