Need Books or Papers 1

[PUMPDESIGNER]

I am looking to buy books or research papers, (will take freebies too), about surveys of large numbers of centrifugal pumps.
I have seen reference made of one such large survey where the author studied a large number of centrifugal pumps and analyzed how their NPSHr, Specific Speed, Suction Specific Speed, flow, and head were correlated.
That one specifically, but also any others.

Please help me find as much reference materials as possible from reliable sources.

Thank you, you guys are always willing to help.

PUMPDESIGNER

 

[vanstoja]

Sounds like you're referring to papers/articles by S.Yeddiah of Worthington Pumps who plotted hundreds of points to get a Suction Specific Speed vs flowrate correlation of S=2580Q^0.182. In Power, June, 1973, pg.78,"Factor Pump Size Into NPSH Comparisons", Yeddiah says "chart is based on calculated values of S for several hundred pumps made by 10 leading companies. In the 1975 ASME Cavitation and Polyphase Flow Forum booklet, pgs.20-21, "Relationship Between the NPSH Requirements of a Centrifugal Pump and Its Performance", he describes these pumps in very general terms, eg. head range of 80-400 feet and says "the author has compiled statistical data relating to the NPSH characteristics of about 600 pumps manufactured by the ten leading pump manufacturers in the world. I don't know if Yedidiah ever published a compilation of these data but his later book, which I've never seen, may be worth a
look.
Other NPSH-related pump data are the British National Engineering Lab(NEL} tested pumps of which 23 designs are tabulated by F.G.Hammitt in the 1975 ASME Cavitation Forum booklet, pp. 12-15, "Detailed Cavitating Flow Regimes for Centrifugal Pumps and Heads vs NPSH Curves". I.S. Piersall also treats some of the same NEL data in IAHR/AIRH Symposium, 1972, Rome pp. I7-1 to I7-13, Cavitation Speed Scale Effects in Pumps".
H.H.Anderson (1977), "Statistical Records of Pump and Water Turbine Efficiences", IME Conference Publication 1977-7, "Scaling for Performance Prediction in Rotordynamic Machines"pp.1-6 (Paper #C172/77)plots efficiency vs flowrate for 15,000 pumps and 310 water turbines. These pumps cover some 200 different designs in Anderson's data alone. Other sources cited are Hitachi, Sulzer, Escher Wyss among the 12 source references cited. One is a book by Anderson published by Trade & Technical Press, 2nd Ed. 1972.
C.S.Martin (1983), "Representation of Pump Characteristics for Transient Analysis", ASME/FED-Vol.6, pp.1-13 tabulates complete performance characteristic data for some 27 different centrifugal pump designs and cites all the reference sources.

 

[PUMPDESIGNER]

I thought you might come through for me on this vanstoja.


PUMPDESIGNER

 

[imok2]

Thought this might interest you:

http://www.pump-magazine.com/pm.htm

 

[vanstoja]

Some other possible sources of many-pumps data are:
Makay,E. & Szamody,O. "Survey of Feed Pump Outages", EPRI Report FP-754, April 1978

Hallam,J.L. "Suction Specific Speed is Important in Pump Failure Rates", Oil and Gas Journal, Jan. 17, 1983.
Both of these are cited by by Don Cavi in "NPSHR Data and Tests Need Clarification", Power Engineering, February 1985, pp.47-50. He says the Hallam data represents five years of experience with 480 centrifugal pumps in a large refinery. Makay's feed pump failure data has been publised in several papers and , as I recall, covers some 580 pumps.
A 3rd edition of H.H. Anderson's book, "Centrifugal Pumps" was published in 1993 by Elsevier Advanced Technology Press, Oxford, England.
What is your purpose in collecting many-pumps data?

 

[PUMPDESIGNER]

vanstoja,
Numerous reasons I need the info.

1
There is a lot of aging infrastructure in this country now. We are purposely trying to focus on refurbishment instead of new, more fun, better pay, more knowledge required. In those cases we are often paid to do surveys and reports first, old equipment no documentation, difficult to find out what existing equipment is.

2
We are involved as consultants on lawsuits. Again, data hard to come by, people uncooperative, etc. And sometimes we are not paid for on site investigation, just a quicky opnion using information they provide, which is little to nothing.

3
Yuck, I have to deal with what we call "water systems" market products. That market is terrible, low cost, cheapo, junk that even the manufacturers have little data on, and believe it or not, they do not publish curves, often just a chart with specific flow rates and pressures, and never power curves or NPSHr curves.

PUMPDESIGNER

 

[vanstoja]

In my opinion, your search for basic performance data on "ancient" centrifugal pumps would be better served by concentrating on finding pump manufacturer's databooks covering their various model numbers for specific applications. The published papers and articles on "many-pumps" data seem to focus on trendlines that are, at best, suitable for rough estimations of expected performance for new designs. The scatter in this data, such as Yedidiah's Suction Specific Speed plots, is so great that the linear curvefits would not pass muster as reliable data if the coefficients of determination (COD) (R^2 values in Excel spreadsheet plots) were shown on the curves. I would never buy into fitted data curve equations that had CODs less than 0.9. Such manufacturer's data may be exceedingly difficult to find and/or obtain. I can only offer one obscure example of such manufacturer's pump data from the 1950s,but suspect that most pump suppliers had such compilations that could be purchased in the pre-1950s era.
I have a book titled "Pump Engineering Data" published in 1954 and reprinted in 1957 by C.H.Wheeler manufacturing Co., Economy Pump Division in Philadelphia,PA that cost all of $3.00 US. It's 417 pages of text include 150 pages of general engineering data including piping friction loss data such as found in Crane 409 or 410. The remainder covers performance, dimensional, price data on Wheeler's 11 categories of pump types including horizontal, vertical, sump pumps, non-clogging pumps, paper stock pumps, axial flow pumps, condensation pumps and caisson pumps. For each model number, flow, head, efficiency, motor horsepower and shafthorsepower are given with pictures, outline dimensions and typical arrangements. A General Information section describes materials, extras and options. Now all you need to find are the comparable contemporary databooks for Ingersoll-Rand, Worthington, De Laval, Foster Wheeler, Allis Chalmers, Byron-Jackson, Peerless Pumps, Bingham-Willamette and their European counterparts to characterize
most of the "ancient" centrifugal pumps operating in the USA. Your best bet may be to locate an "ancient" pump designer from each of these pump companies who would know whether these books indeed existed and how they might be obtained now.

 

[PUMPDESIGNER]

I thank you very much for your opinion, and I do value your opinion over most.
I will for a few days look at what I have already and see what I can do with that. I wrote an Excel spreadsheet and I post up every pump I come across to see if I can on my own find trends and other interesting things. So far I have just over 100 pumps plotted up, and I have "learned" a few things, usualy subtle things that pertain to certain design syles or manufacturing methods. Not easy to quantify and communicate to others though.

You are correct about possibly not having much after I do a lot of work, could be a waste of time, I need to be very careful to make sure before starting.

PUMPDESIGNER

 

[stanier]

Hi PumpDesigner

Streeter and Wylie have an algorithm for determining the moment of inertia of a pump and motor assembly that was derived from numerous case studies. this may be one parameter that you need to look at if waterhammer figures in your applications. the database may also contain other information that you seek. At least you have a starting point. I believe that the authors still are active in the waterhammer market as I come across names at conferences.

 

[smckennz]

If you are trying to squeeze the last percent out of a pump then there a number of non-hydraulic factors that come into play and should be recognised when evaluating historical date.
1) Baseline. This is what the pump designer reckons the pump will do, and has been validated, sometimes only in part (e.g model test or single-size model for range). May or may not have an attached tolerance.

2) Commercial risk. Many specs have a non-performance penalty attached. It normally takes the form of lump sum deduction in lieu of either excess power consumption or lack of flow delivered. The higher the penalty, the lower the curve, in general terms. Where the benefits of advertising or salesmanship are strong but the likelihood of retrospective nonperformance claim is weak, a manufacturer may be tempted to overstate performance.

3) Test Standard. Some tests are quite lenient in terms of measurement tolerance. A skilled pump designer may deliberately undersize a pump because he knows that it will still pass acceptance testing due to the specified measurement tolerances.

So the published numbers aint necessarily directly comparable. 1 & 2 can be knocked out by actual test results. But 3 remains.

Cheers

Steve McKenzie