Use of Fenske equation in steam stripping?

[LaSalle1940]

To do some screening in a very short feasibility study, I was thinking of using the Fenske equation to get a handle on the minimum number of discrete stages in a refining steam stripping operation (specifically, stripping off hydrogen sulfide from an oil stream). Glancing at the equation in any number of texts, I don't see where it WOULDN'T be applicable, assuming one uses a relative volatility derived from water/H2S equilibrium data, and assuming one knows/specifies the mole fraction of H2S in the overhead and bottoms streams.

I'm open to comments one way or the other. Thanks.

 

[hanon]

Have you thought in applying the absortion/stripping factor with nomographs to solve this problem and find the theoretical stages? Maybe it can be applied

A = L/(V·K) L is liquid rate; V vapor rate; K Equilibrium constant y =K·x K aprox= Vapor_Pressure/Total_Pressure

 

[DickRussell]

I agree with hanon's reply. The Brown-Souder chart is good for either absorption or stripping in simple separations. Basically, it is a semi-logarithmic plot of fraction not absorbed or stripped vs. absorption or stripping factor (A=L/VK, S=KV/L). The equations are the same:

(1-f) = (A-1)/[A^(n+1) -1]

where "n" is number of theoretical stages of separation.

For your case, the oil stream flow ("L") is given, and the pressure and temperature also may be defined if the oil stream is attached to another operation. You can look up the approximate K value of the H2S out of the oil on a nomograph (eg. the ones in Hydrocarbon Processing, September 1961). Since a separation like this has a low stage efficiency, perhaps only 0.2, you want to do the separation in 4-5 stages. So you look up your fraction not stripped on the curve for 4 or 5 stages, read off the stripping factor needed, back out the "V" (steam rate). Keep in mind that this just gives you a decent approximation of what will be required. If you need a greater stripping factor for a reasonable number of stages and steam rate, then obviously you need a higher K value of the H2S (higher temperature). The clustering of the curves at various numbers of stages also shows that you won't see much advantage to adding stages if the stripping factor KV/L is less than 1.0. In the end, you can use your hand-method screening of design parameters to refine the answer using any process simulator for doing the rigorous calcs.

 

[LaSalle1940]

Thanks for the assistance. Funny, but I've been unable to locate the Brown-Souder chart in the usual references (Perry; Treybal's Mass Transfer Operations; King's Separation Processes). Have you a reference where I might find it, please? Thanks again.

 

[UmeshMathur]

LaSalle1940:

Generally, the absorption equation is attributed both to Kremser (1930) and Souders-Brown (1932). You'll find it more easily if you use Kremser's name.

A good plot of the Kremser-Souders-Brown equation is available in the 11th edition of GPSA's Engineering Data Book, Fig.19-48, page 19-31 (1999).

The theoretical basis of this equation is described in King, C. J.: "Separation Processes", Chapter 8 (McGraw-Hill, 1971). King also has a graphical plot of the equation that is set up a bit differently (using different parameters on a log-log plot) from the GPSA plot. King also provides the theoretical equation from which the plot was derived, in case you prefer iterative computer solution. However, you're still required to provide the K-value of the each of the components in the gas phase on each stage. This requires going to nomographs of the type Dick Russell has alluded to.

Also, note that the Kremser equation assumes constant L/(K.V) which is often not observed in practice for absorption of rich gases. Also, thanks to large temperature gradients, the K-values from top to bottom can vary a lot; therefore, one must use average values.

It turns out that finding good K-values is a lot more complex computation than solving the Kremser equation. That is why it is generally recommended to solve the problem more rigorously and use a general purpose process simulator. Dick Russell's PD-Plus is an excellent, inexpensive choice (

http://www.deerhaventech.com).

Finally, the theoretical basis of the Fenske equation is not quite appropriate for absorption calculations (constant relative volatility and equimolar flow, which apply more closely in distillation).