Vessel Nozzles Diameter and Location

[shvet2008]

Hi, forum

As a contractor we are in charge of detailed design of a HC vacuum tower revamp. Licensor specified to replace the bottom pump with a new one, but unfortunately pump was purchased with more capacity than required and during some cases we have capacity less than min stable flow. We can do nothing with this pump and shall apply spillback to pump suction - to tower bottom. And we can't make appreciable changes in tower design. We have a problem - from one hand we shall make a spillback, on other hand we shall not make big changes in tower design or we face problems with licensor. Can anyone helps us? Any worth sources will be appreciated.

General figures:
- liquid - gasoil
- pump suction flowrate (normal) - 90 m3/h
- pump suction flowrate (pump best efficiency point) - 196 m3/h
- pump min cont stable flow - 40 m3/h

http://files.engineering.com/getfil...-4f62-ab86-b067276ef506&file=tower_sketch.pdf

Q1: How can we put pumped liquid back to bottom in so way that:
- spillback doesn't produce many bubbles and these bubbles aren't carry by liquid flow to pump suction nozzle and cause a cavitation
- spillback isn't carry to liquid outlet nozzle and heated liquid isn't suck straight back to pump
We can't put a big distributor to tower bottom since it can interrupt bottom tray to work properly. Anyway licensor wouldn't be happy to see any changes but the less changes we make the better. The best it would be just another one new nozzle in the tower bottom.

Q2: (concerned not only this case)
Are there any recommendation/practices/rules etc. for vessel nozzle diameter? Any type of vessel:
- towers
- separators
- storage tanks/vessels
- buffers/depulsators/recievers etc.
Any type of fluid and nozzles:
- liquid/vapor inlet
- liquid/vapor outlet
- two phase inlet

Any advice or opinion would be appreciated - I understood in general forum members are experienced in refinery/chemical processing. But please note that potentially we already have a problem with licensor and any competent sources would be extremely worth in this case. Anything we could use in correspondence.

 

[georgeverghese]

This drawing doesnt show the tower bottoms reboiler return and / or the pump spillback return. A process sketch of the bottoms circuit will also help.

 

[shvet2008]

to georgeverghese

1. Tower has no reboiler - see sketch below
2. Pump spillback return is an issue. Can you suggest best connection for spillback? Tower feed line is not proper because of top tray flooding.
3. A process sketch see attached file

http://files.engineering.com/getfil...6d529685&file=vacuum_dryer_sketch_diagram.pdf

 

[georgeverghese]

Can we not use the steam out connection for the spillback from the product pumps? This nozzle can then be for both spillback and maintenance steamout service. Nozzle location for S1 seems to be well away (some 200mm) from pump suction exit vortex breaker top, so it should mix okay with new degassed gasoil from the last bottom tray.

By the way, what is this DN25 balance line for ?

 

[shvet2008]

to georgeverghese

Q: Can we not use the steam out connection for the spillback from the product pumps? This nozzle can then be for both spillback and maintenance steamout service. Nozzle location for S1 seems to be well away (some 200mm) from pump suction exit vortex breaker top, so it should mix okay with new degassed gasoil from the last bottom tray.
A: Yes we can and we thought the same way. Actually we are plannig to do the same way. The only difference is the velocity - as for me in steamout nozzle velocity is rather high approx 5.5 m/s. We are planning to make a new nozzle on the same elevation as steamout with DN100-150. We don't decide yet.

Q: By the way, what is this DN25 balance line for ?
A: For equalizing. See API 686
"3.1.2.10 For pumps taking suction from vacuum towers or columns, an equalizing line from the pump back to the vapor space in the tower or column shall be provided to vent the pump at start-up"

 

[LittleInch]

Ok,

For me this doesn't start well "We can do nothing with this pump ". Err yes you can, you can purchase a pump which matches the duty and doesn't need a "spill back". You're just choosing not to for some reason and are now trying to retrofit a design and deal with the symptom instead of fixing the root cause. Can't you add a VFD to the pump??

Your other potential option, though often not really liked, is to use the manhole as a large nozzle and modify the manhole flange to incorporate a nozzle in it. You usually need to provide good support to the pipe ( maybe spring hangers) but it doesn't involve a new nozzle per se.

Of course you then need to dismantle the pipe before getting access to the manway, but the tower would need to be drained and gas free before you enter so no big issue there.

If you're in charge of detail design then you need to give the client a set of options.

Start with a new pump and then list all the other options with issues such as the problems with the licensor.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

Okay, so it is a pump startup vent line, not some kind of pressure balance line.

At 5-6m/sec on this S1 nozzle running almost continously, there may be erosion-corrosion concerns. One alternative may be to take up the option @LittleInch suggests, which should be relatively easy. You could also use both S1 and / or some other nozzle higher up the column if available.

 

[shvet2008]

to LittleInch

Q1: For me this doesn't start well "We can do nothing with this pump ". Err yes you can, you can purchase a pump which matches the duty and doesn't need a "spill back".
A1: You are an optimist. As I sad the pumh has already bought. We are just a detailed engineering contractor and don't deal with equipment procurement. This pump is suitable for this service. Yes it has poor efficiency and from time to time work below MCSF but it is still suitable. All I have to do this is a spillback and a question is how I can do this best way with min problems with licensor.

Q2: You're just choosing not to for some reason and are now trying to retrofit a design and deal with the symptom instead of fixing the root cause.
A2: Yes, I do. Beacuse I can do nothing with a root case.

Q3: Can't you add a VFD to the pump??
A3: I can't. VFD is suitable if a pump has excessive head. In my case the pump has excessive capacity but not head.

Q4: Your other potential option, though often not really liked, is to use the manhole as a large nozzle and modify the manhole flange to incorporate a nozzle in it. You usually need to provide good support to the pipe ( maybe spring hangers) but it doesn't involve a new nozzle per se.
A4: Thanks a lot. We suggest this option to client.

to georgeverghese

Q5: Okay, so it is a pump startup vent line, not some kind of pressure balance line.
A5: No. It is a balance line. You need to fill a pump with liquid before start. This line is intended to balance pressure betwen a pump and suction vessel.

Q6: You could also use both S1 and / or some other nozzle higher up the column if available.
A6: There is no other spare nozzle in column. All available options are:
- manhole
- steamout - high velocity
- feed line - tray flooding
- pump suction line - liquid overheating
- a new nozzle
- a new pump - pump has already bought
As I mentioned before for some reasons all really available options are:
- manhole
- a new nozzle

 

[ashtree]

Maybe i missed something in the discussion about the type of pump being used but last time i checked the pump affinity rules apply to both flow and head.
So if excessive head is seen as your problem then fitting a VFD will solve both the flow and the head issue as Little Inch suggests.

Flow is proportional to speed.
Head is proportional to the square of the speed.

Somewhere there will be a speed that will give you the output that is desired. I think everybody would agree that a better matched pump would be ideal but as you suggest you have to work with the pump you are given.

Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[shvet2008]

to ashtree

Q1: So if excessive head is seen as your problem then fitting a VFD will solve both the flow and the head issue as Little Inch suggests.
A1: Repeat "A3: ... In my case the pump has excessive capacity but not head."

 

[ashtree]

Have you got a pump curve to upload?

Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[tickle]

Shvet

I suggest that as an engineer you need to at least identify what the correct solution is as per LittleInch's advice. The first step would be to discuss with the pump vendors. The impellor may be able to be changed, pump speed (VSD/VVVF), the pump displacement, or an orifice on the discharge. It should not really matter that someone else purchased the pump. I would also be investigating the two pump arrangements.

From a solution to the problem I would be putting the recycle line back into the vessel inlet, it is designed for it. Anything else may impact on operation.

 

[georgeverghese]

From your first post, the pump min flow requirement is 40m3/hr while the liquid downflow traffic through the column is 90m3/hr, so it should be possible to tie in this spillback line into the feedline downstream of the LCV.

Must say though that if the BEP flow is 200m3/hr, you would typically have a min flow of 0.33*200 = 65m3/hr for a fixed speed drive, which is a lot more than 40m3/hr, but even if this were to be the case, the spillback tien d/s of the LCV should still be possible ,since min flow < 90m3/hr.

Though this is not the topic of discussion, presume you havent shown the stripping gas feed into the column since this is irrelevant to the topic being discussed.

 

[LittleInch]

Are you sure the liquid will overheat if you go back into the pump suction? This option only works when there is some sort of flow through the system, but given the recycle flow should only be the difference between forward flow and your min stable flow, the amount of extra heat might be acceptable. The key is to decide when you're doing this re-circulation and how you control it.

Would be interesting to find out is a smaller impellor / higher speed combination actually changes the BEP or MCSF. A vendor question for sure.

appreciate you're working within boundaries, but to me the pump is marginal as to whether it's "suitable". At the least put the issue in front of the end client and let them decide. They will need to operate this thing for the next 20 years....

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[shvet2008]

Hi, guys. Sorry for delay. I was ill. Lets back to pump.

Q1: Have you got a pump curve to upload?
A1: See attached.

http://files.engineering.com/getfil...47-4646-9140-13e74c76bca8&file=pump_curve.pdf

Q2: From your first post, the pump min flow requirement is 40m3/hr while the liquid downflow traffic through the column is 90m3/hr, so it should be possible to tie in this spillback line into the feedline downstream of the LCV.
A2: We thought this way. This leads to possible top tray flooding during normal case if liquid can't be pumped to storage. This circumstances can appear for example if operator change FC set point in pump discharge. See attached sketch.

http://files.engineering.com/getfil...5e1a7c&file=pump_discahrge_sketch_diagram.pdf

Q3: Though this is not the topic of discussion, presume you havent shown the stripping gas feed into the column since this is irrelevant to the topic being discussed.
A3: There is no stripping gas. Gasoil vacuum drying does not require stripping gas. Stripping gas only reduces water partial pressure and therefore vacuum dryer efficiency.

Q4: Are you sure the liquid will overheat if you go back into the pump suction?
A4: Yes, of course. Efficiency is less 30% and spillback line is rather short. Pumped liquid heating shall be much more than heat loss in spillback pipe. This is why we consider spillback to suction vessel. Spillback straight to pump suction is possible but online cooler is required. Suction vessel is better choice than cooler.

Lets back to topic. All things we discussed are great and interesting but I asked for help with other. Any competent sourse (codes, practicies, articles etc.) how to back liquid to suction vessel in right way. Anything I could use in correspondence.

 

[georgeverghese]

If the product FC setpoint is reduced, level will rise in the column bottoms sump, and the LC will reduce fresh feed to the column to maintain level, so there wont be a liquid overload on the top tray ?

 

[shvet2008]

to LittleInch

Q1: appreciate you're working within boundaries, but to me the pump is marginal as to whether it's "suitable"
A1: What is the specific boundaries of your "suitable"? Could you please define exact figures when pump is suitable or not (links would be appreciated)? When there are no exact criteria it is some kind of person's authority struggle but not engineering.

to ashtree

Q2: So if excessive head is seen as your problem then fitting a VFD will solve both the flow and the head issue as Little Inch suggests.
A2: VFD can reduce MCSF? Can you share any information how VFD afects MCSF?

to georgeverghese

Q3: If the product FC setpoint is reduced, level will rise in the column bottoms sump, and the LC will reduce fresh feed to the column to maintain level, so there wont be a liquid overload on the top tray ?
A3: OK, lets imagine. Normal case. Vacuum tower (VT) feedrate is 90m3/h. Bottom flowrate is 90m3/h. FC set point is 90m3/h. Once operator change FC set point to 20 m3/h. What happens?
1. Spillback FC controls pump flowrate and bottom flowrate is 40m3/h. This stream is splitted and 20m3/h goes through coolers to unit's battery limit and further to storage. 20 m3/h goes back to VT feed and VT's top tray.
2. Yes LC will decrease VT feedrate up to 20m3/h but how quick? What is feedback of LC. Liquid has to fall down all 5 trays and DCS has to recognize changes and try to control it. All time LC delay top tray will be flooded and IT CAN'T LET PASS LIQUID THROUGH DOWNCOMERS. It means that liquid level in top tray will significantly increase and liquid will be sucked to vacuum system. Vacuum system can't handle large amount of liquid, the best consequence of top tray flooding will be separator overfilling.

If I would be an lisencor's engineering manager I would be in a rage. On one hand we have an usual nozzle in tower bottom. On other hand we have vacuum separators overfilling and possible vacuum system damage.

 

[ashtree]

In answer to your question about whether or not the VFD can reduce the MCSF , I will say yes usually.

From the pump curve you have given the MCSF is not specifically shown but i am going to assume it is probably close to the left boundary of the blue shaded area. This boundary is not vertical.
The pump curves when operating at reduced speed will look very much like the curves for the reduced impeller diameter and the left hand boundary of the prefferred operating range slopes back towards zero as the impeller speed reduces. This implies that the MCSF will also slope back towards zero as the speed reduces.

There will be some practical limits hence my early qualification.

Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[LittleInch]

Well clearly the pump is not suitable otherwise you wouldn't be needing to try all these options for a low flow case.

Suitable has many different meanings, not just numbers. I think that is engineering which is a constant balance between competing issues. Here though the balance is incorrect and all of your issues would be solved by getting a smaller pump able to do the design flow and head as well as the low flow case.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.