Convestion coefficient in a jacketed vessel

[finechem]

I am trying to calculate the convection coefficient for the heat transfer surface in a jacketed reactor. I have completed water tests and have determined overall ht trans coef, but I would like to get an idea for what is happening in the jacket when I change heat transfer fluid properties (visc, k, etc.). Currently I am approximating using equations for annular space (using hydraulic radius, 4Ac/heat transfer area) but on an 8' diameter vessel and 3" jacket space this must introduce significant area. Notwithstanding the transfer in the bottom head, does anyone have an equation, or article reference that will allow me a better approximation along the straight side of the vessel? Thanks for your help - hope this is not too wordy...

 

[unclesyd]

Is your heat transfer fluid in a liquid or vapor form?

This might be of interest to you.

Heating with vaporized heating fluid.
W
e have numerous jacketed reactors that are heated by vaporized thermal heating fluid. Our development people worked for several years trying to model the heat up of these vessels. I was the spoiler in their attempt. When they came down to install some additional thermocouples i had to ask them where they were going to install them and at the same time I pointed out that the three existing couples in the shell were not reading together. I pointed out that the shell heated like a glass tea kettle with vapor condensing at random areas. You could put your hand on the shell next to a TC reading 235C. I had to have insulation removed on one side of the vessel and an IR scan to prove my point about the heating patterns. This was the pattern on all our vessels heated with vaporized heating fluid.

 

[gerhardl]

In your question you stipulate that the geometrical form would perhaps be the largest factor contributing to heat transfer.

'Unclesyd' tells us that heat transfer in such circumstances could be a dynamic more randomized system depending on a number of variables.

Contributing to actual effect would obviously be more factors than in the 'idealized' laboratory and formula set-up, for instance flow and heat distribution both in cap and vessel. For steam heating the condensating rate (where the vaporizing heat is transferred) also condensate level in the shell and condensate removal from shell will give a high influence on effect in heat transfer.

So what are you actually looking for? A theoretical figure or optimizing a process?

I have seen steam-heated reactor vessels with 10 tons capacity run at below 50% actually obtainable effect caused by wrongly constructed condensate removal. The construction could, without large cost, be vastly improved with correct piping, flow control and steamtraps.

 

[Montemayor]

finechem:

Process reactors are not sized according to their heating or cooling requirements. They are sized according to their reaction requirements. Any jacketed reactor is capable of transfering only the amount of energy that is justified by the available shell and head(s) - PERIOD. You can add internal coils and external heat exchangers, but that is another subject.

If you believe you are going to be justified in "optimizing" the amount of heat transfer through the jacket, I would state you are wasting your valuable time. You have been given the fixed amount of surface area that you have. Take it and run with it. You can't improve on it and you will find that it is insufficient for all your reactor heating/cooling needs - whether the reactor is endothermic or exothermic.

I don't believe you can contribute any worth to the heat transfer in a jacketed reactor by studying the film transfer coefficients. You can try to increase the heat transfer effect to the ultimate by increasing the convection, the Reynolds Number, the pressure drop, or the turbulence, etc., etc. And you will have to be satisfied with that.

What you see is what you get. If you want to estimate the available heat transfer capability, use a "U" of approximately 70 Btu/hr-ft²-^oF when using cooling water.