Data-driven discovery of the governing equation of granular flow in the homogeneous cooling state using sparse regression

With the arrival of the era of big data and the rapid development of high-precision discrete simulations, a wealth of high-quality data is readily available, but discovering physical laws from these data remains a great challenge. In this study, an attempt is made to discover the governing equation of the granular flow for the homogeneous cooling state from discrete element method (DEM) data through sparse regression. It is shown that not only the governing equation but also the energy dissipation rate can be obtained accurately from DEM data for systems having different physical properties of particles and operating conditions. The present work provides the evidence that the macroscopic governing equation and the constitutive relation of granular flow can be discovered from microscopic data using a purely data-driven method.