16 replies to this topic

[pawan]

I would like to debate the need of reflux drum in continuous distillation systems. Why it can't be through gravity flow? Simply condenser outlet can have one U seal on both sides & one can go to column as reflux and other one can be used for product withdrawal. Let us start this discussion / debate for vacuum distillation which is more critical in terms of pressure balancing, flows etc.

[Zauberberg]

How would you control the flow - both reflux and the product? With gravity flow there cannot be any control valves in the system. And there's no head available for product rundown.

The drum provides residence time as well. Without such an item, almost any instability in column operation would cretae a snowballing effect.

[pawan]

I am attaching a sketch for clear understanding of basic concept. The head is provided for FT & FV for controls but to the tune of 1 meter max in each.

Now if this configuration is used you are removing reflux drum & reflux pumps saving some of your capital.

Attached Files

•  Sketch.xls   18.5KB   303 downloads

[Zauberberg]

Pawan,

No offense, but this cannot be considered as serious concept in industrial distillation. Even if it might be workable for some applications, just by considering a few possible drawbacks would be sufficient to eliminate this as a design solution. The cost of having a distillation system incapable of making (and delivering) products is much higher than the cost of 10 reflux drums and 20 reflux pumps altogether.

How much would need to be height of the condenser in order for liquid leg to overcome frictional losses, pressure drop across the condenser, and FT/FCV pressure drop? A few dozens of meters, for sure. That already - just by itself - sets the requirement for additional piping costs.

Furthermore, how the product will be transferred to storage, or to a unit downstream? From a standpipe e.g. operating at 1 barg, to transfer the overhead product to the storage 1km far away? Or to a downstream distillation column operating at 10 barg? This is virtually impossible.

Another item to be considered is: control. Based on what process parameter you will set the reflux rate, and the product rate? In common applications, reflux is set on flow control, and product rundown is set based on level in the reflux drum. In your particular scheme, how would you know how much reflux and how much product you can withdraw from the standpipe/liquid leg?

As you can see, there's a lot of common-sense engineering issues which make this arrangement unpractical, difficult to operate and, essentially not applicable for great majority of industrial distillation systems.

[Art Montemayor]

Pawan:

I don’t think you mean to “debate” the need of reflux drum in continuous distillation systems. I believe you mean to discuss and/or analyze the possible application of such a system IN A SPECIFIC APPLICATION.

I state this because I have operated distillation columns operating without a reflux drum or a reflux pump – for many years. These colums produced an overheads product purity of 99.9% mol purity – consistently and “24 – 7” (24 hrs a day, 7 days a week)! HOWEVER, this was a very specific and special application – air separation columns for the production of pure oxygen and nitrogen. In fact this is still being done today – 50 years after I did it. It is still being done because there is NO ALTERNATIVE solution to delivering a cryogenic liquid nitrogen reflux to a Linde Double Column except under gravity flow. As I have stated, this is a specific and special application.

I mention this because I interpret this as what Zauberberg is pointing out. He doesn’t state that it won’t work – or can’t work. He states that it isn’t a practical or viable application under the usual and conventional scopes of work that are applied to most industrial distillation processes. And I believe he is correct when we consider the caveat that he applies.

However, having stated that, I should also emphasize that basically your query asking if a reflux can be effected without use of a reflux drum or pump is a valid one and that the general response is: YES, it not only can be done, it has been done – for many years in the past (over 110 years, in the case of Linde’s Columns). But I believe that this statement should be made only in a SPECIFIC light, not a general case. I say this because in the greatest amount of industrial cases, the employment of a reflux drum and a reflux pump (together with all related instrumentation) is a required necessity in order to fulfill the related scope of operations.

I hope these comments help out in discussing this specific subject.

[DB Shah]

Dear Pawan,

I too have operated column without reflux drum which is working fine since last 20 yrs in Formic acid plant (Methyl format column)and also Linde ASU column, but it is to be considered case specific. Without reflux drum what you save is - some capital cost/ some reflux pumping power.

Agianst that what you loose?

What will be the elevation of the condensor if you do not intend to have a reflux drum?
You will have to place condensor above distillation colum & the elev of the condnesor will be
= column elevation + head loss in FT/CV/lines etc.
At that elevation you will be circulating the cooling fluid (most probably cooling water to that extra height.)

Just an illustration-
Consider a distillation column condnesing 1000 kg/hr vapour of 200 kcal/kg latent heat and liq density of 1000 kg/m3. Height of the column 50 mtrs. and FT/Line/CV losses ~ 10 mtrs

Opex Saving -
1 M3/hr of pumping of (reflux+product) from reflux drum elevation (~ 8 mts) to 50 mtrs
Power saving 1*1000*(50-8)/3600*.75/75 = 0.12 kw

Additional opex due to CW pumping at higher elevation
CW flow with 10°C DT in condenser = 1000*200/10=20000 kg/hr ~ 20 M3/hr
Difference in Pumping head = Elevation condensor - elevation of reflux drum = (50+10)-8+ = 52 mtrs

Addtional pumping = 20*1000*52*.75/3600/75 = 2.9 kW

Net power loss = 2.9-0.12 = 2.78 kw ( this is without pump inefficiency)
If your pump efficiency is 60%, power loss = 2.78/.6 = 4.6 kw

Hence unless process specific compulsion I donot see good reasons to do away with the reflux drum.

[pawan]

Thanks to all three of you.
Actually I used the word DEBATE because I am also in favor of using reflux drum due to following reasons. Yes you are right its a discussion as long as we agree.

1. First & most important function is to provide surge in case of column fluctuations.
2. Second is the control & flexibility in the reflux & withdrawal independently, which is not possible in gravity flow (even if possible they are dependent on each other).
3. It gives flexibility to cope up with the pressure sensitive operations OR variations in the column operating pressure due to changes in composition.
4. It also gives lower opex as indicated by Mr. Shah.
5. Also need minimum 14 meter heigh structure in case of vacuum operation so as to withdraw product through barometric leg. So if column is short there is no point in going to 14 meter elevation.

And some more points can be added....
BUT

in one of the company, the usual practice is to follow gravity system even in conventional organics handling. Apart from opex & capex considerations, which may be minor & are case specific, I want to prove it logically incorrect practice. Since, it is already running, it is difficult to convince the team.

Now....Mr. Zauberberg's message

Today FT & FCV do not require more than 2 meters of elevation, So there is no significant height difference if column is already running to say 11 or 12 meter height (which is the case here).
Further, generally product is stored in day tanks & after quality checks only it is transferred to main storage tanks. So day tanks are not likely to be so far located.

For setting the control, the practice here is to control the overhead product flow & rest goes to reflux. YES This is one of the reasons I do not wish to design for gravity flow. This is already covered in point 2 above.

So, I accept & convinced with your last point of control mechanism.

Now.....Mr. Art's message
Yes Art I agree that we are not saying that it won't work. I also agree that it should be case specific. But here my problem is that it is used as a normal practice here in this company which I also do not agree. so logically I mentioned above points & look forward to have more reasons from others.

Now.....Mr. Shah's message
I agree Opex is much bigger than capex but I am looking for points other than this.

I hope I am on the right track.

[DB Shah]

Dear Pawan,

As we all indicated that reflux drum installation will be case specific. Because the beauty of process engineering is - You donot have A answer to THE question. You have choices, just as we choose leaders in our democracy-choose the least rascal.

If I understand your point, you are trying to convince client for installing reflux drums for all above stated reasons, and the counter part is convinced with his philosopy of gravity reflux.

As I had earlier stated even we had our formic acid plant with gravity relfux control on methyl format column. We operated it for > 20yrs on that philosophy & recently 2 yrs back we installed reflux drum and reflux pump.

Even though I am in favour of reflux drum I can also argue in favour of gravity flow-

1. Head loss requirement
As correctly stated by you top product draw off is measured,controlled where you have plenty of head available. Balance of flow is send back as reflux. So no issue.

2.Control philosophy
As stated in Point-1, product off draw has a control valve, this control valve is a TIC and controls column profile. This gave us very nice control for years. Hence no issue again. Yes you cannot measure reflux flow (enough head not available to install any flow elements), but if I have product flow with consistent purity, why worry about reflux flow measurement?

3.Opex
Suppose due to some other requireemnts in the complex the cooling water system is already designed for high head, then opex will remain same irrespective of condensor elevation. On the contrary, I save power in reflux and product pumping.

4.Capex
If the column top vapour MOC is exotic I will certainly condense it near the column itself rather than bringing it down. For the same mass flow vapour line sizes are higher compared to liquid lines and due to sub cooled liquid I may be able to use cheaper MOC for reflux lines. Also in this case reflux drum will be of costly MOC. So a huge saving in capex if I condense the vapour on the top and use gravity control.

With this arguments, should I not continue with gravity flow ???

[Art Montemayor]

Pawan:

I agree entirely with DB. When a gravity flow Reflux system works, it is a thing of beauty – to enjoy and to relax with. There are no moving parts and, consequently, there is no maintenance! Like the little battery-operated drummer bunny, it just works, and works, and works, and works, ……..!

HOWEVER, there is always the eternal TRADE OFFS looming in the background:

• The column feed flow rate starts to vary – or you require a turn-down or turn-up in capacity and you have to vary the Reflux rate;
• The column feed composition starts to vary - and you have to vary the Reflux rate;
• The column feed temperature starts to vary - and you have to vary the Reflux rate;
• The distillation efficiency starts to vary - and you have to vary the Reflux rate;
• The reboiler starts to foul - and you have to vary the Reflux rate;
• The condenser cooling water starts to heat up - and you have to vary the Reflux rate;
• The product storage has to be moved further away – and you need an overheads product pump.

There are other trade offs to confront, but for the sake of brevity, I won’t go into them. The main point here is that, like every thing else in life, there are choices to be made – and each choice has a trade off attached to it. Everything has a price; there are no free rides or lunches. Gravity Reflux systems are great and the ideal solution – however, they must have a “fit” within the scope of work and the conditions of the process. Otherwise, you have to pay a “price” for incorporating them. Many times – more often than not – this is an unacceptable solution. In my opinion, this is unfortunate, because I would love to have nothing BUT Gravity Reflux systems. However, I have to live in an imperfect world.

[pawan]

Perfect Art & Shah
I fully agree on all points.

However, I see Gravity reflux as primary choice for Batch Distillations and Pumping system for Continuous distillation in general.

For Specific cases it may change.

Just for further debate on DBs points...they are the same points I am arguing with the company...

1. Head loss requirement
In case of vaccum system (as applicable in this case), we need to put Barometric leg to collect the product which means I need a pump for transfer of the product from that Baro leg vessel to downstream process + I need that vessel also - So I am not avoiding anything while compromising a lot on flexibility in a continuous distillation system. ( Most of those points are already mentioned by Art). Even same reflux pump can transfer product to any location - no in-between equipments & instrumentation.

2.Control philosophy
Surges in the feed flow, composition, steam flow etc can not be easily accomodated in gravity system as you cant produce more reflux than vapor flow. Also your product withdrawal may go on off & may affect downstream system. So not desirable. Need to put column under total reflux for accomodating any variation & hence compromising on capacity, energy, losses & sometime off-spec product also if condenser can't acocmodate variations.

3.Opex
This may or may not be - on the contrary higher head will result in higher flow through condenser and will reduce exit temperature which may result in lower losses in vacuum system, better purity, higher capacity etc.

4.Capex
In case of exotic MOCs, it doesn't matter - ultimately you need pump & vessel again as mentioned in point 1 above. Yes, there will be some gain if MOC selection highly depends on temperature.

So why to use gravity flow....?

[breizh]

Hello ,
Sorry if I'm Off the discussion .I've been operating a brand new plant with a train of columns under vacuum,continuous process where the condenser was installed on the top the column , meaning the reflux was an internal reflux and the distillate was withdrawn on the side just underneath the condenser .

Breizh

[pawan]

Breizh

Actually that is another option for different purpose & they are called "hammer condenser". More useful in case of very slight low boiling impurity in the product.

Edited by pawan, 20 November 2010 - 09:07 AM.

[Art Montemayor]

Pawan:

I don't know where they are called "hammer condensers", but I do know they are used quite a lot. The last time I saw a series of these type of overhead condensers was in 1986 at the BASF complex outside of Frankfurt, Germany. These type of condensers - as I keep harping - are SPECIFIC to the application. They are air-cooled (not water-cooled!) units that structurally "sit" on top of the distillation column - exactly like the type that Carl Linde designed and built in 1892 for his air separation columns. This type of overhead condenser is very practical and efficient - but it can only be applied to those condenser temperatures that allow ambient air as the coolant. Again, the application is restricted by the specific nature of the design and the fluids involved.

You continue to discuss and make an argument against the decision to use gravity flow reflux, but you also continue to resort to specifics. I will repeat: the use of gravity flow in a distillation column is CASE SPECIFIC. Your client(s) decision on using this method may be also based on the specifics of the application. We haven't asked you whether the client decision is a general one or a specific one, but now is as good a time to ask it. Otherwise, we'll be bouncing back and forth between generalities and specifics. If we are to discuss generalities, we will be here adding to this thread until Hell freezes over.

I believe we all agree with your general opinion of using a conventional reflux drum and pump - but we haven't seen any specific information or data yet that would discard the use of a gravity flow system. Is there any specific data on the application(s)?

[pawan]

Yes Art.
Let me put some data here. There is one vacuum column of ~6 Meter total height T-901 (packed height only 4 meters) operating at 20 mmHg abs pressure. Overhead is to go to a feed vessel V- 801 (Under atmospheric pressure) which is feeding the product to another distillation column T-901 at 4 barg pressure using pump.

Now if I need to put gravity reflux - I need to raise the elevation of T-901 to use Baro leg for gravity flow so as to withdraw the overhead in V-801.

If any other data is required I can share. Will you recommend gravity reflux in this specific case?

[Art Montemayor]

Pawan:

You haven’t furnished all the basic data, but I now have sufficient information to address your specific assertion and your question.

If you need to install a gravity reflux system, you DO NOT need to raise the elevation of T-901 in orderto use a Barometric leg and withdraw the overhead product in V-801. There are several other options you can use:

You can build a highter structural steel suppor for the overhead condenser. You haven’t mentioned anything about structural, but I should warn you that the structural steel WILL be required if you install the condenser above grade and for gravity flow. You will have need to design it for access, maintenance, and operation. This mean safety hand rails, stairs, lighting, etc., etc. This part of the gravity reflux system could very easily be the most capital expensive in the project – depending on how high you go and how strict your local safety laws and regulations are.

You can also EXTEND the height of the tower by adding a top section that includes a vertical, vapor-in-tubes overhead condenser. Please refer to your workbood which I am attaching with my sketches and explanations.

I cannot recommend gravity reflux for your application because I don’t have all the basic data. I don’t have all the capital costs and capital returns, for example. I can only make general process comments and point to the advantages and disadvantages. I don’t even know the fluids involved. If I were to consider gravity reflux the first thing(s) I would need to know for process design feasibility would be the condensing temperature of the overheads product and the ability to use ambient air. I would also have to accept that the reflux ratio would practically stay constant during the life of the process and that there would be no process modifications in flow rates, temperatures, compositions, and pressures. If that were the case, I would opt for an air-cooled, gravity reflux condensing system as I depict it in the workbook.

I hope this work helps you out.

Attached Files

•  Gravity Reflux System Rev1.xls   209.5KB   158 downloads

[DB Shah]

Dear Pawan
Time and again we are emphasizing that any selection will be case specific. I am neither in favour nor against any of the control philosophy.

Your company is taking a stand for garavity control, surly that is an experience based stand. Let me tell you, you can learn much more from your experienced operator rather than any of us who are not aware about the actual scenario. We in the forum can only give guidance. Details and finer aspects are to be worked by you.

Secondly you cannot and should not expect the company management to understand the process engineering terms. They are not bothered of control philosophy, reflux ratio etc, there only concern is money. Make a comparative statement of both the cases. Try to quantify merits and demerits of each operation.

From the facts given and stand of your company, I pressume that the product is a speciality chemical like pharma/agri or pigment etc which are very high cost products. Here major variable cost is raw material cost. % of Opex is negligible. Hence oprating at safer paramters like higher reflux ratio than required is possible. (I designed one pharma plant of my friend, cost of end product 7000 Rs/kg ~ 155 $/kg as against a popular chemical- Methanol - 0.3~0.4 $/kg)

I have an excel file. I have tried to touch some deeper aspects (which may or may not be applicable for your case).

For eg-
Cause
Reflux pump tripping

Result
1. Column starving of reflux-> bottom tempertaure shoot up -> degradation ??
2. Reflux + Level controller opens wide -> Reflux line and product line communicates -> vacuum breaks -> air ingress from atmospheric tank -> degradation/explosive mixture ?? (NRVs are bound to pass)

This scenario will not be applicable in gravity control. Also note that your column will not be deprived of reflux in gravity control logic. Due to barometric leg, column will not face vacuum break from outer source except vacuum pump failure.

This is a just a hypothetical case for understanding. You should carry out such exercise and suggest a convincing option to your company.

Attached Files

•  Column control.xlsx   23.54KB   170 downloads

[pawan]

Thanks to All for your time & Inputs.