Berlin 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
DY: Fachverband Dynamik und Statistische Physik
DY 19: Machine Learning in Dynamics and Statistical Physics II (joint session DY/SOE)
DY 19.4: Talk
Tuesday, March 19, 2024, 10:15–10:30, BH-N 243
Statistical criteria for the prediction of dynamical clustering in granular gases — •Sai Preetham Sata1,2, Dmitry Puzyrev2, and Ralf Stannarius1,2 — 1Institute of Physics, Otto-von-Guericke University, Magdeburg, Germany — 2Department of Microgravity and Translational Regenerative Medicine and MARS, Otto von Guericke University, Magdeburg, Germany
Granular gases excited by external forces can undergo transitions from the homogeneous to a dynamical cluster state [1, 2], depending on filling fraction, excitation parameters and container geometry. We compare two statistical criteria for the clustering transition, viz. the Kolmogorov-Smirnov Test (KS-Test) on the particle number density profile and the so-called caging-effect based on the local packing fraction [2]. Both criteria are evaluated for various combinations of system parameters in the VIP-Gran experiment [3] and combined into one dataset. This allows us to compare existing clustering criteria and tune them to provide matching clustering thresholds. The aim is to develop improved threshold criteria. Machine learning models are trained with this dataset to predict whether particular parameters lead to homogeneous or dynamical cluster states.
This study is supported by DLR projects VICKI and EVA II(50WM2252 and 50WK2348)
References: [1] É. Falcon et al., Phys. Rev. Lett., 83:440-443, 1999 [2] E. Opsomer et al., Europhys. Lett., 99:40001, 2012 [3] S. Aumaître et al., Rev. Sci. Instr., 89, 2018.
Keywords: Machine Learning; Kolmogorov-Smirnov Test; Caging Effect; VIP-Gran experiment; Granular gas