More Probs - Gas-Solid 2 Phase flow
 

Dear Glenn, Jeff and friends,

I have a difficult task at hand (working from weeks still heading nowhere) and ur suggestions will be of great help. i am discussing this problem with CFX support engineers out here as well but not getting quite right.

I am trying to simulate a gas-liquid reaction in a packed bed (practically its a gas-solid reaction but as i cannot define varying comp solid in CFX, i m specifying it as liquid). Initially, reactor contatins 0.7 volume fraction of liquid (dispersed phase) and 0.3 volume fraction gas(continuous phase). Now through inlet, i m entering some air. air reacts with gasphase and liquidphase and produces some product. Now as per real physics, solid must not leave the system and also solid should not move in the system so in our case, liquid must not leave the system and volume fraction should remain constant.

i have run a few simulations but not getting quite right..

when i use degassing condition, liquid is not leaving the system but the volume fraction in the system is not fixed (its varying from 0 to 1) and also liquid is moving around.....

and when i specify avg static pressure at outlet, liquid is leaving the system !!!!!!!!!!!!!!

so neither case is satisfactory...Is there any way in which i can specify fixed velocity (ZERO in my case) for a particular phase in a multiphase simulation.

Can anybody comment on this how to resolve this problem. Is it possible in CFX 5.6 to fix velocity of a phase in multiphase simulation and solve momentum balance only for the other phase...

Eagerly waiting for some positive replys..

Thanks & Regards, Paresh Jain.

Re: More Probs - Gas-Solid 2 Phase flow
 

Hmmm...

I really think that using the two fluid model may be the wrong way to go. Fixing one of the phases to be a stationary bed of particles is not what that model was meant for at all, although I can see the motives for trying to use it.

If it were my problem, and the bed is truely fixed, I'd use a porous media model to get the pressure drop and an algebraic scalar to keep track of the concentrations in the bed particles. I've even done moving beds of pellets (on a conveyor) in this way.

The two fluid model you are using will try to impart fluid velocities to your solid phase and, even if you can prevent it from leaving any of the flow boundaries, the code will try to "stir" it up because the momentum transfer is two way. The gas imparts momentum on the liquid, and the liquid imparts momentum on the gas.

Good Luck, Jeff

Re: More Probs - Gas-Solid 2 Phase flow
 

Dear Jeff, thank u. Nice to hear so soon from u. but i have one query, if i define my solid phases as additional variables, then their conc will be funtion of gas conc in system, time and conc of that species itself. So say if i specify solid comp A as an algebraic scaler and then use some equation for it, will CFX allow me to use terms like conc of A as a function of A itself. I mean, can i define A as, A = funtion(conc of A, time, conc of O2) ????

and should i define it as algebraic scaler or say specific additional variable ? can i use kg/m^3 as unit of additional variable ?

thanking you once again. Sincerely, Paresh Jain

New More Probs - Gas-Solid 2 Phase flow
 

Hi jeff and friends, well i tried with using additional variables, it seems to work ok. But as it is not capturing my physics properly, i would like to know if there is any other way for multiphase modeing with porous media by which i can solve my problem. please go thru the main string of this issue and suggest me if this problem can be solved by using multiphase models....

Hoping for ur reply.. Thanks and regards, Paresh Jain

Re: More Probs - Gas-Solid 2 Phase flow
 

Paresh,

Since your species conc. in the solid is an algebraic scalar, you need to do the following.

Calculate a source term in the gas phase from your algebraic concentration. Set this as a source term in your gas.

Since you know the rate at which A is being consumed, based on the source term you just calculated, and the time step, dt, you can calculate the new A value as:

A - rate*dt

This should update your A values at each time step.

Hope this helps, Jeff