5 replies to this topic

[Villis]

Hello all!
I'm glad that i just joined your community!
I'm really looking forward to help and to get helped.
Well, this is my first question...
I am a student in chemical engineering and i have to make a feasibility study on an ethylene oxide production plant. The method we use is the direct oxidation process which involves a reaction of ethylene with air on a silver/alumina catalyst. I have a problem with the reactor sizing and i really can't find any solution. The data that were given/calculated to/by me are: the volumetric flow (Q=8,44 m/s), the LMTD (LMTD=79,4), the porosity (ε=0,4), the heat exchange surface (A=2168,9), the overall heat transfer coefficient (U=176 J/(s x m^2 x K) ) the residence time (tr=6 sec), the selectivity (S=6) and the %conversion of ethylene (=20%). I have to calculate the number of the tubes, the tube and the shell diameter and the catalyst and the reactor's volume.
Well, the first thing i thought was to calculate the reactor's volume from the equation Vr (reactor's volume) = tr (residence time) x Q (volumetric flow). Is that correct?
I also tried to assume a tube diameter (d) and a tube length (L) and find the tube number from the equation: A= Ν x pi x d x L
Right after that, i can calculate the tube volume (where the reaction mix is) through the equation: V=N x pi x L x(d^2)/4
But then, i must find the shell diameter and i don't know how. I know that i have to choose a tube layout (for example triangular or square) but what shall i then do?

ps. I just forgot to mention that i have a PFR reactor but the kinetics are unknown and i can't use the PFR's design equation.

Well thanks in advance. I'm looking forward to your replies

[Nikhiln]

Villis,

Your approach seems good enough for a preliminary/feasibility study. A couple of points to consider:

a)The value of Vr you have calculated would be the volume of tubes filled with calatyst ONLY IF the given residence time already accounts for porosity. If not, you need to account for the presence of catalyst by dividing the calculated Vr by porosity.
b)As far as deciding shell diameter is concerned, you need to visualise the reactor as a shell and tube heat exchanger. (An EO reactor is nothing but that, with reaction in the tubes). The exothermic heat duty is known. Once you decide the shell-side coolant (usually BFW to generate steam)and its inlet conditions, calculate the required flowrate. Then assume a triangular-pitch layout and calculate the shell diameter. Adjust the diameter so that coolant velocity is within acceptable limits. Needless to say, this diameter will depend on the initial length and diameter of tubes you assumed.


This should suffice for a feasibility study without involving kinetics and give you an approximate sizing.

[Villis]

Thanks for your reply Nikhiln.
I would like, however, some clarifications on the points you mentioned.
a )Using Vr (reactor's volume) = tr (residence time) x Q (volumetric flow), i can account the tube volume and not the whole reactor's (including the shell) right? And that's because the reaction takes place in the tubes, right? And what about the equation V=N x pi x L x(d^2)/4? Is this the same? Will it give me the same results as the previous one?
b ) Well, (almost) all you mentioned here, are already done. I decided for the coolant-it's BFW- and i know that the reactor is identical to a heat exchanger. I also know the exothermal heat duty and the inlet temperature of the coolant but how can i calculate the flowrate or the outlet temperature? I mean the energy balances should be difficult cause i have no specific heat value for the reaction mixture. This is what i do.I take a standard value for the inlet temperature of the boiling water feed (although i found an outlet temperature too in the literature) and then i calculate the heat transfer area (Q=UAΔT)...
And i really have no equations for calculating the shell diameter and i also don't know the acceptable limits for the velocity (we want the velocity to be specific, so that the reynolds number is within the turbulent flow area??)

Thanks once more for your reply

Edited by Villis, 19 March 2010 - 02:10 AM.

[Nikhiln]

Thanks for your reply Nikhiln.
I would like, however, some clarifications on the points you mentioned.
a )Using Vr (reactor's volume) = tr (residence time) x Q (volumetric flow), i can account the tube volume and not the whole reactor's (including the shell) right? And that's because the reaction takes place in the tubes, right? And what about the equation V=N x pi x L x(d^2)/4? Is this the same? Will it give me the same results as the previous one?
b ) Well, (almost) all you mentioned here, are already done. I decided for the coolant-it's BFW- and i know that the reactor is identical to a heat exchanger. I also know the exothermal heat duty and the inlet temperature of the coolant but how can i calculate the flowrate or the outlet temperature? I mean the energy balances should be difficult cause i have no specific heat value for the reaction mixture. This is what i do.I take a standard value for the inlet temperature of the boiling water feed (although i found an outlet temperature too in the literature) and then i calculate the heat transfer area (Q=UAΔT)...
And i really have no equations for calculating the shell diameter and i also don't know the acceptable limits for the velocity (we want the velocity to be specific, so that the reynolds number is within the turbulent flow area??)

Thanks once more for your reply

Villis,

a) Of course, the residence time applies for the tube side only. I mentioned earlier that Vr would be volume of the tubes. Maybe you need to rename Vr as Vt (volume of tubes, not reactor). And yes, the number of tubes is calculated using the equation you used (thats why I skipped it and went directly to deciding the shell diameter). As mentioned earlier, you need to comfirm if the residence time of 6 sec is including the catalyst porosity or just empty tubes.
b)Well, I sort of preempted this question, because you mentioned some of the data were given/calculated by you.......including area, LMTD & U. You have not mentioned which is given and which is calculated. Anyway, since you mention you already calculated exothermic heat duty, I can give you a few tips on how to proceed. It wouldnt be fair for me to point out everything.

Tip1: The steam generated would be saturated steam, as in a regular boiler. That should help you calculate the water flowrate, based on inlet conditions.
Tip2: If you have access to HTRI,you can input the tube layout to find the shell diameter. In case you dont, the required equation is available in Kern's 'Process Heat transfer'.
Tip 3: The water circulation is basically thermosyphon, so not exactly turbulent as we would expect in a regular exchanger. Which means a pressure drop of 1 bar approx would suffice on the shell side.

These 3 tips should be enough to estimate an approximate value for shell diameter.
Cheers.

[Villis]

First thanks for the a). I'm done with that
Secondly... hmmmm to be honest i've read Kern's book but found no suitable equation. I calculated the cool water's flowrate from the energy balance but can't find the equation to find the shell diameter from the tubes' diameter. I don't have htri but i can use aspen heat exchanger design (not that succesfully )...
Thanks one more for your valuable help!

Edited by Villis, 19 March 2010 - 03:11 PM.

[Villis]

Well, i'm done with everything (that has to do with the reactor)! Thanks one more for your help