piping system curve 1

[terry100]

we have a closed loop cooling water system with an expansion tank before the pump feeding the chiller. should this be included when making a system curve?

 

[Artisi]

Yes, the losses through the tank are an additional loss in the pipe system.

 

[Latexman]

If the return enters the top or top side above the liquid level, yes, include the lift needed.

If the return enters below the liquid level, yes and no. There is no lift needed, but you need to include the entrance loss into the suction piping, which is at the tank, and the exit loss from the discharge piping, which is also at the tank.

Use Bernoulli's theorem and you'll see what I'm talking about.

Good luck,
Latexman

 

[briand2]

You may lose something like 1 velocity pressure on entry to the tank, and 0.5 velocity pressure on exiting the tank. Assuming your pipe velocity is something like 2.5 m/s (probably a bit high, but I'm trying to be conservative with regard to estimation of losses), your velocity pressure is (density x (velocity^2)/2), i.e. about (1000(kg/m3) x (2.5(m/s)^2)) / 2 = 3125 Pa, so 1.5 of these is about 4700 Pa (or 19 inches of water, depending upon your favoured / favored units).

Brian

 

[katmar]

A star to Brian for putting the problem in perspective.

 

[Latexman]

Well done Brian!

If the return to the tank enters above the liquid level then a hydraulic lift is required that is dependent on the size and shape of the tank. This may or may not add significantly more head to the system curve, it depends on the tank.

Good luck,
Latexman

 

[Artisi]

If the system is closed as indicated in the OP - once flow is established the only head involved in the system will be friction and entry losses. Static inlet head equals static discharge head.

 

[Latexman]

Again, it depends where the return liquid is discharged to the tank. If it is discharged subsurface, you are right. If it is discharged into a top nozzle and it falls through an air gap to the liquid surface, the pump has to provide this hydraulic lift according to Bernoilli's Theorem. I've seen many expansion water tanks where the water is "splash returned". This provides some mixing in the tank and cuts down on biological growth.

Good luck,
Latexman

 

[briand2]

Latexman is absolutely correct; I didn't allow for static lift as I assumed from the wording of the post that the system was chilled water and the expansion tank would be totally full, as is typical for the systems I come across in HVAC.

If there is a static lift, it must be allowed for as Latexman states (equivalent to serving a cooling tower, for example).

Regards,
Brian

 

[Artisi]

No argument from me if there is a lift involved - like briand2 I assumed from the wording it was a closed loop.

 

[Latexman]

Could be a terminology thing. I have more experience with cooling towers, but I have seen a chilled water tank in the Middle East whose return was splash filled from a top nozzle. I thought either way was still a closed system. Neither is an open system. To me an open system is the same thing as a single pass system, like taking river water into a heat exchanger and then discharge it back to the river downstream of the intake.

Good luck,
Latexman

 

[RossABQ]

Am I missing something here? On a closed loop system, I would expect an expansion tank to be a closed vessel with a diaphragm divider and air pressure above the diaphragm. I am talking about an Amtrol type vessel. There is no flow into or out of the vessel to speak of, in fact there is only an inlet connection. In which case the only affect on the system curve is that the air pressure above the bladder sets the minimum pressure of the system as a whole. There is no friction loss here other than the Tee where it connects.

 

[Artisi]

Good point RossABQ if it is in fact a diaphram style expansion tank.