Avalanches on a Conical Bead Pile

April 10, 2021   /  

Name: Henry Whyte
Major: Physics
Minor: Mathematics
Advisors: Dr. Susan Lehman, Dr. Niklas Manz

Avalanches are one of the most destructive forces in nature, a dangerous natural phenomenon that pose a risk to snowboarders, skiers and winter outdoor enthusiasts. Under the right environmental conditions, an avalanche can tear trees from its roots, flatten buildings, and consume anything before it, leaving a trail of destruction. At the College of Wooster with the help of a conical pile of steel-shot beads we can now investigate the conditions that may contribute to avalanches. The pile is driven through the autonomous dropping of beads onto the apex one at a time. Environmental conditions that facilitate an avalanche may change with time. Similarly, conditions in the lab change, which can effect the bead pile and subsequently experimental outcomes. The previous steel-shot beads used for this experiment were replaced this year with new beads of the same kind. To ensure the use of these new beads were not drastically changing the results of the experiment, an investigation of the mass data was completed. This investigation included comparing sets of data from the old beads and the new ones. Two runs were completed at different iii iv cohesion levels, where cohesion between the beads on the pile was created using the magnetic field due to a set of Helmholtz coils. From the probability distribution of avalanche sizes, I suggest that the amount of oil used to cover the older beads is greater than the amount on the newer set of beads. We also examined the use of video analysis and the software program PIVlab, which is used to analyze our avalanche videos. This program determines the velocity of small areas of the pile, and the magnitude of this velocity data was compared to that of a program called Tracker. Tracker is another video analysis tool which is well suited to track single particles, whereas PIVlab is better suited for analyzing the motion of whole systems. We found that the average velocity output from PIVlab was of the same order of magnitude as the calculated velocity from Tracker, which provides reassurance of our past and continued use of PIVlab.

Click here to view Henry’s presentation. (NOTE: A Wooster login is required to view this presentation.)

Henry will be online to field comments on April 16:
noon-2pm EDT (PST: 9-11am, Africa/Europe: late afternoon)

9 thoughts on “Avalanches on a Conical Bead Pile”

  1. Henry, congratulations! It’s fascinating to see how MATLAB and a GUI can help to analyze avalanches. Was there anything that struck you as particularly challenging in your work?

    1. I would say the most challenging and time consuming part of my work was file management and consistency of parameters between the GUI and MATLAB program. While analyzing the avalanches recorded, parameters sometimes need adjusting. Those adjustments need to happen across both the GUI and MATLAB programs to keep the velocity results consistent for every avalanche analyzed. When it comes to file management, a single run where we record avalanches takes upwards of 7 days with hundreds of avalanches occurring during that time. Each one of those avalanche videos must be converted to roughly 6000 images and then organized into folders. Once we have those folders organized, only then can we transfer the images over to a different computer for analysis in the GUI and in MATLAB.

  2. Thanks so much, Henry for sharing. Congrats! I have always had a question about the bead pile experiment. I know it is treated as a stochastic process, but with more precise machinery and measurements, does it become completely deterministic?

    1. Thank you Carlos, and I really like your question. Even if we were to have better machinery, one of the cool things in my opinion about the bead pile is its unpredictability. While we are always analyzing and trying to find patterns which can help us better understand what factors cause different types of avalanches, there is no way to actually predict exactly when an avalanche will occur.

    2. While adding beads to the pile at a constant rate, the avalanche is in a critical state. In the case of the bead pile, it has the properties of both a solid and liquid while in its critical state. While in this state, it can avalanche at any time without outside forces or tuning parameters. So while we cannot predict when an avalanche will occur, we can over time come to understand what causes various kinds of avalanches.

      1. Good answer, Henry, and nice job overall!

        I would just add to Carlos, that just because we can say WHY an avalanche happened AFTERWARDS, but that’s different from ever being able to say *in advance* that an avalanche is going to happen on the next bead drop.
        So I would say that it is deterministic in the sense that Newton’s laws are governing it, and in retrospect we can say — oh yes, this bead experienced a force at this angle and so it moved here, so there was an avalanche. But knowing in advance that this particular bead was *about* to experience that particular force and so therefore there *would* be an avalanche is totally different.

        1. Thank you Dr. Lehman! I think a good example of how we can translate this to a real world example is with an avalanche warning. When avalanches occur cannot be predicted, but we do understand conditions that can cause them, which would allow for caution to be taken at a Ski Resort for example.

  3. Henry, very nice results! It is interesting how much cohesion affects this complex system.

    1. Thank you Dr. Leary, and yes, as we add cohesion and the beads have a stronger connection to one another, it is more likely that a given avalanche will be larger in size.

Comments are closed.