low flow pump surging

[davefao]

In their pump control bulletin Sundyne is advising to place the first throttling valve as close as possible to the outlet of the pump to avoid "flow surging" ; the liquid volume between pump and valve comes into an oscillation and placing the first valve as close as possible reduces the amplitude . Can anyone give me a good reference of a textbook or document where this kind
of phenomenon is well explained . In general books concerning vibrations with a hydraulic origin ( hp pumps in parallel ,boiler operation )

 

[Slagathor]

Is this a high head pump (low specific speed)? If so, the curve may be flat left of BEP. This means that flow can vary greatly with the slightest changes in system head. This can set up a surging situation where pressure transients (turbulence, etc) increase the head the pump sees, resulting in a transient drop in flow. This drop in flow can be dramatic with a flat curve. Then the velocity head and friction head in the system drop, and flow increases, setting up the cycle again. Obviously, the turbulence patterns and their frequency can have a large impact on this sort of behavior. I have never seen an analytical model of this behavior, though I have seen it in the field many times.

 

[BigInch]

You are seeing a very high frequency "water hammer", except instead of one or two "hammers" that is usually associated with a water hammer per say, it is appearing as a high frequency vibration or pulsation. At any restriction in the pipe, a valve, or a change of pipe area, a certain portion of any pressure wave is reflected, while the balance continues down the pipe. When a restriction is located at a point where the pressures are reflected in "tune", or at a particular harmonic being generated by the pump, vibrations can increase in amplitude. It is much more commonly known in reciprocating pumps and compressors, where there are a number of consultants that specialize in pulsation analysis. Centrifugal pump pulsation is less commonly known, apparently because it is much less of a problem for most users, however it is apparently still enough of a problem that Gould Pumps has a download discussing this available at,

[COLOR=black yellow]

http://www.gouldspumps.com/download_files/Technews/PumpLines_winter2001.pdf

[/color]

,wich I'm sure you will find of interest.







Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[BigInch]

or witch

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[JJPellin]

This is a problem which is especially prevelent with Sundyne pumps. In their literature, they suggest that the first control valve should be within five feet of the pump discharge flange. They have a chart in the manual for model LMV-331 that depicts the required maximum distance to the first control valve based on the horsepower of the driver and the percent of design flow. There are about 175 Sundyne pumps in our refinery. I can testify that some of the ones that give us the most trouble have the control valve very far from the pump. As noted in the first reply above, Sundynes tend to have flat curves where a very small change in pressure can produce a large change in flow. This characteristic drives the requirment for a close control valve. I am sorry, but I do not have a technical reference to explain this any better.

 

[BigInch]

JJ, I took a look at the curve for the 331 and see your problem. That is NOT a pump that will like running at any speed less than 50 m3/h. The "curve" is absolutely flat.

I don't have this doc, but at the 2002 Pump User Symposium, Lecture 9, "Numerical Simulation of BPF Pressure Pulsation Field in Centrifugal Pumps", seems to be addressing the subject. If its of interest to you, it can be purchased at the Texas AM Turbomachinery Lab website,

http://turbolab.tamu.edu/

Going the Big Inch!

http://virtualpipeline.spaces.msn.com