Designing an Electric-fired biochar reactor

[CadeJohnson]

Experiments and discussions in the first half of 2023, regarding small-scale biochar reactors have made me realize that the pyrolysis of biomass and evaporation of biomass residual moisture are endothermic, but few practical measures of energy input requirements have been made. The concept of burning fossil fuel to provide energy for pyrolysis is ludicrous in the context of CDR, and the concept of burning biomass for fuel to provide energy, while perhaps practical for relatively primitive efforts, is problematic in the context of CDR for two reasons: it is inefficient - some of the carbon in the biomass is consumed in combustion and released as CO2, and it is wasteful of a potentially valuable resource - the pyrolysis gas has potential value as a replacement chemical feedstock as fossil fuel uses are deprecated worldwide. A solar-heated pyrolysis device would be useful, but what energy input is required? An electric-fired biochar reactor can provide data for the design of such a unit.

I sought to measure heat transfer to biochar reactors with external heating, and have realized that accurately measuring all the elements of heat transfer involved is an effort that exceeds my DIY capabilities. However, it may be possible to accurately determine heat input by placing the heat source inside the biochar reactor. An electric heating element inside a well-insulated reactor can be monitored simply - by observing the reactor temperature with a thermocouple, and by recording the electrical power consumption with an ammeter.

Reactor D Design

The reactor vessel will be a one-gallon paint can. The can will be insulated by coating it first with a layer of clay-plaster mix, then with additional plaster-sand mix. The clay-plaster layer will give rise to air voids if the plaster is degraded by heat. The total thickness of the insulating layer will be greater than 3cm.
The reactor vessel will have a lid that is removable, that forms a tight seal when closed, that is well insulated, and that has a vent for the escape of pyrolysis gas. A paint can lid can make a good seal to the paint can, but it requires some pressure to seal, and then some potential deformation to pry open. This is not compatible with clay and plaster insulation. Therefore, I anticipate forming a gasket in the seal-canal of the paint can, made from silicone rubber which is stable to about 300C but may degrade at higher pyrolysis temperatures. Alternative seal materials may be necessary. The paint can lid will be covered by a plaster cap similar to the reactor walls and base.
The reactor heater will be a terracotta rim of a flower-pot wrapped with ni-chrome resistance wire and connected to 120 volt AC power. The power will be regulated by a relay switch.
The temperature sensor will be a Type-K thermocouple mounted through the wall of the reactor vessel approximately 5cm below the top of the reactor
The controller will be an Arduino Uno PLC. It will monitor the output of the thermocouple and control the relay switch that provides power to the reactor heater. The controller will have three modes of control. At temperatures well below a target value, the relay will be engaged a high fraction of the time, possibly 100%. At temperatures near the target value, the relay will be engaged a lower fraction of the time but sufficient to cause continued temperature increase. At temperatures above the target value, the relay will be engaged infrequently to maintain the target temperature.

Components of Reactor D, clockwise from top left:

SSD1306 0.96 inch OLED display
Arduino Uno
modified flower pot, 14cm outside diameter, wrapped with nichrome wire R = 29.9 ohms
relay module - this module has ground and 5VDC terminal which proved to be active; it will power the arduino via the 5V bus. This is NOT a recommended way to power an Arduino, so this terminal should not be connected.
MAX6675 thermocouple and amplifier - the amplifer failed prior to any experiments.

The modified flower pot fits into the one-gallon paint can (I checked!). Biomass will be placed into a metal food can in the center of the modified flower pot so it does not touch the nichrome wires - the remnant base of the flower pot, not shown may be used to insure that the can and the nichrome wires remain isloated.

[CadeJohnson]

I have completed the preliminary control circuit and arduino sketch.

The controller monitors temperature and a set-point input and operates a relay on a 50 second cycle based on PID analysis of the temperature error. The controller displays the elapsed time since power-on and the cumulative on-time for the power relay. For making control parameter adjustments, it also displays the PID response parameters.

The Arduino libraries and devices are described in a separate post.

The desired operation is that the controller will use the PID response 0 - 255 as a signal to apportion a 50 second heater cycle between on-time and off-time. The minimum off-time is one second and the minimum on-time is one second. This avoids “relay clatter”.

When the temperature is approaching the setpoint, the proportional response factor will be insufficient to eliminate the remaining error due to heat leakage from the reactor. Therefore the integral response will be needed to “find” the additional power input required to maintain the reactor temperature. Due to the thermal mass of the reactor, derivative response is not likely to contribute to the control function.

During preliminary tests, it will be beneficial to measure reactor response with varied specimen properties: a low thermal-mass (empty specimen can), and a high thermal-mass (can with measured mass of water). These response profiles can be compared to later biomass specimens to extract characterization of biomass moisture content (early pyrolysis phase) and pyrolysis enthalpy (late pyrolysis phase).

[CadeJohnson]

This is the base platform for the electric-fired reactor. The mounting surface is a scrap of formica countertop (plywood under the formica -flammable) There is an inverted flower-pot dish braced with five black screws. Not visible under this dish is a sheet of metal for heat-shielding. A gallon-size paint can lid is affixed to the flower-pot dish with a single central screw. A ground wire is attached to the head of that screw. Two insulated copper wires pass through the pot dish and lid and connect to the ends of the NiChrome wire heating coil. The holes in the paint can lid are sealed with silicone caulk. Silicone is not rated for the full temperature potential of the reactor, but even if it is partially failed, it may provide some sealing effect. We shall see!

The insulated copper wires and ground wire can be seen emerging from beneath the edge of the pot dish and connecting to the relay module. Connections are as follows: two central terminals are AC supply to the module. A short jumper from the “hot” AC supply connects to the relay common leg. One of the heater wires connects to the relay normally-open leg. The other heater wire (white) connects to the “neutral” AC supply. The bare copper ground wire and the AC power cord ground wire are connected to the unused common terminal of the second relay in this module. The black wire to the heater element has an upward oriented loop so an inductive/clamp ammeter can be used to measure the heater current. The wiring scheme described makes the can grounded and AC power is only present in the reactor when the relay is engaged.

The side of the relay module opposite from the AC connections has a terminal block for the DC connections to the Arduino controller.

Second photo shows the reactor can in place over the heater base. It has two penetrations - one is the thermocouple and one is copper vent tube. Both holes are sealed with metal fittings screwed into the metal side of the can (3/8-inch for the copper tube fitting, 3/16-inch for the thermocouple mounting fitting) both seals are
supplemented with plaster caulk. Wire lifting handles are secured to the can. It will be coated generously with insulating clay-sand mix and a thick shell of plaster-of-paris.

[CadeJohnson]

completed set-up; when the plaster is well-cured (and I have time), I will begin a series of tests.

[CadeJohnson]

partial materials list; plaster of paris (way too much), metal paint can, ceramic pots for reactor base and heating element support - subtotal cost about US$45. Not included in price: copper tubing, arduino Uno with patch wires and misc minor electronic components, breadboard, relay board, thermocouple/amplifier, OLED display, AC power cord (salvaged), mounting board (salvaged).

[CadeJohnson]

Experiments showed that the plaster of paris was not an adequate insulation. Therefore, the plaster was removed and replaced with fiberglass. The fiberglass insulation was “duct board” which is one-inch thick rigid material with an R-factor of 4.3 (specifically UL Listed Air Duct Class 1 (form A), Owens Corning. A single four-foot x eight-foot panel cost about $40 - substantially more material than needed, unfortunately.

This material owes some of its rigidity to paper-foil backing on one side, and some internal bonding with polyester epoxy. To cover the paint can, first removed the paper and foil backing, thenI cut pieces from the duct board the height of the can, and affixed the strips like barrel staves around the can - secured with masking tape. I cut a circular piece to cover the can end and the first layer of duct board, then I affixed a second layer of duct board pieces - covering any gaps that may have been visible in the first layer. Then I added a second circular piece over the can end. I attached the second layer by wrapping with aluminum tape, and I added a hoop of wire ever so often, in case the heat caused tape adhesive failure.

This insulation contains ingredients not compatible with the desired operating temperature, and the first heating caused these ingredients, paper tape and adhesive, and polyester epoxy to pyrolyze and escape as smoke - a nuisance; good ventilation was essential. But afterward, the insulation capacity was still the same.