Development of nanobubble-enhanced catalytic ozonation reactor for water purification based on computational fluid dynamics simulation and machine learning prediction

The Practical application of nanobubbles and catalysts in the ozonation of pollutants has been severely hindered by the difficulty of developing efficient reactors where efficiency is controlled by interactions among multiphase flow, mass transfer, and chemical reactions. Due to the complexity of these interactions, accurately describing them remains a significant challenge. In this study, a computational fluid dynamics (CFD) coupled machine learning (ML) model was developed to design and optimize an integrated reactor for nanobubble-enhanced ozone dissolution and catalytic oxidation. Simulation results indicate that the aeration device effectively increases the nanobubble fraction from ∼16% to 60%, thereby significantly accelerating and enhancing ozone mass transfer from the gas to the liquid phase. The integrated reactor of ozone nanobubbles with catalytic oxidation was utilized for the treatment of phenol solution. Both experimental and simulation results show that phenol and its ozone-reactive degradation intermediates (e.g., p-benzoquinone) can be rapidly abated at the first ozone nanobubbles oxidation section, while for the ozone-refractory species (e.g., formic acid and oxalic acid), they can be effectively oxidized at the second catalytic oxidation section. Finally, the optimal structural parameters and operating conditions for the integrated reactor were obtained by an ML model, which achieved >90% TOC removal and > 98% ozone utilization efficiency. Overall, the results of this study demonstrate that CFD simulation coupled with ML provides a feasible and convenient way for design, optimization, and reaction mechanism elucidation of the nanobubble-enhanced catalytic ozonation reactors, laying a foundation for efficient application of catalytic ozonation processes.