4 replies to this topic

[wai]

Dear Art Montemayor,

I working as engineer in a plant for few months. This plant facing mositure breakthrough problem with absorber units. The adsorber designed by other manufacturer. I have no idea why they deisgn in such a way. So far, there's no calculation available regardings the sizing and cycle time. All i have is the specification data according to their design. Since commisioning in Jan 1997, moisture breakthrough occur, the manufacturer suggest we change the mol sieve 13X to 4A in July 1997, which is more suitable for argon purification. Besides, i have done checking on several aspects. Please refer for the event sequence for details. Yet, the problem still occur. Thus, I believe that abosrber units itself cause the mositure breakthrough problem.

The last inspection on the adsorber was in 2006, where the mol sieve was found in good condition. Also i have a filter after the adsorber, where the DP is good. I assume that the mol sieve don't breakup due to attrition. The adsorber continued its operation until today.

Currently, the adsorber is regenerated using pure nitrogen. It pass through heater during regen and heat up to 200C entering the adsorber bed. The operating temperatures of the regen outlet gas is attached.

How can i make calculation to check the NEMA cycle for adsorption and regeneration? Please kindly advise.

Dear all,
Any commend regardings the moisture breakthrough problem? Please kindly advise.

Attached Files

•  Event_Sequence.pdf   92.98KB   130 downloads
•  Design_specification.pdf   12.04KB   145 downloads
•  Valve_Operation.pdf   15.78KB   88 downloads
•  Simplified_Drawings.xls   91.5KB   137 downloads

[Art Montemayor]

Wai:

Thank you very much for informing us with the necessary, detailed background and basic data. We can now make some direct comments that may possibly help you resolve this unfortunate problem you have inherited. I will study and make direct comments in your workbook in the form of call-outs. I have found that this method is the quickest, most accurate method of passing on engineering information. I will follow the USA conventional method of marking up engineering information:

Green color designates items that don't belong and should be removed;
Red color represents comments and corrections that should be included;
Blue color represents the reviewer's comments and recommendations and are not meant to be included in the final product.

Basically, since you do not have the original process calculations to refer to and study as to how the final sizing and NEMA cycle were arrived at, you are definitely at a process disadvantage. You simply do not know the limitations (or design basis) of the adsorber unit. As such, then, there is little we can do except spot and identify the obvious superficial results and questionable items. The basic calculations required for identifying your adsorption needs have to be generated and this will require additional basic data input as well as a Scope of Work. But it must be done to complete the entire evaluation of your adsorption unit.

As of now, I can pin-point various critiques to your installation:

The adsorption of water moisture from the argon stream should be done in a downward direction through the adsorber beds; the bed regeneration should be done in an upward direction with the hot nitrogen;
The regeneration temperature of the nitrogen is simply not hot enough. For a mol sieve application, I always design for 600 – 700 oF nitrogen and with a bed temperature of 400 oF as a minimum; you are not employing sufficient regeneration to the beds. You should be in direct contact with the mol sieve manufacturer (not the distributor) in order to establish the correct use of the correct mol sieve.
You have not identified the use of 4 additional valves on the adsorbers. These are not needed and simply a source of leakage and bypassing of process gas. They should be eliminated unless they can be definitely identified as to scope of operation.
You should generate a complete graph of the temperature profile of each bed, 100% of the time. Your submitted chart tells us nothing since the generated curves are not identified.

I hope to have your workbook ready with comments in a couple of days.

[Art Montemayor]

Wai:

Attached you will find my comments inserted into your Workbook which has been re-named with a revision number 1. I think the failure to attain your design dewpoint is obvious from the remarks and comments that I have made: you lack detailed design calculations and are not regenerating in a proper manner to ensure that the Mol Sieve beds can give you the desired dewpoint.
 Argon_Adsorption_Dryer_Rev1.xls   327KB   183 downloads

[wai]

Dear Art Montemayor,

Adsorbents often have poor thermal conductivity that limits the rate of heat transfer. Based on my case, it also noticed that during the heat cycle, the temperature differential between the bed inlet (200C) and the outlet (165/135C) is great. Could this also be a sign for poor flow distribution in the adsorber? Where the regen gas distribute unevenly in the adsorber leading to part of the bed is not completely regen.

Anyway, thank you so much for the precious advice. I try to review and perform the adsorption calculations. I think i will start from the aspect of regeneration temperature first, which seems like more straight forward.

Hopefully can solve this problem, breakthrough moisture always choke my cold equipments leading to unecessarily thawing.

[Art Montemayor]

Wai:

I'm afraid that either I failed miserably in trying to explain the major defects I found in your system or you have not understood what I tried to point out in my comments. In any case, it seems that I have not clearly pointed out the main fault in your system. Allow me to try again to explain what I found in the basic data you submitted:

1. You cannot expect to achieve the 1.0 ppmV (or less) of water in the product argon when you are using nitrogen with 4.0 ppmV of water as the regen gas. This is not possible because you are using the nitrogen regen gas in a COUNTER-CURRENT direction as compared to the adsorption flow direction. The only way you can ensure the dew point you are seeking by using the same regen gas is to employ a CON-CURRENT direction. The reason for this is that the regen gas cycle is composed of both heating and cooling down of the adsorbent bed. After heating the bed (which is also deficient in your case), you must cool it with the SAME regen gas (with 4.0 ppmV of water). In your cycle, the cool nitrogen enters the bed at the point where the product argon will exit. As the bed is progressively cooled in a downward direction, the 4 ppmV of water is being adsorbed in the top section of the bed. On the next, subsequent drying cycle for that same bed, the argon gas enters the bottom section of the regenerated bed and progresses upward and into the region where the bed has been exposed to the 4 ppmV of water content in the regen gas. Since this top section of the bed has continuously adsorbed the water content in the regen gas, it now has more water than the argon gas going through it. The adsorption process is now aborted and is thrown in REVERSE! The argon gas picks up residual water in the top bed section just before it exits the dryer and, therefore, gives you a higher water content than it would if the top section of the bed had no residual water in it. This is obviously what is happening and it explains what you experience in your operations. In my opinion, it is impossible to achieve your product gas specified water content of < 1.0 ppmV if you continue to use this type of piping arrangement.

It is possible to achieve your product gas dewpoint specs with this same regen gas – but the piping must be re-arranged to a CON-CURRENT regen flow direction AND you must have sufficient adsorbent in your bed to capture (& retain) the water content in your regen nitrogen). If you cannot do this, then the only way that you can achieve your product gas dewpoint specs is to have a regen gas with a dew point that is much lower than 1.0 ppmV. I have done what you are trying to do, but I used cryogenic-source nitrogen (which has much less than 1.0 ppmV of water) as my regen gas. If you have access to an Air Separation Plant, you may be able to use the dry nitrogen produced there as your regen gas. This is the way that argon is conventionally produced.

2. Your regeneration gas temperatures (especially the bed temperature) are not up to the normal molecular sieve regeneration temperatures required to achieve the level of drying you are seeking. You should consult your molecular sieve manufacturer and you can confirm what I say with them. I can assure you that your regen temperatures are too LOW. You must raise your regen gas temperature and your bed temperatures in order to fully regenerate the mol sieve. Again, confirm this with your sieve manufacturer. Note that I say MANUFACTURER and not supplier (or distributor). You must consult with mol sieve experts about the regen requirements as well as to what practical sorptive capacity you can expect from your mol sieve. You need this information in order to carry out the process design calculations for the unit and establish whether your beds are big enough and your velocities are adequate.

You are wasting time concerning yourself with the mol sieve thermal conductivity. That has nothing whatsoever to do with your deficient bed regeneration. You should be heating the bed up to at least 250 ^oC and you will never achieve that if you only heat the regen gas to 200 oC. Think about it. Heat transfer can only occur from a higher temperature to a lower temperature. Again, you can confirm this with your mol sieve manufacturer to find out that what I state is true. I do not believe that you have bad heat distribution. You simply don't have ENOUGH heat. Make a material balance and convince yourself. The heat that goes in must come out; there is no accumulation. The outlet temperature is low because the inlet is low. Do not forget that in applying more heat, you must have adequate materials of construction. You have not told us the history of your unit (who designed it and who commissioned it), but if you lack the original process calculations, than you certainly also lack the materials of construction identification. If that is true, then you must confirm exactly WHAT materials of construction have been inherited by you. Without this information, you are powerless to decide whether to apply the correct, higher temperatures or not. Your regen heater alone will have to absorb a lot of more temperature – approximately 300 to 350 oC. So be very cautious and careful in making any modifications.

Go back and read all my comments carefully. You have not raised any questions on my comments, so I must assume that you concur. If you don't, then please question them and I can explain in more detail. If you have never calculated or sized an adsorption unit, then obtain the services of an experienced process engineer who has and can be relied upon.

I hope these additional words of experience are of some help. Proceed with safe and careful engineering and good luck.