A Visual Investigation of Criticality: Avalanche Classification on a Conical Bead Pile

Kyle McNickle


The behaviors of an avalanching conical bead pile, driven by the additions of individual steel beads to its apex were analyzed. Two full data runs were collected, each containing about a week’s worth of data. The first run was conducted without cohesion and the second run with cohesion. Cohesion was induced in the beads by subjecting them to a magnetic field produced by two Helmholtz coils centered around the pile. The current passing through the coils was set to 750 mA. In each of the runs, a camera fixed to the apparatus was used to record all avalanches over 50 beads in size. The videos from each of these avalanches were analyzed using a particle image velocimetry (PIV) program to extract velocity data from small cells overlaid on each frame. These velocity data were loaded into a custom Igor procedure to generate a multitude of other data waves for analysis. These waves contained information relating to the amount of activity on the surface of the pile, the durations, the speeds, and the total distance moved on the pile, for each avalanche. It was observed that the average sizes of the avalanches, the total distance moved by the pile, and the average activity of all of the cells on the pile were greater for the cohesion run. The automated duration calculations were only accurate for the non-cohesion run. Large cohesion avalanches and small non-cohesion avalanches had the most uniform levels of activity across the surface of the pile.