20 replies to this topic

[fedor]

Dear fellow engineers,

 

First of all thanks for your great work on this forum. especially thanks to Art Montemayor for his great posts on TEG dehydration. I have learned a lot from your posts.

 

Shell Co. Design and Engineering practice for gas dehydration has mentioned that no more than  35 Sm3/m3  (equal to 4.7 SCFM of stripping gas / gal of TEG circulated) shall be used or stripping gas because more than this value will cause flooding in stripping column.

The vendor who is going to supply TEG package has designed a system using  89 Sm3 / m3  (equal to 12 SCFM / gal ) for stripping gas. the reason the amount of stripping gas is relatively high is that the reboiler works at 1.5 bara so the off gas can be discharged to flare header. The 1.5 bara working pressure compared to atmospheric pressure needs more stripping gas to decrease water vapor partial pressure.

designer claims that flooding will  be solved by using bigger diameter for stahl (stripping) column.

 

should we be concerned that vendor is using 3 times more stripping gas than maximum recommended by shell's practice?

I appreciate if you please advise on this issue.

[Bobby Strain]

So, why do you think Shell has expertise that applies to your system? And, of course, you know how to calculate loading/flooding in a packed bed. That's why you spent so much time in school.

 

Bobby

[fedor]

Thanks Bobby.

 

Do you mean Shell has just written some nonsense in their document?

 

Have you seen stripping gas rates as high as 12 SCFM / gal ?

[Bobby Strain]

There is usually an economic limit to stripping gas use. The requirements are dictated by your dehydrated gas water content/dewpoint specification. I have operated systems that achieved about -25 F dewpoint with cracked gas at about 150 psig. This is quite a bit more stringent than the typical gas spec of 7 lb/mscf.

 

Bobby

[Pronab]

Fedor,

 

In our TEG regeneration system, we are using 200 Kg/hr of stripping gas, for a 14.5 m³/hr of glycol circulation and we are getting 0.12 ~ 0.15% water content in lean glycol.  Our gas de point is -65 to -68 Deg C.  Previously, stripping gas flow was ~ 450 to 500 Kg/hr.  The Stripping column pressure is 0.1 Barg at present and it was 0.5 Barg before.  For stripping gas flow in a TEG regeneration system, it is an optimization and how you want your gas dew point.  I have seen literature where 6 to 8 SCFM of stripping gas flow per gal of TEG flow is recommended.

[Zauberberg]

Is your company obliged to follow Shell DEP? If yes, then you have to talk to the unit supplier. If not, then you can forget about DEP because it makes no sense to follow standards of other companies.

 

The amount of stripping gas is calculated based on the working temperature of the reboiler and the number of theoretical stages in the stripper - with the target to achieve the required purity of glycol. As it can be seen from the attached chart (GPSA, 13th Ed), amount of stripping gas can vary from 0 to 75 or more Sm³/m³, at atmospheric pressure. For higher pressures - such in your case - the corresponding amounts would normally be slightly higher. If stripper column is sized accordingly, it will operate below the flooding limits at the design rates of stripping gas.

 

One additional observation: if stripping gas is not dry, the resulting amount required for the same stripping effect will increase. Perhaps this is your case.

 

Attached Files

•  Pages from M20.pdf   95.12KB   185 downloads

[fedor]

thanks Dejan

we are not obliged to follow DEP, so i  think we will trust vendor.

 

in your attached PDF  N=1,2  is for total  stages (reboiler and sripping column)? or just the stages in the stripping column?

 

in other words, does N=1 means injecting directly in the reboiler or having a stripping column with one theoretical stage?

Edited by fedor, 21 February 2015 - 06:38 AM.

[Zauberberg]

Hi Fedor,

 

Theoretical stages refer to the stripper. For random packing used in these applications, one stage corresponds to ~4ft of packing. Stripper columns I have operated, reviewed, or designed so far have between 0.8m and 1.3m packing layers.

 

You can also ask vendor to provide you with references of similar projects - particularly those with regeneration systems operating at higher pressures. You are the client so you should use this opportunity to obtain as much information as possible.

[RockDock]

I have seen TEG units with that high of a stripping gas ratio. They usually ave a good deal of BTEX in the overheads as a result. As long as that fits with your requirements, you should be fine.

[Propacket]

Stripping gas can go even higher than 12 SCF/gallon. There will be two consequences:

Normally the stripping gas is taken from the dry sales gas. So higher the stripping gas, higher the operating cost associated with sales gas loss.

Stripping column and still column size must be increased to avoid flooding. Try using some good softwares like KG tower and Sulcol.

[Art Montemayor]

Fedor:

 

You have received some very good, experienced, and practical recommendations in this thread.  I can only add a few comments that may be of some interest and also further emphasize what has already been mentioned.

 

A TEG regeneration system operating at 1.5 bara is not far removed from normal field operations.  Some TEG unit fabricators build their units so “tight” (and cheap) that the pressure drop between the TEG reboiler and the atmospheric vent is practically 0.5 bars.  This is due to competitive bidding and unless you take a strong, client stance in your specifications, the supplier will try to economize as much as possible in the unit design.  The result is usually high superficial velocities (resulting in high TEG losses and increased reboiler pressure).  It is possible to operate the TEG regeneration unit at a pressure well below 1.0 barg - which qualifies the fabrication as “non-pressurized vessel design (no ASME or equal stamp).  Additionally, the sizes of the TEG stripper(s) - both the Still Column and the Gas Stripping Column - are relatively small and employing a conservative, small superficial velocity in both columns does not increase the total cost of the regen unit in a significant way.  I always recommend this wherever possible because it makes for a much better operation and efficiency with less entrainment and TEG losses.  As Zauberberg notes, you are the client and you should use your right to specify what you need and what you want in order to obtain a successful, efficient unit.  I would insist that the designer-fabricator furnish the calculations and related curves that show the stripping gas requirements for your operation and include these in the normally furnished Operations and Maintenance Manual for the Unit.

 

The curves for stripping gas requirements were first developed by Steve Worley for BS&B back in the 1960’s and form the basis for what is still done today.  Some additional work in this area has revealed that the benefits of using stripping gas diminish rapidly beyond rates of about 15 to 20 L/L of TEG solution circulated (Hernandez-Valencia, 1992).  All stripping used - by definition - should be DRY (and pre-heated) product gas.  While humid, saturated gas can be also used, this significantly can alter the desired results and limit the extent of increasing the Lean TEG purity.

 

As an interesting point, I am attaching a copy of a paper issued some years back on Shell’s incorporation of TEG regeneration at high pressure (4 bargs).  This might serve as some insight as to what your designer is telling (or not telling) you.  You don't specify what your Lean TEG purity requirements are, so we can't estimate what your stripping gas needs are.   Neither do you state what your product gas dew point should be.  Have you read the paper, “Glycol Dehydrator Design Manual” by Richard Sivalls that I uploaded to our website sometime back?  If not, I recommend you study what it contains.  A lot of the basic design information there still holds for designing and operating a TEG unit.

 

Attached Files

•  Pressurized Glycol Dehy Systems.docx   247.95KB   181 downloads

[Zauberberg]

Fedor,

 

Here you can find the Sivalls' paper provided by Art Montemayor: http://www.cheresour...gn-manual-1976/

[fedor]

Art:

 

thanks for your great advice. the paper you attached gave me good insight.

i have already read these:

GPSA

Manning

Campbell

Sivalls (the file including your comments)

shell's DEP

Kohl & Nielsen

 

the vendor has used HYSYS simulation and as it is mentioned in paper "STATE OF THE ART REVIEW AND RECENT DEVELOPMENTS IN GLYCOL DEHYDRATION FACILITY MODELLING AND OPTIMIZATION" HYSYS overestimates the required stripping gas.

The lean TEG is 99.7% and dew point of dry gas -4 celsius.

 

another strange result from HYSYS is that it shows solubility of hydrocarbons in rich glycol at bottom of contactor and these hydrocarbons will flash downstream of control valve after losing pressure. it shows that half of the volumetric flow downstream of control valve is gas. can this be correct?

[Art Montemayor]

Fedor:

 

As you may suspect, the amount of hydrocarbons dissolved in the Rich TEG depends on the Contacting operating pressure, the temperature, and the quality of the raw feed gas.  The amount of hydrocarbons released at the TEG Flash Drum (which I presume you have included in your proposed design) depends on the process condtions and the Flash Drum pressure (I usually settle on approximately 45 psig).

 

Depending on your raw feed gas analysis, I would definitely employ a Flash Drum in the TEG regen unit.  Please read the attached file I am attaching on the use of a TEG Flash Drum.

 

I personally would not employ HySys as my simulation design package for a TEG dehydration unit.  It has a lot of bugs and has not demonstrated accurate results in the past.   These are just my personal lessons learned with HySys regarding its use in natural gas dehydration.

 Optimize TEG Circulation and Flash Tank.pdf   200.56KB   137 downloads

 

[fedor]

Art:

 

the contactor pressure is 40 barg and temperature is 60 Celsius.

raw gas molar fractions: 68% methane, 16% ethane, 8% propane, ... , 2% CO2, 0.06% H2S, 0.82 % H2O

 

flash drum has been in vendor's scope of supply from bidding stage.

hysys shows slug flow downstream of control valve.

according to hysys:

TEG enters the reflux condenser as slug flow (which we are worried about). after flash drum the flow becomes single phase but as soon it looses pressure (in filters and lean/rich exchanger) the remaining hydrocarbons flash and TEG becomes slug flow again. we are concerned about slug flow entering the filters, heat exchanger and the control valve before still column.

we have commented on vendors report. he has suggested moving the flash drum before reflux condenser but that will only solve slug flow entering reflux condenser.

 

could hysys be correct about these 2 phase slug flow predictions?

what simulation package do you recommend for TEG dehydration?

 

thanks in advance

Edited by fedor, 27 February 2015 - 03:03 AM.

[Zauberberg]

There is no PFD we can comment on, but a common way to eliminate two-phase flow from glycol circuits is to locate level control valve as close as possible to the last equipment in the circuit, meaning:

 

1) Absorber LCV install as close to the Flash Drum as physically possible;

2) Flash Drum LCV install as close to the Still Column as physically possible.

 

This will normally result either in increase of design pressure of the equipment between the source and the LCV, or necessity to provide PSV sized for blow-by case.

[Art Montemayor]

Fedor:

 

Zauberberg is giving you the correct way design engineers mitigate two-phase or slug flow.  It’s common sense that you gain nothing by transporting mixed flow in a pipe except problems.  Therefore, what experienced engineers do is that they allow for the mixed flow immediately at the point where the phases are separated - at the Flash Drum and at the top entrance to the Still.  Both the Drum and Still should be designed for phase separation at those junctions and should do their job accordingly.  Attached is a sketch of how I’ve always piped up the entrance to Strippers or 2-phase separators.  Note the LCV orientation and the use of an expander pipe joint to slow down velocities and facilitate phase separation.  You can use a conventional control valve instead of the 90^o style, but I’ve found that it takes more space between it and the vessel due to the diaphragm operator.

 

40 bargs (480 psig) is not a high pressure as TEG contactors go.  The amounts of ethane and propane in the rich TEG shouldn’t be that much.  I would expect to vent the majority of dissolved gases (90-95%) at the Flash Drum.  What does your simulation predict?  If you don’t have any heavier hydrocarbons, you shouldn’t have a liquid hydrocarbon phase in your Flash Drum - which makes it easier to separate and justifies your using a generous low superficial velocity there.

 Flash Drum.xlsx   13.65KB   110 downloads

Edited by Art Montemayor, 27 February 2015 - 05:59 PM.

[fedor]

Art / Dejan

 

your advice that LCV be as closest possible to flash drum/still column is implemented in the design.

the problem is that simulation predicts that TEG exiting flash drum still has hydrocarbons (CO2, methane, ethane and propane) that will flash after passing through filters and lean/rich exchanger and especially after lean/rich exchanger which is heated to about 150 Celsius and its water will also vaporize.

 

Art

have you seen any real cases of problematic slug flow passing through still column LCV ?

i mean any vibration or hammer shocks? or damage to LCV?

 

Dejan

can you plz explain more on "PSV sized for blow-by case" and how can a PSV help ?

Edited by fedor, 27 February 2015 - 01:39 PM.

[Art Montemayor]

Fedor:

 

The facts are that you will always have some hydrocarbon gases dissolved in the rich TEG and these will form a slight residual even after the flash drum expansion.  This is proven in the concern for BTEX’s in the TEG still’s vent gas stream to atmosphere.  These, in my opinion, in your case will be slight.  What your simulation is showing hasn’t been quantified in a heat and mass balance, so I don’t have any idea of the flow rate of these dissolved gases as they enter the Still.  I can assure you that whatever they may be, they will most certainly not reach the 400 ^oF TEG reboiler.  They will exist out the top of the Still - presumably out to atmosphere or to a post treatment or incineration, all depending on what your process design calls for.  We haven’t seen a P&ID yet or any project specifications or scope of work.  Who is doing the expert process design and what is their experience in TEG dehydration?

 

If the dissolved gases are a “problem”, you haven’t quantified or qualified it.  A typical TEG still should always be designed to separate out these undesirables through the top vent.  In the past these were simply contaminants to the atmosphere.  Today, these are normally not allowed and downstream facilities have to be used or process modifications have to be done to either, recover them using an effluent blower, incinerate the hydrocarbons or to use them as fuel in the direct-fired TEG reboiler.  Again, we are responding without the input of design scope, details, and project description.

 

I have never experienced any TEG field operation with slug flow passing through the still column LCV, hammer shocks, or control valve damage - and I wouldn’t expect it.  Typical TEG circulation flow rates - even those for really large dehydration loads - are relatively small compared with other normal process liquid flow rates.  15 gpm for a 100 MMScf natural gas unit would be approximate and here we would be looking at pipe sizes smaller than 1 inch.  Using larger than required TEG pipe sizes is a necessity in order to have proper piping support and stability in the field.  Under these field conditions surges and slug flow have been the least of my field concerns in the past. 

[Zauberberg]

Art is providing a great advice. Whatever flashing you get downstream of the filters is due to pressure drop across this equipment. The amount of gas should be really negligible, and poses no real concern.

[Zauberberg]

And one more experience-based advice: I always specified the LCV (and the manual bypass valve, if exists in the design) between the Absorber and the Flash Drum, as a severe service valve. This is due to the fact that the valve operates at very high pressure drop, and it handles a fluid which carries solids (gas impurities, corrosion products). Normally you want something very reliable in this application so the unit vendor can supply you with a cage-guided valve with seat/plug made of abrasion resistant material. On many occasions the SS 400 class of steel was more than sufficient, but this should be assessed on a case to case basis.