Dear all,

This is my first posting to this forum. I've read a portion of the posts regarding static mixers. They seem to be regarding batch processes using static mixers in a circulation loop.

I'm trying to redesign the standard biodiesel process to make it faster and easier to automate. My proposed system diagram uses a metering pump directly to a long static mixer, much like a modified plug flow reactor. I'm trying to figure out some decent starting conditions. Does anyone have experience with this?

I am thinking of a linear flowrate of 0.1 ft/second and anticipating total reaction time less than 5 minutes. This is giving me a very long and thin reactor, about 60 feet long. As is done with PFRs, I would enclose the tubes in a housing and line up 30x 2 foot sections with u-joints at the ends. Pressure would be > 100 psi. Does anyone have experience with such a system? I'm interested in any information about laminar vs turbulent reaction yield, pressure drop across the reactor, pressure effect on yield, temperature increase caused by the mixer, pressure increase caused by the metering pump, effect of temperature on the reaction at this moderately high pressure, and any patents that this may infringe on. I would be glad to post information about final design on this site if I get something that works well.

Best regards

Marion

Check out Johno's double pumper and related threads. Here is a search.

HTH.

HERE are the pictures of Johno's very innovative design... A fine piece of work in my opinion!
Let us know what you come up with!
Good Luck,
Jon

I'd be happy to share what little knowlege I have about continuous flow processors that use static mixers. I've had great luck with the mixing, but poor luck with simple metering pumps that use check valves. I'm slowly working on a metering variant I'd be happy to discuss that might fit your concept pretty well.

The question of laminar vs turbulent flow has come up in other applications, but I have no resolution to offer in this case. Do you want to run a back pressure >100 psi, or system pressure that high? High back pressures lead to downstream cavitation inside the static mixer, unless also accompanied by high system pressure or slower velocities. Excessive cavitation can lead to soap formation, under unfavorable conditions.

Greenline uses continuous flow oil, methanol and sodium methylate metered at stoicheometric ratios into a static mixer in a recirculating loop, followed by a settling tank, etc. They may have a patent, which might also be involved in their current bankruptcy case. Instead of a single long static mixer as you describe, they use a shorter one and recirculate through it. The recirculation flow rate around the loop is many times greater than the rate of new liquid introduction.

My static mixer/metered flow experiments showed surprisingly high conversion with very short static mix lengths. If a stoicheometric mix enters the static mixer, there is less need to mix for a long time/length. The reaction among stoicheometricly correct chemicals takes only seconds. Long mix times are necessary to homogenize a batch that has zones of rich and lean ratios. This is not to say that higher conversions aren't possible with longer mixers, only that the incremental improvements become more difficult.


What flow rate are you hoping for? Is this for a homebrew system, small producer, or bigger? Do you already have some experience making biodiesel? What is your feedstock? Inquiring minds want to know.

Cheers,
JohnO

quote:

My static mixer/metered flow experiments showed surprisingly high conversion with very short static mix lengths. If a stoicheometric mix enters the static mixer, there is less need to mix for a long time/length. The reaction among stoicheometricly correct chemicals takes only seconds. Long mix times are necessary to homogenize a batch that has zones of rich and lean ratios. This is not to say that higher conversions aren't possible with longer mixers, only that the incremental improvements become more difficult.

I plan to use high system pressure and a back pressure regulator to keep the pressure high. Based on your feedback and some further research, I will probably not need such a long residence time and the static mixer can be shorter. Maybe I will start with just 30 feet and go from there. I need to calculate the Re still. I will go with a larger Re and target turbulent flow despite some controversy over this in the literature. This is for a system with capacity of 1.5L/min (~750 gal/day). I want to build it so it's fully automated, enclosed and safe, includes methanol recovery and final processing. Then I want to run it for 1/2 a day once a week and then try to get more people to donate oil.

Right now I have access to use a 75 gallon typical batch system without methanol recovery built by my business associate.

Forgot to say -- thanks!

quote:

Originally posted by johno:
I'd be happy to share what little knowlege I have about continuous flow processors that use static mixers. I've had great luck with the mixing, but poor luck with simple metering pumps that use check valves. I'm slowly working on a metering variant I'd be happy to discuss that might fit your concept pretty well.

Cheers,
JohnO

Please share about the metering pump issues. I think I would like a pump that will create thin regular alternating slugs of methoxide and triglyceride, as opposed to a metering pump that creates churn or does a lot of mixing since that's what I use the static mixer for.

I want it to be automated programmable, reliable, durable and safe. I can spend some $$ on it, for those features. OK $$ but not $$$

Thanks again

Here is my static mixer. Used in a recirced. batch. Methoxide introduced slowly at the front of the mixer.






It is made from 2.5 inch SS tubing.

Metering pump issues:
I've used a variety of check valves and given up on them for use in a metering pump for used cooking oil. If it's filtered first, then check valves work ok. I'll try to describe the problem, please ask questions if something isn't clear:
1) on the suction side of a pump, the check valve must open pretty easily.
2) Positive displacement pumps with oscillating pistons must be oriented to flush out air bubbles.
3) an air bubble on the suction side of a check valve will expand in the low pressure zone, displacing an equal amount of oil or methoxide.
4) when the piston pushes the liquid out, the bubble is compressed, delaying and reducing the quantity of liquid flow. It is no longer "metering".
5) the stronger the opening pressure needed for either the suction or pressure-side of a pump, the greater the effect of air bubbles.

OK, that's the air bubble problem. Moving on to the next problem - debris:
1) if the liquid contains any debris (grit, food particles, little pieces of teflon pipe tape, polymerized "chicken skin", rust, undissolved KOH, chunks of reacted KCO, etc), then one of the checkvalves will get stuck open a little.
2) a slightly stuck open check valve will leak as the piston moves. This changes the volume of liquid pumped. It is no longer "metering".
3) to reduce the chance of a sticking check valve, redundant check valves can be added.
4) each check valve increases the pressure needed to open them, increasing the chance and effect of leakage and air bubbles.

OK, that's the two big problems I've experienced. A third is piston ring wear and leakage. When the o-ring seal on the piston pump wears, it starts leaking air on suction strokes. See the air bubble problem, above. Leakage of oil or methoxide "the other way" past an o-ring isn't as big a problem because it's more obvious and doesn't change displacement as much as a bubble.

Well filtered oil and methoxide probably work well in piston pumps with check valves. My attitude is that well filtered oil is ready for better uses than biodiesel, and trying to filter methoxide requires expensive materials and toxic messes. Using sodium methylate should be cleaner, but I haven't tried it.

Parastaltic pumps require flexible tubing that is also compatible with veg oil. I haven't found a suitable hose - they either split and leak, or swell too much. I gave up that approach, but it might be worth further effort.

My remaining approaches use either mechanical valves, or solenoid valves and piston pumps. Matching displacement and materials with piston pumps is easy. The valving is the difficulty. If you enjoy machining, then I'd recommend a rotary valve and piston pump for each liquid. Less machining is needed if you use solenoid valves for both inlet and outlet of both liquids. Slave the inlets together and outlets together, and sequence the motions of the pistons together, and the two pumps will push stoichiometrically accurate quantities into a static mixer simultaneously. This does NOT require a crankshaft or complex linkage, but there are lots of ways to do it.

My current favorite is to position the two cylinders side-by-side, with the ends of the piston rods connected so they move up and down together. Got that image? The 4 valves are at the bottom. A spring pulls (or pushes) the pistons down. A feed pump (centrifugal or any other sort that generates enough pressure) constantly provides oil to the inlet valve. Methoxide at the other inlet valve does not have a feed pump. Got it?

When the inlet valves open, the pressurized oil pushes up the oil piston, which lifts up the methoxide piston. When the inlet valves close, the pistons stop. When the outlet valves open, both liquids get pushed out simultaneously. Putting a switch at the top and bottom of the piston strokes and a relay completes the pumping package.

A variant is to connect 4 cylinders opposite each other in pairs, giving flow with each stroke, instead of flow only on the "down" stroke. This would eliminate the need for a spring.

For safety, air valves instead of solenoid valves can be used, and air-logic valves at end of stroke. Or Hydraulic.

A previous favorite was to link two rotary pumps together with variable v-belt drive. I tried this using roller-vane pumps. It works really well to get "close", but not well enough for true continuous batch processing. The problem is that internal leakage of the pumps is a variable that changes unpredictable. It sped things up when mixing 700 gallons batches of oil with 140 gallons methoxide, but their would always be too much oil or methoxide left when the other liquid ran out. Still, there was a big advantage to running "most" of the mix through a short (18 inch) static mixer on the way to the final mix/settle tank. We'd then circulate the batch for a while (hours) to finish the reaction, but we'd have 100 gallons of glycerin dropped out by the time we were through pumping into the tank.

Other positive displacement pumps are also attractive - gear pumps meant for hydraulic systems are cheap, but I never found a suitable cheap one for methoxide.

More anon,

Cheers,
JohnO