3 replies to this topic

[clarenceyue]

Dear All,

I am a final year student at Nanyang Technological University Singapore, currently working on my Chemical Engineering Design Project, and my group and I have been tasked to size a methanol recovery tower in a biodiesel plant.

We have worked out our flow rates using HYSYS and we obtained the following data:
Mass Flow Rate of Vapour Leaving the tower = 361 kg/hr (~99.9% methanol)
Liquid density (top tray) = 737.8 kg/m3
Vapour density (top tray) = 1.373 kg/m3

We have decided to use a packed column since the throughput for this tower is rather small. I have proposed using the classical Souders-Brown correlation to determine the column diameter based on the top tray quantities that I have stated above. My results from the calculation are as follows:

U = K*(Density of Liquid / Density of Vapour)^0.5, K = 0.06 m/s (using the recommendations from Gael Ulrich's book)
U = 1.39 m/s (max. velocity of the vapour)

Calculating the column diameter based on a velocity of 80% of the maximum velocity:
Volumetric Flow Rate of Vapour = 361/1.373/3600 = 0.0730 m3/s
D = (0.0730/(0.8*1.39)/(0.25*3.142)^0.5 = 0.29 m = ~0.3 m

I would like to ask any of the esteemed members here if:
1) the S-B equation is applicable for packed beds
2) the column diameter is reasonable for the given flow rate

Thank you for any advice you would be able to provide me with.

[Marchese]

What is 0,25 in the denominator of the last formula utilized?

[clarenceyue]

What is 0,25 in the denominator of the last formula utilized?

Area = 0.25*pi*D^2 = 0.25*3.142*D^2

This is where 0.25 comes from.

[katmar]

The Souders-Brown Correlation is for entrainment, and is the basis for practical correlations like Fair's (See Perry 7 or 8 ed) for trayed columns. To a first approximation it is applicable to all types of trays - bubble cap, sieve or valve - because it is looking at the vapor space above the liquid.

Although liquid entrainment by the vapor is the most important sizing factor in most columns, the possibility of liquid flooding should be investigated as well. High liquid flows can cause high liquid gradients across the tray (resulting in maldistribution of the vapor) or flooding of the downcomers.

Having said all that, I would not use S-B or Fair for sizing a packed column. There are specific methods that have been developed for packed columns, such as the Generalized Pressure Drop Correlation by Eckert and Leva. This is also very well covered by Perry. A Google search on "Generalized Pressure Drop Correlation" will get you thousands of hits.