Multipoint Feeding Strategy of Aluminum Reduction Cell Based on Distributed Subspace Predictive Control

With the continuous development of large-scale aluminum reduction cells, problem uniform distribution alumina concentration in cell has become more and serious for process. In order to achieve concentration, a data-driven distributed subspace predictive control feeding strategy is proposed this paper. Firstly, divided into multiple sub-systems that affect each other according position port. Based on method, prediction model whole identified, output expression sub-system deduced by decomposition. Secondly, controller designed subsystem, input information can be exchanged between through network. Thirdly, under consideration influence subsystems, subsystem solves Nash-optimal quantity, so realizes feeding. Finally, simulation results show that, compared with traditional significantly improve current efficiency cell.