Odd behavior of chilled water circulator w/ high elevation gain

[TTM2]

See attached PDF for general piping layout. The idea is that the chiller unit cools the incoming water which is then pumped ~100ft in the air to cool a piece of equipment at higher pressure. My return line is filled with bubbles and unpredictable behavior which I have never observed in lab settings.

I had run this system in a shop before, all components tabletop with minimal elevation gain, with water. I was able to tune the bypass needle valve to hit the flowrate I needed, while staying within our pressure ratings of the pump, and there was no funny business in the water return line (bubbles, cavitation, noise etc).

Now it is installed in "real life" and must move 50/50 water+glycol up 100 feet to cool a piece of equipment in 3/8-1/4 copper tube. The return line has the flowmeter (standard rotameter with clear plastic so bubbles are easy to see). My return line is filled will all kinds of nonsense like chaotic bubbles, jumps, whooshing. Very inconsistent.


At first I thought the bubbles were trapped in the line and working their way out, but after running ~4 hours they are as bad as ever. Is there cavitation in the line, or lowering of the vapor pressure on the return line downwards forming bubbles? I need to nail a certain flowrate and I need to use my bypass needle valve to tune it.

http://files.engineering.com/getfil...40-9b29-efdf9ebb6deb&file=chiller_diagram.pdf

 

[TTM2]

What kinds of effects would the 100ft elevation gain introduce which I didn't consider in the shop setting?

 

[Compositepro]

Your vented reservoir must be at your highest elevation or pressurized if it is lower. Currently your return pipe is 100 ft. tall and vented to atmosphere at the reservoir at the bottom. Atmospheric pressure can only support a column of water that is 32 ft. tall, not 100 ft. Water is cascading down the pipe.

 

[TTM2]

Thank you for the reply. A problem which i suspected but couldn't articulate.

Let me change the scenario on you. We have a twin system which is acting just about right, maybe such that the siphon analogy doesn't fit..?

System #2
Same plumbing, but the low pressure drop part at height (100ft) is exchanged for a high pressure drop device (many small capillary tubes for radiant heat dissipation).

The outlet pressure is ~175 Psi with flow rates ~.5 L/min, fully closed bypass valve. Same vented reservoir. The flow back to the chiller is more uniform, lacking bubbles. Is a cascading chug-chug-chug happening, but damped out before reaching the rotameter?

If we are in a pinch, it seems to me as long as we can say there is orderly flow to our target device then we can feel confident heat is dissipated (although precise flow rates may not be known - delta T over device IS known however.)

 

[TTM2]

Last question - is it reasonable to assume nothing will be damaged if the cascading falling action is allowed to continue through copper return tube, back to the refrigerator reservoir?

 

[MikeHalloran]

Put the rotameter in the 'up' leg, going >to< the 'item to chill'.

Plug three of the four atmospheric vents you've shown; I'm not sure it matters which three, but I'd certainly plug the two at the top.



Mike Halloran
Pembroke Pines, FL, USA

 

[TTM2]

I wouldn't say the at-height vent valves qualify as an atmospheric vent in the same way the reservoir would. See these items:

http://www.mcmaster.com/#air-vent-valves/=pgvzfg

 

[LittleInch]

As composite pro says, you either needs to place the reservoir at the top of your circuit or make it a pressurised loop. Your pump will then probably be oversized s all you need to pump against is friction and not the static height it is currently working with.

An alternative is to place a back pressure valve at the foot of your riser to maintain a pressure equivalent to 100ft, about 3bar, 45psi. This will fill your pipe and maintain it above atmospheric pressure. Long term running in this mode will cause damage, fatigue of bends and equipment and is generally a bad idea.

By the way, your pump looks like it is operating way off the end of it's curve, but at that power probably won't matter too much.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[1gibson]

What diameter is the pipe that drops 100ft down? If it is small enough then it is possible the falling fluid can pull a vacuum near the top of that pipe, and the fluid will boil at relatively low temperature. Similar to (and could be argued that it is a form of) cavitation.

 

[TTM2]

its 3/8'' copper dropping down to the chiller.

Ii suppose the best question would be: is there an off the shelf component to prevent this action which doesn't involve a pressurized reservoir? LittleInch mentions a back pressure valve - does this go on the return side before rotameter?

Photo of rotameter attached.

http://files.engineering.com/getfil...-e830-4983-8c3e-f9e440844235&file=bubbles.JPG

 

[zdas04]

The minimum flow rate for 3/8 tubing to run full is something over 2.4 L/min. For 1/4 tubing it is 1.2 L/min. Your flow is 0.2 L/min. The line is not running full. The impact of this is that the falling liquid tries to expand to fill the void space, with some of it flashing, all of it going really fast. You need smaller pipe or bigger flow rates (and associated bigger dP) to fix this.

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

 

[1gibson]

Would it help to zig-zag the copper tubing (45* bends, 1 ft linear tube length between bends) to limit the velocity and reduce the "free fall" of the fluid? Maybe 45* every ft in is excessive, but strike a balance between eliminating the free fall, and adding too much restriction.

 

[zdas04]

Not a bit. The result of that would be to accumulate slugs that would tend to launch with every bit of force a couple of ounces of water can contain (not huge, but troublesome).

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

 

[TTM2]

Speaking with chiller manufacturer tomorrow morning about our issue, also ran it past our pipe engineer.
They both have beef with flowmeter placement and air vent valves. Its a little troubling that neither seem to be particularly worried about the 32 foot water column limitation. I might have made some incorrect assumptions about how the chiller and reservoir are vented.

Developing....