Swarm intelligence-driven performance optimization of multi-shaft stirrer reactors: leader shaft control strategy

The growing gap between production demand and capacity has placed increasing pressure on the efficiency and automation of chemical production processes. However, multi-shaft stirred reactors, inspired by the swarm intelligence of geese, have emerged as a promising solution. This work first addresses the critical issue of the “leader shaft” and determines its optimal position. Using high-precision simulations of the flow field at multiple scales, revealing that improving mixing performance depends on understanding the variation of flow dynamics and energy demands. Additionally, the torque attractor fractal dimension is identified as a key metric linking operational conditions to energy demands. Building on this, a multi-agent intelligent control system, focused on the leader shaft, is developed and validated through Belousov-Zhabotinsky reaction experiment. Results demonstrate a 46.7% increase in reaction rate compared to conventional optimal operating conditions. This control strategy significantly enhances reactor automation and has the potential to substantially improve industrial production capacity.