ME107B_1 Car Roll: An Experiment in Buidling Numerical Models

Scott Moura, Jennifer Hisoire, Daniel Hon, Kenny Langston, Tuthien Le, Matthew Rosenbrock, Timothy So
February 15, 2006

Mechanical Engineering Laboratory (ME 107B)
Kip Morris (Graduate Student Instructor)
UC Berkeley
A predictive simulation is developed for the car/track system in Hesse Hall by performing a minimal number of experiments. The system characteristics are crucial to the modeling process. Most notably, the track contains a center guiding rail to constrain the car onto the track. The effect of this center guide rail is significant to the dynamics of the car itself. Force balance and moment balance equations serve as the basis for our model. Several assumptions concerning friction and drag forces are made to simplify the system. The validity of these assumptions is confirmed through experimentation. In total, five experiments are performed to validate those assumptions and evaluate specific unknowns inherent to the system. A genetic algorithm is used to optimize the unknown constants against the experimental data. Our final simulation results are reasonably accurate for plastic center sections, but not thin aluminum. The carís motion is especially hard to predict near the center hump.