How to turn plastic waste into diesel fuel cheaply

[Col]

Sizing of retort

How should you determine the scale for your reactor based on the amount of fuel you want to produce per day or per hour?

Hi Vast.

Here are some notes I made some time back while reading the comments from fellow contributors, and some YouTube videos, might help you in sizing the retort. Probably only relevant to smaller systems though.

Guys with 2kW heaters are getting their 19-20L retorts up to 400C in 1-2 hours, or 45mins to 350C.
BBD—2kW heater at base of retort, then extra heater bands and line wrap, 19-20L, 2hrs to heat, gets 1L/hr from start to finish, but nothing in the first hour while the temperature rises, so probably 2L/hr during the peak cracking/distillation period. 11hr batch time. 1L consumes 100W.
YouTube - an oil burner under a gas bottle was getting up to 1 gallon per hour but found it was too fast and hard to control so dropped it back to 1 gallon per 3 hours. He gets a gunky oil off the bottom and a good oil/diesel condensed off the top. Another guy seems to run at a similar speed with a similar sized system. Don’t know about residue left behind.

I have some information on instant evap v boiling pot methods. I'll dig 'em up and post shortly.
Col

[Col]

Small retort, fast distillation

Pfaulder has been doing this for a long time, Short path, high vacuum distillation in wiped film evaporators. At pressures of 0.1mmHg they evaporate up to 85% of the used oil between 260C to 301C. The water and light ends are separated before the WFE process. The remaining slurry can be used as a road base extender. The benefit of conducting the process under high vacuum is the lower temperatures don't cause degradation of the hydrocarbons. The short residence time also plays a part here, the heated evaporation plate is very near the condensers so the molecules are only 'hot' for a short time, also reducing the chance of degradation. They maximise volatile extraction by agitating/progressively wiping the feedstock down the heated wall of the retort.
While their website makes for interesting reading we are looking for principles we can apply in a DIY application. The principles here, as you have identified, are high heat, short residence time (if we could apply a strong vacuum we would go for a lower heat and short residence time). We can achieve this, as Asad said some time back, by using a wide, shallow depth of feedstock, over a series of smaller flames/heat source (with a mixer/wiper for further improvement). To size the retort you could take Excalibur's figures of fuel used : oil distilled, over the lapsed time and work from there.
Col

[Col]

reducing bush

A plate welded at the desired dividing point between the two chambers would have a center port hole bored. Reflux temperature might be controlled by the insertion of a "reducing bush", so in this way the heat can be manipulated.

Hi Excalibur.
What do you mean by a 'reducing bush'? I have made my splash plates out of 2 circle plates. The retort is 800mm diameter. The lower splash plate is a 700mm diameter, 5mm thick circle plate, centred in the retort, 50mm in from the retort walls. The second, higher plate is a large 'washer' which meets the retort walls at the circumference but has a 100mm hole in the centre. This way the vapours must pass from the liquid surface to the outer edge of the first plate, between the two plates which are about 50mm apart, then to the inner hole of the second plate, then up to the mesh pack in the reflux. I figure the vapours will pass easily but any splashing liquid won't reach the mesh packing. Is my second plate what you mean by a 'reducing bush'?
Col

[Excalibur]

I was wondering after reading Excalibur`s account of his efforts, if you want to create a continues feed system, should you reduce the amount of feedstock in your retort to get a faster reaction from processing the feedstock?

Will it not be better to have less feedstock with temperatures that can almost instantly convert your continues feed into gas than to heat the feedstock up to the correct temperature because the feedstock level is to high, thereby when introducing new colder feedstock dropping the temperatures in the retort and slowing down the process?

If one would want to create a system that can deliver 50 liter diesel per hour, how big should the reactor unit be to allow for such a system?

If the feedstock level can be kept lower but the pyrolysis takes place on a faster scale, will that allow to have a smaller reactor unit or will that defeat the purpose to try and create a chamber where the reaction can take place faster in a shorter time period? A smaller reactor means a faster time to achieve the heat requirements, less feedstock in the chamber, can also reduce boil over of feedstock, but that also means that maybe you can`t reduce the size of the reactor?

How should you determine the scale for your reactor based on the amount of fuel you want to produce per day or per hour?

This is a very interesting question that I agonized over when I built my current "Orion" retort. One thing has become clear, is that one present limitation is the amount of heat that can be poured in. If incoming feedstock isn't matched by heat applied, then the it doesn't matter so much whether the retort level is high or low. The temperature can only drop. I'd need to slow down the feed to match heat or increase heat to match feed. Another possible is to increase the heat of the feedstock itself. I have a huge amount of heat I can generate, it's just that I'm not applying it in the most efficient manner.

To size the retort you could take Excalibur's figures of fuel used : oil distilled, over the lapsed time and work from there.

I would just caution, those figures will be distorted because of the unknown quantity of retort gases used to make heat.

Hi Excalibur.
What do you mean by a 'reducing bush'?

Think of this terms of a simple method of reducing the throat size between the two chambers.

[Beyond Biodiesel]

Came upon this though could be useful here.
Pyrolysis of grasses and leaves to produce bio oil.

Thank-you, mercedes 308, for posting this link. A tube furnace might be just what is needed for the person interested in a continuous process system.

In response to further questions and comments here, I have been exploring vacuum distillation for the last 2 years. I do it in a 5-gallon (20L) system. I reason that I can easily adapt that system to a continuous process flash evaporator when I wish to move forward; however, distilling 5-gallons (20L) of WMO every day is as much fuel as I need at this time.

[Col]

Auto levelling of oil for a continuous system

As for maintaining a certain level in the retort maybe this would work - set a horizontal pipe at the desired oil level, passing from inside the retort out to another tank. This tank has a removable lid for checking on the level, and a quality floating valve similar to those used in a toilet cistern, but all metal. This tank can be made as small as is practical to fit the float valve in, 40L? The float valve is fed from a bulk oil drum which can be topped up as required. As the vapours pass from retort to condensers, oil flows from intermediary drum to retort and oil feeds from bulk drum to intermediary drum via float valve. The oil finds it's own level between intermediary drum and retort, this level is set by the float valve.

This idea I suggested some time back has whiskers on it - if the oil feed malfunctioned/stopped for any reason (bulk supply ran out, feed pump to header tank faulted, float level jammed) and the retort kept going the oil level would drop until there was no longer an oil/liquid seal and air would enter the retort - KABOOM!
Someone has suggested measuring the mass of the distilled/cracked products in order to match the feed to the output. That could be done I guess except for the off gases, and there would be a delay between the oil cracking and the vapours condensing resulting in fluctuations in oil level, which may not be a major problem in a retort filled deep with feedstock, but if we aim for a shallow retort how do we maintain a steady oil level? I expect the bubbling will be too violent for a float level to work, even if it could handle the harsh conditions. Perhaps the only way to reliably do it is to run a few feed rates through it and monitor conditions, when the feed rate gets too high you've found your maximum rate.
A thought - I expect the heat supply will be one of the more difficult things to control if using a liquid fuel. If you settle on a good heat supply and just keep it steady you could manipulate the feed rate up and down to control the temperature. If you have enough temperature readings in the right places you wouldn't need to know the oil level, you could tell what was going on inside the retort by what the temperatures are doing.
I would appreciate anyone's thoughts on this.
Thanks, Col

[Excalibur]

On Ebay I picked up a copy of The Petroleum Handbook compiled by Royal Dutch Shell, 3rd edition 1948. I paid $1 and with shipping another $28. The book is a bit battered but perfectly legible.

I highly recommend the book for studying distillation and cracking of hydrocarbons. Some of the principles of refining are quite difficult to understand though this book delves into the subject, explaining the methods, procedures, etc in ways that a layman can grasp with relative ease. I'd read portions of this book on loan a couple of years ago.

The book is old but the basic fundamental principles have not changed. It might even be argued that the book is more relative to the DIY enthusiast because of the straight forward approach without use of super exotic materials, modern technologies and computer controlled programming.

658pages
Of interest is smell removal methods for fuels, naming it mercaptans, a hydrosulphide.

[Col]

tube furnace for a continous system

Thank-you, mercedes 308, for posting this link. A tube furnace might be just what is needed for the person interested in a continuous process system.

References are made to rpm but it doesn't seem to say that the tube is rotated. I assume the rpm references are for the cylinder rotation as some people are doing for tyre pyrolysis. What is the benefit of this method for a continuous system? I guess it's one way to spread the feedstock over the heating walls and keep the feedstock mixed. Any other benefits?

[Excalibur]

Someone has suggested measuring the mass of the distilled/cracked products in order to match the feed to the output. That could be done I guess except for the off gases, and there would be a delay between the oil cracking and the vapours condensing resulting in fluctuations in oil level, which may not be a major problem in a retort filled deep with feedstock, but if we aim for a shallow retort how do we maintain a steady oil level? I expect the bubbling will be too violent for a float level to work, even if it could handle the harsh conditions. Perhaps the only way to reliably do it is to run a few feed rates through it and monitor conditions, when the feed rate gets too high you've found your maximum rate.
A thought - I expect the heat supply will be one of the more difficult things to control if using a liquid fuel. If you settle on a good heat supply and just keep it steady you could manipulate the feed rate up and down to control the temperature. If you have enough temperature readings in the right places you wouldn't need to know the oil level, you could tell what was going on inside the retort by what the temperatures are doing.
I would appreciate anyone's thoughts on this.
Thanks, Col

Col, good on you for giving this aspect some proper focused thought. Yes, matching feed to output is cumbersome. While I haven't measured, the off gases probably amount to a few percent. Yes, if feed matched output and the heat input could be stable, that would be magic. In practice it hasn't been so easy. Originally, I thought I'd be able to crank the heat up high, then apply the feed till a state of equilibrium was reached. One upgrade I'm planning is to use a needle valve instead of ball valve for improved flow control of feed. A ball valve is on or off with little adjustability in between.
I've wondered about a float in such harsh conditions too. The froth and frantic bubbles might affect buoyancy. A float would need to be made from a reasonably thick gauge metal. Could the float be filled with something other than air, perhaps a substance giving good buoyancy but less affected by retort activity?
One surprise happened that I'm attributing to a low oil level in the retort was that when feed was applied, it immediately flash-boiled. This resulted in dark red output with particles, presumably carbon.

[Col]

Float level

I've wondered about a float in such harsh conditions too. The froth and frantic bubbles might affect buoyancy. A float would need to be made from a reasonably thick gauge metal. Could the float be filled with something other than air, perhaps a substance giving good buoyancy but less affected by retort activity?
One surprise happened that I'm attributing to a low oil level in the retort was that when feed was applied, it immediately flash-boiled. This resulted in dark red output with particles, presumably carbon.

I don't think I would want a float filled with air inside the retort come to think of it. If not air, then what? A large, disc-like float might be good, it would 'absorb' a lot of the violence inside the retort and give a more steady read, but the bigger the float is the more it impinges on the surface area of the oil which we need for evaporation.
If we had a pipe off to the side of the retort (to get away from the splashing etc) with a steel ball floating in it, and made the pipe of borosilicate or some other non-metallic material, we could use the floating steel ball to actuate a series of magnetic relays which would then tell us the oil level - like this...

Uhz 517c11 High Temperature Liquid Water Level Sensor With Transmitter From Water Meter Manufacturer Max 450 Centigrade 150lb - Buy Liquid Water Level Sensor,Water Meter Manufacturer,Pool Water Level Sensor Product on Alibaba.com

Udm-10 With High Quality Liquid Level Measurement Sensors - Buy High Quality Liquid Level Measurement Sensors,Fuel Level Float,Plastic Water Level Indicator Product on Alibaba.com

Liquid Level Sensor - Buy Liquid Level Sensor Product on Alibaba.com

The level readings from all three could be assisted by putting baffles made from perforated sheet around the level sensor. You could also program the controls for the feed pump to wait for the sensor to stabilise over a period of a few seconds before actuating the pump, just to be sure.

Apart from the colour and sediment, what was the quality of the product produced by flash-boiling? Still ok?

[Col]

steady float level

The level readings from all three could be assisted by putting baffles made from perforated sheet around the level sensor.

Actually, you could just put a sheet wall around the level sensor so long as it was open at the bottom and top to allow the oil level to rise and fall. The smaller the allocated enclosure the less turbulence.

[Col]

Petroleum handbook

...The Petroleum Handbook compiled by Royal Dutch Shell, 3rd edition 1948. ...I highly recommend the book

Thanks, just bought one!

[rozier56]

Hi BB,Sorry been off-line for a few day's. when i say the length is 850mm[85 cm] and the radius of the retort is 333mm[30,3cm].Sorry we are metric system here,
thks

[Excalibur]

I don't think I would want a float filled with air inside the retort come to think of it. If not air, then what? A large, disc-like float might be good, it would 'absorb' a lot of the violence inside the retort and give a more steady read, but the bigger the float is the more it impinges on the surface area of the oil which we need for evaporation.
If we had a pipe off to the side of the retort (to get away from the splashing etc) with a steel ball floating in it, and made the pipe of borosilicate or some other non-metallic material, we could use the floating steel ball to actuate a series of magnetic relays which would then tell us the oil level - like this...

The level readings from all three could be assisted by putting baffles made from perforated sheet around the level sensor. You could also program the controls for the feed pump to wait for the sensor to stabilise over a period of a few seconds before actuating the pump, just to be sure.

Apart from the colour and sediment, what was the quality of the product produced by flash-boiling? Still ok?

How about an inert gas, perhaps argon or nitrogen for the float. Some study needed on this.

Sensors for the level would be great for automatic feed. It could also be connected to an alarm system.
Brainstorming a couple of months ago, we came up with a DIY idea for a float level device. Get a pneumatic air cylinder with a magnet type piston in the length of stroke equivalent to the max depth required for the level. Mount the cylinder vertical so the spear protrudes into the retort with float attached to its' end. Ensure the spear is free to move and the cylinder to retort has a perfect seal. Attach proximity sensors to the cylinder to sense the magnetic piston position. Remove any seal on the cylinders' piston and plug the air in/outlets.

Product from flash boiling was mixed in with the other 180 liters or so but all looked well enough. The feedstock hitting the hot zone may have cracked the oil to some degree but the important thing is the free carbon wasn't bonded onto the oil any more.

Yes, parts of that book are right up our alley. I was "like a dog with a bone" when I first read it.

[Excalibur]

Hi folks,How is this idea?
My retort is 850mmlengthx333mmwidth radius.=296metres.
Proposed retort to be placed directly on top of retort is 1meter x 12mm width radius =45 meters.
How does that sound keeping in mind that 1/4 of 296 =74meters.
Another point is that the retort temp in my case is always higher than my reflux temp.Temp is measured in retort open area and reflux at the top of the reflux

Hi BB,Sorry been off-line for a few day's. when i say the length is 850mm[85 cm] and the radius of the retort is 333mm[30,3cm].Sorry we are metric system here,
thks

There are still some errors in your earlier post with dimensions, etc
Please review and correct these.
Is the reflux radius 12cm not 12mm??
Is the retort 296 liters or .296cubic meters (not 296 metres)??

Edit: Ok so I think I worked out what you are saying. The proposed reflux is maybe too small being only 45 liters when the 1:4 recommendation is 74 liters. If it's what you have and especially if it's already made, I'd go with it. Consider mounting it a bit further away with a short length of pipe to compensate for the lesser volume.