EPJ DS - Data Science

EPJ Data Science - Countering crowd control collapse

EPJ Data Science - Countering crowd control collapse
© Angel Herrero de Frutos, iStockphotos, 138179229

Understanding crowd dynamics can prevent disaster at cultural or sports events.

Physicists investigating a recent crowd disaster in Germany found that one of the key causes was that at some point the crowd dynamics turned turbulent, akin to behaviour found in unstable fluid flows. The study, led by Dirk Helbing from the Risk Center at the Swiss Federal Institute of Technology ETH Zurich, Switzerland, is published in EPJ Data Science.

Never before have crowd disasters been studied by relying on a qualitative analysis of large public data sets. These include media and public authority reports, YouTube videos, Google Earth maps, 360? photographs, and other Internet sources. Based on this approach, the authors shed some new light on the crowd disaster that occurred at the Love Parade in Duisburg, Germany, in July 2010, leaving 21 dead and over 500 injuried.

The study focuses on the dynamics occurring when the density of people becomes very high. Physical interactions then inadvertently transfer forces from one body to another, similar to the pressure in dense granular materials. Under such conditions, Helbing and his colleague found that forces in the crowd add up and vary greatly. This makes it hard to avoid a domino effect when people fall. The forces are so high they can become life-threatening. They cannot be controlled by external police efforts. A collective dynamic called ‘crowd turbulence’ is created.
Contrary to previous thinking, crowd disasters are not always due to crowds becoming uncontrollable because individuals panic. Instead, the authors conclude that amplifying feedback and cascading effects lead to instability in the crowd. This results in a failure of crowd management and control attempts.

The authors also introduce a new scale to assess the criticality of conditions in the crowd designed to help implement preventative measures before the crowd reaches a critical state.

Editors-in-Chief
Frank Schweitzer and Alessandro Vespignani