NPSH Margin in Vertical Turbine Pumps

[dhlinva]

I'm doing a design review for a seawater system using vertical turbine pumps in located in sumps within tanks. The system is to be designed to pump the capacity of the tanks in 9 hours. To do that requires 5000 gpm to be pumped. The BEP of the selected pumps is 3800 to 4000 gpm, but the builder determined early on (before I became involved) that it was possible to operate these pumps at the end point of the curve until we identified possible NPSH constraints as the tank levels decreased. Their solution was to install VFD's, but slowing down the pumps blew the time constraints. The builder then proposed increasing the maximum speed of the pumps 200 rpm to gain a minor amount of additional capacity on the high end.

In order to run the pumps at maximum capacity (to meet the 9 hour constraint) the NPSH margins being used in their simulation are almost non-existent - in the neighborhood of 1 to 3 %. ANSI/HI 9.6.1-1998 states that "NPSH margins are not normally a consideration for most standard vertical turbine pumps, since they generally have a Low Suction Energy and cavitation noise is not an issue." I calculate a suction specific speed of 12,800, but to tell the truth, based on what I've read, I'm not clear on whether this makes it high energy or low.

I would prefer a 5000 gpm pump operating at BEP. Then we wouldn't be looking at NPSH. The situation is complicated by fact that the pumps have been purchased and delivered. I would like an NPSH margin ratio of 1.2 to 1.3 and minimum of 5' (based on Hydraulic Institute recommendations), but I will admit that this amounts to a factor of safety because I'm not sure of what I'm dealing with when operating so far outside "normal" parameters. I'm not sure sure how to intelligently determine what NPSH margin is required without expensive, detailed analyses that might be required only due to a poor selection of pumps. I can't seem to get the attention of the builder with my technical concerns.

My concern is obviously for the longevity of the pumps. I'm not sure what indications of cavitation will be available to the operators with these pumps when located 55' down in a tank (other than loss of discharge pressure). Unless something similar to Goulds PumpSmart® in installed, I think the pumps will be routinely operated in cavitation.

There may be many things I've gotten wrong here. NPSH is not something I generally need to get involved with and have little experience with the finer details. I'd appreciate any input you guys might be able to provide. Thank you.

 

[BigInch]

Can't hardly think about allowing or even having any NPSH problems with a vertical pump in cool water service, especially at those low heads anyway.

The higher the specific speed, the more NPSH is usually required. You don't mention the rpm now, but an increase to 200 rpm isn't very high, but depends more on the percentage over the original design speed.

I would suggest, if changing the pumps is not practical, that you approach the problem by evaluating the validity of the time constraints and increase the allowed time if possible. Also the time that these pumps will be running at marginal NPSH might not be very long, so see if those ideas might give you some reasonable alternatives to rejecting the existing pumps.

http://virtualpipeline.spaces.msn.com

"We can't solve problems by using the same kind of thinking we used when we created them." -Albert Einstein

 

[hbw]

Buy a 5000 GPM operating at BEP without the VFD. Provide enough submergence to satisfy the Manufacturers NPSH and vortexing requirements. Arrange the sump to have nearly lamimar flow to the pumps. See Goulds Pump Manual or this webpage:

http://www.gouldspumps.com/pag_0007.html

 

[dhlinva]

The time constraints are fixed. 1770 rpm is the normal rating - 1950 is the increased speed. NPSH is a problem only because of trying to operate at the endpoint of the curve. We're at 150% of BEP with these pumps. With 55' of head we don't have problems initially, but as the level drops the NPSH available drops significantly. At BEP (4000 gpm) the NPSHr is 18'. At 5500 gpm, it's 44'. This requires the slowling of the pump. Obviously, a pump with a BEP at 5000 gpm is what we should have. No VFD is then required. That's what I would prefer and would hope I would have been insistent upon if I had been involved at an earlier stage. The pumps have been purchased and the builder wants to use them.

I was looking for advice on NPSH margins using the existing pumps. Solid NPSH requirements will preclude the use of these pumps. They can't use them and meet the 9 hour requirement unless they operate at no margin. I feel a margin is necessary, particularly given the unknowns. I know they're the wrong pumps. I'm looking for unshakeable arguments to get rid of them when getting rid of them is going to cost the other party big dollars. If anybody has input regarding necessary NPSH margin, I'd appreciate it.

 

[BigInch]

Was there a margin mentioned in the purchasing specification for that pump? I doubt it. If so, get the mfgr to state MINIMUM NPSHR and the allowable margin, if any. In the end, if its not operated according to the mfgr's recommendations, all warranties will be voided regardless of wether its operated within some NPSHR margin mentioned by HI practice or not. It must be operated within the ranges permitted by the mfgr and that's that. Its also your stick; hit them with it.

http://virtualpipeline.spaces.msn.com

"We can't solve problems by using the same kind of thinking we used when we created them." -Albert Einstein

 

[Artisi]

Trying to run the pumps and the extreme end of the curve is looking for BIG problems even under the best hydraulic conditions and will be further exacerbated with an increase of speed - although the curves are extended out to this point operating there is doubtful.

What has Flowserve had to say about operating at end of curve, at the end of the day they are the ones to consult with for clarification.

 

[dhlinva]

Further exacerbated by the increase in speed and that's just when the pumps are brand new. Wait until they start to get beat up.

I agree that it might not be clear exactly where the pumps will actually operate. Our SimSmart model indicates they will operate somewhere close to the end points. There's not a lot of backpressure.

The Flowserve rep is the one who initially recommended and is supporting their continued use under these conditions. My feeling is he probably didn't have a 5000 gpm pump so he sold what he had. This isn't the only instance of an inappropriate pump, just the worst one. There's no way this pump should have been considered.

I'm usually pretty good at finding my way past the sales reps to get in touch with technical people, but I haven't had any luck getting contact information for anybody other than sales reps at Flowserve. I don't know how to get past the guy who is representing Flowserve to us. That rep has no problems with the intended operating points. He's the one who suggested increasing the speed of the pumps and said we could operate right up to the NPSHr curve before any cavitation occured. That's just patently false. I don't know what his qualifications are, but how do you question someone's credentials and wind up with any cooperation at all? I'm just at my wit's end. I don't know how to decisively shut this subject down. Everything I've read points to - like you said - BIG problems.

 

[hbw]

Here are a few Flowserve contacts:

Gary Bakke, Customer Service Engineer
Phillipsburg, NJ
Tel: 908-859-7049
Email: gbakke@flowserve.com

Brook Campbell, Application Engineer
Houston, TX 77008
Tel: 713-803-4438
Email:

 

[BigInch]

Best not to blame the pump company. They are not responsible for acceptance or the application; only that their pump runs in accordance with their published curve.

http://virtualpipeline.spaces.msn.com

"We can't solve problems by using the same kind of thinking we used when we created them." -Albert Einstein

 

[stanier]

http://turbolab.tamu.edu/pubs/Pump21/P21pg018.pdf

is worth a read.

Another reference is Cavitation and the Centrifugal Pump Aguide for Pump USers. Another resource is

http://www.mcnallyinstitute.com

 

[Artisi]

Did the manufacturer / representative guarantee an operating point in accordance with any spec issued for the project, if they did, bear-in-mind that this is the only point at which they are responsible for performance and not some other arbitrary point chosen after the event.

I agree that you need to get to someone other than a sales rep. so that you can get this sorted out.

 

[dhlinva]

The system is built to a performance spec. 9 hours to pump out the volume of the tanks which by extension, requires 5000 gpm. There was no spec language regarding pump selection criteria although the system design must be technically supportable. It must function properly.

I don't think I'm blaming the pump company rep for anything other than recommending - what I consider - the wrong pump to a builder that didn't know better. I don't think a person well-versed in the technical issues would consider this a good fit for the application. The rep has been active in supporting the continued use of these pumps. Who then is responsible? The builder ultimately, but they are being supported by the person who recommended and sold them the pumps. He's not suggesting that this is a poor application. Not even close.

All I'd really like is to know what NPSH margins Flowserve might think these pumps need to operate successfully over time. I don't think I'm getting qualified information from the only company person I have available to me. If they're not the proper source of information for their product, who is? Seriously, I don't know who to ask.

If the builder can technically support the operation of the pumps as presently intended, I'll listen. At this point, I haven't received any real technical support for this operation other than the fact that the only contact the builder has at Flowserve is saying that they can operate with almost no margin and that it won't be a problem.

With enough NPSH margin, I tend to agree. So the question to me comes down to a difference of opinion as to how much margin is required. The face of Flowserve says operation up to the NPSHr value is acceptable. I'm certain that's not correct, but I don't know exactly what a good number is. The question in my mind is still - how to determine a reasonable NPSH margin for the operation of these vertical turbine pumps. With enough margin, these pumps won't make 9 hours, but they will operate successfully. That's the whole point.

 

[gepman]

You are in a bind because if you demand that the builder change the pumps before they do not work properly you (the owner) will have to pay the costs. If you wait until after the system is installed and fails then you (the owner) will have to pay at least the incidental costs (lost production and other damages) unless you have a protracted legal case.

I have been in the situation (unfortunately) several times as a project engineer for large food processing companies. I would:

Ask Flowserve to guarantee the performance and operation over the range of operating conditions of the pumps. If they won't do that I would put the builder on notice that the system is not properly designed. You then have the option of paying for the changes yourself (if you feel unplanned failure will cost more than replacing it now) and suing the builder for breach of contract OR waiting until they fail.

You can also hire an independent engineer (perhaps one with the joint approval of the builder) to examine the system and determine if it has been properly designed. You can then proceed based upon his recommendation.

Either of the above will help you in case of a lawsuit later. In the food processing industry we really could not tolerate unplanned (major) downtime and there were several times when we changed the design. We would give the contractor additional money for the incremental cost of the better equipment but nothing for returns or demolition of work that was already done. Usually the contractor and the supplier worked things out so that there was as little monetary loss as possible since both wanted to continue working at the facility.

 

[BigInch]

You are seeming to regress to my original idea, extend the time. Even though you say the design pumpout time is fixed, your statement "With enough margin, these pumps won't make 9 hours, but they will operate successfully. That's the whole point." says the time isn't so fixed. Revise the time limit and everybody's off the hook. Apparently you must effectively do that anyway in order to make it possible to accept the pumps at all. Just change the paperwork to match and its all over.

http://virtualpipeline.spaces.msn.com

"We can't solve problems by using the same kind of thinking we used when we created them." -Albert Einstein

 

[hbw]

Flowserve is right. A pump can operate with no margin, where NPSH available is equal to NPSH required. There will be some cavitation at this condition as NPSH required is determined at a 3% drop in head which is caused by inception of cavitation.

Following are good centrifugal pump selection criteria:

NPSH margin = 4 feet minimum
Suction Specific Speed, Nss = 11,000 maximum
Rated Capacity = 80% to 110% BEP
Curve Stability = continuously rising to shutoff
Head Rise to Shutoff = 10% minimum t0 20% maximum
Installed / Max Impeller Diameter = 95% Maximum
Maximum HP at runout= 95% Motor Nameplate rating

Nss of your pump is 12800 which indicates that it could have maintenance (bearing, seal life)issues in the future due to internal recirculation at the impeller eye

 

[dhlinva]

The whole point is that the system operate well over the long haul. With enough margin the pumps will operate successfully, but won't make 9 hours. This doesn't met the performance spec and will necessitate their replacement. If the builder can come up with sufficient technical justification for their use with no margin, I'll have to pay attention. I don't think it exists.

Extending the time is the biggest no brainer available. If that wasn't the first solution considered, I'd have to be an idiot.

 

[dhlinva]

I'm working on contacts within Flowserve - the Houston number appears to be in the right direction. Thank you

http://turbolab.tamu.edu/pubs/Pump21/P21pg018.pdf

is a good read. I'd found that before and was one of the indicators that we might have a problem.

What I've read indicates that there is a problem with recirculation at the impeller eye only with flows below that of BEP - in the neighborhood of 50%. An interesting issue, but another subject I think.

I understand that operation at zero margin is not always a problem and doesn't have to be a problem, but we're not operating at BEP or even 110% of BEP. We're at 150% of BEP. My problem is that I don't know and don't have the experience to operate outside normal operating parameters. The greater the deviation from established guidelines, the more attention has to be paid to the details. It doesn't mean you can't go there - you just have to know what you're doing. If we were at BEP or close to it, I wouldn't even have raised the issue.

I appreciate the input and the support, guys. Thanks a lot.

 

[hbw]

One more item.

NPSH calculation are made with absolute pressure not gage pressure. Atmospheric pressure is part of NPSHa. A vertical wet pit sump pump has 14.7 PSIA, 34 feet water column NPSHA to begin with from atmospheric pressure acting on the water surface. To the 34 feet, you add the static head from the water surface to the eye of the first impeller. If you have 4 feet of submergence, NPSHa is 34' + 4' = 38'

NPSHr can not exceed NPSHa. If it does, the pumping system will be a failure resulting from poor design.

Ask Flowserve for a performance curve drawn at the VFD speed they propose. Require the 12 month warrantee.

 

[RossABQ]

If I'm understanding your situation correctly, that you need to pump down all the way on these tanks, I would use the verticals to bring level down quickly and use a submersible (or two) to get the last couple feet at least. Do you have room in the sump to drop one in? These are much more tolerant of low NPSHA, and would allow you to drop the RPM on the big boys without penalty.

 

[Artisi]

Let's face some facts here, these are the wrong pumps for this duty, they are undersized for the application, low on efficiency and operating far to right of BEP, they have insufficient NSPHa margin and will end up giving you problems from the day you start operations until the day they are replaced. The only way out is to reduce the demand which has already been pointed out by others.

Wishful thinking or looking for someone to say "yes you have enough NPSH margin" will not overcome the poor pump selection.

As you are responsible for a design review and not getting any response from your client to your concerns re the poor pump selection, prepare your report based on your concerns of poor selection and the long term viabilty of the pump units, the assurance of the Flowserve rep that the pumps will operate as required and your job is complete, present it and your job is completed. You cannot be responsible for other peoples assurances / mistakes or your clients inabilty to be concerned with your findings.