Modelling Suction Pressure Under Low Temperature Conditions 4

[KimWonGun]

Considering suction pressure of an oil pump in temperatures below -20F, test data show suction pressures that fall short of the 14.5 psi threshold. But my attempt to model this predicts cavitation because the suction pressure is calculated to be at 14.5 psi.

I am using the equations f = 64/Re and P = f*[(q/a)^2*rho*L]/(2*D). What am I overlooking?

 

[Latexman]

What kind of pump? Surging flow? Acceleration losses?

Good luck,
Latexman

 

[BigInch]

Are both of your pressures psig, or psia.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

 

[ione]

kim,

The formula you've provided in your OP accounts for pressure losses due to friction (Darcy-Weisbach), which are just a term involved in the NPSH available calculation and which enter the formula with a minus sign.


As suggested above by Latexman, if you are dealing with a positive displacement reciprocating pump you have also to take into account the acceleration head. Fluid inside a PD pump undergoes a sequence of acceleration-deceleration cycles as consequence of the reciprocating motion, which produces a pulsation phenomenon on the suction side, and when calculationg NPSH available you've to include also the head required to accelerate the liquid in the suction line.

You can give a read here to go deeper:

http://www.pump-zone/pumps/centrifugal-pumps/acceleration-head.html

 

[KimWonGun]

The pump is a constant displacement, gear-type.

The instrument used to measure suction pressure is absolute. I am unclear what BigInch is asking with "both of your pressures."

But even if the pump were a positive displacement reciprocating type, wouldn't that just contribute to the likelihood of cavitation in my model?

Just in case my description is unclear, my test does not suggest cavitation at temperatures below -20F while my model (with the Darby-Weisbach frictional pressure loss) does predict cavitation.

What am I still overlooking?

 

[ione]

What’s the vapour pressure of your fluid at -20 F?

....and yes an understimate of the acceleration head when calculating NPSH available could lead to cavitation

 

[KimWonGun]

The vapor pressure is less than 0.1 mm Hg at + 68F. Unfortunately I do not know the vapor pressure at -20F.

 

[BigInch]

Suction pressure and vapor pressure; are they both absolute pressure readings.

Vapor pressure is quite low and will most likely go lower with a lower temperature, so probably no mistake there.

You haven't said enough of anything for anybody to tell you what, if anything might have gone wrong.

First, why do you think you should not have cavitation. Your temperature is low enough to give you a viscosity for a lot of oils high enough that you will lose a lot of pressure in even a small length of suction piping, thus I might expect you would have cavitation problems at those temperatures, unless you heated the oil up to at least 25F

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

 

[KimWonGun]

Suction pressure is an absolute measurement. Vapor pressure at +20F was taken from the oil manufacturer's data sheet.

At this point I am only trying to reconcile my experimental data at temperatures below -20F with modelling equations. My test result indicates that the suction pressure at -20F is over 60 mm Hg absolute. The measured discharge pressure also does not decline. Using the Darcy-Weisbach equation, my calculated value for suction pressure at -20F is under 1 mm Hg absolute. The calculated discharge pressure correspondingly drops.

I am probably overlooking the obvious and welcome any suggestions.

 

[Latexman]

Are you using something like:

P_s = P₁ + Z₁ x [ρ] /144 - [Δ]P₁

where Ps = Suction pressure, psig or psia.
P1 = Pressure above the liquid level in the suction vessel, same units as Ps.
Z1 = Elevation of the liquid level above the suction connection to the pump, ft. If the liquid level is below the pump suction, this value is negative.
[ρ] = Density of the liquid, lb/ft³.
[Δ]P₁ = Frictional losses including contraction and enlargement losses in the suction vessel and line.

Good luck,
Latexman

 

[ione]

Are you using consistent units for all the terms used to compute the NPSH available. What I mean is that in your previous posts you were talking about mm Hg for suction pressure and vapor pressure, have you used the same unit for the elevation of the liquid level above the suction connection and for friction losses?

 

[BigInch]

Well ... there are better equations to use for low pressure flow too.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

 

[ione]

Kim,

Have you taken into account the change in density due to the actual temperature (-20 °F)?

 

[katmar]

Problems like this are always much easier to understand if you start off with a drawing. At this stage we have no idea what your suction piping looks like, and that would be the very first thing that I would look at - before I tried to do any calculations.

Just for the sake of argument let me assume that the oil comes from a tank and flows through a simple pipeline to the pump suction. If the oil level in the tank is 3m above the pump centre line and the oil's SG is 0.85 then the absolute pressure at the start of the pipeline would be about 950 mmHg. You say that the measured pressure at the pump suction flange is 60 mmHg abs and the calculated pressure is 1 mmHg abs.

In other words, your measured pressure drop is 890 mmHg and your calculated pressure drop is 949 mmHg. This is a difference of about 6%. I doubt very much that anyone can calculate the pressure drop to within 6% for a suction pipe taking into account a valve or two, entrance effects, some pipe fittings that may include a reducer and so on. I would not back myself in this case.

Substitute your real numbers for my hypothetical example, tell us what the suction piping and supply looks like, and then we might be able to help you.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[KimWonGun]

I hope the attached sketch helps illustrate the set-up sufficiently.

The reservoir is unpressurized.

I double-checked my units and believe all relevant variables to be consistent.

Because the oil density changes slightly, I had assumed a constant value at room temperature. But even though the effect is small, I will incorporate it in my model.

The oil level in the reservoir is only 1.5" above the pump center line. The oil's s.g. at -20F is about 1.026. The absolute pressure at the start of the pipeline would therefore be about 750 mm Hg. The measured pressure at the pump inlet is 60 mm Hg abs and the calculated pressure is 1 mm Hg abs. Therefore the measured pressure drop is 690 mm HG and the calculated pressure drop is 749 mm Hg, an 8% difference.

Again, my objective is to reconcile test data with simulated data based on the Darcy-Weisbach equation for pressure loss due to friction. Experimental results at temperatures below -20F do not indicate the liquid pressure dropping below vapor pressure while the simulated results suggest the opposite.

Any additional suggestions would be appreciated!

 

[ione]

You've not told us the flow rate. And why should pressure be 750 mmHg instead 760 mm Hg if the tank is atmospheric?

 

[katmar]

The information you have now given exactly confirms what I had calculated before. If you work backwards to calculate the pipe ID you will see that to get a pressure drop of 690 mmHg instead of your calculated 749 mmHg the pipe ID would have to be 1.02" (everything else being held constant). This sort of variation is quite likely unless you have had your pipe made up by a gun smith.

Really what I am saying is that it impossible to get a more accurate prediction than what you already have and you should just accept the measured value as being correct.

The one parameter you have not given so far is the viscosity of the oil. Based on some guesses on the viscosity I am getting rather high velocities in the pipe. Perhaps you need to give full details on the problem.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[KimWonGun]

Thanks for catching my error; I should indeed have used 760 mm Hg.

Based on a measured pressure drop across the (square edge) orifice of about 35 psi using a discharge coefficient of 0.74, the flow at -20F is about 20.1 in^3/s.

The kinematic viscosity is about 2,504 cSt at -20F.

My test velocity is 3,600 RPM.

Before drawing any conclusion that the comparison between test and simulated data for pump pressure at temperatures below -20F is satisfactory --- despite test results indicating that the pump pressure remains high and the liquid pressure does not fall below vapor pressure while the calculated result predicts that the pump pressure drops dramatically due to cavitation --- I wanted to confirm I did not overlook anything critical to affect the result.

 

[ione]

With your data I got:

v = 0.65 m/s
Re = 6.59345
f = 9.706
friction losses = 939 mmHg (which is approx a 25% higher than what you've calculated).

 

[ione]

Kim,

Are fluid characteristics at working temperature of -20 °F measured values or extrapolated values? What I mean is that an actual viscosity lower than that you’ve used (2,504 cSt) will obviously lead to lower friction losses.