Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

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

Suction design, parallel operating pumps, bellmouth elbow

Status
Not open for further replies.

rocketscientist

Chemical
Aug 19, 2000
86
We have 2 identical pumps, with one operating spare pumping from the sump of a water scrubber. The pumps draw from separate suction lines. These suctions are drawn from bellmouth elblows. These are els where the suction line makes an elbow drawing from the bottom of the tank. The clearence "C" is only about 200 mm, about 8 inches, between the opening of the 90 elbow and the tank bottom. The flow is about 1900 gpm per pump, simultaneously, through two 14-inch suction lines. Unfortunately, the D, the opening of the of the bellmouth is identical to the suction line, about 14-inch, i.e., about 350 mm. Also, because the tank is fiberglass, there could be deflection in the bottom of the tank. Often the packing breaks up, and this could be collecting in the bottom of the scrubber. The diameter of the scrubber is only about 3400 mm, i.e., 134 inches, or about 11.2 ft. The pump suctions fill about 180 degrees of one side of the vessel.

With two pumps running, with a combined flow of about 3800 gpm, there is severe cavitation. The cavitation is worst with 2 pumps running but still present with a single pump running. This makes me suspect that the bellmouth may be the problem. There is noticable vibration throughout the system. The vibration is passed to the tank wall. Below the liquid level, there is a loud thumping.

I suspect that it may be possible to treat this as a high volatile liquid issue and cutting the bellmouth may increase the available area for the suction.

Earlier, we ran at nearly 5000 gpm with the pumps. I designed two orifice plates to reduce the vibration and flow and this worked well but we still have hot bearings and cavitation. We are operating on cold water. Once we get to 60 C temperatures it will be much worse.

Any ideas on how I can reduce cavitation? How does cutting the bellmouth sound?
 
Replies continue below

Recommended for you

A sketch of the set-up would tell a lot more.

Do you really know what cavitation is?

Where is the pump/s operating in terms of flow and head in relation to BEP on the pump curve.

It may well not be cavitation, could well be disturbed flow into the pump from the inlet conditions - hence the sketch will help.
 
Funny, I thought I was pretty specific. A drawing would probably help if I had one handy.

You are correct, I am applying cavitation loosely. I, too, think it's more likely a flow disruption, but it does continue throughout the process, which seems unlikely. The scrubber is about 60 ft above grade and the water passes through a heat exchanger. It seems unlikely that a mere flow disruption would continue through this network.

I was thinking more along the lines that having two pumps running in such close proximity is causing a more severe initial flow disruption than expected. Keep in mind the suction velocity is about 0.4085*(1900)/14^2 = 4 fps. Of course, the limit is 6 fps for suction flow. The piping is fiberglass, of a large diameter, so that probably rules out joint problems. Following Crane's Hydraulic handbook for submergence we need 1 ft of submergence (above the bellmounth opening) for every ft per second of suction velocity: 4 ft above the top of the nozzle pipe (not centerline). We see our "flow disturbance" worsen below about 55% but it is even severe at 80% level.

The more I think about it, the more I think a solution I saw once in a KO pot may work. It was the same principle, but in reverse. The idea was to slow the liquid down to allow it to disperse easily. They did it by cutting "teeth" into the bellmouth. This might work if we did not add a disturbance. It would be easier if I had more than 200 mm clearance to add an expansion section on the suction. But, maybe, the teeth might work, or a shape that would cause less disruption in the flow. Of course, if I cut the teeth and it is a problem of pump interference, we could make things worse not better.

Any thoughts.
 
If I understand your detailed description, these are horizontal pumps drawing from a sump via a pipe with an elbow and a bellmouth at the bottom of the vertical run.

If so, then the sump design will have to follow all the rules that apply to a vertical pump drawing from a similar sump with respect to minimum submergence, height of the bellmouth off the floor, distance to the adjacent pumps and walls, and velocity of the fluid approaching the suction(s).

Since horizontal pumps wouldn't necessarily have a "minimum submergence" nor sump design requirement as would a vertical pump, you are going to have to determine what submergence is needed to prevent vortexing (which is what I suspect here.)

We would need a lot more details about your sump configuration to be able to give any tips regarding your situation.

rmw
 
Any issues that could affect a pair of vertical pumps in a common sump could affect these pumps. The main issues that occur to me are vortexing, pre-rotation, acoustic resonance, structural resonance and flow disruption from the proximity of the intakes to the floor, walls or each other. Sump design should follow a number of rules as already stated. But even a sump that meets the minimum requirements of the Hydraulic Institute can have unforeseen problems. There are a number of good articles on the subject published in the proceedings of the International Pump Users Symposium over the years. I don't have access to my collection to reference specific articles. Can you see the intake to look for visible vortexing? You could add a simple vortex breaker if one is visible or of it is impossible to check. This could also help to break up pre-rotation. Have you analyzed the vibration to see if it is characteristic of cavitation?

Johnny Pellin
 
I agree with JJPellin. Vortex related problems seem to be the most likely for your system as I understand it.

Are there any indications that when the two pumps are operating in parallel that they may be shifting flow dominance back and forth? I am presuming that these pumps are both powered by ordinary induction motors and not some sort of VFD.
 
Thanks for your input. I am at the thinking about the experimentation stage. There's a couple of options:

1. Raise the C, i.e., the clearence, the distance from the
mouth of the bellmouth to the vessel bottom.

2. Cut slot in the side to raise the area available for
entry into the bellmouth el.

3. Move the outboard bellmouth els further apart by
extending them into the tank.

4. Installing temporary dividers to isolate the els from
each other. These should be like a chinese hat vortex
breaker ---- with a gap at the bottom for some flow.

Anything to add to the list?

We are in startup mode so there may still be time for some iniative.

We're still on water now but will be on acid soon.

 
You may want to consider welding a flange (with a well rounded edge) to the bellmouth to further isolate the acceleration of the flow velocities into the bellmouth. This will allow keeping the bellmouth entrance low which helps minimize vortex formation.

If your flow rate varies rapidly or substantially, high flow rate periods can initiate formation of vortices that persist to cause trouble during periods of reduced flow rates. Low velocities approaching the inlet and gentle acceleration of the flow into the piping system are the keys to avoiding vortex formation.
 
I would strongly suggest that you collect vibration data on the piping and the pumps and have it analyzed by an expert. They should be able to distinguish between cavitation, acoustic resonance and structural resonance based on the frequency content of the vibration, the magnitudes, locations and dominant directions.

Johnny Pellin
 
Simple question asked initially, where are the pumps operating on their curve in relation to BEP.
At the moment you don't seem to have any idea what the problem is - if you don't know what it is so how can you fix the problem?
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor