Pump Performance Cure Shifting

Centrifugal Pumps and Viscosity

If the pump is being tested, then there is no need to do anything else other than wait for the results.

Sir,

thanks for your response.thing is we've test data, for some 2 different temperatures. But for a wide range of variations in temperatures instead of testing (which is costlier and time consuming) we are trying to do theoretically(analytically).our flow medium is aviation fuel(JetA)

You should expect that as temperature increases your pumps internal leakage rate will increase. Fluid has a higher viscosity at lower temperatures. As you increase the temperature, the viscosity of the fluid will decrease, allowing it to pass through small opening/clearances easier, and at a higher flow rate, thus increasing the leakage in the pump, and decreasing pump efficiency.

That's the reason that the motor oil in your car should be matched to the ambient temperature you plan to run your engine. Too cold and the oil has a hard time getting through tight clearance spaces, too hot and the oil becomes very thin, and does not lubricate properly.

I advise to simply wait for the performance curves.

Sir,

thanks for your response.thing is we've test data, for some 2 different temperatures. But for a wide range of variations in temperatures instead of testing (which is costlier and time consuming) we are trying to do theoretically(analytically).our flow medium is aviation fuel(JetA). please provide me some analytical approach.

The trail-blazing testing being carried out now will enable the <...analytical approach...> required to be actually determined. However, rather more than two test temperatures would be needed to obtain anything of significant meaning; all that two data points produce is a straight line.

In this scenario, extrapolation beyond the test range is unlikely to produce an accurate projection, particularly if the test temperature is close to the boiling temperature of the fluid at the pressure being achieved.

It is actually the viscosity (flowability) and not the temperature that determines the close clearances.

As is well known liquid viscosity generally decreases with increasing temperature. Hence, you need to have the viscosity characteristics / viscosity curves of the fluid (plotted against temperature) being pumped and interpolate the data accordingly.

There is no direct correlation or extrapolation data available for temperature scales. It is only through viscosity that the two are connected.

Satish

Ref various posters' comments about viscosity - you haven't said what fluid you're pumping. If it's water, there is a useful formula for viscosity as function of temperature.

μ = A*e**(B*T/(C + T)) where μ is centipoise, A = 1.77858, B = -4.20645, C = 128.317.

I don't know whether there is any simple formula relating pump performance to viscosity.

Cheers........Codey

sir,

thanks for your response.our flow medium is aviation fuel(JetA). i've the plot between viscosity and temperature for this

Yes, for known changes in viscosity , performance curves can be plotted wrt to the original curves.

How? If you have a pump curve at a given speed and want the curve at a different speed, r x original, for each point Q, H plot new points Q_new = Q*r, H_new = H*r².

Do you know of something similar for viscosity?

Cheers.......Codey

Have a look at these posts from a recent thread on centrifugal pumps. #11 & #14 . It is not a straightforward formula, but it is normal to calculate and draw a viscous curve when selecting a centrifugal pump on a viscous application. Irshad is right, you would calculate performance for viscosity not for temperature, and there are also charts available giving viscosities for different fluids at different temperatures.

sir,

thank you.could you explain in details.

sir,

thanks for your response.our flow medium is aviation fuel(JetA). we've the known viscosity at different temperature.i've viscosity curve plotted against temperature.(ex. @25⁰C - viscosity=1.5 mm²/sec & @80⁰C=0.71 mm²/sec). for example i've pump performance curves at 25⁰C only. how to plot it for 80⁰C. please explain in detail.

I suggest that you follow this thread and pay especial attention to pages 5 through 7. The Specific Gravity of a fluid has a great deal to do with the TDH of a centrifugal pump as well as the viscosity. The temperature of the fluid will affect both the S.G. and the viscosity of the fluid. In most cases, for typical fluids including water, the S.G change is more significant than the affect of the internal leakage or increased internal losses from viscosity changes. The selection of a pump must be made from the predicted operating conditions which would include the S.G., the delta pressure required and the operating temperature in order to properly select the pump capable of the head requirement.

http://www.peerlesspump.com/System%20Analysis%20for%20Pumping%20Equipment%20Selection%20B4003.pdf

"The Specific Gravity of a fluid has a great deal to do with the TDH of a centrifugal pump"

No. SG has nothing to do with head. SG affects the pressure and absorbed power, but not the TDH. Read again your link; at the bottom of p5 it states "Specific gravity affects the liquid mass but not the head", which is correct. Also see this recent thread.

As SG goes up motor amp draw rises by the same ratio to produce the same head.

Amp draw goes up because the larger S.G means that the differential pressure is going to increase hence the BHP requirement is going to increase also.The higher S.G. means more weight is being pumped (volume will be virtually same but weight throughput is increasing). Technically unless there is sufficient reserve HP in the driver a significant increase in the S.G. can cause the driver to slow down and at a lower RPM the head will decrease until overloads kick out.

Head is a unit of pressure expressed as a column of liquid, in Feet, instead of the
usual unit of pounds per square inch gauge(PSIG). A pressure in PSIG is simply
converted to its equivalent column of liquid. That column of liquid develops the
same pressure in PSIG at the bottom of that liquid column. The length of that
column, in Feet, is referred to as the head.

There are many reasons why it is preferable to express a pressure in terms of its
equivalent head in Feet, or column of liquid, rather than in PSIG:

• A pump performance curve is usually plotted based on its flow rate in
  gallons per minute (GPM), and its head in Feet. The unit of head is used
  because a pump always develop the same head regardless of the specific
  gravity of the liquid.
  

If the unit of pressure in PSIG is used, instead of the head (in Feet), then
the pump performance curve will change every time the liquid specific

You are correct that "head" has nothing to do with S.G., however, since I'm pretty sure that the above test is not measuring "head' but is measuring delta pressure across the pump and that S.g. at the test temperature has a lot to do with the results being obtained. A conclusion that temperature affects head would be an error without knowing how the temperature affects the S.G. of the liquid being pumped.

A pump impeller is designed to produce a given head at a given RPM and at a given diameter and is affected by internal losses of the pump itself. All of the test setups that I have been witness to measured pressure differentials, temperatures and used water as the test medium since it was the industry standard and its properties and S.G. at the test temperatures was the most predictable. The gpm, kw and rpm were also factors that were readily measurable as part of the test procedure. No one ever directly measured for head as far as I can remember.

It was in that context that I was referring to needing S.G. as opposed to the purist use of the the term "head'. If they are using the delta pressure then then need the S.G. in order to compute "head".

"SG has nothing to do with head"

(rider) in a momentum based pumping methodology

Sorry, I don't follow (I'll use jetlag as an excuse). Please explain.

In a positive displacement pump the density of the media is not and 'balanced out' as it is in a momentum based 'impeller' system.

For instance if you were pumping mercury.

I can see what you are driving at, but I don't see the relevance. If anybody wanted to convert the output of a positive displacement pump from pressure to head, then yes, you would have to account for the SG. But I have never ever heard anyone refer to the output of a positive displacement pump in terms of head. It doesn't make sense since it is pressure that matters, and I can't think of any reason why anyone would want to convert it to head. The only reason that 'head' is used for centrifs is because that is what they produce; a fixed head, according to the pump curve, irrespective of SG (whereas a positive displacement pump gives a fixed pressure, according to the pump curve, irrespective of SG).

Besides, the thread concerns centrifugal pumps, and I was commenting on a reference to 'the TDH of a centrifugal pump'.

I'd be inclined to say a positive displacement pump gives a fixed flow, irrespective of SG and pressure (up to design limit). Pressure is determined by the system resistance. The pump curve is a vertical line in theory, not quite vertical in practice.

Cheers........Codey

A good point, Codey. I can't argue with that. Although for PD pumps where slip is involved (eg. gear pumps) the pump curve often ends up looking not dissimilar to a centrif pump curve. The axis' are usually reversed, and flow drops as pressure increases.

But my point was to do with the effect of SG on pump performance, and a better way of putting it might be to say that SG has no influence on the pump curve.

And this all stemmed from Spinco's misuse of the term TDH. I'm sure he knew what he meant, but I don't like to see head and pressure confused.

More a comment for the viewers who often go away thinking this is a rule for all pumps.

Not disagreeing, or attempting to argue, just mentioning a 'rider'.

Yes, it's probably not a bad thing to point out the difference.

I do not know how to attach afile. I have gor some material regarding this issue. The file is pdf but text can not be selected.

Make arrangements with the original poster to exchange documents via e-mail. This is best done using the private messaging facility within CR4 and not in the main forum, which attempt would run contrary to Rule 7.