21 replies to this topic

[cjre2]

Hello there! I am new here in cheresources. I am currently troubleshooting a problem on the overhead condenser system of our reformate splitter. The overhead vapor of our column goes to two (2) parallel but different air fin coolers (see attached for the configurations and the details of the exchanger). A butterfly valve was installed for flow distribution to these exchangers. By the way, the reason why we have this configuration is that the BED contractor failed to identify that existing Exchanger A is limiting, hence, EPC contractor has to install a parallel exchanger due to time constraint.

I have checked the sufficiency of the surface area and motor/fan power and found it to be adequate. However, we are still losing about 150-230 barrels of product to the flare (estimated based on flare seal drum level changes). This is observed in spite of fan pitch are set at the angle for maximum air delivery for both, motor and fan rpm are per design, no excessive air recirculation as evidenced by the actual air inlet temperature of 35 deg C vs 38 deg C used in the exchanger evaluation, no significant air side fouling, and tubeside fouling is minimal as this is in clean service. At present, we are changing the distribution between the two exchangers by adjusting the butterfly valve and observe its effect on flare slop generation to determine the optimum opening of the butterfly valve. Nonetheless, we can not totally eliminate the flare slop. What other factors should I look at in troubleshooting this issue? Initially, I am doubtful on the size of the outlet of Exchanger B. Can you share to me on how to assess or provide design guidelines for condenser outlet line sizing? Also, the outlet line of Exchanger B has a longer horizontal run compared to Exchanger A since the former is located away from the drum. Both exchangers are at more or less the same elevation and equipped with siphon break line and NC vents. Appreciate expert advice on this. Thank you.

Attached Files

•  Book1.xls   105KB   103 downloads

[hasriabdan]

Hi,

is the press is 7 kPag, which is too low with the outlet temp at 68 degC. Please revise these parameters especially the system pressure - i.e, need to increase I believe.

I don't see any problem on the design as it is adequate enough to cater the current process.

Regards,
Hasri Abdan
Melaka Refinery

[USR]

Dear,
I strongly feel that a hot vapour condenser bypass line could have given you much flexibility of operation.

[cjre2]

Dear,
I strongly feel that a hot vapour condenser bypass line could have given you much flexibility of operation.

I also think this is a good idea but the changes may be more complicated given that we have two different exchangers in parallel. We have considered this but this may not be doable in April 2011 due to the large valves needed. Do you have other idea on things that we can consider? Appreciate your assistance. Thank you.

[DB Shah]

Dear cjre2

For parallel condensers you will have to finely carry out - Hydraulics and thermal balance. What I mean to say is if exchanger A is designed for 75% load and B is designed for balance heat load of 25%, vapour flow should also be distributed in the same proportion.

From your problem description & schematic, I feel that even though exchanger B is designed for 25% heat load, due to higher pressure drop in the B exchanger loop, exchanger A is receiving higher flow rate.

Reason-

1.Heat load of exchanger A is 75% & B is 25% hence vapour flow distribution should be 75% in A & 25% in B.
2. If we go by pressure drop co-relation DP is a direct function of length, direct function of square of ratio of flow and inverse function of fifth power of line diameter
(Loop A vapour line dia 20", Loop B dia 12" - from your attached sheet.)
If the pressure drop in A loop is x psi for 75% flow, pressure drop in B loop will be

DP in loop B = x * (Lb/La) * (25/75)^2 * (12/20)^5 psi

Where La & Lb are line length of vapour lines of A & B condenser.
25 & 75 are the flow ratio for both the condenser
20 & 12 are the vapour line size.

Assuming length of vapour inlet to both loop is same.
DP in loop B = 1.4 times of loop A for required flow distribution

Apart from above you are also throttling butterfly valve on loop B and line length is higher in loop B.
Ideally you should size loop B for less DP than loop A and then provide a valve on loop B to control
flow distribution.

I had installed one condenser in parallel in 2002 and I was under the same dilemma. Fortunately I was able to discuss the matter at length with Mr Andrew Sloley before designing. You can view the detail conversation at -
http://www.distillat...question015.htm

DB Shah

[cjre2]

DB Shah:

Thank you for your inputs and for the your previous discussion. These will be helpful. I will study what you have done in your previous design work and see what we can apply here.

I have done simulation and found that the delta P is more concentrated on the Loop A primarily due to the 2 pass flow configuration of Exchanger A. Based on the simulation, we can position the butterfly valve betwee 40-45 degrees in order to balance the flow. We will try this tomorrow.

[Zauberberg]

Another thing worth checking is, calculating vapor fraction of the overhead product at 7 kPag and at the actual condenser outlet temperature. A common flowsheeter would be sufficient for such purpose, and that way you can estimate product losses. Then you will confirm if 68 degC is sufficient for condensing the product, and also you can calculate at what pressure the overhead product is fully condensed at the actual condenser outlet T.

If there is a margin in available reboiler duty, you could increase the tower pressure (do it gradually, in several stages) and achieve higher condensing duty since LMTD will increase as well. Based on my experience, tower operating pressure can be raised safely up to ~1 bar below design pressure, for low pressure columns. You probably have a system designed for 3.5barg? As Norm Lieberman suggests in his handbooks, tower pressure should not be considered as something "cast in stone" - particularly where ambient temperature or degradation of heat transfer equipment by aging cause reduced condenser performance. We have made similar changes in the operation of our SR Naphtha splitter long time ago, with excellent results.

[cjre2]

Sir Zauberberg thank you sir for this suggestion. Currently, we do not have any means to increase the pressure since the overhead receiver is floated to flare. We can implement in April the use of N2 push-pull system as a means to vary system pressure.

By the way, I was wondering if blow-through in condensers happen similar to steam blow through in reboilers wherein the reboiling capacity is lost as a result. I am just wondering if this is possible due to the fact that we have two different exchangers with 2 pass and 1 pass bundles. In addition, I also noticed that after closing the non-condensible vent on Exchanger B, the outlet temperature dropped. What could possibly explain it? Do NC vent needs to remain open all the time?

[Zauberberg]

The vent is there for maintenance purposes, usually it is kept open during steaming/steam-out condition, and also during startup (for a very short time) in order to vent accumulated air and avoid airlock/vapor pocketing which diminishes heat transfer in normal operation. For sure you must not keep it open all the time - it's a hazard.

If there is no option to vary tower pressure, than the only remaining option is to distribute flows properly, as suggested earlier in this thread. Relative flows need to be equivalent to relative duties of the exchangers. It would be a nice thing to setup a spreadsheet which calculates total overhead pressure drop (excluding the hand valve) at e.g. 40, 70, and 100% of design throughput, and based on that calculation to try to estimate valve position in order to get as good distribution as possible.

One good field indication of flow distribution is - condenser outlet temperatures. If the flows are ideally distributed these temperatures should be equal. The deviation of outlet temperature will show you how far you are from proper flow distribution, and whether something further can be done. You can also think in that direction for future: modifying the hand valve into HCV, and change its position based on the condenser outlet temperatures also shown in the DCS. It is relatively inexpensive solution.

Good luck,

[DB Shah]

"In addition, I also noticed that after closing the non-condensible vent on Exchanger B, the outlet temperature dropped. What could possibly explain it? Do NC vent needs to remain open all the time?"

NC valves are provided on total condensing exchangers to remove NC as & when required. To give you an eg-
In methanol refining column of Methanol plant, methanol vapour is total condensing. Hence ideally column top temperture and condensed methanol temperature at condensor outlet should be same(single component condensation). However methanol does have some dissolved CO2 in crude Methanol. This starts accumulating in the condnesor top. But as inerts starts to build up in the condensor the partial press of the condensible drops and condensing temperature also drops. We monitor the column top temperature and condnesor outlet tmeperature, (a temp diff alarm is given in DCS) At certain differential temperature operator opens the vent manualy for 5~10 minutes and discharges the inerts to the flare. Hence as condensor is now free of inert condensed methanol temperature also increases and equals column top temp. What you are observing is the same phenomenon when you close your NC.

In your case if you open the NC completely you are bound to have product loss. Your higher loss after taking B exchanger in line may be due to the reason -initially only vent of A was open, after installing B exchanger, vent of B is also kept open by you. Higher loss due to two vents open( I assume you installed hex B afterwards & initially were operating with A only.)

On the other hand if you closed the NC, inerts accumulates and have temperature drop as observed by you.

If this is the problem-
Note down temperature of A & B condensed fluid say X°C when NC are open and no inerts accumulate. (This temp will not be same as column top temp if you have multicomponents in vapour- as dew pt & bubble pt are different)

Now close/adjust the vents of A & B exchangers such that condensed fluid temp is just below X & remains steady.
Intermittently vent out the inerts as and when the condensed fluid/reflux drum temperature drops.

To summarize-
If exchangers are designed for given load and not performing -
Adjust hydraulic loads in both the exhangers as described earlier in my reply in this thread. I hope you have carried out hydraulic cal in any simulator like aspen/hysis/cc for proper distribution. Just a word of caution-when you vary the flow in simulator, vary the Dp of exchanger accordingly. Calculate exchanger DP in a good software like HTRI/HTFS and input the calculated DP in the simulator. I was lucky as aspen plus & aspen tasc (HTFS) have dynamic link. The DP of exchanger is calculated in TACS & corrected in Aspen during run time environment.

If losses are due to excessecive NC opening, adjust vents as described in this reply.
Hope this helps.

DB Shah

[cjre2]

DB Shah:

i will surely check this out. I will first open the vent of Exchanger A as this has been in service with vent closed. NC could have accumulated already that could explain the observed reduced condensing capacity. I have checked the hydraulics at the design load and found that the flow will be balanced if the butterfly valve is 50% open. At present, the temperatures from the exchangers are the same validating that the condensing duty are balanced as desired.

[cjre2]

I forgot to ask, is vapor blowthrough in condensers (similar to steam blowthrough in reboilers) happen if the outlet line size is too large? What is the guideline is sizing the outlet line of condensers? Appreciate if you can provide me design practice for this so I can check the configuration in our facility. Thank you.

[cjre2]

Hi DB Shah,

I tried opening and closing the vents of Exchanger A and B, but I did not see any changes in the outlet temperatures. I think there is no NC accumulating in these exchangers. Do you have an idea on vapor blow-through in condenser? Could this happen if the outlet line is improperly sized (larger)? Do you have any design practice to check if the outlet line is properly sized?

[Zauberberg]

The question is not clear. What do you mean by "properly sized"? There is always a continuous vapor phase between the column and the overhead receiver (reflux drum), and the liquid simply drains from condenser into the drum. If the condenser outlet line was to be greatly undersized, you would probably encounter liquid backup into the condenser, followed by slugging of liquid once when sufficient head/pressure is created upstream of the slug, and all this manifested by pressure cycling in the distillation column.

[cjre2]

Hi Sir,

Sorry if my question is vague. The purpose of my question is to validate whether the 8" outlet line from Exchanger B is adequate (not oversized) that is why I was looking for design guidelines/calculations that I can use to confirm.

Will a continuous vapor phase exist from the column and overhead receiver if the condenser is supposedly total condensing? Are you referring to continous vapor flow or the line is not completely liquid filled that is why there is a vapor phase in the line?

[DB Shah]

(When you reply you can delete the post to which you are replying to avoid the repetition of quote in your post)

Vapour blow through
If I understand your query correctly, you are worried that the 8" condensate outlet of B is over sized. In this scenario what you envisage is 8" line has condensate flow as well as vapour.
You are throttling the butter fly valve and getting good heat transfer in B. If there was a vapour blow through, condensate line will be hotter at B outlet. Also not like a reboiler, condenser is a sort of sucking equipment. (BTW is your line dipped in reflux or do you have a goose neck arrangement?) I do not see chance of vapour bypass through exchanger B to flare through reflux drum.

NC
You do not get instantaneous temp drop as NC takes time to accumulate. You had said earlier that you observed temperature drop after you close NC valve. This confirms the presence of Inerts. Only thing is do not keep NC open continuously, watch condensate temp, at a certain drop of temperature you can open NC intermittently to purge out Inerts. Close the valve again.

Secondly if you have no non condensable, where are you loosing the product from. Is it that pressure you maintain is very low - Zaubeberg mentioned it in his very first reply to validate your pressure indication v/s temp in any simulator, hope you have confirmed it.

Your root problem is product loss. I do not have much info regarding the components involved, operating pressure of column. I will suggest to one by one eliminate the cause of loss.

You can loose product from –
1. Vent valve on reflux drum
2. NC valve on exchanger A
3. NC valve on exchanger B
Any other _ _ _

Back calculate the hourly rate of product lost. You already indicated 150 to 230 bbls. Divide this flow by the period in which you observed the above loss.

Hence loss = 230 bbl/time *rho-liq= X kg/hr
X kg/hr * rho-vapour = M3/hr of vapour.

Confirm in a simulator the pressure drop in your vent headers. Your pressure at reflux drum is 7 kpag. With this upstream pressure try to validate if the vent line is sufficient to allow such flow. Calculate same with NC lines on exchanger A & B (Pressure = 7 + DP of exchanger).

Also try to establish thermal balance. When you get 230 bbl of liquid it means the vent line is capable to condense such rate of vapour. Confirm this with Q = 230 bbl * latent heat /time = Y kcal/hr (manage your units). Now try to see if your vent line can condense this much amount of vapour.
Q = U A Dt
A = area in m2 = pi .D. L (D & L are dia & length of vent line)
DT calculate considering avg ambient temp.
Ud for bare pipe - take a value of ~ 5 kcal/hr-m2°C
And verify that Q by ambient heat transfer is matching with Y kcal/hr

[cjre2]

By the way, how can I check if the fan efficiency or air flow is still as designed? I am thinking this may be contributing to the observed limitation in condensing capacity of Exchanger A and B. Appreciate your inputs... Thank you..

[cjre2]

We have done thorough inspection of the air fin coolers (Exchanger A and and we found significant fouling of the air side with dirt and dust. This could be the reason why we are not getting sufficient cooling in spite of the adequate surface area and motor power and balanced flow (same outlet temperatures). Appreciate if your guys can share the effective means of cleaning the air side of air fin coolers. Thank you.

[DB Shah]

Dear cjre2

Nice you identified the problem.

You mentioned in your first post "no significant air side fouling" and after couple of days you say that there is signigficant fouling on air side after "thorough inspection" SO what did you inspected first time? OR you never went to the field. Dear cjre2, not to be offensive, but without proper homework you misguide many of the members.
Next time you post a query do your homework "Thoroughly" to make sure you do not (knowingly or unknowningly) misguide the members. Many a time veterans like Art put in lot of efforts and time to give the best of their experience.

Again no offence intended.

[cjre2]

Sorry about that sir and do not worry I am not offended in any way. Initially I checked the fin/air side condition looking at the top tubes and found that the air side is clean. Yesterday when I was trying to check the actual fan motor installed ( i was suspecting the the actual motors were not as designed as I was running out of things to check >)) while using a flash light, I accidentally saw the underside of the tube rows and boom! I saw the fouling.

I would like to apologize for this and I also would like to express my gratitude and appreciation to those who have spent their precious time trying to help me.

[DB Shah]

All is well that ends well.
Hope you regain the condendser performance very soon.
Cheers.

[Art Montemayor]

cjre2:

One common method that I have used in the past to clean the air side of air-cooled tube bundles is to have the maintenance crew use a protective suit and secure plastic face shields while employing a long extended steam nozzle on the tube bundle surface. The steam jet does an excellent job of cleaning the outside of the finned tubes.

It is a tedious and neck aching job, so make sure your maintenance crew prepares appropriately and switches personnel frequently. I would use a 4-man crew as a minimum to do the job. A saturated steam pressure of 75 psig is usually enough for the cleaning. Your maintenance crew may already know of theis steam cleaning method.

I hope this experience helps you out.

P.S.:
I hope you take D.B. Shah's comments to heart. He is trying to really help you out by emphasizing that you have to be serious when you inspect equipment while troubleshooting. When we say "THOROUGH", we mean THOROUGH - and not just a quick "look-see". In order to eliminate possible causes, other people including yourself depend on an accurate and detailed inspection.