This study introduces an innovative model for simulating fluid flow in multi-impeller tanks using Computational Fluid Dynamics (CFD) and Artificial Neural Networks (ANN). The effects of blade number, impeller diameter, and rotational speed on material distribution within the tank are investigated. The results show that an optimized design, featuring multiple blades and appropriately selected impeller dimensions and rotational speeds, significantly improves mixing efficiency and uniformity. The optimal configuration (three impellers, D/2 blade diameter, 80 rpm) reduced mixing time by an average of 21.6% compared to single-impeller systems, achieving near-ideal conditions. The study underscores the importance of strategically selecting impeller type, blade number, diameter, and rotational speed to improve mixing quality in industrial applications. The ANN is also evaluated as a surrogate model to reduce reliance on repeated CFD simulations during parameter screening. The ANN model shows strong predictive performance with a mean relative error of 2.45%, confirming its accuracy for this application. While CFD provides high-fidelity results, the ANN offers a lightweight alternative suitable for fast evaluations in design processes or software integration. This work provides valuable insights for optimizing mixing processes with significant implications for various industrial sectors. While the single-impeller CFD simulations were validated against experimental data from the literature, the multi-impeller configurations have not yet been directly validated. Future work should include targeted experiments to confirm the observed trends and enhance generalizability.
