Pump head requirement: Closed vs Open loop

[rollingstein]

If a pump pumps water to a heat exchanger at a certain height, say, 20 m is the head required indeed lower for a closed loop system than an open loop system?

The link below says so, but I wanted to verify. Assume pipe friction losses are the same in both cases. Can one really take credit for the static head developing a pressure at pump suction in a closed loop system? It makes sense from the basics but intuitively I wasn't sure.


http://www.baltimoreaircoil.com/english/resource-library/file/552

 

[Baluncore]

So the link is to a pdf.
It is correct. So long as the pump inlet pipe is full of water it will provide pump inlet pressure that will cancel the head at the pump outlet. The pump presents a pressure difference that only needs to overcome head loss due to flow.

 

[Travis_King]

To expand on what Baluncore said, you'll recover your static head loss (that is, the 20 m of head the discharge must overcome is, when the system is full of water, recovered as suction head). In a true closed loop, the pump is serving to overcome only the line losses due to friction and pressure drops over valves, fittings, and equipment (like heat exchangers).

However, with heat exchangers it is often difficult to have a true closed loop system as the heat exchanger becomes less and less effective as the cooling water heats up. You can't circulate the same volume of water continually, or else your heat exchanger will lose it's ability to cool. You'll need to employ a method of cooling the water after it leaves the exchanger, which typically means opening up that closed loop. I've done this before in a semi-closed system, where my pump passed water through a heat exchanger and into an elevated holding tank where cold makeup water was being introduced. The water mixed, cooled, and was sent back down via gravity to feed the pump. Very little static head allowance was required for that pump.

 

[rollingstein]

I've done this before in a semi-closed system, where my pump passed water through a heat exchanger and into an elevated holding tank where cold makeup water was being introduced. The water mixed, cooled, and was sent back down via gravity to feed the pump. Very little static head allowance was required for that pump.

Interesting use case! Thanks!

OTOH in your case, whatever you saved on this pumps static head requirement you did have to pay via the other makeup water pump's head spec? That one has to supply the full static head, right?

 

[Travis_King]

In a typical case, yes. System savings must be looked at holistically. This application was, however, a simple and cheap solution to replace the need for a several hundred thousand dollar glycol chiller loop. The elevated tank was existing, for plant process water, and so the pumps, too, already existed. The savings there was significant.

A typical heat exchanger closed loop could be similar to this:

Pump->Heat Exchanger->Glycol Chiller (+makeup water)->back to pump

 

[rollingstein]

In a typical case, yes. System savings must be looked at holistically. This application was, however, a simple and cheap solution to replace the need for a several hundred thousand dollar glycol chiller loop. The elevated tank was existing, for plant process water, and so the pumps, too, already existed. The savings there was significant.

A typical heat exchanger closed loop could be similar to this:

Pump->Heat Exchanger->Glycol Chiller (+makeup water)->back to pump

Or if you had a short grade level open loop cooling tower with exchange to a closed loop for pumping to high equipment. Essentially saving on pumping costs.

Not sure if the economics works out though.

 

[Reive]

Interesting use case! Thanks!

OTOH in your case, whatever you saved on this pumps static head requirement you did have to pay via the other makeup water pump's head spec? That one has to supply the full static head, right?

In an open system such as the one with the open tank on a roof and the pump in the basement, is the dynamic head loss of the pump only the discharge run to the roof? The return line is fed via gravity to the inlet of the pump and all dynamic losses in that side should not be felt by the pump. Correct?