Dual Robot Log Handling

From Mindworks
Jump to navigation Jump to search
Sponsors
Team Name RoBoLoG
Duration Fall 2019 - Spring 2020
Faculty Adviser Steve Beyerlein
Mentor Nikki Van Rooyen
Client Idaho Forest Group
  • Jeremy Fromm
  • Chad Kosmiki
  • Alex Gomez
Team Members
  • An Le
  • Deryk Ahner
  • Jongin Hwang
  • Sam Malinowski

The goal of the project is to design end effectors and program two 6-axis robots to pick up, rotate, and manipulate a log through a band saw to cut a log into a cant. The end effectors will bolt to the 6th axis of each robot and secure the log by penetrating the wood at each end. The cutting planes for each log will be determined by an optimized scan and communicated to the robots via a Programmable Logic controller. In the first year of the project, our goal is to complete the end effector design and program the robots to cut a single face of the cant.

Project Value[edit | edit source]

In modern sawmills, logs are cut into lumber through a massive automated assembly line of hydraulic equipment, which is expensive to maintain and requires extensive computing power. This could be replaced by a two-robot system which can cut logs with greater efficiency and reduced overhead, maintenance, and computing costs.

Product Requirements[edit | edit source]

End Effector Programming General
  • Grip the log on its ends, preventing any slip
  • Member should be at least 6 inches long
  • Member should be less than 3.25 inches wide and 1.5 inches tall
  • Spike should penetrate the log no more than 0.5 inches to minimize waste
  • Resistant to wear at the points of contact
  • Have a high stiffness spring mechanism to absorb shock, prevent the robot from detecting a collision during log penetration
  • Strong enough to avoid fracture due to bending
  • Economical to produce, install, and uninstall

PLC:

  • Accept data from optimizer as input variables (log length, diameter and height at each log end)
  • Convert variables to be readable by robots and output to robots
  • The PLC is programmed via Studio 5000 software.

Robots:

  • Program to pick up a log, move it to the initial cutting position, and manipulate it through the optimized cutting path
  • The robots are Fanuc R2000iC 210L Robots simulated via Fanuc's robot simulation software Roboguide.

HMI:

  • Start and stop the robots and display errors
  • Display log geometry
  • The HMI will be developed with Ignition software.
  • Log diameter ranges from 5 to 12 inches
  • Log lengths up to 8 feet
  • Log mass does not exceed 350 kg (larger logs would be at risk of exceeding the maximum robot load of roughly 2 kN).
  • Angular orientation of the log is accurate up to ±2°
  • One face of the log is cut in ≤ 30 seconds


End Effector[edit | edit source]

The biggest challenges in the end effector design is implementing shock absorption along the log longitudinal axis without compromising bending under the gravity load of the log and while remaining within the strict size requirements. We constructed a rapid prototype to test the shock absorption method of mounting compression springs around the mounting bolts. This design would require 2 plates with the springs located in between each plate. The prototype can be seen below.

Rapid Prototyping
Rapid Prototype Description
Rapid Prototype1.jpg

This prototype was assembled with wood laser plates, wooden dowels,
wood glue, nuts, bolts, plastic bushings, and conical compression springs.
We simulated loading by grabbing the back plate with one hand and the end
of the dowels with the others and pulling down and pressing in on the
dowels. This gave us a good idea of the the viability of mounting springs
around the bolt pattern.

Rapid Prototype Problem.jpg

After simulating the load of pressing into and holding a log and
conducting a mathematical analysis of the compression force on each
spring, we found an issue with the design. This can be seen in the
table below. To balance the moment created by the weight of the log, the
spring pattern needs to provide a counter force. Since this force is
significant relative to the expected pressing force, the lower springs
will compress more than the upper springs. This would cause a vertical
deflection of the end effector tip, compromising the design concept
and send us back to the drawing board.


Design Solution
Design Diagram Description
End Effector Brainstorming Diagram.jpg

To solve the vertical deflection problem, we brainstormed a design which
includes a shaft and a shock absorption mechanism within a central housing
member. This shock absorption can be accomplished with a standard compression
spring. As the spikes are pressed into each end of the log, the shaft will
slide along the housing member and compress the spring. The member can be
mounted to the plat via a flat head screw, and the spikes are mounted to the
linear motion shafts. The spikes and shafts are connected through the spike
mounting plate which bolts to the housing member. This design will solve the
problems illustrated by the rapid prototype and will allow simpler fabrication
and installation.


End Effector Spike Penetration Testing
3D Model View Description
Spike Test Pic.PNG

To select the appropriate compression springs and spike tip geometry, we needed
to develop a model describing the force required for the spikes to penetrate the logs.
To accomplish this, we designed a compression test using the University's load frame.
We obtained 6 inch long samples of a pre-cut cant and designed and fabricated a
fixture and spikes to secure the samples into the load frame drive the spikes into
them.


End Effector Spike Geometry
60 Degree Spike 90 Degree Spike 120 Degree Spike Description
60 spike1.PNG

90 spike1.PNG

120 spike1.PNG

We planned to drive each spike a maximum of 0.5 inches
into five log samples. This would provide us with an average
maximum force required, for each spike geometry. Then, we would
decide the optimal tip angle and select a compression spring
according to the maximum force found.

After fabricating the spikes and test fixture, we no longer had access to the university's load frame machine due to the outbreak of COVID-19. Since we were not able to obtain these test results, we needed a new method to select spike tip geometry and compression springs. Idaho Forest Group had pre-manufactured spikes which we decided to incorporate in our design. We found research online regarding the force required to extract a nail from wood, and we used this model for spring selection. This led us to find a list of 6 springs which could fit within the housing member, withstand the maximum expected force, and deflect approximately 0.25 - 0.5 inches. Any of these springs will be suitable for our final design, which can bee seen below.

End Effector Design
3D Model Views Description
Final Render.JPG

After receiving the model for Idaho Forest Group's spikes, we
modified and finalized our design seen to the left. This is a
rendering of our final design. The circular plate mounts to the
robot's sixth axis via 10 bolts and 2 dowel pins. The design
consists of 2 spikes to penetrate each log end for increased grip.
The housing member is 6 inches long, 3.25 inches wide, and 1.4
inches tall, allowing it to fit within a 2 x 4 inch board and
protect the robots from being damaged by the saw.

Top View Annotated1.png

The image to the left illustrates all the key features of the
end effector design. flat head screws mount the housing member
to the circular plate. The spikes mount to outer shafts which
connect through the rectangular plate to inner shafts. The
rectangular plate bolts to the housing member, and the outer
and inner shafts slide along oil-embedded sleeve bearings
to reduce friction and prevent corrosion from steel-on-steel
sliding. The inner shafts press into compression springs which
absorb the impact of penetrating the spikes into the log ends.
The deflection of the compression springs is equal to the
displacement labeled in the image. The displacement is controlled
by physical interference of the outer shafts and the rectangular
plate. This controlled deflection ensures that we always know the
exact location of the logs relative to the robots. This design
minimizes vertical deflection under the weight of the log.


3D Model View Description
Spike Test Pic.PNG

To select the appropriate compression springs and spike tip geometry, we needed
to develop a model describing the force required for the spikes to penetrate the logs.
To accomplish this, we designed a compression test using the University's load frame.
We obtained 6 inch long samples of a pre-cut cant and designed and fabricated a
fixture and spikes to secure the samples into the load frame drive the spikes into
them.


End Effector Spike Geometry
60 Degree Spike 90 Degree Spike 120 Degree Spike Description
60 spike1.PNG

90 spike1.PNG

120 spike1.PNG

We planned to drive each spike a maximum of 0.5 inches
into five log samples. This would provide us with an average
maximum force required, for each spike geometry. Then, we would
decide the optimal tip angle and select a compression spring
according to the maximum force found.

After fabricating the spikes and test fixture, we no longer had access to the university's load frame machine due to the outbreak of COVID-19. Since we were not able to obtain these test results, we needed a new method to select spike tip geometry and compression springs. Idaho Forest Group had pre-manufactured spikes which we decided to incorporate in our design. We found research online regarding the force required to extract a nail from wood, and we used this model for spring selection. This led us to find a list of 6 springs which could fit within the housing member, withstand the maximum expected force, and deflect approximately 0.25 - 0.5 inches. Any of these springs will be suitable for our final design, which can bee seen below.

End Effector - Robot Interfacing

As seen in the images below, the end effector bolts to the end of each robot. Each end effector grips one end of the log as the robots maneuver the log through the cutting paths.

Robot Mounting System View
Robot End Effector Zoom Render.png

Robots W Log Render1.jpeg


PLC Programming[edit | edit source]

The PLC's role is to communicate data from an optimizer to the robots. Prior to reaching the robots, each log is scanned by an optimizer which generates unique cutting paths to maximize the economic value of each log. For each log, we need the following parameters: length, angular orientation, and height at each end. The PLC will receive these variables as each log is scanned, convert them to coincide with the robot coordinate system, and output them to the robots.

Optimizer Data Description
Optimizer Data.PNG

This image shows an example of what an optimizer
scan looks like. It provides an optimal log breakdown
geometry, communicates the data with the PLC. The PLC
then sends the data to the robots to cut the log.


Robot Simulation[edit | edit source]

In order to perform move each log through unique cutting planes for each log, the robots need to be able to receive the PLC data and automatically account for log geometry. This requires a baseline program which cuts one face of a perfectly cylindrical, infinitesimal log. Embedded in the program are offsets which become filled by optimizer data for each log via the PLC. With these offsets, the robots can perform the following motions to cut a log into a cant.

Simulation Setup Description
Robot Setup1.png

The log will be initially positioned on the horse bench.
Each robot (with mounted end effectors) will press the spikes
into each end of the log. Then, the robot will pick up the log,
rotate it to the proper orientation, and move it to the band saw
at the initial cutting position. The robots will then move the
log through the band saw in a straight path from the initial to
the final cutting position. The robots can then rotate the log
90 degrees and repeat the process for each face of the log. Once
this is complete, the log will be a cant.


Human Machine Interface[edit | edit source]

The purpose of the HMI is for an operator to interact with the robotic system to ensure the operation is running smoothly and act accordingly if it is not. The operator will also be physically watching the system to detect any potential safety concerns or operational errors.

Human Machine Interface Description
HMI1.png

The operator can view the log geometry which includes
log length, height on each end, and clockwise rotation.
If needed, the operator can edit these values to change
the PLC values communicated with the robots. The operator
also has control over starting, resetting and emergency
stopping the robots motion.


Design Validation[edit | edit source]

All design validation testing is to be completed at the automation facility in Athol, ID. Due to COVID-19 and time constraints, our team was not able to travel to the facility. Our clients with Idaho Forest Group will need to perform all onsite testing as they continue into the next phases of the project. The two areas needing validation are the mechanical design of the end effector and the system software integration.

To implement the end effector into the robotic system, the fist step is to calibrate the tool center point. This will be done by mounting the end effectors to each robot, screwing in the calibration rod, and moving each robot so the rod is touching the saw blade where the logs will be cut. Then, these robot positions will act as the geometric origin for the robots. Once the end effector is calibrated, the shock absorption mechanism needs to be verified. This will be done by programming the robots to drive the spikes into each end of a log and lift up. If the robots do not detect a collision and shut down in the process, then the shock absorption mechanism is working as designed. lastly, the end effector should be able to maintain the log securely as the robots perform basic linear motions. To verify the end effector grip, Our clients should program these basic motions, run the program, and (when the robots are stationary) grab the log and attempt to move it. If all goes well, the log should not move significantly as a human applies loading.

Once the end effectors are mounted, and their design is verified for operation, the system software integration can be tested. The operator (one of us) will input the appropriate log dimensions into the HMI to communicate with the PLC. Then, the PLC should send the data to the robot Teach Pendant, and the robots should maneuver the log to cut a single face with the saw blade.

Team Members[edit | edit source]

Sam.jpg

Name: Sam Malinowski
Major: Mechanical Engineering
Hometown: Boise, ID
Email: mali6321@vandals.uidaho.edu


Jin.jpg

Name: Jongin Hwang
Major: Mechanical Engineering
Hometown: Seoul, South Korea
Email: hwan8622@vandals.uidaho.edu


Deryk.jpg

Name: Deryk Ahner
Major: Mechanical Engineering
Hometown: Vancouver, WA
Email: ahne5837@vandals.uidaho.edu

An.jpg

Name: An Le
Major: Mechanical Engineering
Hometown: Lewiston, ID
Email: Le9187@vandals.uidaho.edu


Additional Documentation[edit | edit source]