Distillation - mechanical recompression 1

[fvincent]

I was wandering about the technical and economical feasability of using top vapours stream of a rectification column of isopropanol as heating agent of its own reboiler, after a mechanical compressor.

I need some help in order to find out who supplies such equipment and who has already implemented such modification in an existing column. Costs? Power?

By the way the vapour stream leaving the column is about 11,000 kg/h at 80oC.

Thanks

fvincent

 

[DickRussell]

The concept has been around for a while. I read an article on doing just that in Hydrocarbon Processing, October 2004 I think, just last night. I've also done such calculations myself in the course of daily work. Any process simulator can be used to model the proposed configuration either way, to get duties and compressor power for costing purposes. Basically, the idea is most likely to be economical for a close-boiling system, with fairly low pressure drop over the column. This makes for a low temperature difference between top and bottom and thus minimizes the compression ratio needed to get the overhead vapor to condense at a hot enough temperature to drive the reboiler. Condensing will be hurt (need more pressure) if the overhead vapor has a light fraction as well as the main distillate constituent. A trim condenser, using cooling water, could be used, but that adds equipment. The compressor will be expensive, but the payback must come from utility savings (power vs. steam and CW). Even if having a column reboil itself isn't economically feasible, you still may be able to use vapor recompression to reboil an adjacent column to obtain savings. In the right situation, even the compressor may not be needed; that would be called cascade reboiling, and that also has been presented in the journals.

HTH
Dick Russell

 

[25362]

Dick Russell is right. Another possibility of heat integration could be using the same vapors to heat the feed to the column.

I've just seen a site that deals with this type of integration:

http://tx.technion.ac.il/~dlewin/054402/LECTURE_09.pdf

that may interest you.

 

[fvincent]

Thanks DickRussel and 25362

I think the preheating of the feed is more efficient with the bottom stream, isn't it. In my case, I am considering increasing the effectiveness of this preheater, since bottom stream leaves the column at nearly 100oC and the feed after the preheater is only at 62oC (isopropanol at 60%w)

As to the compressor, I have read a lot too, but actual costs and power are still missing. I have asked proposals for the compressor but no answer came.

That is why I was wondering if someone could give a hint about the equipement to pressurize 10.5 t/h of an azeotropic mixture of isopropanol and water which leaves the top of the column at 80oC and at 0.94 bar absolute. I think I should pump the vapours at least up to 3.2 bar abs in order to obtain a 15oC temperature difference between condensing IPA steam and bottom temperature.

Thanks for the comments and suggestions you gave.

fvincent

 

[epoisses]

Just a thought... if you do want to use electricity rather than steam to reboil that column, wouldn't it be more interesting to install an electrical heater as a reboiler instead of the compressor, and improve both the energy efficiency and maintenance cost of your project? Or am I missing something?

 

[fvincent]

yes, you are missing the fact that the power is related only to pump heat from a lower temperature source to a higher temperature source, so that power is only a small fraction of the total energy transferred. Just like in an air-conditioner or refrigerator...
regards



fvincent

 

[25362]

Besides, the heat available in the vapors goes to water or air in a condenser and is probably lost.

 

[DickRussell]

Assuming your "tons/hr" is metric, a compressor polytropic efficiency of 72%, and a condenser pressure drop of 0.15 bar, the compressor and condenser look like this:

COMPRESSOR BLOCK PC101 Overhead vapor compressor, PC-101
GAS HORSEPOWER = 443.4
POLYTROPIC EFFICIENCY = 72.0 %
DRIVER POWER = 330.6 KILOWATT AT 100% DRIVER EFFICIENCY
INLET VOLUME = 6922.8 CU.M/HR, LIQUID FRACTION =0.0000
INLET CONDITIONS OUTLET CONDITIONS
STAGE TEMP(C) PRES(BARA ) Z TEMP(C) PRES(BARA ) Z AVG.K
1 80.0 0.9400 0.9854 142.2 3.200 0.9677 1.105
POLYTROPIC HEAD PER STAGE (METERS): 8323.


HEAT EXCHANGER BLOCK TT101 Condenser/Reboiler, TT-101
DUTY= 2.1664 MM KCAL/HR
HOT SIDE INLET TEMP= 142.2 C, L/F=0.0000, FLOW= 10500.00 KG/HR
VAPOR DENSITY= 4.471 KG/M3
OUTLET TEMP= 110.9 C, L/F=1.0000
LIQUID DENSITY= 716.87 KG/M3

The efficiency will affect just the discharge temperature, not the condenser approach. Assuming the bottoms of the column is just water at 1 atm, or 100 C, and you wanted a 15 C approach, then you would need a discharge pressure of 3.63 bar(a), and the results look like this:

COMPRESSOR BLOCK PC101 Overhead vapor compressor, PC-101
GAS HORSEPOWER = 492.2
POLYTROPIC EFFICIENCY = 72.0 %
DRIVER POWER = 367.0 KILOWATT AT 100% DRIVER EFFICIENCY
INLET VOLUME = 6922.8 CU.M/HR, LIQUID FRACTION =0.0000
INLET CONDITIONS OUTLET CONDITIONS
STAGE TEMP(C) PRES(BARA ) Z TEMP(C) PRES(BARA ) Z AVG.K
1 80.0 0.9400 0.9854 148.6 3.630 0.9648 1.105
POLYTROPIC HEAD PER STAGE (METERS): 9239.


HEAT EXCHANGER BLOCK TT101 Condenser/Reboiler, TT-101
DUTY= 2.1582 MM KCAL/HR
HOT SIDE INLET TEMP= 148.6 C, L/F=0.0000, FLOW= 10500.00 KG/HR
VAPOR DENSITY= 5.009 KG/M3
OUTLET TEMP= 115.0 C, L/F=1.0000
LIQUID DENSITY= 710.99 KG/M3

HTH

RAR

 

[fvincent]

Now you're talking!!!
Thanks a lot
you deserve my seldom given star!

fvincent

 

[DickRussell]

It was fun. A nonideal VLE method was needed to get the condensing temperature right. Being an azeotrope, the condensing curve will desuperheat quickly, then drop only slightly due to pressure drop. I looked up the UNIQUAC interaction parameters for the binary in Dechema VLE Data Series (Vol. 1, Part 1, p. 334 I used). A different data set gave the azeotrope as 68.13 mol% isopropanol. After the first run, I added a controller to tweak the discharge pressure to home in on the 115 C HX outlet temperature. I suppose it would have been simpler to do it by trial and error, but I thought you might come back with revised data to run through.

The pressure you gave is slightly in vacuum at the top. Why not make it atmospheric to avoid air leakage? Air leakage will hurt the tail of the condensation, requiring higher pressure and a vent condenser (which you likely have now anyway).

To expand on your earlier response to a question on the use of electrical work to do reboiling indirectly, the output shows the "air conditioning" effect; expending 0.316 MMkcal/hr of electric power input "lifts" the latent energy of the overhead vapor to a temperature/pressure where it can be extracted usefully, flowing "downhill" (in temperature) to boil the column bottoms. The higher the "lift", the more energy it takes and the lower the economic benefit.

Of course, the thermal condition of the feed will affect the relation between condenser and reboiler duties. If the feed is subcooled, you'll need more reboil duty than will be available by condensing the (compressed) overhead vapor. If the feed is partially vaporized, you will need a trim condenser to handle the excess vapor energy. A feed/bottoms exchanger could help improve on the first situation.

 

[fvincent]

Well, in fact the column is installed in Sao Paulo where barometric pressure is 0.92 bar. So 0.94 bar is in fact slightly above external pressure.

The feed preheating will be improved. Currently the exit stream leaving the bottom tray preheats the feed to only 66oC. Since bottom temperature is approximately 100oC (only water a little below 1.0 bar abs), it is possible to recover additional heat, at least preheating the feed up to the inlet tray temperature (near 90oC).

The column operates today with direct steam. So, if the heat pump is installed, the indirect boiler will be required. The total bottom stream will obviously reduce (only a small direct steam condensate will exist). Feed preheating has to be re-evaluated, of course.

I had used a simpler approach for calculating power. I considered (in fact - omitted) that the azeotropic point of isopropanol+water would not change with the pressure elevation. Furthermore, according to the saturation curve of pure isopropanol, the pressure corresponding to 115oC would be 3.16 bar abs. You showed that the pressure had to be higher.

I have also considered fully adiabatic compression instead of polytropic and finally - since it was a short calculation - I assumed a Z=1.

I thank you very much again.
Now I have to look for costs. I don't have the slightest idea of the price of such compressor (which has to be oil free). Some suppliers I contacted are really late...

I wished they could give the numbers as promptly as you gave..
regards



fvincent

 

[DickRussell]

Methanol-water is non-azeotropic, ethanol-water just makes one, but isopropanol-water is quite nonideal (68 mol% isopropanol). By the time you get to C4 or higher alcohols, there is so much nonideality that there is a region of immiscibility (heterogeneous azeotrope). So, yes, the nonideality of the VLE did have a considerable affect on the condensing temperature vs. ideal.

I presume you are just roughing out whether or not the concept is at all in the ballpark of profitability. If it does seem so, be sure to engage the services of a good engineering outfit to do a more detailed evaluation and cost estimate. The detailed engineering costs alone could be considerable.

 

[fvincent]

oh yes.. it is only a general study on energy saving alternatives for a nitrocellulose plant (non-military purposes, by the way...)


fvincent