Rise Up Robot
a beam climbing robot
Stanford University, ME104: Mechanical Systems Design, 5 week team project
Challenge
In this project, I collaborated with team of 5 to design and 3D model a robot that carries a payload up a 12ft. long angled I-beam as quickly as possible. Because of COVID-19, we completed this entire project remotely while managing time differences between the east and west coast.
My Role
Engineer, designer
Skills
CAD (SolidWorks and Fusion360)
Finite element analysis, back-of-the-envelope analysis
Mechanical design
Hardware analysis (motors, gears, batteries, power transmission, efficiency)
Teamwork
The Team!
Sidney Stevens, Rachel Wallstrom, Amanda Tu, Jenna Wang, Jabreea Johnson
Ideation
There were several challenges to this project. First, the robot had to stay on the I-beam. Second, it had to hold the payload that was located off the side of the I-beam. Third, it had to make it up the I-beam as quickly as possible. Given these constraints, we came up with many solutions for each problem in order to explore our options.
Preliminary Design
After brainstorming and doing free-body-diagrams of our strongest ideas, we decided on a design that best reduced the stress on the arms holding the payload. Given the angle of the beam and the placement of the payload, we angled the arms such that the forces lined up axially, allowing them to experience minimal bending. This way, the arms could be thinner and lighter to reduce the total mass and increase the speed that the robot could climb. We hoped to connect them with a tension cable to reduce the mass even further.
To keep the robot from falling off the beam, we decided on a lower and side wheel design that would use the side and bottom of the flange to keep the robot from tipping.
While the preliminary design was promising, there were some issues that we would discover later that caused us to alter it slightly.
Optimizing Power
This was definitely the hardest part of the project for our team. Even though none of us were very strong in electrical engineering or power optimization, together we pooled our knowledge and by the end we were all motor and gear experts.
Because our robot’s design wasn’t solidified yet, we had to work with estimates of robot mass and wheel placements. At first, we incorrectly assumed that our robot would be in steady-state, which gave us wildly incorrect calculations.
After realizing that a speedy robot would not reach steady-state, we altered our power flow to take this into consideration. Starting with our ideal shaft torque of the motor (determined by optimizing the power output of the robot), we could calculate the power loss and efficiency at each step of the transmission and iterate through different voltages to find a battery that would support both the amperage and not overheat the motor. To see our final calculations, see the final report here.
A snippet of the many calculations that we did to determine the maximum voltage we could use without overheating the motor
CAD Part 1: The First Design
Because we were working remotely, we decided to use Fusion360 in order to better collaborate in designing the robot. Although the design wasn’t completely finalized, we were able to create a mock-up of the general idea, giving us the ability to see our robot in 3D and catch any errors that we hadn’t seen before modeling it. See rationale behind the preliminary design here!
CAD Part 2: Coming Together
After creating the first model, we realized that there were some issues with the design. First, the two-arm design was great in theory, but if we were to actually make the robot, it had a high chance of swinging out of plane if the robot was tipped even a little bit. Second, we had no idea how we were going to attach the side wheels to the platform while restricting all degrees of freedom. And third, we needed to figure out where to place the wheels and where to attach each component to the platform to optimize speed and power.
With these things in mind, we went back to the drawing board and came up with our final design.
First, we added a third arm to the bracket assembly, creating two arms that would straddle the drive wheel to optimize the normal force on the wheel and reduce the chance of it swinging out of place. A cable looping around the compression arm kept it from sliding out.
Next, we created a new side wheel assembly that used two shoulder screws to keep the whole fixture from spinning out. We decided that both wheels would be ball bearings to reduce friction as well as keep the overall radius small.
Finally, we redesigned the platform. The positioning of each wheel was determined through back-of-the-hand analysis in order to optimize the normal force on the drive wheel, therefore increasing the overall power and speed of the robot. The shape was designed to be reminiscent of a potato, a playful way of calling back to our professor for the project who once described our robot as potato-esque.
We also ran a finite element analysis to ensure that it would be able to withstand the forces and not hit the beam due to deformation.
The Final Robot
We calculated that our final robot would ascend the 12ft. long beam in 4.82 seconds. See the full report here!
While we did find a small error in our calculations that caused the wheels to slip in the final test, the design was very close and if we were to build it in person, a few small tweaks would have solved the issue. Overall, this was a great lesson in mechanics, the importance of checking your work, and accepting that there are always improvements to be made!