Currently doing a hydrodynamic model of an underfeed stoker. An approximate model of the bed is used to determine the velocity profile of the underfeed stoker...

Although it is not part of my problem at the moment, I am curious if we can possibly model the fixed bed combustion using FLUENT... This would involved two phase reactions (solid heterogeneous reactions and gas phase reactions). To reduce the complexity of the problem the bed can be model in 1 dimension (i.e. no radial reaction)...

Actually, 2 phase reaction model is possible for pf combustion using DPM... I am not sure if a UDF similar to DPM can be created to model fixed bed combustion...

Regards, Stanley

I would like to hear the opinion of the specialists in the world, too.

My opinion is :

I do not think that it is possible to create UDF for fixed bed combustion including hetrogenous reaction by generalized CFD code. First of all, two time scale are entirely different. For gas phase, its residence time in the combustion chamber is couple of seconds. Whereas the residence time of soild fuel in the fixed-bed is couple of hours. My question is, how generalized CFD code can handle two different time scale in the same fluid region simultaneously ?

I myself had spent fairly much time to check whether it is possible or not. And I concluded that it is impossible.

For your reference, I know one research group who are doing that kind of simulation. They created 1-D fixed bed combustion 'in-house' program and they are coupling their in-house program with Fluent. But they are using 'manual coupling'. That is, they calculate gas region by Fluent and fixed-bed region by in-house program. Next they updates the information of interface(Fluent's boundary) by manually. Their result for fixed-bed type incinerator is, I think, very good. But coupling(or iteration) between each program is being done by manually.

I hope to hear the opinion of CFD specialists, whcih will be very helpful for my future analysis, especially for the simulation of stoker-type incinerator.

Sincerely, Jinwook

Its certainly possible to write UDFs to calculate fixed bed or generalised gas-solid reactions within Fluent - I have working models for Fluent 4.5 and 5.4. Although at the moment I assume the solid is stationary - though I'm working on using a simple 1D type model to calculate solids flow - so I can get settling behaviour etc.

The question of time scales is certainly the tricky bit as Jim-Wook Lee has pointed out. In my model, I assume a pseudo-steady state - ie at each point in time I calculate the gas-phase velocities, concentrations and system temperature using the steady state model - and then use the calculated solid consumption rates to advance the solids distribution for the next time step - then repeat. While an approximation, it has been used in many gas-solid systems in the past (eg. underground coal gasification). The other alternative is to solve a fully coupled transient system - which then requires small time steps - this becomes far less practical in 3D as compared to the 1D models of the past.

Regards Greg Perkins

Dear Greg Perkins

Thank you for your reply. Your reply is very helpful to me.

I would like to know some more information for your modeling.

1. Is your model 1D or multi-D ? I think that 1D is fairly simple and we can get the 1D code fairly easily. However, 2D(or 3D) is another problem.

2. Is your model for batch reactor or continuous-flow(and soild fuel) reactor such as moving-bed or stoker-type incinerator ? My primary interest is modeling of stoker-type incinerator with couple of assumptions.

Your kind answer will be very helpful for me.

Sincerely, Jinwook

My model is built into Fluent through UDFs so it can be used in 2D or 3D. The current model assumes a stationary solid (set of spheres) that is consumed due to reaction of a shrinking core - in this the solid volume fraction decreases as reaction proceeds.

I am extending the model in a couple of ways at the moment, since my main interest is in underground coal gasification. In this situation I have a) reaction at a sharp front between a cavity and porous solid b) reaction of particles in a moving bed which fall from the roof as the cavity develops.

To address a) I am hoping to use a simple version of my current model, but which assumes a receding solid slab at the solid/void interface.

To address b) I wish to implement the shrinking core model with simple solids flow to simulate. In this case there will be a bed region in which I assume a constant bed porosity and use mass conservation to determine the vertical flow of solid at each point. While the CFD will be a fully 3D model, solids flow will be restricted to only the vertical direction. The expectation here is to have a reacting bed, and model the consumption of char in the bed and the settling of ash to the bottom, and also the permeability distribution of the bed, which has a major impact on the flow of reactants in the system.

As I mention I envisage this model as part of a global undeground coal gasification model. However, I could be used to model moving bed reactors etc.

I am just about to start implementing this new model . . .

Greg

Dear Greg Perkins

Great !!! And thank you for your detail explanation. Please let me know the citation of the paper when your research result is published. I would like to read the paper(Journal paper or conference paper, whichever would be good).

Sincerely, Jinwook

Hi Greg,

I would like to ask you about the citations you've found on the pseudo-steady state model.

I am having the same problem in my project, the time step can go to 10-5 s, and I have to model until 30 hours of process.

Thanks

Arturo

Dear Jin,

How about considering the model as pseudo-steady state where time function are immaterial???

Is this research group you are referring to belongs to L.D. Smoot of BYU???

Cheers.

Stanley

Hi Greg,

How did you consider your bed model??? Does your bed model assume a continuum phase or of several single particles of solid lumped together with various interactions (i.e. gas-solid, solid-solid)???

I notice that you are modelling an underground coal gasification... how did you model the devolatilisation phase of the coal??? Did you use a simple Arrhenius rate or did you use other type of model (i.e. competing reaction model or network models for example FLASHCHAIN)...

In my opinion, gasification and solid bed reaction differs in bed height... luckily for gasification, you normally don't have to bother the effect of bed height and particle diameter... Unfortunately, most bed combustion are usually involved a shallow bed wherein effect of bed height against particle diameter is significant... Do anyone of you have an opinion on this one...

If you don't mind, do you have any papers of your work which I can refer to???

Cheers, Stanley

Hi Jin,

What type of Stoker model you are dealing with... Is it an underfeed or an overfeed???

Regards, Stanley

At the moment I consider a stationary set of particles - these are represented in a Eulerian-Eulerian continuum model representation.

I haven't got the full solids flow working yet - that's what I'm trying to implement now.

As for devolatisation - I use a simple Arrenhius rate expression.

I didn't quite understand what you mean by the difference for gasification and combustion - I assume you are talking about the 'steady-state' results - since particle diameter will also change for gasification as conversion proceeds.

Greg

Well in the underground coal gasification literature many models adopt this approach since you need to simulate up to several months of real time. I haven't seen it mentioned much in the usual fixed bed combustion/gasification literature. Most of these models are 1D and often steady-state so its not so critical.

I think there's a reference to it in the book by Bishoff/Bischoff (196x) - its quite a famous book for chemical engineers.

But if you non-dimensionalise the gas and solid mass conservation equations and then compare the time constants for the gas (tg) and solid (ts) you get:-

tg/ts (approx.)= rhog/rhos = 1/1000

where rho is density of gas/solid.

so changes in the solid are 1000 times slower than changes in the gas - so on the time scale of the gas I consider the solid to be fixed - ie not changing. Hence the pseudo-steady-state approximation.

Greg

Dear Stanley

My modeling is for side-feed incinerator. And, at this time, my modeling is gas-phase combustion model.

For gas phase combustion, it is fairly conventional approach, but slightly different from previous works. In summary(with simple form),

For combustible matter combustion, considering that there might be many reactions, first approach might be two material, like

CxHyOz)1 + a O2 ---(k1)--> b CO + c H2O +Q1

CxHyOz)2 + a O2 ---(k2)--> b CO + c H2O + Q2, (Q2=Q1 but k2 not equal to k1)

CO + 0.5 O2 --(k3)--> CO2 + Qco

For water vaporization,

CxHyOz)w_H2O + d O2 ---(k4)--> e CO2 + f H2O +Q(=0.), with very large k4, because vaporization is prior to combustion.

Then, it is fairly easy to consider heat and material balance and adjustment with experimental result. I have published couple of papers in the Korea domestic journal but not in the international journal.

It is fairly experiment-dependent approach, so that its theoretical background is fairly weak. But we can design the combustion chamber pf the incinerator by following steps.

First adjust the model and the model coefficients by comparing with experimental results. Next, we use the model for the modification and scale-up of the combustion chamber of the incinerator. The incinerator is going to be constructed and if the performance is good, I will let you know.

Sincerely, Jinwook

Dear Stanley

I am not refering to BYU's approach because they have particle gasification/combustion model and 1D fixed bed model. As far as I know, they do not have incinerator model. If they have it, please let me know.

For researcdh work, I think that psedu steady-state or fully transient model is clearly better than steady state model. However, I think that, sometimes, steady state model, with some simplication, would be enough for engineering purpose.

Anyway, thanks for your advise.

Sincerely, Jinwook

Hi Jin,

Your work is very interesting... Do you got the paper of Shin and Choi [(2000) Combustion and Flame Vol.121 pp. 167-180] I think that this paper is very related to your work and I hope that this will be helpful to you...

I presume that you have written your paper in Korean language... I would like to know if you got the English version of it...

Regards, Stanley

Hi Greg...

Thanks for your information...

Actually, what I'm pointing out in my previous posting is the difference between modelling a fixed bed gasification and a fixed bed combustion. Unlike gasification, most fixed bed combustion involved a very shallow bed (fixed bed gasification normally got a deep bed of 1 to 2 meters or higher while fixed bed combustion is typically less than 1 meter). So in this case, we need to consider the effect of bed depth, bed diameter and particle diameter in relation to the chemical reaction kinetics and transfer properties of the bed...

What I am inquiring at is did you consider these parameters in your bed model? From the literature of fixed bed reactors, the effect of bed height is negligible for a deep bed reactors and small particle...

Regards, Stanley

Well as yet I haven't considered these aspects - and since I'm modelling gasification (underground coal gasification) I don't expect to consider this.

One question is - what do you think would change?? a) Heat and mass transport correlations? b) reaction kinetics etc. c) other??

I would have thought if the model is fundamental you could apply it to both without any significant changes and expect to get reasonable results. In my model I will use intrinsic reaction rates etc. so it should be fairly fundamental.

Have you seen the paper by Cooper et al., Chem Eng Sci Vol 55 (2000) p4451-4460. This is a combustion model.

Greg

Dear Stanley

Thank you for your reply. I know Shin & Choi's research activity very well. Their approach is fairly similar to mine except that I include complex chemical reaction than their's. I know that complexity does not always guarantee good result. But I think that complex chemistry can be extended to the good model in the future. They are now extending their work to include bed(grate) combustion.

Unfortunately, I do not have English version.

Sincerely, Jinwook

Hi Greg...

Actually, I have no concrete idea on how these parameters would affect the combustion model... Most works done on this aspect are used in catalytic reactors and not yet on combustion...

Oh yeah, I got that paper... that paper is quite a good one... very interesting indeed...

Regards, Stanley

Dear Stanley,

I already have one-dimensional or two-dimensional transient or steady-state codes for fixed-bed reactor/combustor.

Cheer!

Ray