Optimal Reagents Control for Flotation Processes: An Adaptive Dynamic Programming Approach
Zhongmei Li, Mengzhe Huang, Weihua Gui, Yangchun Hua, Jianyong Zhu · 2019
The reagents control with reliability and robustness guarantees plays a substantial role in improving the system stability and reducing reagents consumption during the flotation processes. Due to the inherent complexity of flotation process, it poses a great challenge for optimal controller design. In this paper, a data-driven adaptive optimal feedback control approach using adaptive dynamic programming (ADP) is presented without knowing the flotation process dynamics. In particular, a learning-based control scheme is obtained using policy iteration (PI) method. It is a key strategy to formulate the flotation reagents control as a two-stage optimization problem. By solving this problem on-line, the concentrate grade and tailing grade satisfy the desired flotation performance with a minimum reagents consumption. The numerical simulation shows that the ADP controller can not only realize asymptotic tracking but also suppress the disturbance by employing the online production data.