Crumbler Head Temperature Sensor

The purpose of this project is to develop a method of monitoring the surface temperature of a biomass grinder drum to notify the operator when product could potentially become damaged.

Background
Forest Concepts LLC. is a company that manufactures and maintains Biomass grinders for use around the country. Biomass is produced by taking chips of a variety of materials that can range from wood chips to corn stalks and running them through a grinder to reduce the chips into a size that can be rapidly burned for a heat source. Forest Concepts came to the University of Idaho seeking a design that could measure the surface temperature of the two rotating drums that grind up the material to monitor for overheating which could potentially damage their product.

Deliverables
Forest Concepts has asked for the following deliverables:
 * A reliable method of measuring temperature to ensure the product is not damaged
 * A cheap sensor array for a permanent fixture that can be implemented on all crumblers or
 * A removable sensor array that can be easily swapped between grinders if need be
 * A sensor output that is readable by industry standard PLC's
 * A method of verifying the temperature reading is accurate

Specifications
After communicating further with Forest Concepts, the following specifications were determined:
 * Maximum allowable temperature of 200 degrees Celsius


 * Common running temperature of 100 – 150 Degrees Celsius


 * Accuracy of +/- 10 -15 Degrees Celsius
 * 4-20mA industry standard for Sensors


 * 0-5 DC volts works for our operating system with their PLC


 * The main products are Wood Chips, and Corn Stock


 * The ash from these products is usually less than 3%


 * Crumblerhead is made of A2 Tool Steel


 * The teeth of the Crumblerhead are a Carbide Tip


 * Must have a method for the verification of effectiveness of our Temperature Sensor System


 * Prototype Sensor system will be tested on Forest Concepts Research Head

Emissivity
Emissivity is the ratio of energy radiated from a materials surface to that radiated from a black body. So, a black body would have an emissivity of 1.

Using the Stefan-Boltzmann law, we can determine the energy emitted by our drums as:

𝑀= 𝜀𝜎𝑇4

Where M is the flux per unit area of the source, 𝜀 is the emissivity of the object, 𝜎 is the Stefan-Boltzmann constant (5.67∗10−8𝑊𝑚−2𝐾−4) and T is the surface temperature of the object.

From this equation, we can derive the temperature for our infrared sensor which is called a Lambertian Source.

This sensor treats the entire surface area of the drum as a plane. Our sensor calculates the radius of the surface area projection with a constant Spot Ratio of 1:1 (1cm away = 1cm diameter of circle).

For our IR sensor type we can actually adjust the emissivity to match that of the material that the IR waves are being pointed at. The only major factor that could hinder our project would be the possibility of dust in between the IR and the crumblerhead.

Common values of emissivity based on other materials include:


 * Oxidized steel - .79
 * Silicon Carbide - .83-.96
 * Gravel - .28

Thermocouples
Thermocouples are composed of two dissimilar metals the generate an electric potential as one wire heats up faster than the other one based off of the material properties. For our experiment, a K-type thermocouple would most likely be the choice for measuring the temperature of the outgoing product if the customer so decides to go forward with that plan.

Degredation of Product
The entire reason of monitoring the crumbler head temperature is to ensure that ForestConcepts product is not being damaged by overheating. The three main components of wood that can actually melt if raised to roughly 450 Fahrenheit are Lignin, Cellulose, and Cellulose-lignin. If the temperature of the lignin is raised to this threshold, the normally-rigid protein will actually behave as a more fluid medium, causing permanent structural damage to the product as well as a loss of energy content in the burning during biomass fueling.

#1: IR Sensors
Design Idea: With each crumblerhead there are two cylindrical teethed rotating drums. So, for each of these drums we would have 5 IR sensors of Forest Concepts approval which would be evenly spaced out lengthwise along the drum. Each of these sensors will pick up a temperature for each section of the drum and whichever sensor is reading the highest temperature of the 5 will be displayed ion the PLC of the crumblerhead. The 5 sensors will be housed on the outside of the metal plates surrounding the crumbler head on either side of the drums for a total of 10 sensors per crumbler.


 * Pros


 * 1) Is a Permanent fixture to the system
 * 2) Cheap/easy maintenance
 * 3) Accurate to Forest Concepts wanted Specs
 * 4) Simple and effective


 * Cons


 * 1) Possible errors when high ash in-between the sensors and the drums
 * 2) Could have coding errors when trying to create functions to show max temperature with the sensor array

#2 Thermocouple on end product
Design Idea: At the end of the conveyor belt where the product is eventually drop into a container or bailer, we would have a thermocouple rod that as the product falls by it or passes by it on the conveyor it makes contact and the thermocouple would readout that temperature to the PLC.


 * Pros


 * 1) Directly measuring the product rather than the crumblerhead drum
 * 2) Simple to setup and install
 * 3) Cheap
 * 4) Is a permanent fixture on the system


 * Cons


 * 1) Hard to make sure your temperature readout function is accurate
 * 2) The time between the product leaving the crumblerhead and where the thermocouple is at may give the product time to cool down, providing inaccuracy.
 * 3) Possibly not very durable

Chosen Design
For our design, we have chosen to use the infrared sensor arrays on both drums of the crumbler. These arrays will consist of 6 sensors per side for a total of 12 sensors per unit.

Sensor Housing


Part of the dilemma with the environment in which our sensor will be working is the amount of moving parts and particulate debris that will be circulating around our field of view. With that being said, we needed a sensor that was:


 * rugged enough for use without maintenance
 * fixed to the existing system
 * easily inserted with minor modifications to existing side plate

The initial housing design consisted of a rapid-prototype housing with flanges to be inserted into a keyway in the side plate. However, after further discussion with our clients, keyways are much harder to machine and much more expensive so we proceeded with the route of making the sensor cylindrical with a fixture on the outside of the plate. We have decided to fix the sensor housing to the plate wall via magnets due to the fact that they will not come loose with time as well as getting an IR-visible Lexan cover to keep dust out of our sensor housing.

Arduino Housing
One of the other challenges in our design was constructing a safe place for our Arduino to be protected from the elements while still being accessible for connections. Our initial idea for an Arduino Housing is shown to the right. For prototype testing, the box was cut using the UI laser cutter out of 1/4" wood for ease of production. This box ensured that the Arduino and sensitive components would be protected from the elements during prototype testing.

Preliminary Data Acquisition
In order to being determining the emissivity of our sensor, ForestConcepts has asked us to create a Labview Program for use with our sensors. This can easily be done using an Arduino to control and gather information and use a serial communication with Labview in order to decipher the data. In order to do this, the file:Arduino sensor code.pdf must first be loaded into the arduino. This code cycles between one sensor every 1.5 seconds and is configurable for any 3 pins for Clock, Data, and Acquire pins.

Next, the Labview VI must be constructed to communicate with the Arduino via the VISA add-in. The back window (or block coding of the program) can be seen to the left.

The arrangement of coding seen in the first picture of this section is the coding aspect of the Labview program. To the left, the user interface while using the application is shown. This is what anyone recording data will see while data is coming in. For preliminary data acquisition, there are 4 graphs: ambient and infrared readings for both sensors. Also, to debug any errors, there is a window to view the incoming strings through the serial port.

For the Serial communication from the Arduino to Labview, we have set up strings using comma delimiters so Labview can sort our code for usable information. For example, one string coming in could be:

Ambient,1,79.21

By using commas in our code, we can tell Labview to take the following steps for sorting the data:
 * 1) Look for the first comma, take the information to the right of it which is the sensor number (in preliminary testing either 0 or 1 for sensor 1 or 2) and make a branch for the sensor number
 * 2) Once the sensor number is found, take the data to the left of the first comma, either ambient or infrared, and create another branch for each sensor
 * 3) Labview then looks for the second comma and places the numerical values in the correct graph

Experimental Setup
The sensors were shipped to Forest Concepts to test the viability of the design. Initially, there were only 2 IR sensors sent with the coding to take data. Data was taken over the course of several hours, using a thermal imaging camera on the crumbler heads to compare the temperatures the infrared sensors were reporting. Forest Concepts duct-taped our sensors onto the feed surface of the crumbler to avoid having to make alterations to the pre-existing prototype research crumbler.

Prototype Data Results
The graph to the left displays the reading of one of the sensors over the course of the testing. Mostly, the test was just to ensure that we were getting accurate readings with no hiccups in the code or data presentation. These readings yielded an accurate temperature within 10 degrees Fahrenheit of the thermal imaging camera, well within our required specification. The detailed experimental setup can be found in the archives of this page.

Design Implementation
For the final chosen design, we will make a 12 sensor array to attach to the crumbler and use shielded wires to run data to the Arduino Mega. Once the Arduino gathers the data, it will be output to Labview in a serial format for prototype purposes.

Arduino Code
For the final arduino code loaded onto the Mega, there was a modification made to allow any number of sensors to be connected and still run. The Arduino cycles through all sensors in the Void Setup stage, and then proceeds to loop through any sensors that are plugged in.

Labview Code
For the final labview code, NI VISA now sifts through the strings and populates an array with the temperature values. This array is then entered to a waveform graph, using a shift register to only update the changed values in the array every cycle.

Sensor connections to Arduino
Initially, a 5 wire shielded cable was going to be ordered and custom cut to length for the crumblers. However, after further discussion with Forest Concepts and their plans to use the sensors on an array of different crumblers, our team decided to use ethernet cables and connectors to help the Arduino communicate with the sensors. Ethernet cables are a very practical application for a variety of reasons: Ethernet cables have provided a very effective and easy to use connection method for our sensors.
 * They are shielded, helping cancel out any potential noise that could disrupt a signal
 * They are cheap
 * They come in a variety of lengths, leaving Forest Concepts plenty of room for different sensor configurations

Arduino PCB Shield


In addition to the Ethernet ports that must be connected to the Arduino to communicate with our sensors, the circuitry required to connect all 12 sensors posed a huge problem as far as packaging. Feifan designed a PCB using the Eagle softward by AutoCAD to design a shield that plugs directly into the Mega and is easily solderable for the female Ethernet ports on the shield itself.