4 replies to this topic

[babucher]

Greetings,

This is my first post on the forum. My questions concern purging a distillation column prior to startup.

I intend to use a packed column to batch distill (multistage) a methanol/water mixture, and want to remove the noncondensables, especially the oxygen, before charging and starting the unit.

From Kister (Distillation Operation) I've found this info:

* Sweep or Pressurize/Depressurize
- the advantage of pressurize/depressurize is that there are no dead spots that a sweep may miss
- the disadvantage is the pressure requirements of the proposed column

* If Steam, then Inert
If we're going to steam out the air, Kister recommends following it up with an inerting blanket because condensation of the steam as the column cools can cause a vacuum that might pull in air. It might even cause physical damage if the column isn't up to snuff.

* Venting
This I'm not really sure about. Kister says that after purging with the inert gas, the process gas is used to purge the inert. If this is the case, then I'm assuming the noncondensable is removed via a vent in the condenser at the top of the column. Also, I would think the process gas that mixes with the noncondensible is condensed and recycled to the feed tank. In the end the column should be slightly pressurized with process gas and ready to receive the feed charge.

A question arises because some autoclaves are set up to use gravity displacement of the air inside. The air and steam (which is introduced in the top) are mostly separate because of stratification, and the air descends to the bottom because it is heavier.

Questions:

Is the stratification due only to thermal differences? Does molecular weight difference have anything to do with stratification? I wouldn't think so, but even stratification itself is somewhat baffling to me because it seems to be multiple gas phases.

Is this ever accounted for in the purging of columns by putting vents at the bottom of the column, or is it something that can be ignored because of convective mixing or other reasons?

Any other thoughts you think might be important to bring up? Especially if you know of any references on column startup that are more detailed/expansive than Kister's section in Distillation Operation.

Thank you,
Brian Bucher

[Art Montemayor]

Brian:

First, I’ll address your 2 questions:

1) Stratification within a static distillation column is not due only to thermal differences. A lot has to do with the shape, design, and specific characteristics of the many internals that each column has. The effect of stratification is impaired when you have “dead ends” or blocked in internal volumes where free, natural convection cannot take place. And I think you can agree that this can be so – perhaps more so in a tray column than in a packed one. Even packing has some areas that channel fluids more freely than other areas – and this causes the characteristic of “channeling” and reduces the effectiveness of a given bed. In order to induce stratification (the natural separation of two fluids with different densities) you must be at (or close to) a static condition in order to allow the different fluids to seek their respective locations relative to each other. Flow, or turbulence, discourages the effect of stratification.
2) I’ve never seen this accounted for in the purging of columns; I doubt if the effect could be noticed. When you are purging you are causing a flow, and not necessarily a pure laminar one that could foster stratification. The moment you effect a purge you are inducing convective currents that cause mixing and require more purging in order to reduce the amount of residual contaminant within the column.

I’ve read Kister’s book in libraries and I haven’t been very impressed with his “hands-on” recommendations and opinions. I believe he generalizes too much and he also doesn’t reveal a deep knowledge of the mechanical equipment he describes. It may be that he approaches the subject as a scholar or that he simply has not operated or fabricated such equipment. Allow me to comment on what you state he says:

Sweep or Pressurize/Depressurize
There certainly are dead spots that a sweep of a gaseous fluid may miss. It is virtually impossible to obtain a 100% plug flow of gas through a vertical column – whether the flow is down or up. However, when you pressurize the column you are merely diluting the contents with the diluent (if it be nitrogen). As you depressurize, you leave behind a diluted amount of fluid within the column. That’s why you pressurize/depressurize several times in order to reduce the remaining contaminant.

The pressure rating of the column is not necessarily a deterrent for pressurization/depressurization. Don’t forget that you can always use other methods to assist, such as a suction blower that evacuates the majority of the initial column gaseous contents and assists in maintaining a positive flow outwards. You can also simply flood the column with water and follow through with a desirable inert such as nitrogen. What you fail to mention is that the first order of importance in purging a column is to know all the mechanical design information about the equipment (and everything connected to it): the Maximum Allowable Working pressure (MAWP) and the Maximum Allowable Working Vacuum (MAWV). If you have a full vacuum (FV) design, then you would be smart to consider an evacuation of the initial contents down to as low a value as you can. This can be done overnight, prior to startup.

Steam Out
I heartily agree with the dangers involving a steam out of equipment. If you don’t have a FV design, then you should steer clear of this method. I don’t worry about pulling in air; I would be more concerned with collapsing the entire column with an un-expected vacuum condition.

Venting
I would evacuate the column with a blower prior to introducing nitrogen – as explained above. This saves nitrogen and is a faster inerting method. You are wrong in stating that the process gas that mixes with the non-condensable is condensed and recycled to the feed tank. If this is an atmospheric distillation (which you haven’t stated) then the overheads condenser is vented to the atmosphere; if the distillation is done under pressure, then all non-condensables are automatically purged to the atmosphere through the overheads condenser. I don’t see any problems or anything different from conventional distillation methods. There should not be a requirement to pressurize the column with process gas prior to receiving the initial feed charge. You can receive the initial feed charge into the column while the column is under slight vacuum, go into total reflux until your desired column temperatures are achieved and then proceed to adjust your reflux and overheads make. You have stated that this is a batch distillation, so you have no concern about steady-state feed to the column. Your L/D can be your control together with your reboiler duty and condenser duty.

Have you done research and calculations to verify that you have cause for concern in a possible methanol combustion/explosion taking place during your startup? Have you generated or have you a written startup procedure in hand for this operation? If you have concerns that you can substantiate, then I would hold a PSSR (Pre-Startup Safety Review) and possibly a Hazop with my team members prior to startup of this distillation operation.

I hope that these thoughts and recommendations can be of help to you. Good Luck.

[babucher]

Art,

Thank you for your detailed response. It's great to get feedback from someone who is highly experienced. If I may, I have a few followup questions.

--- quote ---
In order to induce stratification (the natural separation of two fluids with different densities) you must be at (or close to) a static condition in order to allow the different fluids to seek their respective locations relative to each other.
--- end quote ---

This I understand when it comes to immiscible fluids, like oil and water, but what about stratification of gases? Until I thought more about it recently I had just assumed that there would be no 'natural' stratification in the gas phase since gases aren't immiscible with each other due to the nature of the gas phase. Now I'm reconsidering that. I guess at this point my natural inclination is to understand that gas stratification is not an equilibrium condition, but that it's a very slowly changing non-equilibrium condition that still heads towards equilibrium. Am I off in thinking this, and that there actually _are_ (stable) equilibrium conditions in which multiple gas phases exist?

If so, then I could see why it would be ignored in purging because of the convection increasing the rate toward a single homogeneous gas phase.


--- quote ---
What you fail to mention is that the first order of importance in purging a column is to know all the mechanical design information about the equipment (and everything connected to it): the Maximum Allowable Working pressure (MAWP) and the Maximum Allowable Working Vacuum (MAWV).
--- end quote ---

Yes, I totally agree that these are necessary, but I don't have them yet. The things I'm bringing up are more conceptual questions that will help me understand the entire system better. Be assured the mechanical design information is very important to me!

--- quote ---
You are wrong in stating that the process gas that mixes with the non-condensable is condensed and recycled to the feed tank. If this is an atmospheric distillation (which you haven’t stated) then the overheads condenser is vented to the atmosphere; if the distillation is done under pressure, then all non-condensables are automatically purged to the atmosphere through the overheads condenser.
--- end quote ---

Here I'm a little confused. What happens to the process gas during the purging of the non-condensables? If it's not collected and recycled to the feed tank, then it's either thrown away as waste (which seems to be a waste) or it is all vented to the atmosphere, which isn't the most environmentally thing to do. Have I misunderstood?

--- quote ---
I don’t see any problems or anything different from conventional distillation methods. There should not be a requirement to pressurize the column with process gas prior to receiving the initial feed charge. You can receive the initial feed charge into the column while the column is under slight vacuum, go into total reflux until your desired column temperatures are achieved and then proceed to adjust your reflux and overheads make. You have stated that this is a batch distillation, so you have no concern about steady-state feed to the column. Your L/D can be your control together with your reboiler duty and condenser duty.
--- end quote ---

In this scenario, without a purge step, I assume we are venting the non-condensables to the atmosphere until they have been almost exhausted, then allowing the column to come to steady-state, all at total reflux. Is this what you are saying?

Instead of:
1. Purge air with nitrogen
2. Purge nitrogen with methanol/water vapor
3. Charge pot and begin total reflux
4. Reach steady state and GO

We have
1. Purge air with nitrogen
2. Charge pot and begin total reflux
3. Vent nitrogen through condenser until exhausted
4. Reach steady state and GO


I haven't done UFL and LFL calculations yet or written a startup procedure, but these things will be done.

Oh, do you (or anyone reading this) have any references you consider better than Kister regarding distillation operation and, especially, startup?

Thank you very much for your helpful insight,
Brian

[Art Montemayor]

Brian:

You are correct in considering that gas stratification is not an equilibrium condition. There is no clear demarcation line where all the heavy gases are at the bottom and all the lights are at the top. To rely on this to happen is to wait forever.

In a distillation process (look at the attached Excel Sketch I’ve generated for your review) the vapors that exist the overheads should be theoretically all condensables. However, we know that that there are always some non-condensables in this stream – to some degree, more or less. At startup, in your case, there is a bunch: all the nitrogen you introduce for safety will start to rush out through the overheads and the total condenser you have. The condenser can’t condense the nitrogen, so it is vented out while any methanol vapors that increase differentially in concentration are condensed and dumped into the reflux drum where they accumulate and are ready for reflux and product purposes. I don’t know what you call “process gas”, but I suspect you mean the overheads vapors mixed with non-condensables. At steady state conditions I doubt you will have any more non-condensables to handle or vent out of your system. Essentially all that will be going up through the tower will be methanol+water vapors and these will be rectified into pure methanol vapors as they approach the overheads outlet of the tower (the top tray). At this point, you have your purest (methanol+water) mixture and your product – and it should be all condensable.

Kister, as I have said, doesn’t impress me very much with his hands-on experience – at least not on what he writes about. However, I have to raise my hat off to him for discussing the problems and hazards of troubleshooting and starting up a process. He is one of the few authors who has demonstrated the courage and foresight to give these important functions the importance that they deserve. And in that regard he has my utmost respect. I would hope that more Chemical Engineering authors take his lead and continue to develop these themes and subject matters to the point where they become common discussion among students and young graduates. Not enough can be written about the troubles, hazards, dangers, and problems facing an engineer during startups and shutdowns. These are historically the most dangerous times of any process and should be studied and rehearsed until all participants dominate the procedures.

Unfortunately, I cannot recommend any writings on this very important subject. Perhaps some of the people reading this can suggest or refer us to some knowledgeable book or author.

Attached Files

•  Atmospheric_Methanol_Column.xls   19.5KB   178 downloads

[babucher]

Art,

I just realized I hadn't replied with a Thank You for your last post. So, thank you!

Regards,
Brian