Variable Methanol volatilities

[ccaissie]

I had read in a post by member 25367 that the relative volatilities of Methanol/water change with purity. In designing a methanol column for a client, a packing vendor had stated some very high stage heights, saying methanol is different.

By my own calcs, VLE's etc., I have come up with a height of 20-24", and a second vendor with experience in methanol says that is correct, between 18-24" for Pall Rings.

The column is 6" diameter, feed is 70-90% methanol, and the top purity is 97-98%. I figure 7-8 stages. I include a quick sheet calc to ilustrate what I'm doing.

Is there basis for the larger height at these purities?

Thanks,
CSC

 

[25362]

Perry's Chemical Engineers' Handbook ed. VIII (McGraw-Hill) gives on table 13-2 the coefficients for each of the Margules, Van Laar and Wilson's equations to enable estimating the value of the activity coefficients [γ] for water and methanol at various mol fractions. There is an appreciable increase in [γ]_w at high alcohol %.

I have old notes taken from Shinskey's Distllation Control (McGraw-Hill), stating that relative volatilities vary between 9 at low methanol % to about 1.1 at high purities. Therefore he suggests to use the geometric average of 3.15 rather than the arithmetic average of 5.05.

IMO one should better act with prudence and resort to commercial experience for this separation.

BTW, are you sure the a sheet calc was attached to your message?

 

[katmar]

I also cannot see your calculation sheet.

I would not use any sort of average relative volatility for this calculation. The short-cut methods using these averages may work for ideal mixtures of HC's, but they do not work with mixtures of alcohols and water. In fact, anything plus water is usually a problem. Apart from being widely available, water is a very peculiar substance.

You must do proper tray to tray calculations. Let's see your missing calc sheet and then we can comment. But on the face of it you are on the right track. Years ago I had to design a methanol water separator and although I can no longer remember all the details there were no hidden traps and the column worked well. In those days we did not have access to anything beyond McCabe Thiele, so it can't be too difficult a problem

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[ccaissie]

Thank you, 25362 and Katmar.

I have Perry's 3rd and had not found the constant tables. I had not delved into the principles deeply, and as 25362 noted, this is the province of the pros.

I also noted that the attachments weren't there. I reattach. They are the canned software graphs.

Back to the math, I will do a stage by stage analysis, and also trust the experience of those who are in the business.

97-98% purity would be considered high purity?

Once this little unit is up and running, I'll go back to telescope making.

Thanks,
CSC

 

[25362]

A quick check using the Fenske-Underwood-Gilliland correlations for the upper column section (using low relative volatilities) may suggest more theoretical stages are needed.

The HETS of dumped packing may be quite large. Good liquid entry-flow distributors may be of help to minimize actual HETS by reducing the bypassing of liquid and vapor countercurrent flows.

Remember, contaminants may affect performance.
Besides, an MSDS on methanol may always be helpful.


Thread798-231432 may be of interest.

 

[ccaissie]

Much appreciated comments.

As I am looking at 97% purity at the top, and moderate volumes in the process, the responses regrading pervaporation are maybe not relevant.

The links to the biodiesel sites is an intuitive leap. Yes, it's biodiesel, but this process is for the waste methanol and chemical water from an esterification stage.

Based on VLE, Methanol seems to behave ok at these purities, so the reasoning behind picking an average volatility is well taken, and I have assigned a higher relative volatility, say 4.0 in some rough calculations for the process. I can see where at .99+ it becomes very important.

This column design has evolved to some fairly safe numbers mostly because we can choose our feed point and our feed temperature, the unit is relatively small, and it is a side-stream process at a biodiesel plant meant for recovery of waste methanol-water from an esterification stage. The diameter and stage calcs are converging with the recommendations of the packing manufacturers, a good sign.

My study of Kister's "Distillation Design" confirms my basic design, and has given me confirmation for some assumptions, and adds a good foundation for designing and operating this monster. Not a whole lot on rebolier duty, though.

I am interested in the comments of member MYRDALE who has an operating methanol still, much larger than this unit. Although scaling is a risky method, his setup does confirm that I am within the tolerances for this separation. His/Her quoting of energy duty was very useful.

I am somewhat agog of the rigor of this sort of design process and am on my knees in front of professional chemical engineering.

My recent challenge is refining the heat requirements to drive this beast. I've got some basics, but I will post a query separately.

Respectfully,
CSC