Abstract
Flow interferences occur in the dual-impeller stirred tank between paddles as well as between paddles and baffles and the tank wall, leading to inefficient utilization of the stirring energy. To address this issue, this study investigates the flow characteristics within the mixing tank using Euler-Euler numerical simulation and the particle image velocimetry (PIV) experimental method. The three-dimensional nonconstant flow characteristics are analyzed to optimize the critical stirrer geometry. By employing the Sobol method, an approximate model is established for sensitivity analysis to identify key parameters affecting the solid-liquid dual-impeller stirred tank's performance. Numerical simulations demonstrate that the optimized stirred tank exhibits a significantly improved solid-liquid suspension capacity and considerably reduces flow losses near the wall and baffle areas. Under the designated conditions, the cloud height is increased by 8.7%, and power consumption is reduced by 15.6% compared to the prototype. PIV tests performed on the stirred tank before and after optimization confirmed the reliability of the obtained optimization results. The primary objective of this study is to enhance mixing efficiency and homogeneity in solid-liquid mixing tanks while concurrently minimizing energy consumption and cost. These results validate the feasibility of employing a multiobjective optimal design approach that combines the RBF agent model with the Sobol method. The findings offer valuable insights for the design of similar mixing tanks.