New Minmax Linear Quadratic Fault-Tolerant Tracking Control for Batch Processes

Ridong Zhang,Renquan Lu,Anke Xue,Furong Gao
DOI: https://doi.org/10.1109/tac.2015.2504181
IF: 6.549
2016-01-01
IEEE Transactions on Automatic Control
Abstract:A new design of linear quadratic control (LQ) is proposed for batch processes with partial actuator faults and unknown disturbance. The autoregressive model built through input-output process data is transformed into a new non-minimal state space (NMSS) model in which the process state variables and tracking error are combined, then a new LQ control scheme is derived based on minimax optimization. Unlike LQ based on traditional NMSS model (TNMSSLQ), the process state variables and tracking error can both be regulated in the proposed method. Monotonic situations of optimal performance cost are further derived to yield sufficient conditions for both nominal and robust system stability. Finally, an injection velocity control example is illustrated to show the effectiveness.
What problem does this paper attempt to address?