Take the 2-minute tour ×
Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. It's 100% free, no registration required.

I'm in the business of purifying used cooking oil. Normally, I heat the oil up and let it settle for a couple of days. Water and solids settle to the bottom and cleaner oil remains on top. I'm trying to accelerate this process.

So, the idea is to take a vertical tank 5ft diameter, about 10ft tall and put a paddle mixer on top with vertical blades that would extend to about 3" from the wall. If I turn on the mixer, it will spin at about 60 rpm and spin the liquid with it. According to my calculations this should give me a force of 3.5g and thus decrease settling time by about 3.5 times.

Please let me know if this can work or if I'm missing something. Your input will be highly appreciated.

share|improve this question
    
How about an axial cyclone: kirkprocess.com/products/services/how-it-works May become problematic with the solids though. –  Bernhard Apr 1 '13 at 7:44
    
How about having a spinning drum inside the tank? Like a washing machine drum, with holes for the solids to go through? That way, with luck, you can get the fluid inside the inner drum to spin uniformly and the solids will leave through the holes. The fluid between the drum and the outer tank will be turbulent (hence mixed), but that doesn't matter because you can take clean oil from inside the drum. I don't know if this will actually work, but it seemed worth mentioning. –  Nathaniel Apr 2 '13 at 9:08
    
@Nathaniel, this is what I meant by Couette flow. –  aditya kp Apr 3 '13 at 4:29

3 Answers 3

This approach will not work; since it will result in mixing of the oil and particles, not a settling/separation. This is called a "stir-tank reactor" in chemical-engineering jargon. If you look into this properly, you will realize that this configuration will REDUCE THE MIXING TIME, and in fact, deter separation. Refer: http://en.wikipedia.org/wiki/Continuous_stirred-tank_reactor for a more detailed explanation. It is a model for PERFECT MIXING.

Spinning the whole thing is not really an option. The tank is not well balanced and when full weighs about 15000 lbs. Potential for a huge mess is tremendous.

I suggest thinking about a Couette flow instead (http://en.wikipedia.org/wiki/Couette_flow) since centrifugation is not possible. The RPM of the tank should be low enough, so that turbulence isn't generated, (Reynold's number: http://en.wikipedia.org/wiki/Reynolds_number) since turbulence causes increased mixing.

share|improve this answer
1  
How would Couette flow help speed up separation? (I'm not doubting you, just curious. It isn't obvious to me.) –  Nathaniel Apr 2 '13 at 8:56
1  
A Couette flow is steady because of the absence of irregularities; which are present in the flow around impellers (viz. the blades. these make flow around impeller unsteady). Unsteadyness will hamper separation, in the way of disturbing the particles and not letting them get settled. thus, couette flow is better off. The Couette flow was just a suggestion, since centrifugation is not an option. (Doubting is not an issue at all! such questions keep us going, dont they?!) –  aditya kp Apr 3 '13 at 4:27

It sounds like a good idea, except if the wall is stationary and the fluid is moving, there's going to be turbulence at the wall, which could prevent settling. I wonder if you could make the whole thing spin? (Easy to say, I know...) Sounds like some experimentation is in order, maybe on a smaller scale. (I can picture pails of fluid suspended on opposite sides of a rotating arm, kind of like those centrifuges used in biology labs.) http://en.wikipedia.org/wiki/Centrifuge

share|improve this answer
    
Spinning the whole thing is not really an option. The tank is not well balanced and when full weighs about 15000 lbs. Potential for a huge mess is tremendous. I also thought about adding a channel that would be attached as a large spiral on the inner wall of the tank. This would act as a scraper and direct all debree that was thrown outward down to the bottom. –  Greasehauler Apr 1 '13 at 1:09
    
@Greasehauler: I think you have good ideas, and the only thing I might contribute is what I tell my kids and what I always told my programming students: Do small experiments. That way you can try things, and the failures will be educational and not very expensive. –  Mike Dunlavey Apr 1 '13 at 13:54

As other answers pointed out, stirring the tank will create turbulence and large scale mixing and prevent settling. Typical solutions for this separation process are:

  • Settling tanks - what you have

  • Baffled tanks - If you wish for a continuos operation, you can equip a tank with vertical baffles (with clearings at the top and bottom) and outlets at the top and bottom. As your oil slowly flows through the tank, it can separate, the baffles ensure a calm, laminar flow. The speed of separation will be the same as before. The Volume will have to be something like (amount per day) X (Separation time in days).

  • Hydrocyclone - Basically a funnel where you inject the oil tangentially and have two outlfows, for light and heavy material. You will need a pump suitable for the "dirty" oil & Sizing may be tricky.

  • Centrifuge - There are quite a few designs, for continuos and batch operations. At one Biodiesel factory that also were equiped to handle grease I've seen quite a few disc-stack centrifuges

For a cyclone or centrifuge, You will want to talk to a vendor of specialized equipment. Even if you choose to go another route, they may be able to help you understand the problem more completely. A baffled tank could be homebrewed (If you don't want to spend much money).

I would suggest to ask the same question in a dedicated engineering forum, as you will get answers from people who deal with the same problem.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.