Closed system and NPSHa 1

[tikitime]

When selecting a pump for a closed loop hot water system, how do you calculate the NPSHa?

NPSHa = Static Height - Inlet Friction - 3' Safety Factor

Is the static height calculated from the highest point in the system?

For inlet friction, how far back in the pipe line do you go to calculate this?

Thanks!

 

[sailoday28]

For the closed loop system, find the pump inlet flange suction pressue. Having this and the suction flow will allow you to calculate p/rho +U^2/2 -pvapor/rho. That is the NPSHA.
Regards

 

[Artisi]

As this is a closed loop system and without being too penantic - the NPSHa is the discharge pressure from the pump less all the losses in the system back to the inlet.

 

[quark]

Artisi's reply is precise, simple and excellent.

Yes, static head is considered from the highest point in the system.

For inlet friction, the pipe starts from where you considered the static head.

3' in a closed loop - Better make it vapor pressure of water at the given temperature.

 

[sailoday28]

Artisi (Mechanical)quark (Mechanical)
Comapre my previous response with that of obtaining pump disch pressure. How is mine different?

Regards

 

[BigInch]

Sailo, you didn't mention the pump.

Simply and elegantly put, but not very specific in what one must do to calculate NPSHa.

I suggest,

to conservatively calculate the minimum possible NPSHa, (including startup and maintenance situations if applicable) take the fluid's vapor pressure at the point of lowest possible free fluid surface drawdown in the loop flowing towards the suction flange, add static head, then subtract flow losses from that.

Otherwise (if no free liquid surface can ever exist) take the minimum absolute pressure in the whole system, add static head from that point to the suction flange and subtract vapor pressure.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[quark]

Sailoday,

The procedure is correct if you get the total pressure at suction flange (by that, you are already including frictional losses).

However, the equation is dimensionally incorrect. If you express the NPSHa in meters or ft(because P/rho is in meters or ft) then velocity pressure should have acceleration due to gravity in the denominator. i.e it is actually U²/2g.

It is, often, difficult to get the suction pressure. That is why I said Artisi's reply is precise (it is easy to measure discharge pressure).

 

[TenPenny]

If you have a closed system, then take the notion of static height and throw it away. You need to know the pressure at the suction flange of the pump; since you have a closed system, the height is but one minor component of this, and will needlessly add too many complicated estimates and variances. (density of fluid, piping losses, etc etc)

You need to know the suction pressure at the pump inlet, in absolute terms, less the vapour pressure. That's your raw NPSHa, from that you can subtract your margin. Keep it simple, go to the source.

 

[sailoday28]

quark (Mechanical)Yes, I agree about being incorrect, in that divison by g is necessary.
With regard to difficulty, the designer in order to obtain pump disch pressure would have to be able to calculate suction pressure.

Artisi, Using discharge pressure, you should also include V^2/2g and elevation at pump outlet flange.

BigInch (Petroleum)I did mention "pump inlet suction flange"-=With quarks comment about g, my input is correct.

Regards

 

[BigInch]

Yes certainly. I see them as totally equal statements. One simply the reverse equivalent of the other.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[BigInch]

10P,

I don't follow how you say you can neglect static head and calculate the suction pressure, unless your system is flat.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[Artisi]

Lets get back to basics, the question was "how do you calculate NPSHa for a closed loop" For me there is no real need to calculate the NPSHa for a closed system as it becomes a function of the pump discharge head minus the friction losses for the overall system.

Assuming we are given the job to size a pump for a particular installion - we can accept that the flow and static head are known together with any head losses thru' heat exchangers etc.

If it was my job to make the pump selection, I would calculate the total head to be imposed on the pump, which in a closed system is all friction head -the static head is not involved once the system has been filled - I would then look for a pump unit that is capable of the flow at the friction head with an NPSHr less than the total head capabilities of the selected pump minus the friction head being imposed (the NPSHa for the system)
You may need to evaluate the static head to ensure the system can be filled - but this is a secondary consideration as you can always fill the system using another pump unit.

 

[sailoday28]

Artisi"As this is a closed loop system and without being too penantic - the NPSHa is the discharge pressure from the pump less all the losses in the system back to the inlet."
I would accout for the temp change thru the heat exchangers to account for vapor pressure.
regards

 

[TenPenny]

"10P,

I don't follow how you say you can neglect static head and calculate the suction pressure, unless your system is flat. "

I didn't say "calculate" the suction pressure, I said MEASURE the suction pressure. Why calculate NPSHa from elevations, when all you have to do is measure the suction pressure and determine the vapour pressure of the liquid? A simple physical measurement. If you measure the suction pressure, it doesn't matter what your system curve looks like, you're measuring the real physical deal. I've spent too many years trying to figure out why people calculate one thing, when the system is doing something else. If at all possible, measure the actual pressures.

 

[sailoday28]

TenPenny (Mechanical)You still have to convert the suction flow to velocity head and thus have "total suction pressure"
Regards

 

[25362]

It appears there are two NPSHA situations as BigInch indicated: before starting up the system, and when it is in operation.

 

[BigInch]

10P, I can't measure the suction pressure in a system that has not been constructed yet.

The real problem as I see it is, that without considering minimum static head, a closed loop system is an indeterminate system. The only way to arrive at a common and consistant method to do this is to specify what minimum (absolute) suction pressure is needed (static head, -flow losses - vapor pressure > NPSHr) and provide at least that. Otherwise, you could be talking about any of an infinite number of possibilities, with suction pressures ranging from vapor pressure to infinity.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[TenPenny]

I was thinking of an existing system, not a new design.

Since it is a closed system, all the designer has to do is pick one point in the system where the pressure is known, and work from there.