At work, a lot of the machines use chilled water loops to carry heat to and from chillers, which then regulate process temperatures. Yesterday, one of these machines decided that it couldn’t quite cool to the process setpoint. The tech working on it noticed that the chilled water flow was set really high on this machine, so he dropped the flow. After this, the machine was able to run through the shift without any problems. FWIW, I remember doing the same thing to another machine.
My question is, can flow affect heat transfer, and if so, how? The process was set to room temperature +10 degrees and the chilled water loops flow at only a few GPM, so I doubt that the chilled water was cavitating inside the machine. My best guess is that reducing the flow decreased the turbulence inside the pipe*, and this gave us just a little more heat transfer. That, or it was just a coincidence. Does anyone have a (more educated) opinion on whether flow rate would noticeably affect heat transfer?
Question #2:
Skin effect = turbulent flow, for flow in a straight pipe …right?
[sub]*No, to point out the obvious, I have never taken a fluid dynamics course. I’d like to, but…[/sub]
Certainly flow rate affects heat transfer; in convection, fluid velocity is the most important variable. But usually it goes in the other direction - higher velocity equals higher heat transfer rate. Turbulent vs. laminar flow has an effect too, but turbulence also tends to increase heat transfer.
Cavitation is a reasonable idea, but you’d have heard a loud noise like marbles going through a blender, from the gas bubbles collapsing. It’s possible that the flow rate was so high that there was a lot of air entrained in it at its source - the heat from the process wouldn’t transfer into the air bubbles as well as into the “solid” water. Can you see the intake or discharge of the water and check that?
I keep re-reading your question, and have decided I can’t tell the following: was the flow on the loop between the chiller and the machine what was reduced, and was it the temperature of this loop that subsequently was lower? This is sort of trivial, because if you lower the flow in this loop and measure it where it leaves the chiller or enters the machine, of course you’d expect it to be cooler. But I couldn’t rule it out as I understood what you wrote…
Napier,
We reduced flow in the chilled water line. This is part of the incoming facilities and is used to carry heat away from the chiller. After reducing flow, the chiller was able to control the process.
Shiva,
I’ve never really had a reason before this to look at the flow, but I think it runs in the range of 2GPM to 4GPM. It’s a 3/8" chilled water loop, which should give about 100-200FPM, if I got the math right.
ElvisL1ves,
If the water was hot, like the coolant in an engine, or flowing at a ridiculously high rate then I could see it cavitating. There are a few 90 degree elbows in the line, and the tubing makes a few tight 180 degree bends inside the exchangers. Something might be happening here which we can’t hear (considering that this stuff is in a factory, it’s a real possibility.) I still find it hard to believe that the pressure in this loop could drop to a point where tepid water would boil.
I would guess that flow through the chiller is unregulated and by reducing the flow through the hot machine, it reduced the flow through the chiller.
This allows the water more time to get cooler in the chiller, making the chiller the more efficient machine. Efficient in the way that it takes heat out of the water better than the hot machine puts the heat into the water.
Is the chiller controlled by a temperature sensor, and where is that sensor? Is it the same sensor that told you the machine is overheating?
There are some chillers with the temperature sensor (thermostat) on the coolant return pipe. The assumption is that the coolant coming out of the machine is close to the temperature of the machine. But if you increase the flow rate, the coolant warms up less as it passes through the machine, so the chiller thinks the machine is cooler than it really is. It would compensate by slacking off and sending warmer coolant to the machine. I think this would result in the machine temperature going up. That’s my theory, anyway.
Is it also possible that the water pump adds some heat when the flow rate is higher?
Cornflakes, re cavitation, it’s a very local phenomenon occurring in regions of very low local pressure (like on the back sides of pump impeller vanes or boat propellers). The gas bubbles implode and return to liquid as soon as the flow pressure recovers. For that to be the problem with apparent low heat transfer into the water, the process would have to depend largely on dumping heat into the restricted areas you describe, not into straight or large-radius runs of pipe. I doubt that’s what you have.
Sorry, but I think I left some confusion regarding how the chilled water loop runs (I think that PCW is a standard acronym, mind if I use it?) The PCW run does not directly cool the process. Process heat is dumped into a separate loop which runs between the chiller and the machine. The chiller controls the flow of heat from this line into PCW.
honkytonkwillie and scr4, the tech dropped the flow in the PCW run, not in the process loop. Your explanations would work if the chiller was cooling the PCW for use in the process, but since the PCW is acting as a heat sink, any loss in heat transfer per unit volume should be made up for by the increased volume, right?
I can’t rule out that there’s something else wrong with the machine. This chiller is at a pretty mild step in the process and it might take a while for things to go out again.
Also, thanks for the interest and replies. Somehow, I forgot to say that before now.