Pump Cavitation Issue

Increase the NPSH to the pump! Enough and the bubbles will go right through.

One thing to look at would be the suction piping layout, which should avoid any "pockets" in the upper part of pipes or fittings, where bubbles could accumulate and coalesce. Notably, any reducers should ordinarily be oriented flat side up (though there are somewhat uncommon exceptions).

Rather than pumping the liquid out, what scope is there for displacing it out of the reactor by pumping a gas into the space above the slurry, bearing in mind the gas properties and pressure, the maximum working pressure of the reactor and the pressure in the downstream equipment?

'... Apart from changing the pump to a positive displacement type, could we drain the air from the pump casing?...'

.

If the air takes a sufficiently long time to build up to a noticeable level then intermittently shutting down to vent the air out of the casing might be a tolerable option....but I doubt that is the case.

.

other options include:

.

- pressurizing the reactor

.

- if the reactor can't be pressurized then two holding tanks in parallel each alternating; one tank being filled by reactor effluent, isolated from the pump, and the other isolated from the reactor and pressurized to provide greater NPSH to the existing pump.

.

- reducing the pressure of the destination

.

- significantly reducing the temperature of the slurry.

.

- relocating the current pump to reduce the height of the pump relative to the reactor. In addition, if the pump can be positioned such that any bubbles would exit the pump housing downstream when the pump is turned off, intermittent operation could be used to clear bubbles.

.

- adding a booster pump between the reactor and the current pump

.

- applying sufficient suction to the casing of the current pump.

...

That is all I can think of at the moment.

You might find this helpful:

http://www.pumpschool.com/applications/NPSH.pdf

IF you have the room, install a tee (branch side up) two pipe sizes larger than the suction line between the tank and close to the pump inlet.

Install with the increaser/reducer flat on bottom. Install at least 12 inches of pipe to the tee branch, finished with a cap and a 1/2" tube from the end of the cap to the top of the tank- no valve or normally-open full port ball valve at the pipe cap and no kinks in the tube.

IF you have ENOUGH room, install at least a foot of pipe after the increaser and before the tee. Ideally, the increaser should be right at the tank outlet.

The low velocity should give the bubbles enough time to sufficiently "burp" the fluid to stop the problem. The inlet at the "bottom" of this assembly should see essentially "pure" slurry.

Another option is to install the increaser and a 90 looking up directly at the tank outlet, then another 90 looking toward the pump. Connect the horizontal 90 to the branch of the tee, with the same 12 inches and cap/tube on the upper "run" and the reducer on the bottom "run" connected to a 90 looking at the pump.

Either option should make "all better".

I forgot to mention that- unless the discharge flow is mass-sensitive, you can install a VFD on the pump as well to give the vent even more time to work and increase the residence time more.

Sir,

Check to make sure your flow rate and discharge pressure are within the range of the pump curve.

Since you are pumping a slurry, I would suggest you look at the impeller and the probable need for it to be with larger passages. Normally this would mean a larger motor also, but see what I did:

I had an application once, with positive suction head, where it was way off the curve for the centrifugal pump. Bought an impeller and volute sized for a much larger horsepower and had the impeller turned down to keep the actual horsepower within the nameplate of the motor. This eliminated my cavitation problem. Definitely was not a standard application, nor something the dealer was considering as an approach to my problem, but worked perfectly (took about 3 tries with increasing impeller sizes and amounts of turning before I got it right; verified load with a clamp-on ammeter).

-JMM

You really haven't told us what type of pump it and the basic design i.e. vertical split API OH2 pump, horizontally split double suction between bearings or whatever. Some pumps are designed to be self venting such as the OH2 and some chemical pumps with their center of casing discharge.

If there is a casing vent on the pump could you run a continuous bypass from the vent back to the vessel to above the liquid level to allow a continuous venting stream?