Relocating a chilled water expansion tank

[packdad]

We have a chilled water system with a closed, pressurized expansion tank located near the mid-elevation of the system. We want to relocate the tank above the topmost elevation of the system so that it may be vented to atmosphere and not require air overpressure.

The current tank ties into the system near the pump suction. The new tank would tie in on the return side of one of the uppermost air handling units.

A comparison between the current tank overpressure and the elevation of the new tank indicates there would be little change in system pressures as a result of the move.

Are there any other concerns associated with a change like this?

 

[mauner]

packdad,

The major concern is a loss of NPSH for the chilled water pumps.

The elevation head is applied at the tee where the tank joins the system. The absolute pressure at that point would be 33.9 ft + height of water above connection. You then subtract the head loss, based on flow velocity, fitting K-factors, pipe length, etc all the way down to the pump and add the pressure due to the elevation change from the connection point to the pump. This is the system pressure used to calculate the NPSHa.

Right now the headloss for the connection at the pump suction is essentially zero(unless it is upstream of an isolation valve). I would expect the return header pressure loss to be anywhere from 20% to 50% of the pumps total developed head. If you assume the head loss is 15 psid it will essentially cancell with atmospheric pressure on top of the tank. This will result in a system pressure = height of water in tank (~5ft) which can be compared to 86 ft if the original tank was pressurize to 20 psig. The change in available system pressure may result in pump cavitation. A new pump may need to be purchased to meet the new calculated NPSHa. As an alternative, you could run a new pipe from the old connection up to the tank at the new location. The new pipe would not be part of the flow stream and wouldn't have any headloss.

 

[quark]

I don't see any process change in the alternative, unless you are pressurizing the system to an extent than required, at the moment.

In the obsence of an expansion tank, the pressure at the highest point of a return header is generally atmospheric and pressure starts to decrease further as the water goes towards the pump suction. This negative pressure can be negated by either placing a pressurized tank in the system or an atmospheric tank at the highest point.

I don't have a clue why the pressurized tank was placed in the midway of the return piping. My obvious choice is, always, to keep the tank at the highest point.

 

[packdad]

In this case, the friction head loss between the new tank tie-in point and the pump suction will be more than compensated for by the static head of the tank, so we are OK on pump NPSHa. Calculated dynamic losses between the new tank and the pump are around 22 FT whereas the elevation difference will be over 60 FT. NPSHr is only 9 FT for our pump.

What I haven't done yet is gone step-by-step from the tank to the pump suction to make sure there isn't a (high) point in between where we might encroach on the fluid vapor or air dissolution pressures, but I highly doubt this could be the case.

There wouldn't be any strange dynamic effects as a result of moving the tank, would there? No surges into or out of the tank as the pumps starts or stops? I'm just trying to think of anything subtle that might trip me up later.

I have no idea why the system was designed with a pressurized surge tank, either. From talking to ex-Navy folks, this is apparently a typical design for subs (or anywhere you don't have access to a high point, I suppose). But this is a building!