Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations IDS on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Option to resolve acceleration head loss

Status
Not open for further replies.
aak7

aak7

Petroleum
Apr 15, 2012
6
Hi,
I am new here and need some guidance.

A new pump skid needs to be fabricated and fitted in a ready location at my gas platform. From what I heard, the old pump never worked - did not know a lot of details. Did a calculation to determine if NPSHa > NPSHr so there won't be repeat incidents. (suspect pump didn't work because of cavitation).

This pump skid is for well equalization purpose. Liquid service is diesel, from a 2000L tank, routed through a long piping to a pump (2 decks down), to be pumped to individual well to equalize the SCSSV.

I considered pressure of the tank (which is atmospheric), static head (fluid elevation from tank to pump suction), friction loss and acceleration head. from calculation, if taking into account the acceleration head, NPSHa will be < NPSHr, which mean that heavy cavitation occurs.

Acceleration head, Ha, formula I used:

Ha = LVNC/gk, where

L=actual suction pipe length (metres)
V=mean flow velocity in suction line (m/s)
N=pump speed (rpm)
C=pump constant factor. I used C=0.016
g=9.81 m/s2
k=liquid factor (a constant)

I want to explore ways to resolve the acceleration head loss issue, to ensure NPSHa > NPSHr. Options:

1. Reduce actual suction pipe length, L by reroute whole piping to make it shorter/more straightforward between pump suction and tank. But this will be a tremendous work, possibility of welding (which means potential of production downtime)

2. Increase suction pipe diameter to reduce suction piping flow velocity, V but maintain piping routing configuration. More reasonable option but need major piping work, which may cause more time

3. Install a small tank (100L) near the pump suction. But need to see if space constraint is an issue.

I am more inclined to Option #3. However, my question is, by putting in that small tank as intermediate tank, will that negate the actual suction pipe length from big 2000L tank to pump suction, and L is now considered to be length between 100L to pump suction instead?

In short, if original L = 25 metres, but by installing the intermediate tank, L= 2 metres? Thus reducing the acceleration head loss?

Need guidance. Thanks so much!
 
Replies continue below

Recommended for you

Sort by date Sort by votes
Where does this formula come from? The acceleration head is always (in my experience) calculated by

Ha = v2/2g

where Ha = acceleration head in metre of flowing liquid
v = velocity in m/s
g = gravity in m/s2

Generally the acceleration head can be disregarded in NPSH calculations because it is recovered in the pump.

Whatever you decide to do you are shooting in the dark if you do not find out why the original pump did not perform to design.

Katmar Software - Engineering & Risk Analysis Software

"An undefined problem has an infinite number of solutions"
 
Upvote 0 Downvote
Is the term acceleration head loss is valid for the NPSHa calculation? I remembered my process engineering course about calculating NPSHa which only considered static elevation, pressure of suction vessel and friction loss. No acceleration head.

I was referring to the website that ione was talking about, as well as the research paper here, dated 2008. Please refer page 54:

 
Upvote 0 Downvote
I don't have understood which kind of pump are you talking about. For positive displacement pumps, the acceleration head plays a role in NPSHa calculations.
 
Upvote 0 Downvote
Yes this is for positive displacement pump. What i meant, does the acceleration head play a role in NPSHa calculation for a positive displacement pump? In which you answered yes.

Are you able to give opinion on the option #3 I asked about?
 
Upvote 0 Downvote
Sorry - I assumed this was a centrifugal pump. If it is a positive displacement pump then yes you must take acceleration into account. In a centrifugal pump you don't have to worry about the v2/2g acceleration that I was talking about because that is recovered in the pump, but with a reciprocating pump the line contents are re-accelerated on every stroke of the pump, so you must include it in the NPSH calculation as pointed out in the paper you referenced.

I have used dampers on the delivery side to smooth flow from reciprocating pumps for process control reasons, but never on the suction side. However, I think having one on the suction side (your option #3) would be a good choice for you because your suction line is relatively long. The friction loss would have to be calculated over the entire suction line length but the acceleration losses would be reduced as per the recommendations in your paper.

Katmar Software - Engineering & Risk Analysis Software

"An undefined problem has an infinite number of solutions"
 
Upvote 0 Downvote
The use of a dampener fitted in the suction side will attenuate the head losses due to mass acceleration and deceleration and will improve the NPSHa.
 
Upvote 0 Downvote
But if you've had problems before, go ahead and provide the full head using the LVNC equation. If it later on it still doesn't work, be sure you've left some space to pop in a dampener.

What would you be doing, if you knew that you could not fail?
 
Upvote 0 Downvote
Hi ione,
Using formula Ha = LVNC/gk, which expression will be affected if dampener is used? Appreciate your comment.
 
Upvote 0 Downvote
Hi BigInch,
What do you mean by 'provide the full head using the LVNC equation'? Does it mean, reduce the acceleration head loss by reducing L (Option 1 and 3) or V (Option 2)?
 
Upvote 0 Downvote
Try to provide the head required by the equation. Do not immediately resort to providing any lower head than required by that equation, unless you have no choice. If you still have a problem with it not working, then you can try adding the dampener.
Best first solution in this case is to try increasing the available head, probably by making the suction pipe diameter larger, or ... shorter.

What would you be doing, if you knew that you could not fail?
 
Upvote 0 Downvote
not "or", larger diameter AND shorter.

What would you be doing, if you knew that you could not fail?
 
Upvote 0 Downvote
I would try a suction side pulsation dampener in your situation.

Regards,

SNORGY.
 
Upvote 0 Downvote
The formula to calculate the acceleration head is ha = LVNC/gk as you already wrote. You chose 0.016 for k. That is a low value. What kind of positive displacement pump do you have?

To size a suction side flow stabilizer you need different formulas. With a properly sized suction flow stabilizer you can practically eliminate acceleration head completely.
 
Upvote 0 Downvote
The other thing you might try is to see if you have room in the pump itself to re-cylinder and go to larger diameter plungers. If you are limited on motor power you might be able to reduce operating speed (is the motor inverter-duty rated?) to compensate. Slower pump rpm will help towards reducing cavitation damage.

Regards,

SNORGY.
 
Upvote 0 Downvote
The attached article by Ione nails this topic. However: if I was in your situation, I would:

1) Contact a GOOD pulsation dampener company that REALLY knows what they are talking about. 95% of them do not truly understand what they are doing. A good company is Performance Pulsation. Their owner (John) is a pulsation wiz. John recommends a "maintenance free" flow-through style dampener. He also claims that other technologies may dampen pulsations on the suction, but they will not assist with regaining lost NPSHa. I never asked him to elaborate on why, so I could not tell you. NOTE: The article Ione attached disagrees with this concept.

Dampener to be mounted as close as possible to pump suction.

2) Lighten the spring rate on the suction side of the pump. The pressure difference between the suction and the cylinder pressure is what actuates the valve. There must be a sufficient pressure difference to open it.

3) Of course, run the pump slower as a result of #2, but also to assist with NPSHa issues.

4) It is interesting to notes that all fluids are slightly compressible. Therefore, if you have a 100[ft] of suction line then the equation you use should not account for all 100[ft]. It has been awhile since I have done this, but I seem to think that it is uncommon to see pulsation 30+ [ft] ahead of the pump suction.
 
Upvote 0 Downvote
hi micalbrch,
Sorry, I was mistaken. The k should not be 0.016. I have modified my k to be 0.2 (equiv to duplex single acting pump type).

0.016 was used by our vendor, which is using another formula to calculate acceleration head loss: Ha = 0.016LQN/d^2. This was suggested by Milton Roy pump but I have never seen this formula used other places in other papers or internet.

What was really strange is that both formulas, a) Ha = LVNC/gk and (b) Ha=0.016LQN/d^2 , gave vastly different results.

Using equation a) will give NPSHa > NPSHr, ie no cavitation. But b) will result in NPSHa < NPSHr , ie heavy cavitation will be exhibited.

I read in some papers that equation a) will begin to lose validity if length of suction piping >50ft. (In my case, the suction piping is 82ft). But how valid is eq b)? In the end we had a discussion and concluded that there will be no cavitation. But I am still a bit weary.
 
Upvote 0 Downvote
You should work with formula a) because b) seems to be a formula for a certain pump typ (triplex?). Formula a) works with the velocity while b) works with flow and pipe diameter. That seems to be o.k. as one can be expressed by the other. But 0.016 considers pi plus a certain pump factor (C) which you do not know. So formula a) must be yours.

0.2 is o.k. for your pump type. What acceleration loss do you get when using formula a)?
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

SPKR
  • Question
Friction loss for Flanged Concentric reducer
Replies
1
Views
77
TenPenny
TenPenny
ER_Azza
  • Locked
  • Question
Dead head pressure when the density is not water
Replies
11
Views
809
georgeverghese
georgeverghese
jkate
  • Locked
  • Question
Pump Head Calculation - 1 Pump pumping to 2 locations
Replies
13
Views
651
Snickster
Snickster
vij36
  • Locked
  • Question
Circulation pump head calculation in a pressurized closed loop
Replies
2
Views
364
Geo Lu
Geo Lu
RegiSTR
  • Locked
  • Question
Pump head in closed system 1
Replies
7
Views
604
RegiSTR
RegiSTR

Part and Inventory Search

Sponsor

Back
Top Bottom