Bottlenecks and traffic jams

When a granular media is discharged from a silo, spontaneous arching or vaulting over the orifice training causes intermittency in the flow that can lead to clogging. To restart the flow, an external disturbance is applied to the system, typically a vibration. In our team we investigate both the problem of training clogging and the problem of clog destruction by vibration. We know that the training of jams is a Poissonian process in which the probability that a stable arc is generated is constant in time. In contrast, arc breakup is a process with report in which the probability of breakup decreases as time progresses, giving rise to an anomalous statistic in the distribution of breakup times. This dual behavior has been used to define a jamming diagram based on the fact that the breakup time distribution may not converge if the hole size or vibration intensity is sufficiently small. We are currently investigating the physical origin of the anomalous statistics found for the breakup time distributions.

Another interesting question that remains to be resolved is whether, in a static silo, there is a critical orifice size above which clogging is not possible. Recent work in this field seems to suggest that such a transition to a fluid state does not exist, implying that even enormously large orifices will clog if one waits long enough. Despite this, there is no fully conclusive physical explanation of this issue so we are working in different directions to try to find one. One of them is to decouple the geometrical and dynamic contributions to the jamming. In addition, we have been working on the role that several variables play in the jamming process such as the width of the silo, or the surprising reduction in the probability of jamming found if an obstacle is placed on top of the outlet orifice.