I’ll be filling up the “My Projects” category with the projects I’ve worked on to create a portfolio, so I wanted to kick it off with my first big project back in college. At the University of Idaho, I participated in the Formula SAE program for three years. Though I worked on many of the car’s sub-systems, I always seemed to end up with the frame design and build.
My first year, as an undergrad junior, I was less involved with the design and more involved with the fabrication. Here, I learned how to TIG weld thin walled chromoly tubing and I picked up a ton of lessons on proper jigging, tube cutting, tacking, and finishing. Here’s the finished frame from 2003:
My second year, as an undergrad senior, I was in charge of the frame design and fabrication. Taking my lessons from the previous year, I incorporated a number of changes in the design and tried to focus on minimizing weight while maximizing rigidity. I also took the fabrication to a new level by producing drawings and cutting jigs for the individual tubes in order to speed up production. Here’s the frame design from 2004:
My last 1.5 years as a grad student gave me the opportunity to pursue my frame design. My major professor was the leader of the FSAE program, so he asked me to find a way to optimize the frame design through the use of software. Having minored in computer science, I decided to further my programming and structural analysis knowledge with additional courses such as evolutionary computation, artificial intelligence, discrete optimization, and matrix analysis.
Once I had to necessary tools in my bag, I wrote my own FEA software and a genetic algorithm optimization program specifically for the size and shape optimization of frame and truss structures. Why did I write my own FEA routine? Because it needed to be fast for the genetic algorithm to work in a feasible amount of time — running the static analysis tens or hundreds of times each second.
The final test for my new optimization software was to set it loose on the FSAE frame and see what it could do. With a base design given to the program, it gave us several possible variations with different amounts of weight reduction and stiffness. After comparing the options, we chose a single design and ran with it. And here’s the design from 2005, plus an animated frame build at the bottom:
The optimization software and my thesis can be downloaded at SourceForge.net under the ESOP project folder. “ESOP2″ is an updated version of the software done by another grad student to include shear panel capabilities.