Calculating NPSHa

[BoomerSooner7]

Hello all,

We recently installed a high volume (3000 GPM, water) closed flow loop to test multi-staged centrifugal pumps which includes a booster pump being fed by a tank. We are using a Labview program to plot the measured performance curve against the catalog curve for the multi-staged pump being tested, everything is kosher there. However with this booster pump I want to add fuctionality to the program by calculating and displaying the NPSHa at the booster pump on the program. My inputs to the DAQ card are a suction pressure transmitter (30inHg - 30psig), an RTD and a flowmeter. My question is this; I need a formula based on any or all of these inputs plus the calculated vapor pressure of the fluid at the specified temp. via the RTD to display the NPSHa in feet. Note: the pressure transmitter at the suction impeller is in psig. Any thoughts?

 

[BigInch]

If you're OK with assuming that NPSHa is equal to suction pressure when converted to head, i.e. suction pressure is close enough to the pump to indicate NPSHa accurately enough, just do this

NPSHa = (Ps + Pa - VP) * 144 / Water_Density_pcf + E

Where,
Ps = suction pressure at the gage, psig
Pa = atmospheric pressure, psia
VP = vapor pressure, psia, roughly 0.4 psia
Water Density = roughly 62.34 pcf

E = elevation difference between suction gage elevation and the pump centerline elevation, measured in feet. If the gage is above the pump centerline elevation, the sign of E is positive, so NPSHa at the pump will be increased, so add that elevation difference. If the gage is below the pump centerline elevation, the sign of E is negative and NPSHa at the pump is reduced by that elevation difference.

If there are any fittings, valves, or long pipe sections between the suction pressure gage and the pump, the NPSHa must be reduced due to frictional flow in each of those items, if they are significant.

If you need more accuracy, you'll need the temperature of the water to determine the exact water density and its vapor pressure.

**********************
"The problem isn't working out the equation,
its finding the answer to the real question." BigInch

http://virtualpipeline.spaces.live.com/

 

[ione]

BigInch,

Perfect approach. Just one thing. The OP has got a RTD to measure the temperature, and this parameter is sometime vital when dealing with NPSHa.

 

[BoomerSooner7]

Thanks for the comments guys.

The pressure transmitter is actually mounted on the suction flange of the booster pump and it is on the pump centerline, that is as close to the impeller as we can get. And yes we have an RTD to measure the temp. of the water. Having that said, can I assume the following:

NPSHa (in feet) = (Ps (psig)+ Pa (14.7 psia) - VP (calculated by temp. from RTD))*144/Water density (62.34)

I am more or less looking at this as a reference number to warn the operator when conditions are close to cavitation. We will add a safety margin for warning levels and possibly shutdown when we get close to NPSHr

 

[ione]

BoomerSooner7,

This should work, but please be careful with this stuff, as cavitation can happen even at NPSHa values well above the NPSH required by your pump.

 

[BigInch]

Ione, As you say, if you add a even relatively very small margin for a NPSHa of 5%, there's probaby no point in doing exact calcs for VP and density. I'd add a nice margin conservative enough to include those small effects and be done with it.

**********************
"The problem isn't working out the equation,
its finding the answer to the real question." BigInch

http://virtualpipeline.spaces.live.com/

 

[BoomerSooner7]

Based on your guy's experience and the information I gave you on what we have to work with what would you consider a "safe" margin? 10% above NPSHr, 15%?....Taking into consideration like BigInch said this is a calculated value.

 

[BigInch]

Its not a safety margin for calculating NPSHa, its a safety margin for not knowing exactly when cavitation starts to affect things. Part of that, might be due to an elbow or some other fitting right in front of the pump suction that puts a spin on the water before it enters the pump, but most of using a safety margin is just not knowing precisely when cavitation starts. For room temperature water, there's not too much worry and 5% is probably good. For water at 95C, maybe 15% isn't enough.

What's the water temperature there?

**********************
"The problem isn't working out the equation,
its finding the answer to the real question." BigInch

http://virtualpipeline.spaces.live.com/

 

[BoomerSooner7]

The usual water temp. at startup of a test is around 65-68 F. Testing 1500 HP pumps at BEP for say a 2 hour period, I've seen the water reach 155 F. Our limit is 180 F due to fiberglass tanks and piping.

 

[BigInch]

That temperature difference does make a difference then.

VP will go from 0.4 to 6.8 psia.

http://www.engineeringtoolbox.com/water-vapor-saturation-pressure-air-d_689.html

Density from 62.34 to 60.58 pcf

http://www.engineeringtoolbox.com/water-density-specific-weight-d_595.html

At 180 F, VP=6.8 psia and reduces NPSHa by 16.2 feet.

If you needed 22 ft NPSHr and NPSHa was 40-16.2 = 23.8 ft, I'd be tempted to set the alarm at least at 25 ft, so ya, that's 15% or so. Best would be to get hold of an ultrasonic monitor and check if there's any indications at that level, if you can. If you get cavitation noise, raise it to 20%.




**********************
"The problem isn't working out the equation,
its finding the answer to the real question." BigInch

http://virtualpipeline.spaces.live.com/

 

[Artisi]

What is the "normal" operating temperature for system?

 

[BoomerSooner7]

Our ambient temp. is around 68 F. When we start a test it starts there and through the "normal" duration of our tests, the water will get up to around 80 F.

 

[Artisi]

Seems the argument is getting too academic - if the normal operating temp. is 80F then that is the point you should consider basing your calculations on.

 

[BigInch]

I don't think he should do that, if he can somehow reach 180F while running. He'll be cavitating w/o any alarm. He should either set the alarm on the 180F head, but that would make the alarm too low for normal operation, or to have continuous protection he must calculate the NPSHa for each temperature.

The alternative would be to set the alarm for NPSHa at the normal high temperature of 80F with a shutdown right above that.

**********************
"The problem isn't working out the equation,
its finding the answer to the real question." BigInch

http://virtualpipeline.spaces.live.com/

 

[Artisi]

My understanding is the temp will only reach 150 during an extended test run - which seems an extremely long test period in any event and how often will such extended testing be undertaken.

Sure have an alarm set point for above the normal operating temp. but monitor the "cavitation" manually during any test period which is fairly normal proceedure when testing large pumps.

 

[BoomerSooner7]

My goal for the display is a near "real-time" calculation of NPSHa since I have an RTD to measure what the temp. of the fluid is in real-time. In theory, I should be able to run the inputs from the pressure transmitter and RTD to a calculator with an algorithm that will calculate the vapor pressure and density of the fluid at any given temp. (I will have to search for the coefficients for the equation(s)) and then run the output from the calculator to the display. Now that I have a general equation (NPSHa = (Ps + Pa - VP)*144/Water density), now it's just a matter of calculating the VP and water density at the given temp, the Ps will come directly from pressure transmitter and the Pa will be a constant 14.7. I know there will be some variation and that this will not be an exact figure, but I think with enough safety margin we will be protected. Thanks for all your comments, good stuff!

 

[BigInch]

If you plot the data points for density and VP given in the above eng-toolbox links in Excel you can pull out those coefficients.

**********************
"The problem isn't working out the equation,
its finding the answer to the real question." BigInch

http://virtualpipeline.spaces.live.com/