Sundyne Model LMV-341 Question Revisited

[JJPellin]

I would like to revisit a question I asked a long time ago. I have still not been able to resolve the problem. The original post was this:

thread407-188371

I will briefly repeat my summary. We have two identical Sundyne pumps, model LMV 341. One of them has had three failures where the inducer came loose and water leaked under the sleeve directly into the gearbox. We also have a problem that may be totally unrelated. Both pumps are experiencing rapid cavitation damage despite the fact that they are running at exactly BEP and have an NPSH margin of more than 80 feet. Since my last reply to my original post, the spare pump has failed. That pump had even more extreme cavitation damage in the impeller eye than we had seen in the main pump. We have not had any more problems with loose inducers or water leakage into the gearbox.

I am now focused more on the cause of the cavitation damage. Two points have occurred to me that I would like to explore. First, anyone familiar with Sundyne pump knows that most models have the suction and discharge flange in line with one another and centered on the impeller center line. There are only two models that deviate from this pattern. In the direct drive (no gearbox) models, Model LMV-806 has the suction and discharge flanges offset to one side of the centerline of the impeller. Since they are direct drive, they are limited in speed to 3600 rpm (60 Hz) or 3000 rpm (50 Hz). But the model of our new pumps has this same feature. A colleague suggested that pre-rotation into the inducer could be the cause of the cavitation. When I began to analyze the installation, I noted the flange orientation and a piping issue. The flanges on this case are oriented such that the incoming flow would tend to pre-rotate in a counter-clockwise direction because of the offset of the flanges. The impeller rotates counter-clockwise. This is the same issue as with the LMV-806, but in the LMV-341, there is the possibility of shaft speeds as high as 22,000 rpm. Our pumps run at about 19,000 rpm.

The second issue was piping arrangement. Our suction line has a horizontal elbow about 60 inches before the suction flange. But for a 4 inch line, this is allowed by our specifications. It may be coincidence, but this horizontal bend is also oriented such that it would tend to pre-rotate the fluid in a counter-clockwise direction.

I am beginning to believe that pre-rotation could be the central issue in the ongoing cavitation problem. I also wonder if this could have played some role in the problem with inducers coming loose.

I am interested in your opinions or experiences that may be relevant to pre-rotation into an axial flow inducer. Could this negate the affect of the inducer? Could the pre-rotation result in cavitation even with an extremely high NPSH margin? Could a piping bend more than 10 pipe diameters before the suction be contributing to a pre-rotation problem? Any experiences, opinions or suggestions will be greatly appreciated.


Johnny Pellin

 

[ardeIII]

johnny
any possibility of entrained gas in your suction stream, perhaps from nitrogen/gas blanket or other gas/liquid interface in the process? Entrained gas can cause cavitation symptoms and damage even with adequate NPSH margin and near BEP operation.

An issue we have seen with Sundyne units relates to the distance from the pump discharge to downstream control valve. depending on control valve distance from pump, pressure pulsations can result in failure of pump shaft and or locking device. Attached the paper for your reference, see section 2.4.

another interesting article from Sundyne website on impeller cavitation damage with adequate NPSHa:

http://www.sundyne.com/Files/Sundyne/Global/US-en/field_engineering_bulletins/40.2.95_Rev-Orig.pdf

as far as loosening nuts, we have seen this on lower energy end suction pumps and it seems to be due to off-BEP operation and/or cavitation, the forces associated with this resulting in impeller bore fretting, loosening locking devices and in some cases breaking shafts. We spent much time pulling hair out investigating assembly, locking devices, locknut torque, impeller/shaft fitup etc. before finding this and making adjustments that helped solve problem. Unfortunately none of this appears to be your issue.

best of luck

 

[chief]

I note from your previous post, that you were fitting a suction gauge to your pump. What was the reading when you ran up the pump again.
You have an 80ft head available, but with such a high rpm pump and flowrate, you may be having a suction delvery problem.
How much pipe length are you from the tank suction point and what is the pipe dia. How many bends in the suction piping.
Are you taking suction from a large dia manifold or is the suction pipe the same dia as the pump inlet.
What is the position of the suction pipe into the tank. Is it a side position, underneath or a vertical elbow. Is it in the tank corner or close to a sidewall.
If vertical, is there a 'bell end' to prevent vortexing.

Apologise for all the questions, but sometimes it is best to eliminate the basics, before becoming side-tracked

Offshore Engineering&Design

 

[Artisi]

johnny
Sorry I can't be more specific but like you, think the problem could well stems from pre-rotation. As you are aware, pre-rotation results in a mis-match of flow onto the impeller and this could cause "local" cavitation as the flow hits the impeller at other than the ideal direction - especially at the high rational speeed. Cavitation could also be causing some fretting of the inducer leading to it becoming loose.

Can you check for pre-rotation by use of a pressure gauge in the inlet line and if you there is evidence of pre-rotation it maybe worth looking into some form of flow straightening at the pump inlet.

 

[JJPellin]

Thank you all for your replies. I don't believe that pressure drop in the suction piping is the source of our problems. We were able to fit a pressure gauge to a low point drain on the suction line. I don't have the exact data with me, but it confirmed what our piping models suggested. We are only getting a foot or two of losses between the suction vessel and the pump suction. We don't have a restriction in the suction piping.

I have previously considered dissolved gasses as an issue. The blanket gas is fuel gas (ethane and methane). I have read technical articles that describe the affect of dissolved non-condensable gasses on cavitation. I agree that this can result in cavitation even if the NPSH margin would suggest otherwise. But some of the literature has gone on to suggest that this form of cavitation is more likely to result in loss of efficiency or capacity than physical damage to the impeller. Some even suggest that the presence of non-condensable gas in the collapsing bubbles can have a cushioning affect that can actually reduce the damage to the impeller. Some have suggested deliberately injecting gas into a pump suction to get this benefit. Because of testing done with a similar pair of pumps in another service, I don't believe that dissolved gas is the core problem.

I need to go back and reexamine some of my old notes on this problem. I had previously explored the control valve location and decided it was not the problem. But, this is worth a second look. I believe the control valve is relatively close to the pump, but can't rule out the possibility of pressure pulsations. I will dig into the again with an open mind.

I had already checked the Sundyne article on eye diameter to tip diameter ratios. According to the chart included in the article, our impeller should be in the range where no damage would be expected.

I have arranged for Sundyne to bring in two of their factory engineers for a series of meetings in a couple of weeks. I am going to challenge them to help me work this through to resolution. If this cannot be solved, we may be forced to replace these pumps with a competitor’s product. And I will be very reluctant to purchase any similar pumps from Sundyne in the future. With the current projections for expansion projects at our plant, they should have an incentive to help us solve this.

I do not know how to check for pre-rotation using a pressure gauge on the suction line as Artisi suggests. Can you be more specific about the method you have in mind? If I can confirm pre-rotation caused by the piping, I will either re-route the piping or add appropriate flow straighteners. But, if the pre-rotation is caused by the case configuration and internal geometry, I will need a more complicated solution. I will consider new cases that have the nozzles located along the center line if this is necessary. If Sundyne will not help me with a new case design, I can use some aftermarket shops that have done similar work for me in the past.

Johnny Pellin

 

[Artisi]

Johnny

I can't put my hand to any info. on establishing pre-rotation using pressure gauges, but if memoery serves me, it is addressed in the ISO and Australian pump test code.