To address the issues of low solid–liquid mixing and mass transfer efficiency and difficult real-time regulation in the resource utilization of non-ferrous metal smelting slag, this study constructs a research framework integrating Computational Fluid Dynamics-Discrete Element Method (CFD-DEM) coupling models and machine learning. The framework systematically investigates particle motion characteristics and mass transfer laws in stirred tanks and enables an intelligent prediction of key parameters. Through a CFD-DEM two-way coupling simulation, the study quantifies particle dispersion characteristics using relative standard deviation (RSD) and calculates the mass transfer coefficient (k) based on the Hughmark model, revealing the effects of particle size and impeller speed on mixing and mass transfer efficiency. For parameter prediction, particle motion and mass transfer data are used to train a multi-model prediction library, with model performance evaluated through comparative experiments. The results show that increasing the rotational speed shortens the particle mixing time, reduces RSD values by 25–40%, increases the coupling force, and decreases stability during the circulation phase. Different machine learning (ML) algorithms exhibit varying performances in the time-series prediction of particle motion characteristics and real-time prediction of mass transfer coefficients. Notably, GA-BP achieves a fitting degree R of 0.99 in both predictions, meeting the requirements for the structural optimization and intelligent regulation of stirred tanks. This research provides theoretical support and technical pathways for the structural optimization and intelligent control of stirred tanks, offering engineering application value in fields such as hydrometallurgy and solid waste resource utilization.

Coal mine gas with methane concentrations below 8% cannot sustain stable self-combustion, posing significant challenges for safe utilization and greenhouse gas mitigation. To address this limitation, we developed a large-scale industrial square rotary regenerative thermal oxidizer (RTO) capable of high-efficiency oxidation under ultra-low methane conditions. This work integrates multi-scale computational fluid dynamics (CFD) modeling, laboratory and pilot-scale physical experiments, and multi-physics coupled simulations to capture the complex interactions of fluid flow, species transport, and thermal response in regenerative ceramics. Compared with conventional circular or three-bed RTOs, the proposed square rotating design achieves 13% higher heat storage utilization, 15% smaller floor area, and enhanced spatial uniformity of the temperature field. Multi-scale simulations reveal that increasing methane molar fraction (CH₄) from 0.012 to 0.017 raises the peak temperature from 1280 K to 1350 K, reduces the burnout height from 1.18 m to 1.15 m, and, under constant oxygen supply, extends the high-temperature zone to 1450 K with a stabilized burnout position at 1.06 ± 0.01 m. Incorporating a 15° conical expansion combustion chamber increases local turbulent kinetic energy by 17.4%, accelerating oxidation while maintaining methane removal rates > 98% within an optimized bottom blowing time of 30–90 s. This study not only provides validated design thresholds for ultra-low concentration methane oxidation—such as temperature windows, buffer zones, and switching cycles—but also offers an engineering framework for scaling RTO systems to industrial coal mine applications. This advances both energy recovery efficiency and methane emission control, demonstrating clear advantages over existing RTO configurations.