Improving cooling jacket efficiency

[KangCy]

Hello,



I have a mixing tank 1.7 m3 half covered with plain cooling jacket (heat transfer area 2.44 m2).
I tend to use the mixing tank for a reaction that is exothermic (ca -10kW), for that reason I want to check if cooling duty of chiller and jacket is sufficient.
We carried out some tests and the results showed that jacket had a cooling rate of 2.3 oC/h dropping temperatures of the reactor content (400L) from 15 to -10oC in 11hours. That seems extremely slow.

Notes:
[ul]
[li]Reactor content: 400L of 30% IPA[/li]
[li]Reactor size 1.7 m3 and filled to 0.4 m3.[/li]
[li]Tank contents were mixed during cooling and process was batch.[/li]
[li]Jacket was insulated, pipes of 4 m were not insulated.[/li]
[li]Coolant: 50% glycol[/li]
[li]Coolant rate 38 m3/h[/li]
[li]Coolant initial temperature -2 oC and cooled down down to -20 oC.[/li]
[li]Coolant buffer 1 m3 was used. The coolant was circulated from buffer to chiller to jacket and back to buffer.[/li]
[li]Jacket coolant velocity is 0.2 m/s (seems low).[/li]
[li]We found out that chiller has a cooling duty of 12 kW at temperatures input of 15oC and drops to 5 kW at temperatures input around -5 oC. It showed that cooling duty decreases as temperature input to chiller decreases. Kind of makes sense.[/li]
[/ul]

Then we tested the cooling efficiency of the jacket, showing very bad results, of cooling duty ranging from 2 kW (at temperatures around 15 oC) to 0.3kW at temperatures below 0 oC. This duty was calculated measuring the temperature content of the reactor over time, taking into account the mass and specific heat capacity of reactor content.

The temperatures input and output to the jacket were measures but due to calibration errors they were not showing any reasonable temperature difference to be used to calculate duty. In order to get such a low cooling duty the temperature difference should be around 0.01 oC.

In order to understand if this is reasonable, I calculated the U-value of the cooling jacket. I used the following tool: CheCalc ‐ Jacketed Vessel Heat Transfer and I came up with a U of 140W/m2 K (quite reasonable for a cooling jacket).

The effective heat transfer area of the jacket was measured to be 2.44 m2 (the liquid in the reactor was touching half of the cooling jacket. With an estimated LMDT of 10-12 oC (that is the difference throught the tests between cpolant input temperature and reactor temperature), the estimated duty was calculated at 3.42 kW. So I don't understand why in reality the cooling of mixing reactor is a lot less efficient.

My questions are:
[ul]
[li][/li]
[li]What do you think are the reasons of this inneficiency and how can I improve it on an existing plant without changing the design.[/li]
[li]Is the LMΔΤ for jacket of a batch reactor correctly assumed that is the difference of inlet temperature to reactor content temperature?[/li]
[li]To me it seems that bigger chiller won't make any difference. What do you think?[/li]
[li]Do you think that low coolant velocity in the jacket is affecting this? Higher flow though won't give higher DT because it's already too small difference.[/li]
[/ul]

 

[LittleInch]

Could you pipe the cooling fluid through a set of small pipes in the reactor contents itself instead or as well as the jacket?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Compositepro]

A blast of cold air on the liquid surface in the mixer is surprisingly effective in increasing cooling rate in viscous materials. While air is not great at heat transfer, the liquid surface film cools rapidly and is then incorporated quickly into the bulk fluid, resulting in good overall heat transfer. I have also heard of dry ice pellets being poured into mixers for rapid cooling. One caution though, is that dry ice may contain traces of oil from the compressors.

 

[KangCy]

How is the mixing looking like from the video?

 

[Compositepro]

That does not appear to be high his viscosity at all. why are your blades pumping upward? There appears to be a frozen or crystallized white mass on the bottom of the mixer, under the clear liquid.

 

[KangCy]

What do you mean by pumping upward?

 

[Compositepro]

The blade angle and direction of rotation of the shaft are such that they are pumping the fluid upwards in the vessel. This is unusual. It will result in less fluid velocity and turbulence on the bottom surface. The motor is almost certainly three-phase, and it may have been connected in the wrong phase rotation. This is a very common problem.

 

[EdStainless]

Is it just me or is that a very small agitator for the vessel size?

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Latexman]

Compositepro - are you sure? I saw the video a few days ago, and it's archived now and I cannot re-look at it, but my recollection was it was turning clockwise viewed from above, and the P4's were pumping down in the center and up at the tank diameter. We use Chemineer agitators, and that's their standard set-up.

Ed - OP said agitator diameter/tank diameter was 0.6, and when I looked at it a few days ago that looked right. D/T = 0.6 for a low viscosity application is on the big side in my experience.

The biggest issue I saw was the agitator was only turning 30 RPM. Maybe that was done just for the movie, but 30 RPM in this case is slow.

Good Luck,
Latexman

 

[KangCy]

Latexman, if we agitate faster, we will add more heat to the system right? Do you think it will solve this issue? Because the solution is anyway non viscous.

 

[Latexman]

The mechanical equivalent to heat, 1 Btu = 778 ft-lb[sub]f[/sub] or 1 g cal = 4.184 J, results in lots of work for very little heat. I ignore the heat from mechanical energy in my agitation calculations. I've never seen the mechanical equivalent to heat get significant, compared to the heat removal required from a polymerization reaction, but I'm not familiar with your system. Do the calcs; numbers is hard.

Good Luck,
Latexman