Injecting CO2 1

[birkATO]

We have a process where we inject about 1600# of liquid CO2 into a reaction in a 2000 gal reactor. The way it is set up right now is we have a circulating line running from teh bottom of the reactor to the top and there is an injection tube in the line where we inject CO2. The injection tube is a hollow tube about 0.5" in diameter with several small holes drilled in it to disperse the CO2. There is a check valve and a globe valve on this injection tube. We have problems getting the CO2 into the reactor quickly enough (it takes about 1 hour to get 1 cyl=~400# into the reactor). Are there CO2 injection systems out there that might work better than the one I am working with? If I open the globe valve all the way the CO2 turns to dry ice and blocks the line.

 

[Montemayor]

birkATO:

I know I can solve your problem - and probably in a very simple way. I've worked with CO2 systems for a large part of my career and never had any problem moving tons/hr. However, you've got to furnish specific, basic data. You haven't stated, for example:

1) What conditions you are injecting the liquid CO2 (LCO2) under; LCO2 is nomally distributed in two ways - high pressure liquid (1,100 psig & 80 oF) in cylinders for small consumptions and low pressure liquid (250 psig & -10 oF) in 6 to 12 ton vessels for large consumptions.

2) What pressure and temperature is your reactor's recirculation line running at?

3) Are you using the LCO2 storage vessel (or cylinder) pressure to be the driving force to inject the LCO2 into the recirc line? Or are you using a pump?

You should have no problem injecting the LCO2 - with the appropriate driving force. That's why you must furnish the identity of the LCO2 system that you are working with. An injection system that only yields 400 lb/hr for a 2,000 gal reactor is either plugged up or badly designed. You could be plugging up with dry ice if the system is not designed correctly and if you have too much pressure drop across the orifice in your injection tube.

I can't comment on your injection system's merit if you haven't described it specifically. There are no LCO2 injection systems in the market place specifically for your application. Each one is designed for the application. You can design, fabricate, and install your own very easily. However, you must have the basic data for the application.

I would normally expect you to be pumping the LCO2 with a PD pump, using a 10- 12 ton LCO2 tank as the storage source. This is no major challenge or problem. If you furnish more specific details, we may resolve this problem on this Forum. Having run similar batch reactor operations, I would not tolerate this very slow injection time unless there were process or safety constraints.



Art Montemayor
Spring, TX

 

[birkATO]

We are using Liquid CO2 cylindars that usually have 150-300 psi in them. The cylindars hold about 400# or so. We are using the pressure in the cylindar as the driving force. We hook a cryogenic SS hose up to the CO2 cylindar liquid port and shoot it straight into the circulating line. I am interested in learning how to design an aparatus to get about 800-1200#/hr into the reactor if that is possible without a pump.

 

[Montemayor]

birkATO:

You haven't identified the pressure in your reactor recirc loop, so I can't tell what pressure drop your injection point is going through.

Why do you use such small LCO2 cylinders? A small, 6-ton tank at 250 psig should be more in line with what you are doing. At 150 psig you are at -35 oF, below the normal rating of common carbon steel tanks and piping - unless you're using stainless steel construction. Also, at 400 lb/cylinder, you need to manifold 4 of these cylinders together to inject your proposed 1,200 lb/hr. Otherwise, you spend/waste a lot of time on changing over cylinders. With a tank you can have the LCO2 delivered directly to the tank by a tanker-truck with a pump on it. This is what is normally done with industrial accounts.

The lower the pressure in your LCO2 storage cylinders, the colder it will be and the greater the amount of dry ice (solid CO2) formed upon subsequent injection. You haven't said what the temperature and pressure conditions in the recirc loop are, so I can't judge how the expansion can be done without plugging. At least give me a clue.

When you are relying solely on the tank head pressure for the driving force to inject the LCO2, your hands are a little tied in controlling the injection rate and conditions. I would rather employ a small PD pump like a Viking gear model. But we need more info before making a definite recommendation.

Art Montemayor
Spring, TX

 

[TBARBill]

The reactor pressure has got to be quite low if they are able to get most of the CO2 out of those cylinders if they only start out at 150 to 300 PSIG, even liquid-filled. Also, for dry ice formation, the pressure must be below about 75 PSIG. A pump is really the best way to do this, and a larger volume/mass CO2 source would be a big help, too.

How sensative is your reaction to temperature? Increasing the CO2 rate will cool the reactor off, which would increase dry ice formation (plugging) and might well have adverse affects on the reaction itself.

CO2 is fun stuff to work with. I'm on a project a bit larger scale than yours - we inject about 845 tons per hour (350 MMSCF/D) into an oil reservoir for tertiary recovery. Distribution pressure is 2200 PSIG (supercritical) and we boost recovered CO2 from as low as 7 PSIG to full field pressure. We boost into the supercritical with recip compressors and use multi-stage centrifugal pumps for final injection boost. Got to be carefull with new engineers - 1500 PSIG CO2 is a *LOT* different than 1500# natural gas!

 

[chela]

Dear BirkATO,
I would summarise a solution for you:
In the first phase, go for a manifold with double or triple the required number of cylinders connected say 8 or 12- immediately, you will get a higher flow and higher pressure. You can also heat the cylinders with cooling water so that the pressures dont drop.
In the second phase, go for a receiver and pump system if your economics permits it.

 

[birkATO]

Sorry for not providing the info needed. We run the recirc loop at 10-30 psig. I know a bigger tank would help, but we only used about 2000# of LCO2 for each batch and only run about 3 batches per year, so investing in a distrobution system might not be worth it. My boss just seeps to think we can shoot the LCO2 cylindars in in less than an hour the way we are doing it and I can't seem to get them in that fast. I don't know much about LCO2 so I thought if there was a way I could speed it up without designing a new system, then it would satisfy my boss. We are injecting this into a stainless steel recirc line in a stainless tank.

 

[Montemayor]

birkATO:

Now that you seem to be putting out more basic information about the application, I know I can inject 2,000 lb/hr of liquid CO2 into your system using only the cylinder force as the driving force. I've done something very similar in the past that allowed me to get over the inherent blockage problem that exists with this LCO2 expansion. I can explain it to you here only generally; details would take the usual engineering drawings and sketches. Also, in order to finalize the solution, I would require ALL the basic data; you've only doled out pieces at a time and haven't answered all the questions. The answer is as follows:

1) You'll have to have approx. a 3-4" nozzle opening; and it will have to be a flanged connection;
2) You employ a 1.5" -2", 90 degree globe valve; I've designed and built my own in the past.
3) The idea is to expand the LCO2 across the valve in such a way that the resultant snow exits immediately and easily through a connected "horn" (or straight concentric expander) and into your flanged connection. This way the snow is easily deposited and any tendency to formed a plug will be easily pushed out and into the recirc line by the cylinder pressure since the plug formed is conical with the high pressure side being the smaller diameter.

The above is much better understood when seen in a sketch. If you have understood my explanation, you're home free. I can doubly assure you that the system works because I've used it in dry ice manufacturing where I've expanded LCO2 from 125 psig down to 60 psig at a rate of 3,000 lb/hr continuously, 24 hrs/day; 7 days/wk - and without any plugups. I've made this a feature of the patent I presently hold on an automatic dry ice pellet producing machine. One machine has already been built and it has worked easily, although at smaller flow rates because of the smaller design.

So there you have it; as I said, it should be no problem. I still would prefer a pump because it offers much more precise control on the flowrate - but, what the heck!

I hope this helps to solve your problem.

Art Montemayor
Spring, TX