quote:

Originally posted by bio-LOGIC:

Can you use a centrifuge to dewater WVO?

My simple answer (opinion). No.

My extended answer (opinion).
1. The filter does a great job of removing solids and "heavy" liquids by condensing them inside the bowl. The heavy liquids I have been getting are very thick. Nothing compared to water.
2. The filter is not set up to easily handle the liquid difference. When you stop the filter from spinning, the bowl spins to a stop. Liquid can drain out of the bowl when it stops (there is a slight vaccum in the bowl). IF it did seperate water, you would have a hard time "catching" it. I place a jar under the filter when I open the bowl to catch the syrup/bd mixture when it comes out.
3. As it does a good job of filtering, if you filter the raw oil to remove any solids, you may aid the dewatering process as a lot of moisture can be in the food particles. I have not tried this as my oil is pretty clean and water free.

Just my .02, but according to my MSDS for KOH from univar, the other percentage is water. ie 90% koh 10% water.

I have heard the rumor about it being naoh, but I've never seen any documentation to that effect. Judging from the way the flake looks I'd say that they spread a thin paste on something and then dry it. Then, depending on how much they dry it out, the purity is determined by how much moisture is left. Of course this is all conjecture.

bayshorecs, I'm interested to see your final posts about how well this works when you get it set up the way you want it. If it works well for you I think I'll be ditching the wash and dry tank and going this route, no more emulsions, no more time consuming washes. It really sounds to good to be true.

Just another thought,

About 50% of the oil I used in these 2 batches came from a "county fair" source. All of the oil seemed to be from funnel cake cooking oil. Could the high amount of suger in the oil be causing this to occur?

My next batch (hopefully this weekend, I have run out of physical space to store the BD I have made doing these two "unscheduled" batches) will be with my normal supplier oil of clear soy, titrate between 1-1.5. If I get the same results, I should be able to rule out foreign matter in the oil assisting the soap fallout.

I will also be doing some pHLip tests on the next batch of fuel as well. I plan on running the pump again tonight to see if the soap count will get much lower. I have no doubt it will.

When I finish spinning it and call it quits, I will try a 3/27 test on it as well (never did one of those before).

I agree, seems too good to be true. All I know if that I knocked 20 cents per gallon off these last 40 gallons by not using magnesol! Down to $.81 per gallon now.

quote:

Originally posted by RickDaTech:

quote:

Do these number actually work out or is it just a freak thing? As I have most of the soaps out (down to 121ppm) and the 600ml is a "guess" (could not physically capture 100%, I DO have about 500ml), it seems like according to the math, it should be about done. And the soap test of the BD shows it is about done.

Aint it grand with the math works out like it's supposed to?

That will work kinda nice then. If my soap concentration is about 25% and I should end up with about 600ml of syrup after the spin wash is finished on a 20 gallon batch, no need to test for soap amounts during the processing until I get the correct physical amount of junk in the jar.

That will save some time just testing the soap count.

Bayshorecs,

It looks like you do have something interesting here. However, I'm wondering about the practicality of what you're doing. Hope you don't mind a bit of well-intentioned skepticism...

It looks to me as though you're having to process this for many hours to get this to work. I'm using Magnesol and getting much the same result in much less time. Water washing doesn't take as long either, from what I understand. What makes the centrifuge method much better?

George,

Not to answer for Byshorecs.

Each method of washing has it's own set of compromises. This method has a unique blend of not needing water or magnesol and with the potential of full automation. These centrifuges come in many different sizes, so the constant cleaning can be eliminated by going just a little larger. Going just a little larger also increases the volume throuput. My interest in this method is that it has the potential for dramatically decreasing hands on time as well as in the area time. It also eliminates the need for a final filter. The final filters (1-5 micron with water drains) I've looked at are all expensive and need to be replaced with expensive cartridges.

So it's potential strong points are 1) no water, 2) no consumables, 3) reduced labor, and 4) eliminates final filtering.

It's potential weak points are 1) increased equipment costs, 2) process time, and 3)the need to periodically clean the centrifuge

My guess it that it would work faster when you use NaOH as a catalyst.

Thanks Rick.

The other point is that my current setup of 1 tank is causing the wash time to be extended as I always dump clean fuel in the dirty tank. With the new setup, I bet my wash time will be around 2-3 hours.

I have made an animation for what I am building.

2-tank setup

This setup should allow for full automation of the wash process.

I also found out you can get viton o-rings for the filter (part number: 70826). I just never bothered to get any yet. My o-rings are a little warped, but still usable.

This sounds pretty good, however any guesses why commercial producers are not then simply flashing off methanol then running through a centrifuge? It is my understanding that many large-scale centrifuges can even dump solids periodically while running.

I actually have a Barrett 'Clarifuge' I bought via ebay a while back, which is from what I can tell very similar to a "spinner II" but larger. It is a bowl type centrifuge marketed to clean cutting oils to roughly 5 micron or so.

http://www.barrettinc.com/clar.htm

hmmmmmmm

Rob,

I think there are some commercial operations using centrifuges. There is a company in Germany that sells into the commercial biodiesel production market.

I have heard of issues when trying to do liquid/liquid separations like soap and glycerine removal. Centrifuges tend to be finicky and sensitive to changes in feedstock specific gravity. Those that use them have reported they need to tune the speed specifically for each batch depending on the feedstock.

The centrifuge used in this thread is specifically designed for particulate removal. So my guess is that highly filtered feedstock oils will have a tough time pulling out the soaps. Different people may also experience differing results using this type of oil cleaner. I really like the price on them. only a few hundreded dollars for this style vs $10K+ for the commercial versions.

There may also be a "sweet spot" on mine for the certain blend of oil I used. When I run my next batch using my normal source, I may run into problems and need to dial in the speed.

I hope to be able to do another test this weekend.

I do recall seeing this result in the past when filtering magnesol. I just assumed the fine particles of the magnesol produced this result. Just not.

I wish I would have tried this approach earlier.

Rick, if you use NaOH and have solid soaps, you may want to rethink the type of pump to use. Mine is a positive displacement (like a car oil pump) with interlocking gears. A diaphram pump may be a better pump to use. They cost a few $ more though.

I forgot one other thing. Most of the commercial applications were one pass processes. byshorecs is using multiple passes or polishing techniques that don't readily lend themselves to high speed commercial applications.

I guess I was thinking of using the clarifuge just like the bayshores' spinner. Flash off methanol, then recirc. via clarifuge until bowl (2 gallon) is full of heavier material. Determining cleanout interval will take a little experimentation, but based on PPM count and biodiesel qty. a rough estimate of soap volume should be reasonable to estimate.

bayshorecs,

These spinners have specific pressure & volume demands, so It's going to be a matter of getting a pump that matches the specs on the spinner.

With your multi pass process, I think the only real tuning would be to the temperature of the biodiesel. I'm not sure if warmer or colder is better for filtering out soap. When filtering out particulates, warmer is definately better.

Since I am using a DC motor, I can control the speed of the motor and the amount of pressure put against the filter. The more pressure, the more RPMs of the bowl, the finer the filter. Since I am filtering a liquid, faster may not be better. I believe the filter has a bypass to it as well and may have a terminal velocity of the speed of the bowl. Not sure though.

The pumping process itself adds friction/heat to the fuel. In my 60F workshop, the fuel is about 85F after running through the pump for a period of time.

One of the engineers at the company I work for specd out the pump to the filter. He could have used a 1/2hp motor, but had a 3/4hp motor laying around so he put that one on the pump. The pump itself has a bypass where you can set it to fully unload into itself in case you get a blockage.

Pump specs:
Dayton Heavy-Duty Cast Iron Gear Pump (close-coupled)

Pump Head

Decided to stop on this batch until I have the two tank setup. Final soap count was 6 drops of .01N HLC = 87ppm

I also did my first 3/27 test. Don't think it passed though.

pic

Come to think of it, for this batch, I mixed in the reactor for only 1 hour and moved it to the uninsulated settling tank. This was a new process for me. I used to circulate for 2-3 hours, let the glycerin settle in the processor overnight, remove 75% of the glycerin and move to the settling tank to get the rest.

I wonder if:
1. Only reacting 1 hour
2. Moving to the settling tank right away thus cooling the BD down faster

This may have caused my incomplete conversion. My next batch will go back to my standard method. I wonder if the 75% of "other stuff" in the syrup is concentrated, unreacted oil?

Talking with Jim from FryerPower.com sparked another idea. He wants to try filtering his KOH BD for soaps in low temp to try and achieve the as results as using NaOH.

I wonder if you install a coil in a freezer to pass the BD through (try to get down to 40F) then send it through the filter....

As the pump adds heat to the fuel, you would need to:

1. Pass through the heat exchanger after the pump
2. Have a B.A. heat exchanger to remove as much heat as possible
3. A deep freeze or frig to place the coil in and cut a few holes in the side of.

Oooo. Better yet. Get a flat plate heat exchanger like the SVO guys, recirculate cold water through the exchanger and pass the BD through it. If they can bring 70F BD up to 180F in a single pass, you should be able to remove the heat from the fuel just as easily.

The brewing industry calls it "cold filtered beer" right. So would that make it "cold filtered BD"?

I am still having a hard time grasping the leaps and bounds this board has made since I joined last december....

Placed the jar in the fridge last night to drop the temp down to 37F. The syrup got very thick and when you tilted the jar on it's side, it took about 10-15 seconds for it to catch up to the movement.

The question was brought up about what the remaining liquid is. If I have 600ml of liquid at 25%, that should only give me 150ml of "stuff". I wonder if the remaining amount is left over KOH from the reaction. Just thinking if you have a liquid and add powder to it, the liquid typically starts to thickin up, like gravy. Anyone no how to test the concentration of KOH in the sample?