Machine Learning Assisted Experimental Characterization of Bubble Dynamics in Gas–Solid Fluidized Beds

This study introduces a machine learning (ML)- \nassisted image segmentation method for automatic bubble \nidentification in gas−solid quasi-2D fluidized beds, offering \nenhanced accuracy in bubble recognition. Binary images are \nsegmented by the ML method, and an in-house Lagrangian \ntracking technique is developed to track bubble evolution. The MLassisted \nsegmentation method requires few training data, achieves \nan accuracy of 98.75%, and allows for filtering out common sources \nof uncertainty in hydrodynamics, such as varying illumination \nconditions and out-of-focus regions, thus providing an efficient tool to study bubbling in a standard, consistent, and repeatable \nmanner. In this work, the ML-assisted methodology is tested in a particularly challenging case: structured oscillating fluidized beds, \nwhere the spatial and time evolution of the bubble position, velocity, and shape are characteristics of the nucleation-propagationrupture \ncycle. The new method is validated across various operational conditions and particle sizes, demonstrating versatility and \neffectiveness. It shows the ability to capture challenging bubbling dynamics and subtle changes in velocity and size distributions \nobserved in beds of varying particle size. New characteristic features of oscillating beds are identified, including the effect of \nfrequency and particle size on the bubble morphology, aspect, and shape factors and their relationship with the stability of the flow, \nquantified through the rate of coalescence and splitting events. This type of combination of classic analysis with the application of \nthe ML assisted techniques provides a powerful tool to improve standardization and address the reproducibility of hydrodynamic \nstudies, with the potential to be extended from gas−solid fluidization to other multiphase flow systems.