Pump dividing walls in sump 4

[orlandobill]

I have an existing sump (open pool) with three (3) mixed flow, vertical pumps drawing from it. Each pump has a wall between it and the adjacent one forming a three-walled vault around each pump. At any time only 2 of the 3 pumps are running and the combinations are rotated in order to give each pump a day or two off. We are starting to see flow rate decreases in two of the three pumps. We are looking at several things and the sump dimensions is one of them. The pump vault dimensions meet the recommendations from the Hydraulic Institute except that there is no "break-thru" between the vaults. The walls are structural for the walkway above it, so removing the walls is not an option. Should I expect that modifying the existing vaults ($) to add "break-thru" will provide a significant change to the flow rates? Are there any suggestions on how to increase the flow rates? Thank you.

 

[dtn6770]

Good description for the most part, but are you witnessing anything specific to the vault that makes you think your proble is some how related to it, vortexing, drastic level reductions, etc...anything that implies the water (assumed) isn't getting to the pump suction fast enough?

What is the service and how long has the installation been in place? Have any of the pumps been pulled and inspected for wear? Does the vault need a good cleaning?

If the flow reduction is a new thing to an existing (proven) installation I'd pay more attention to the pump and piping unless there's enough crud built up in the vault the may be causing problems.

 

[orlandobill]

Thanks for the reply. The installation is about 10 years old and I'm told the reduced flow has there since the beginning.

There are vortices visable on the surface of the water and the pump currently being rebuilt shows cavitation wear on the impellor.

The pumps are generally rebuilt each year, but no specific inspection reports have been kept on previous rebuilds.

We are doing a pipe inspection next month to look for obstructions and/or scale. We will clean the vault at this time also.

I should also have mentioned that we utilize trash screens in the sump as well as strainers on the inlets. Is this redundant? Should the mixed flow pumps handle anything that passes through the trash screen?

 

[RossABQ]

Trash screens are pretty course, not that pump inlet screen are exactly fine. Neither is inteneded to prevent dirt or even rocks from going thru the pump, just to keep plastic trash bags, tumbleweeds, and styrofoam out of the system.

You have not explicitly stated how far the pumps' inlets are off the bottom of the sump, or where the water enters the sump. Is it flowing into the sump parallel to the three vaults, or perpendicular? Is there any difference in pump performance characteristics between the 3 pumps/vaults?

 

[dtn6770]

"...starting to see flow rate decreases..." is different than what I would term as under-performance from day 1. One indicates a change and the other (the latter) indicates a potentially bunk design.

Honestly, my pump trouble shooting can be documented on a sicky note but it sounds like an NPSHa vs NPSHr situation. Lowering the pumps or increasing the liquid level may help more than trying to get more flow to the pumps.

Vortices aren't good but a litte swirl is better than a full blown "whirlpool" that lets air into the suctions.

Pulling the suction strainers certainly wouldn't hurt as long as the trash sreens are effectively "filtering" out the size of junk that'll give the pumps grief.

Combing over the original design package and comparing it to what you actually have would be a good start. It would be worth a call to pump manufacturer to see what they have to say.

 

[orlandobill]

The sump is roughly a 100' square.
a
-------
| <-|
b| |d
|123 |
-------
c

The flow enters from side d (at the arrow) parallel to side a in a CCW flow. The pumps are lined up along side c in the corner by wall b. There is some natural circulation that we will be addressing with flow straighteners. Pump three performs close to rated (14400 gpm). Pump two is just a little worse and pump one is down near 70%. I'm sure it's not a coincidence that it is the one in the corner.

Sorry for the confusion about the flow loss. The decreases in flow are noticed just after reinstallation. We get close to rated flow after reinstallation and then the flow will drop off steadily. Lately the level has dropped so low that we leave pumps 2 and 3 on constantly and don't even use 1.

 

[Artisi]

It sounds like inflow problems which can be numerous. A detailed sketch with dimnsions would certainly help to get some meaningful discussion underway.

Phitsanulok
Thailand

 

[rmw]

If the flow in the sump is not approaching each individual pump chamber equally, then your pumpage will not be equal.

Your explanation would make it seem like the flow comes in on side D "at" arrow with references to some circulation. If I understood it correctly, then that is at the root of your problem.

Also, if you don't have enough submergence to prevent any or all vortexing, then you have a problem with your submergence. Have you checked your pump manufacturers requirements regarding the minimum submergence, distance from the floor, and distance from each respective wall in the individual pump sumps? I suspect that something might be off there too.

What about sedementation? Is that a factor? Build up can change chamber dimensions.

Camerons Hydraulic has some information if you have a situation where your pump mfg. has disappered or is unable to help you.

I was once associated with the old BJ pump co before they were absorbed and they would go postal over the parameters I mentioned above. I have seen them ready to walk away from business rather than put their pumps in a poorly designed pit and have to fight the problems the rest of the pumps life.

rmw

 

[JJPellin]

The types of pumps you are describing are commonly used in cooling towers in the refining industry. We have five cooling towers with pumps like this. They range in size from 10,000 gpm to over 24,000 gpm each. We have had the sort of problems that you describe in four of the five towers. We have found a number of problems, some of which are very difficult to solve. Many very good points have already been made in the replies above. But allow me to go through the steps we used. First, evaluate the NPSH for the pumps. We found that two of our towers had inadequate NPSHa even at the highest possible levels. This was a design mistake from original construction. Next, evaluate submergence. One of our troublesome towers had marginal submergence. It is worth noting that even if the NPSHa is adequate, you could still have inadequate submergence. The presence of visible vortices is a sign that this may be relevant. Lastly consider if you have a pre-rotation problem causing suction recirculation cavitation. We found this in one tower, even though the sump was properly designed according to Hydraulic Institute criteria. If NPSH is the problem, you are probably stuck and may need to replace the bowl assembly with a different design. If submergence is the problem, you may be able to install a baffle to break up the vortices. If pre-rotation is the problem, then a vane grating basket that opposes the pre-rotation may solve the problem. Pumps of this type, especially very large ones, can be tricky. They can cavitate even if the design looks perfect. They can experience suction recirculation cavitation even at best efficiency point flow, or higher. Designs that would work beautifully with a horizontal pump may not work with a vertical of this type. And to make the problems more likely, buyers often buy these pumps with no shop test since testing facilities that can run these big verticals are somewhat rare. You should also make sure that the pumps are being installed and adjusted correctly. Pumps that look identical on the surface could have drastically different lift setting requirements. We have pumps in our towers that are supposed to be set at 0.125" lift and others that are supposed to be set at 0.030" lift. Open face impellers, closed impellers with radial wear rings, and closed impellers with lateral wear rings have different requirements. We even found one set where the outside pump shop was leaving out the lateral wear ring because they didn't understand what it was for and didn't think it was necessary. Anything that makes the pump inefficient can contribute to accelerated degradation.

 

[Artisi]

JJPellin (Mechanical)

Certainly a star award answer.
I would add thaat the presence of sub-surface vorticies can also be a major problem as they are not always evident from the surface but if there, can be a major source of low performance and vibration and can be difficult to eliminate

Phitsanulok
Thailand

 

[orlandobill]

I agree with Artisi. Excellent post, JJPellin. Thank you everyone for your expert input. I will look into these and report back with findings and fixes. Thanks again.

 

[rcooper]

When we have been asked to look at pumping stations which have been operating poorly for no apparent reason, we have the sump model tested, typically at 1/3 to 1/5 scale. Similarly, if we are building a new large pump, we have the proposal model tested to ensure we don't end up with something which does not perform as planned. A recent example of an existing pumping station which was model tested was a 3100l/s wastewater pumping station. The pumps were failing within 6 months of being installed and the repair costs were astranomical. The model revealled floor and wall vorticies entering the pump which contributed to the early failure of the pumps. The model was used to test some baffles and vanes which eliminated the problem. When these were installed the problems with the short life of the pumps vanished.

With a new pumping station, we proposed a sump with a plan area half that achieved with a standard design, significantly reducing the construction costs. From these experiences, I would heartily recommend model testing if you can find a good modelling consultancy.

 

[ccfowler]

orlandobill,

The above posts contain much useful advice and should provide a good start for you.

You obviously have an expensive problem on your hands. If you can see any evidence of vortices at the surface, you can be quite certain that the pump inlet conditions are far from ideal.

If I had to deal with this problem, I would want to first consider the full economic implications of the existing situation compared to that of a properly running pumping system. I would look at the differential power costs, costs of disrupted operation, costs of excessively frequent rebuilds, etc. It would be no surprise to find that the potential savings can provide funding for significant study and modification efforts to correct this problem, but this is not a certainty. Your economic analysis could show that although not ideal, the costs of living with this problem may be less costly than fixing it. Since it is likely that significant energy savings may be involved, there may be some additional economic incentives available.

If the economics are not sufficient to justify a complete fix, you may be able to implement some measures to mitigate the problem. Paying closer attention to the details of future pump rebuilds, and some modest revisions to your intake structures may pay for themselves very quickly. Fully correcting an existing, operating system can be difficult, expensive, and physically almost impossible.

I would pay close attention to potential means to minimize vortex formation not only at the pumps but upstream of them, too. Once formed, vortices can be very persistent, so it is best to avoid their formation well upstream of the pumps.

 

[RRiche]

I participated in a pump model test done by the experts in the field at Clemson Engineer Hydraulics in Anderson SC. They pointed out how pre-swirl, vorticies, and non-uniform flow affects the performance of pumps. Basically they added flow dampeners in the intake structure to slow down the velocity of the water and flow straightners to reduce the tendency of pre swirl (pre swirl occurs when water enters the chamber at an angle and begings to swirl and causes forces of the impellar which will lead to non uniform flow through the pump which will lead to failure of the bearings). They also added a curtain wall in the upper section to eliminate surface vorticies. They also added fillets and splitters to the floor below the pump to help direct the water to the suction so that water feeds the suction bell from all directions (including the back section near the wall).

Flow into the suction bell of each pump has to be uniform, laminer, and free from vortices to prevent any problems. If you see surface vorticies, you definately have a problem with submergenace and probably getting some air induction and possible caviatation. If you are shutting off one pump, and flow is entering from non symetrical directions, you will likely need straighting vanes installed. I recommend you contact someone like at Clemson Engineering Hydraulics to physically model you intake and they will definately solve your problem.

 

[orlandobill]

Here are some more details:

@14500 gpm:
NPSHa = 42.32 ft
NPSHr = 29.76 ft

NPSH does not seem to be an issue right now.

 

[ccfowler]

orlandobill,

I disagree that NPSH is not an issue. At NPSHa = NPSHr, the pump is already suffering from performance deficiencies due to cavitation. The cavitation free NPSHr is normally substantially greater than the nominal HPSHr. Determining the precise cavitation free NPSHr is a messy matter, but you can find considerable discussions on this topic with a modest search effort. These discussions can give you some guidance on this subject as it applies to your application. Don't be surprised to find that the cavitation free NPSHr may be in the range of 2 to 4 times the nominal NPSHr.

You should consider that the two most common standards for determining the nominal NPSHr are based on the point at which the pump's performance is suffering by either 1% or 3% due to cavitation.

Probably the most important thing to do to minimize vortex formation is to minimize abrupt accelerations in the flow anywhere upstream from the pump inlet. Once formed, a vortex can persist within the flow to cause trouble at the pump.

Depending upon the actual dimensional constraints of your situation, you may be able to ease submergence problems by adding a device to further control the acceleration of the flow approaching the pump inlet. From information above, I assume that there may be some potential problems with solids in the flow, and the implications of these should be thoroughly considered. For a situation such as yours, I would consider attaching a fairly wide flange to the suction bell and make modifications as necessary to assure that the floor of the sump beneath the area of the pump and the attached flange is as near to perfectly level as possible. Also, I would consider installation of a small cone on the sump floor along the pump's axis to help redirect the converging inward flow toward the pump inlet. The practicality of such a device can depend heavily upon the proximity of the adjacent walls.

 

[JJPellin]

This is a problem that could consume many hours of your time, many hundreds of thousands of dollars and many years to correct. The more I read, the more I agree with many of the comments others have posted. Vortices and pre-rotation seem likely. Extensive sump modifications may solve the problem, but the time and expense will be considerable. There may be a cheap and quick modification that you can make that will improve the situation (or possible solve it). The last set of big vertical pumps like this that we purchased came with vane-grating baskets of a particular design that are supposed to oppose pre-rotation and break up vortices. We purchased the pumps with these baskets because we were installing larger pumps in an existing sump that was undersized. We have had no problems at all with the new pumps running for almost a year. The baskets were purchased with the pumps from Rhurpumpen. If you contact Rhurpumpen, they may be able to help.

Are your pumps rotating counter-clockwise. If so, the inlet configuration you sketched out above looks like it would encourage pre-rotation.

 

[orlandobill]

Thanks ccfowler. Here's a FAQ from the HI website:

[red]Q. Our plant is operating pumps on cooling tower service which are experiencing cavitation damage to the cast iron impellers. The pumps are operating close to the best efficiency point and the NPSH available is slightly higher than the NPSH required by the pump. Should there be cavitation damage in the pump impellers under these conditions?

A. There is a common misconception among pump users that if the NPSH available from the system is greater than the NPSH required by the pump, that the pump will operate free of cavitation. This is not so. In order for a pump to operate cavitation free, the NPSHA must be from 2 to 20 times greater than the NPSHR of the pump. By definition, NPSHR is measured when the pump total head is reduced by 3% due to cavitation. For satisfactory operation, some NPSH margin over NPSHR must be provided by the system.
The Hydraulic Institute has recently published a new standard on this subject namely, ANSI/HI 9.6.1-1998 Centrifugal and Vertical Pumps for NPSH Margin. According to this standard, for cooling tower service, NPSHA should be 1.3 to 2.0 times NPSHR, depending on suction energy level, which is also defined in the standard.[/red]

I will look into this new standard and see where my suction energy level is. FYI, my ratio is 1.42.