discussion Not Computers , but possibly applicable.

[BrainEater]

Most of you know I work in a brewery.
I have an 18,000 litre tank of water to cool.I have a 'glycol loop' that runs at -5c.

I use a liquid-liquid heat exchanger like this :

http://img.directindustry.com/images_di/photo-g/67364-2813741.jpg

------
So .

With -5c glycol , it's possible to freeze the heat exchanger solid and destroy it . I know this because we just did it.

The system in place uses a 'flow-sensor' and a thermocouple on the water out, to try and prevent freeze, but I have issues with the design/software...I'm thinking pressure sensors would be better because they would detect the pressure drop from freezeup.

Thoughts ?

I'd love to find some concise math on liquid-liquid Hx operation too ...still lookin , if anyone has seen anything like this.....

 

[xilocient]

It would be helpful to have a diagram with all the control valves and sensors as well as how the fluid flows through the heat exchanger.

When cooling the water, are you pumping the water through the heat exchanger at full blast? If not, why not? this will cool the water down quickly and will minimize the risk of freezing. If you are still running in to freezing problems, the heat exchanger is oversized or you've left the cooling loop on too long. A PID controller on the control valve or pump on the water to the heat exchanger should solve that problem.

In any case, the coldest part of the water side of the system is on the outlet of the heat exchanger. If you can measure the water temperature leaving the exchanger and stop glycol flow when the temperature drops below say 10C, I would expect that no freezing would occur. Using a pressure sensor at the outlet might show an increase in pressure as ice crystals form, but due to the close spacing in the plate style exchanger you posted, the damage may already be done.

 

[clshades]

He needs the coldest water possible. 10c won't work for him. I would suggest flow sensors to stop the glycol in circuit with a heat blanket or heat wire to engage to protect the heat exchanger from freezing. The wiring should be pretty straigt forward as long as the flow sensor is a 4 wire for signal to the heat wire/blanket.

Thermostatic mixing valves would work to control the temp of the glycol which obviously shouldn't be lower than 0°c but we've talked about this before.

Flow sensors can be sensitive enough to detect low rates rather than too late I'm already frozen no flow. I personally think the heat exchanger needs to be protected at all costs or a different style exchanger should be considered.

You may consider a dual exchange setup to keep the glycol from the main system from mixing with the water incase this happens again... glycol aint cheap.

 

[xilocient]

You bring up a good point about a heater to recover the exchanger before it freezes out completely. I picked 10C out of thin air. You should be able to get pretty close to freezing as long as your instrumentation and control system can respond fast enough.

This post may also be helpful.

Another thing to consider is when starting up the system, start the water flow first, then the glycol. Similarly when shutting down, stop the glycol system before shutting the water flow off.

 

[Frank Encruncher]

The margin of what works and what cause damage is very small and may become very costly to over come.
Instead of straight water you might have to switch to a brine or food grade heat transfer fluid.

The brine we use on our tanks contains calcium chloride. So when your temps. drop below freezing you'll be able to keep your flow going.
There is also food grade mineral oils that you might consider like:[URL="http://lubricants.petro-canada.ca/en/products/376.aspx"/URL]

Because you're cutting it so fine, within a couple of degrees, the mechanical fixes and fine tuning to get the system to work may cost ten's of thousands of dollars to work and a couple of years. So you might want to look at changing your cool medium.

 

[clshades]

The water he is cooling is being consumed I believe... mixing in other ingredients may ruin the beer they are making. A 10,000 dollar fix is a drop in the bucket for this brewery but still an expensive adventure heh.

 

[clshades]

8.4 Freeze protection - SWEP

This is more about evaporation but the same theory applies in my opinion. If you read down a little you'll see some mention of heat blankets to protect the exchanger aided by flow sensors. You would need to build a bypass for the water under these conditions possibly with an electric valve to open an close as needed... this would definately need some engineering and deep thought.

Its too bad this system couldn't be presurized to increase the freezing point a little. If I get time i'll draw a couple picture options that may work for ya. I still think a dual exchanger system would work better to minimize the glycol wasted when things break.

 

[BrainEater]

Thanks for all the input guys !
I'll answer in order.
----

xilocient : It's got a pid controller that modulates based on output temp.It's a small part of the plc control system I helped install.Unfortunately , I did not help design it.heh.I am redesigning it !

Deano : yep . Technically the plc programming should operate as a thermostatic valve.Nice work on the copper , solid !

Frank Encruncher : You misunderstand , I'm using a glycol loop to cool a potable water tank.My glycol runs at -5C.I need to cool the water down to 1C , without freezing.

----

Here's where this sits :
Everything is peachy.I'm cooling 18,000 liters @ about 1.5C/hr !
I've already designed the 'anti-freeze'.
As a heat exchanger freezes up , the input pressure (water side) will increase.A simple 15$ pressure sensor will prevent the 10,000$ failure.

whodathunkit !!

:thumb:

 

[clshades]

I figured as much. Higher pressure = lower freezing point. Which equation is that again?