|
| |
Note: Eco Sustainable Village Foundation is studying
these units for their applicability in their modules
Journey to Forever 90-litre processor
This is a cheap,
simple, safe and very effective bio diesel processing set-up that you can easily
build yourself. It's easy to make high-quality biodiesel this way.
We've used 90-litre kerosene water-heater tanks, but any similar or bigger tank
with a tight-fitting removable lid would do.
|
From right: 55-gal (200-litre) WVO storage tank (optional) with white
plastic methanol/methoxide carboy on top; biodiesel processor; holding tank;
two washing tanks. The holding tank and extra washing tank are optional, but
they effectively double the production capacity.
The methoxide is pumped into the processor from the carboy by air-pressure
from a small aquarium air-pump -- no exposure and no fumes. See
Methoxide the easy way;
see also
Adding the methoxide.
The processor lid is airtight, held in place with eight toggle latches.
|
|
The blue pump is a 1" clear water pump with a capacity
of about 2,000 litres per hour.
The pump does the mixing for the process and also transfers the finished
biodiesel from the processor to the holding tank, or from the processor or
holding tank to either of the two washing tanks.
Use standard plumbing parts for junctions and fittings. Use clear reinforced
(braided) PVC tubing. We used hose barbs at the outlet end of the transfer
hose and at the inlets in the lids of the holding tank and two wash tanks so
the transfer hose could easily be disconnected and reconnected as required.
Other hose fittings are short lengths of 3/4" steel water pipe threaded at
one end and with deep scores cut into the other end with a metal cutter, all
connected with stainless steel hose clamps.
The main expense with this processor set-up was the pump (US$35). Most other
parts were salvaged from junkyards and so on, including all the valves -- 12
of them altogether, which would have cost about US$100 if bought new. The
temperature gauge was also salvaged from a dump, found lying face-down in
the mud, but it works and it's accurate. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The pump is stepped down from 1" to standard 3/4"
plumbing fittings. In the photograph on the right, the fitting on the right
of the tank is the 1.5kw electric heating element used for maintaining the
temperature during processing. Before the process the oil is pre-heated (see
below) and then transferred to the processor. |
|
|
|
|
Pump, heater and two aquarium air-pumps plug into a
4-point extension box fitted to the processor stand. |
|
|
Inside the
processor
Clockwise from top right, the 1.5kw electric immersion heater, outlet to the
pump, the temperature gauge -- no need for a thermowell. All are sealed with
silicon.
The processor tank has a well-shaped bottom (so to speak), rounded, and then
further depressed with the outlet at the lowest point -- see below. But the
top end of the welded-in outlet pipe was raised, spoiling the effect. We
poured in some polyester resin (for fibre glass) until it was flush. This
works well.
|
The four tanks are 90-litre kerosene tanks for the kerosene-heated hot water
systems popular in Japan (especially in rural Japan). They are easy to find. Any
similar tank will do. We didn't use the more usual 200-litre (55-gallon) oil
drums because they're too big and heavy for us: using the holding tank we get
almost the same production capacity, and the processor and one washing tank can
easily be loaded into our Toyota Town-Ace diesel van so we can take them
anywhere to demonstrate how easy it is to be fuel-independent and make your own
high-quality biodiesel.
However, it would work just as well with a
55-gal drum, the open-top kind with a clamp-on lid -- or take a closed-top drum,
cut the bottom out, turn it upside down, use the small cap to fit a drain and
the bigger cap to mount a heating element, and find a cinch-clamp lid that will
fit the now-open top. Or, as a Biofuel mailing list member did, make a wooden
lid, glue a sheet of plastic on the bottom, fit inlets to it, cover the rim of
the drum with silicon piping slit lengthwise down one side, and secure the lid
with toggle latches. Other fittings can be cut and welded in place on the drum.
We had to cut the lids off our 90-litre tanks, but if you're careful about it
you can do it so the lid keeps its lip and the tank keeps its rim for a tight
fit. Use silicon for a seal and, for the processor, toggle latches to hold the
lid down tightly (it takes eight).
The tank stands are welded together from scrap steel angle-iron, free for the
taking from a junkyard.
All four tanks came ready-fitted with bottom drains, either 1/2" or 3/4"
standard plumbing fittings. The processor tank had a rounded bottom (see right),
but the other three were flat-bottomed. Once you've got the lid off, you can
stand the tank on a couple of bits of 2x2" lumber with the drain in between.
There's a rim, the bottom is recessed, so the rim rests on the lumber. Place
another length of 2x2" down inside the tank with its end over the drain, and
hammer it down with a heavy hammer. This will give the tank a slightly
round-bottomed effect, with the drain at the lowest part, maybe 1cm lower, which
is enough for good drainage. The drains are positioned towards the side of the
tanks -- lean the tanks forward towards the drain to assist with draining.
The processor and holding tank both have a T-section and two valves fitted to
the bottom drain, one valve for draining off the glycerine by-product, the other
for transferring the biodiesel to the washing tanks via the pump.
We use a couple of short lengths of PVC waterpipe narrowed at one end to fit
inside the outlet drains. The length (height) of the pipes is calculated to be
higher than the depth of by-product in the bottom of the tank after settling.
This "stand-pipe" means the biodiesel can be pumped out to the washing tanks
without getting any by-product in it. Later we remove the stand-pipe and drain
off the by-product.
Other than the outlet drain at the bottom and two caps in the lids, the tanks
did not come fitted with the various inlets and outlets, fittings for
temperature gauge and heating element etc required. We used pop rivets where
possible, for the methanol inlet valve and the toggle latches, sealed inside and
out with silicon. Otherwise holes had to be cut and female threaded plumbing
fittings welded into place.
If you don't have hole saws or drill-mounted cutters for holes this size, use
the plumbing fitting to mark out the size of the hole where required, and
carefully drill a circle of small holes, maybe 2mm, next to each other just
inside the circle. Cut through the metal between the holes, remove the inner bit
of metal, and then use a half-round file to file the edges of the hole smooth
until it fits the plumbing fitting. Then weld the plumbing fitting in place.
Just to make sure, we sealed the welds with a good layer of a tough epoxy resin
with good resistance to biodiesel chemicals, with a layer of silicon on top of
that. It doesn't leak.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The processor lid |
|
|
|
|
Right: The methanol/methoxide
inlet valve -- the inlet is 4mm I/D, which seems to match the strength of
the aquarium air-pump which pumps the methanol out of its carboy. The
methanol is added slowly: it takes 4-6 minutes for each stage with the
Foolproof process
(4.8 litres plus 7.2 litres), or about 10 minutes for single-stage (12
litres). |
|
|
|
|
|
|
|
|
|
Above: The inlet from the pump -- the mixture doesn't
plunge back into the processor in a solid 3/4" jet, which would cause a lot
of splashing and unnecessary agitation. All the agitation required takes
place inside the pump, not the processor. We fitted a 4-inch rose from a
Chinese watering can
inside the lid, with 150-odd holes, drilled out to 2mm, from whence the
mixture droppeth as the gentle rain from heaven upon the place beneath,
covering the whole surface. This works very well. |
|
|
|
|
|
|
|
|
|
|
|
Left: The lid has an extra cap, which previously housed
a fill-gauge. We replaced it with a snap-on lid with three different caps:
the second one, pictured, has a tube which vents to the ouside, fixed and
sealed inside the cap with epoxy and silicon; a third cap has a funnel, for
use with the acid-base method. |
|
|
|
|
The top of the lid: inlet from the pump top-left,
methanol/methoxide inlet valve on right, tank cap in the middle and toggle
latch in front (one of eight). |
|
|
|
The
silicon seal between the inside of the lid and the rim of the processor.
Apply a generous helping of silicon, allow to set for 6-8 hours (depending
on the weather) until it isn't sticky to the touch anymore but is still
soft. Then place the lid carefully in position on top of the processor and
weigh it down with something heavy, making sure the weight is spread evenly
around the edge. The silicon takes the shape of the rim but doesn't stick --
no need for a release agent. |
We have the idea that the level of the mixing tank outlet to the pump relative
to the bottom of the tank is quite important. With our processor the mixing
outlet is set somewhat higher than the level the glycerine by-product will
settle at. That level will vary according to the oil (and process) you use.
Especially with the rose at the inlet in the lid, there's very little agitation
in the tank, mainly just circulation. The necessary agitation takes place inside
the pump. The pump pulls the oil in from the tank outlet, but that's not very
violent. As a result, a lot of the glycerine by-product settles out during the
processing, which is a Good Thing. We think almost half of it settles out before
the processing's finished. This does also remove some of the excess methanol,
which is dissolved in the by-product.
Aleks Kac's
Foolproof two-stage
acid-base process has an optional step of draining off some of the glycerine
during the base stage. ("The process runs fine without this step," he says.
"It's just a twitch to get higher yield" if your processor has a bottom drain.)
He said this about it:
|
We'd have used this tank
for the processor if we'd found it earlier: 100 litres, and taller and
thinner than the others -- tall and thin is a better shape for processors.
But the current processor works well, no need to change it.
|
"The process is running on the smallest sensible
volume of alcohol. While removing a small portion of it with the by-product
would seem to slow the reaction down, the rather large mass of removed
by-product will tip the scale toward ester production."
We find that's the case -- the settled out glycerine by-product more than
offsets any methanol removed, with very satisfactory results.
So this could be a consideration in setting the height of the tank-to-pump
outlet.
Having said above that about half the by-product settles out during the process,
in fact much more than that does, but some of it is constantly being
recirculated (along with its methanol content), and this seems to be about
optimal. We tried to figure all this out in advance before building the
processor. Presuming about 100% production (which we get, using the Foolproof
process as standard), and 20% total methanol v/v WVO, the amount of by-product
will be 20%, or close to it. So if the depth of WVO being processed in the tank
is, say, 100 whatevers, the total depth of by-product that will eventually
settle at the bottom will be 20 whatevers, or close to it. We decided to centre
the tank-to-pump outlet (1" OD) at 30 whatevers from the bottom. It was kind of
random -- we didn't know how much the pump would pull nor quite a lot of other
things, but by dumb luck it seems to be about right, we don't want to change it.
The electric heating element is set at the same height.
This might not be very useful though -- it works for this 1" clear water pump
with this particular size and shape of processor. It's a 90-litre tank, the
diameter is 17.5" (44.5 cm) and the height 23" (58.4 cm), and we process
60-litre batches.
|
|
|
|
|
|
|
|
|
|
|
|
The pre-heating tank |
|
|
|
|
Another 90-litre
kerosene tank, with removable lid after being cut (right). The conveniently
long temperature gauge was rescued from a defunct rice-drying machine and
simply drops into a cap on the lid, no need for fittings. The tank has a
bottom drain set to one side and hammered lower, fitted with a valve. Right:
We cut the bottom out of a frying pan, cut the handle off, turned it upside
down and bolted it to the bottom of the tank to contain the heat from the
burner. The bolts are sealed with tough epoxy resin around the threads and
under the washers. It doesn't leak. |
|
|
|
|
|
|
|
|
|
|
|
|
We use either of two
burners to pre-heat the oil. Left is a kerosene pressure stove, made in a
local factory in India, common throughout the 3rd World, often called a "roarer".
This one was very kindly sent to us by our friend Ramjee Swaminathan in
Bangalore. It burns pure biodiesel very happily (right). It's efficient and
economical, using only about 60-70 ml per hour, and with no smell or fumes.
It takes a bit more than an hour to heat up 60 litres of WVO for processing. |
|
|
|
|
|
|
|
|
|
|
|
|
|
On the right is our
"Turk"-type burner, which burns raw by-product from the biodiesel process.
It burns very hot! It's made out of a sawn-off fire-extinguisher, 4" in
diameter, a stainless steel mug (the wick), and a curry can. The fuel
reservoir is salvaged from a dead kerosene space heater, the "squirrel-cage"
fan from a dead kerosene water heater. It take less than an hour to heat 60
litres and uses 700 ml of by-product to do so. There are basic plans for
Turk burners at Steve Spence's site:
|
Using the
processor
Test the WVO for water content and
dewater if necessary.
Titrate the oil (see
Biodiesel from waste oil,
see also
Better titration).
Prepare the methoxide -- we get methanol in 18-litre cans and use an aquarium
air-pump to transfer it to the methoxide carboy. Once that's done, add the lye
and proceed according to the
Methoxide the easy way
method. We use KOH, potassium hydroxide, and it's thoroughly mixed in 30 minutes
at most. If you use NaOH, sodium hydroxide, it will take longer.
 3.
Decant 60 litres of WVO into the pre-heating tank. We do this manually, using a
20-litre steel-bucket oil drum -- the same as our
5-gallon processor. The
"girth rib" about 4" from the top marks the 15-litre level so it takes four
buckets. We pour the WVO straight from the 18-litre cans we collect it in
through a 0.5 mm steel-mesh strainer into the bucket and then into the
pre-heating tank.
4. Heat the oil to processing temperature.
5. Check that all the valves on the processor are in the correct position -- all
closed except the outlet to the pump, and the outlet from the pump set to
mixing, not transferring.
6. Transfer the heated oil to the processor. You could use a pump for this, but
we found the 1" clear water pump doesn't work very well when it comes to lifting
oil, so we drain it back into the steel bucket and transfer it manually. It's
quick and simple.
7. Check the temperature on the temperature gauge. Secure the processor lid in
position with the eight toggle latches.
|
Methoxide transfer
The transfer lid is on the left: the pipe goes all the way to the bottom of
the carboy, curving round to the bottom left corner. Tip the carboy up on a
wooden block, as in the photo at the top of the page, and all the methoxide
is transferred. The air-pump inlet lid with air line is on the right. The
lid is fitted with a short length of copper pipe to fit the air line,
extending about 1/4" down into the carboy. The air-pump pushes air into the
top of the carboy, forcing the methoxide down to escape via the outlet pipe
at the bottom, through the inlet valve in the processor lid into the
processor.
|
8. Fit the
methoxide carboy with transfer lid and pipe and air-pump inlet lid with air
line, position it above the processor. Connect the outlet pipe to the inlet
valve in the processor lid, open the valve. Connect the aquarium air-pump to the
air line. Fit the vent pipe leading to the outside to the second cap in the
processor lid.
9. Switch on the 1" clear water pump.
10. Switch on the aquarium air-pump.
11. Keep an eye on the temperature gauge as the methoxide transfers -- if the
temperature falls below processing temperature, switch on the electric heating
element. (The element can be fitted with a thermostat or rheostat, but manual
control is simple -- BUT don't go away and forget that the heater's on!)
12. When all the methoxide is transferred, switch off the air-pump. Replace the
outside vent pipe with a closed cap. Leave the methoxide valve in the lid open:
methanol evaporates inside the processor, most of it simply condenses on the
underside of the lid and drips back in. The remainder can vent back into the
methoxide carboy, where it will condense.
13. You can check how the process is progressing via the braided PVC pipe from
the pump outlet to the tank inlet, which doubles as a sight tube.
14. Proceed with processing according to the method you use.
15. When the process is completed, switch off the pump. Turn the pump outlet
valve from mixing to transfer. Close the valve from the tank mixing outlet to
the pump. Open the tank drain valve (the one on the T) and the corresponding
valve to the pump inlet. Make sure the transfer hose is connected tightly with
the hose barb to the inlet in the lid of the holding tank. Check that the
holding tank drain valves are closed. Switch on the pump and transfer the
mixture to the holding tank, both the biodiesel and the glycerine by-product.
16. Allow to settle for 12-24 hours.
17. Switch the transfer hose from the holding tank to one of the washing tanks.
Use the pump to transfer the settled biodiesel to the wash tank. Close the
holding tank outlet valve. Remove the lid of the holding tank, remove the
stand-pipe, drain the glycerine by-product. Allow to settle, then syphon off the
few litres of biodiesel on top for washing with the next batch.
18. Wash the biodiesel as usual. Dry, use.
80-litre per batch closed-loop processor
|
|
|
|
|
|
|
|
|
|
From left to right :
 | Oil drum = Feedstock Holding Tank (FHT) |
| 20-litre drum = Filter Unit with vacuum pump below.
|
| Mixing pumps, condenser behind, liquid trap below
|
| 45 kilogram propane bottle = Reactor |
| Oil drum = biodiesel bubble wash and dry holding
tank, switch gear above. |
|
20-litre drum = Methoxide Mixer
|
|
 "I've
used the system once a month since last June and had no problems. But the drums
are beginning to rust heavily, especially on the methoxide side of things.
"I had my Land Rover emissions test done last year, on standard diesel and the
100% biodiesel.
"Very interesting.
"The average reading for standard was 1.22 (whatever scale they use). Biodiesel
was 0.62!!!
"The bloke doing the check couldn't believe it. They usually sell an expensive
additive for people with smoky diesels or if they're just over the MOT emission
limit."
Ian
April 2002
How it works
Click on the images one by one for
graphic presentation of the processor at work
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1. |
|
|
 |
|
|
|
|
 |
|
|
|
|
 |
|
|
|
|
|
|
|
|
3. |
|
|
|
|
|
|
2. |
|
|
|
4. |
 |
|
|
5. |
 |
|
|
6. |
 |
|
|
7. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
 |
|
|
|
 |
9. |
|
|
 |
|
|
|
8. |
|
|
|
|
10. |
 |
|
11. |
 |
12. |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Above:
Methoxide Mixer showing air tool for mixing, cap for adding caustic or
manually feeding methanol. Top gate valve for sucking in methanol from
source via transparent sighting tube.
Right: Bottom of
Methoxide Mixer showing valve to control feed to reactor and sighting tube,
which is also used as methanol pickup pipe. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Detail of stirrer |
|
|
|
|
Detail of bearing and
coupler |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Left:
The pump assembly
Right: The water trap |
|
|
|
|
|
|
|
|
|
|
|
|
I use three of these for
bubble washing. They are pneumatic exhaust silencers. They don't rot,
they're cheap and they work very well. |
|
|
|
"Top Hat" inspection hatch
and Sulphuric Acid funnel (if using Aleks's Foolproof method) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Inside the FHT showing
pickup filter and a bit of the immersion heater |
|
|
