chiller pump problem

[HARRY 7]

This is the drawing of the circuit:

[HARRY 7]

this is the drawing of the system:

[HARRY 7]

I believe i did it now

[nike123]

Relocate your expansion vessel to roof near pumps inlet and your problems will disappear.
Pump your expansion vessel (air side) to 1,5 bar and ensure that pressure when system is inoperative and cold at lowest temperature is about 1,6 bar at pump inlet (water side).
Than operate system and check that at suction of pump pressure is about 1,5-1,6 bar. If not, add water.
Than measure pressure at point where you refiling system and adjust refiling regulation valve to that pressure.
Purge system of air and you are done.
Check that pressure doesn't exceed admissible pressure of weakest components.

[HARRY 7]

So you beleive the expansion vessel does the
problem in the system??

[nike123]

So you beleive the expansion vessel does the
problem in the system??

Yep!
...........

[HARRY 7]

Can you expain me technically the reason?

[nike123]

http://www.refrigeration-engineer.co...7&postcount=20

[HARRY 7]

I will try it nike and i will let you know.

[Sandro Baptista]

Hi Harry. How it's going?

[HARRY 7]

Hi Sandro.I am going to put
the vessel and the valve in the roof.
When i finish i will inform you!

[49ershvac]

sound to me there is a leak somewhere.do you have a water make -up tank

[HARRY 7]

There is no leak somewhere!the client has left the system works as it is!

[DTLarca]

Can you expain me technically the reason?

The pump creates static pressure sufficient to overcome the frictional pressure drop of the water circulating the system.

Water is incompressible. Also we do not expect the water to expand - there should not be room in the piping for water to expand. So we cannot expect the pump to drop the pressure of the water at its suction side.

Lets imagine a chilled water system where the water flows around a circuit which at all its points is at the same elevation. This means, say, that when the pump is not running the pressure right around the system is 1 bar.

Then when the required water flow rate is happening we might have an evaporator pressure drop of 0.5 bar, a piping pressure drop of 0.2 bar to the fan coil then a pressure through the fan coil of 0.5 bar and a pipe pressure drop back to the pump of 0.2 bar.

The total system frictional pressure drop is 1.4 bar.

When the pump starts it sucks water in from behind it. But the water can't expand to drop its pressure there - if it wanted to expand it would have to fight against the rigidity of the pipework. So in fact the pressure at the suction of the pump should be 1 bar both when it is off and when it is running.

Pressure at pump suction when off is 1 bar and when on is 1 bar.
Pressure at evaporator inlet when pump off is 1 bar and when on is 2.4 bar.
Pressure after evaporator when pump is off is 1 bar and when on is 1.9 bar.
Pressure before fan coil when pump is off is 1 bar and when pump is on is 1.7 bar.
And so on.
But at the inlet of the pump the pressure should always be 1 bar - for this hypothetical project.

Because the pressure should always be the same at the pump inlet we should connect the pressurization unit to that point so that it's pressure sensors and settings can have a nice neat narrow dead band.

But why connect the expansion vessel at the pump inlet?

Well, the vessel is sized to accept the thermal expansion of the water in the system. If it was, for example, located on the discharge of the pump then when the pump starts and so the water pressure there increases it will flow into the vessel compressing the bladder thus removing some water from the piping circuit and thus it would cause the pump suction pressure to drop when normally it doesn't change. So then the pressurization unit kicks in to raise pump suction point pressure. Now the bladder is half compressed meaning before the system's water has risen in temperature we have already lost half of our expansion space which means that later when we do need the expansion space we end up instead blowing the relief valves.

[DTLarca]

When I say we do not expect the water to expand I mean we do not expect the water at the suction of the pump to expand in the same way we expect air to expand on entering a fan or like helium would when being released into a deflated balloon. I'm talking about when the pump starts - not when the water temperature changes.

[DTLarca]

There is another reason why we connect the vessel to the suction and not the discharge and this reason is similar to the reason why you need larger compressed nitrogen vessels for heating circuits as you do for chilled water circuits. In fact in most designers minds this other reason is more important that the first reason I gave above - anyone know what it is.

thinking about that would lead onto considering how you select a vessel size wise - the formula unfortunately involves an iterative process but understanding the theory still takes a person a very long way down the road of understanding chilled water system commissioning and service works.

[Gary]

I increased the pressure from 1.5 bar to 2 bar in the roof from the water pressuse adjustment in the basement.

Keep adjusting the water pressure in the basement until there is .5 bar suction pressure when the pump is running.

And yes... the expansion vessel should be connected to the suction side of the pump.

[HARRY 7]

Thanks for the explanation DTLarca!
Gary i have 5 bar in the basement when the system is working,you tell me to increase more?

[Gary]

The basic problem is that the feed water and expansion tank should be on the return (suction) side, not the discharge side.

I'm wondering if the entire problem could be solved by reversing the direction of the pumps, pumping out of the chiller instead of into the chiller?

[HARRY 7]

We made an offer to the client to remove the feed water and the expansion tank to suction side.He has not accepted it yet!

[Gary]

In your drawing, the feedwater and expansion tank are teed together and then connected to the discharge. If that tee was instead connected to the return, the problem could be resolved without moving anything. This might be the simplest solution.