Overall heat transfer coefficient variation

[lostragio]

Hi everyone,
I've a flooded tubular heat exchanger inside a tank for the production of cold water (0°C): inside the tubular HX flows ammonia at -6 °C. Knowing nomila heat power and effective LMTD, I could calculate U*S (where U is overall heat transfer coefficient and S is the surface): I don't know neither U nor S
For energy efficiency reasons, I want to put a pre-refrigerator (falling film hx) to chill to 1°C the water and then send it to the tank, where it has to exit at 0°C (approx): with this kind of HX is possible to rise the ammonia T to -2.5 °C. Now, my question is that: how U of flooded hx change with the rising of ammonia T? I believe that, even with a lower LMTD is possible to chill the water because of the Surface until the new U*S is lower than the "design" U*S (S is constant while U can only lower with the rising of ammonia T). Am I wrong?
Also, if I rise a bit the Velocity of water (rising the flow rate), I think that the U rises a little bit, but the reduction due to the T rising is grater than the eventual rise due to higher Velocity.

Can someone help me?
Also, can someone link me some articles or publications?

Thank you so much
Please tell me if something is not clear.

 

[StoneCold]

I think you have some basic problems with your idea.
The first is water freezes at 0C. Having ammonia at -6C will cause ice to freeze on the tube bundle.
Falling film heat exchangers are mostly process evaporators.
Heat exchangers in tanks are a poor heat exchange method.
If you want 0C water you need to add a freeze point suppression material to the water. Glycol, or Alcohol, or salt.
You would be better off with either a standard shell and tube heat exchanger or better yet a brazed plate heat exchanger.

Just some food for thought.

Regards
Stonecold.

 

[lostragio]

Hi StoneCold,
there is no mistake, maybe just some misunderstanding:
- ice tank are used in food industry to "store" some ice and avoid over-size HX. during low load period (night) the ammonia compressors work to cool the water to 0°C and accumulate ice on the external wall of the flooded HX. During peak-load the ice is used as "cool tank" to absorb peak request of refrigeration. So, therefore, water in this kind of tank is at a temperature of approximately 0°C (equilibrium between liquid-solid phases)
- this tank is existent (this is the link to see the tank, but there i no link for the HX:

https://www.baltimoreaircoil.eu/products/TSU-C-D-1050).

As explained before, this kind of HX, even if very low efficient, are very used in certain application because it is possible to have a separation between production and consumption of "fridge power": I can use a smaller (and cheaper) HX to satisfy my peak load than I would do if I choose a direct HX (like shell&tube).
So the matter is not to use it or not, but how to make it more efficient: one of the answer is to put a falling film BEFORE the flooded HX. inside the FF it will flow ammonia in phase transition (so, it is an evaporator) and on the eternal side it will be water chilling' from 4 to 1 °C (nominal condition). This kind of technology uses a FF as pre-refrigerator: in this case the flooded is way over-sized and can operate (I suppose) with smaller LMTD. As a matter of fact, FF HXs allow to work with very little cold-hot fluid temperatures differences.
So, again, I'd like to understand what happen to the U coefficient when I reduce the LMTD and if I'm right that the only thing I've to consider is to have a lower U*S product (lower than the design one);: if so, it is possible to make the refrigeration of the water.

I hope now I made myself more clear
Glad to have this technical conversation

Some other infos:

http://www.omegathermoproducts.nl/en/products/engineered-products/falling-film-chiller/

http://paulmueller.com/Refrigeration/Content/LiteratureDocument/136