Pump design 4

[Ndimo]

On a slurry line I have a pump with the following design data :
Design @ duty : 10 000 gpm
BEP : 20 000 gpm
Actual flow max : 8 500 gpm
I'm experiencing frequent failures ( 150 mm shaft breakage)
My questions are :
- Should I blame the designer?
- Could I say that the failures result of a bad design?
- What should I do to return this pump near the BEP?

 

[BobPE]

Should you blame the engineer? Hard to say, there is a lot more information you need to research as to what the original design conditions were, what is going on now that may be different that during orininal design, etc.

Can you say failures are a result of bad engineering? Again, information is the key. I would say though that the failures you indicate can most likely be attributed to the pump running to the left of BEP.

What can you do to return the pump to BEP? An easy fix would be to install a high side by-pass return to the pump suction somwhere. The best fix, get a new pump as the one you have probably could not be refit to run at any sort of efficiency you could tolerate at those flows. The return on investment based on energy consumption would most likely make the new pump choice a good investment.

Did you have the system engineered by an engineer? If so, then asking those questions directly to that engineer should return you answers that you can work from.

BobPE

 

[sprintcar]

The FIRST thing to do is contact the pump manufacturer - supply them with your ACTUAL duty conditions: head, flow, SG, etc. to make sure the pump is correct for your application. For example, did the motor power or RPM change after the initial pump installation? What is the tonnage change from the quoted specification?

Meantime, check your system - do you have large solids (maybe broken screens) that jam the impeller against the casing? What about shut downs where the solids in the slurry can settle out in the pump shell causing extra load on the shaft at start up. Some slurries can pack just like cement if left long enough.

Where is the shaft breaking? If it's at the drive end, check your alignment or belt tension - cog belts tend to be misapplied in pumps and don't slip like v-belts. If it's breaking at the impeller end, is your suction plate/impeller gap correct or do you have parts rubbing?

Is the gland packing and flush water correct or is this contributing to excess load?

You mentioned Frequent Failures - that would indicate a problem that has existed for some time and should be brought to the attention of the pump maker.

Fix the PROBLEM, not the blame.

Keep the wheels on the ground
Bob

 

[Ndimo]

Thanks to all for your good inputs.
What I did today, is to determine a dead head and a running head. It comes out to be 25 ft lower than expected and the running head is quit similar to the dead head. I'm sure that the impeller size and the speed are as design (39", 590rpm). The pressure differentiel is 90 Psia.
On the other hand, looking at the suction plate, I found out that it is rubbed.
I will continue to get the system curve and then talk with GIW the pump maker.
Regards

 

[Artisi]

My first question would be - why is the duty point on such a large pump 50% of BEP - made worse by your actual duty of 8500 gpm.
Like BobPE has said - blame the engineer not the pump -

Naresuan University
Phitsanulok
Thailand

 

[Artisi]

Where is the shaft broken?
I can understand you seeing rubbing of the impeller on the wear plate probably a result of shaft deflection just before failure.



Naresuan University
Phitsanulok
Thailand

 

[Ndimo]

This pump has a very low NPSH required, I think that is the reason they choose it, dispite the fact it's running far from the BEP.
The shaft is broken at the impeller side.
What means, the dead head is far lower than the supplier curve?

 

[checman]

boubacar, Is there a reason for not bypassing some flow back to the supply and put the pump operating condition on a better part of the curve? In some cases the agitation can help keep solids in suspension and be of some benefit to the process.

The low NPSHR of the impeller may result in a high sss number which would tell you that you might need to run at 50% BEP as a minimum continuos flow to avoid low flow cavitation and its damaging vibration.

In many cases I have found that shaft breakage at the impeller is from sudden load changes such as a fast opening valve or slugs of product going through the pump. In slurries I have seen the discharge line partially plug then blow open causing a surge in the discharge line and the pump. I agree with sprintcar you need to determine why the shaft is failing.

Regards checman

 

[Artisi]

For a problem of this nature you should be doing a number of things at the moment and not applying "shot-gun" therapy to the problem - forget the dead-heads, the speed, the NPSHr/a etc - this is all meaning-less at this point, what you need is a systematic study to get to the root cause. Most of the GIW pump range are built like battle-ships to account for all the upset conditions encountered with slurry pumping - your application being no different to hundreds of others out there in the real world.

Therefore:

1. Get a full hydraulic study undertaken on the application - what are you trying to achieve and what is actually being achieved.
2. Contact GIW with the problem together with the hydraulic study.
3. Get a qualified analysis carried out on the shaft failure - this will tell you the mode of failure - from here you maybe able to work backwards to establish what operating conditions can cause this type of failure.

Concentrate on the cause not the result.



Naresuan University
Phitsanulok
Thailand

 

[Artisi]

For those fortunate enough to have fortune-telling skills and who can predict or all ready know that the shaft failure was either torsional failure, a bending failure or a manufacturing fault etc etc. without looking into it are certainly lucky. This allows them the benefit being able of pin-pointing the problem with certainty and offer immediate solutions.
I only wish it was that easy in the real world away from the desk or the text books.

Naresuan University
Phitsanulok
Thailand

 

[Artisi]

are you using OEM pump shafts or pirated parts?

Naresuan University
Phitsanulok
Thailand

 

[sprintcar]

boubacar - GIW has a new tech service line - 1-888-techGIW (832-4449) 9am - 4pm eastern time. Call them with your pump serial number and I'm sure they can help you

Keep the wheels on the ground
Bob

 

[Ndimo]

Thanks at all again.
We are talking with GIW at this moment.
On the other hand, we are not using the OEM pump shaft, but the part is not pirated as our supply is not in Thailand but a very well know, sitting in the USA.

 

[sprintcar]

FYI -
Any 'non-oem' part is generally termed 'pirated' in most industries...'aftermarket' is the automotive term.

You need to get the actual specs on the material from your vendor so the pump maker knows what to look for.

Most highly stressed parts (such as pump shafts) will be designed with a specific material and an engineering factor of safety. Once you change material or process, you have deviated from the original design and the resulting failures are self inflicted...

....and you wanted to blame the Designer!!

Keep the wheels on the ground
Bob

 

[Ndimo]

We are using those parts since 2001. Furthermore we have 4 pumps using them, but we have faillure only on one of them ( 2 shafts broken in 3 weeks). This later has commun suction and cummun discharge with another one, which experienced no problem.
How can I blame the part?

 

[Ndimo]

Please find below some facts listed after the shaft faillure.

1. Radial (Impeller end) cover broken
2. Bearings found to be in satisfactory condition
3. Stuffing box bolts worked themselves loose. Some of the bolts were found loose.
4. Shaft broken outside of RA.
5. Broken impeller plate.
6. The pump ran continuously between the 19th and 26th January 2005
7. The pump was stopped on the 26th Jan 2005 for 1 hour, between 9:18PM – 10:12 PM. Approximately 54 minutes.
8. The pump then ran from the 26th Jan 2005 at 10:12 PM to failure at 4:09 AM on the 30th January 2005.
9. The shaft currently used is provided by Quality Machine and is made of 4140 Steel. These shafts have been used since 2001.
10. The BOT level gradually increased from 32% - 89% up to the point of failure at or around 4:09AM. This occurred over a period of 8 minutes.
11. The current draw increased over a period of 9 minutes from 235A – 393A
12. Suction plate scored.
13. Impeller face badly scored.
14. The pump operates to the far left of the BEP point, which contributes to excessive shaft deflection, and may affect the alignment of the impeller in the volute chamber.

 

[Artisi]

The latest information is what we have been looking for -

" 11.The current draw increased over a period of 9 minutes from 235A – 393A
12. Suction plate scored.
13. Impeller face badly scored.
14. The pump operates to the far left of the BEP point, which contributes to excessive shaft deflection, and may affect the alignment of the impeller in the volute chamber. "

It certainly looks like the shaft has failed from excessive radial loading by operating so far left of BEP.
I would suggest that what you consider to be "normal" operation is marginal and the shaft is being fatigued from excessive deflection during operation, the 9 minutes of excessive current draw is the final stages of the shaft failure as the impeller rubs on the casing.

A close metalurgical examination should tell you exactly what has happened but I wouldn't be suprised to see a typical fatigue failure starting at one point and increasing over time with a final torsional break as power exceed the remaining shaft strength.

Next step is to establish why the pump is being forced to operate under such duress and how to rectify the problem.



Naresuan University
Phitsanulok
Thailand

 

[checman]

boubacar, I find it strange that the pump is running and drawing more power the last 8 minutes but not pumping. Even if the impeller is deflecting into the suction plate it should still pump something if it is turning close to rated RMP. Is the wear on the suction plate mostly towards the top or is it some what even? If some what even it may suggest the pump is running with no flow either due to a plug or vortex (trapped air) at the suction or a plugged discharge. Causing the fluid trapped in the casing to heat up in the pump which heats up the pump shaft and impeller until they expand into the suction plate. There are many ways to wreck a pump not knowing all the specific information about your system one can only make guesses at where the problem with your system is. Is it the pump ? Is the problem a severely unbalanced impeller? Is it something in the system? Is it your operating condition? It maybe any one or any combination of the aforementioned or maybe something that has not yet come to light.

Regards Checman

 

[sprintcar]

boubacar
Is it possible that since you are running 2 pumps in parallel from the same suction point that you are reducing the NPSH available (for the subject pump) below what is required, thus causing severe cavitation and excess shaft loads? If subject pump is being supplied thru an elbow off the pipe to the other pump, this too can restrict suction.

If you have 4 similar pumps running and only one has recently become a problem, this is not a design flaw but rather something that has changed in your flow system or physical mounting of the pump itself. When was the last alignment check? Since the shaft has been replaced recently, it is possible that a different bearing housing (barrel) was installed and it has a different center height or alignment feature that was not realigned? If the housings were not tightened down properly, could the impeller have pulled it forward and caused the suction plate contact?

What material is the OEM shaft made from compared to what you have now? Where is the failure point - at the plug thread (small diameter) or next to the bearing on the large diameter?

Inquiring minds want to know...

Keep the wheels on the ground
Bob

 

[abeltio]

what is the L^3/D^4 ratio?

http://www.mcnallyinstitute.com/11-html/11-06.html

for metric use <2.0

what is the suction specific speed?

http://www.mcnallyinstitute.com/09-html/09-12.html

check these industry accepted parameters to see where your pump is operating.

saludos.
a.