Researchers from Nanjing University of Science and Technology and Westlake University have made significant advances in accurately modelling the movement of irregularly shaped particles through fluids, a complex challenge in fluid dynamics vital for industries including chemical engineering and aerospace. Their study, which integrates machine learning with advanced numerical methods, has been published in the International Journal of Mechanical System Dynamics.
The research addresses the longstanding difficulty of predicting the drag coefficient of non-spherical particles. Traditionally, precise prediction has been limited to spherical particles, leaving a gap in understanding how irregularly shaped particles behave in fluid environments. This complexity arises because single shape factors have proven insufficient in capturing the nuanced influences on particle movement.
By employing the discrete element method (DEM) combined with the lattice Boltzmann method (LBM), the research team built a high-accuracy dataset. This dataset served as the foundation for developing four machine learning models meant to predict the drag coefficient of polygonal particles. Of these, the genetic algorithm-artificial neural network (GA-ANN) model excelled, achieving a prediction error of less than 5%. Such precision marks a pivotal development in fluid dynamics research.
Professor Cheng Cheng, one of the study's lead researchers, underscored the significance of these findings, commenting on the transformative potential of machine learning in solving intricate fluid dynamics challenges. "By leveraging numerical simulations and artificial intelligence (AI)," he stated, "we've reached an unprecedented level of accuracy in predicting the drag coefficients of polygonal particles. This could have far-reaching impacts in both academic and industrial settings."
The implications of this research are expansive, offering benefits to sectors like chemical processing, environmental engineering, and aerospace technology. The enhanced accuracy in predicting drag coefficients can lead to improved processes such as sedimentation, filtration, and propulsion, ultimately increasing the efficiency of various systems. This breakthrough is poised to influence future designs and optimisations of fluid-particle interactions across numerous industrial applications.
Source: Noah Wire Services