Adaptive terminal iterative learning for batch process with batch-varying parameters

Adaptive terminal iterative learning control algorithm based on process model updated along with batch index is proposed for final quality control, in which process parameters are batch-varying. Firstly, the final quality model in batch process is built based on the online least square support vector machine (online LSSVM), where the moving-window technique is introduced to update the process model along with the batch changing. Then, an adaptive terminal iterative learning control algorithm according to optimal cost function is given. Convergence and stability analysis are derived based on Lyapunov energy function. Finally, the control algorithm is applied to the typical batch polymerization process, and the simulation results show that the proposed approach has well control performance, monotonous convergence and adaptive ability.