Predictive State Observer-Based Set-Point Learning Control for Batch Manufacturing Processes With Delay Response
Tao Liu, Shoulin Hao, Youqing Wang, Jing Na
Abstract
A predictive state observer (PSO)-based iterative learning control (ILC) scheme is proposed for industrial batch manufacturing processes with delay response suffering from nonrepetitive uncertainties and disturbances. Combining with a delay-free output predictor, a PSO-based feedback control structure is first presented to improve set-point tracking and disturbance rejection performance against nonrepetitive process uncertainties and disturbances for the initial batch run. Then, an ILC law is introduced to update the closed-loop system set-point in order to improve the tracking performance from batch to batch. A delay-independent sufficient condition is established to ensure the convergence of output tracking error along the batch-direction, based on double-dynamic analysis. Moreover, another delay-dependent sufficient condition is constructed by linear matrix inequality to assess robust stability of the proposed two-dimensional control system along both the time- and batch-directions. Finally, an illustrative example and a real application to the batch temperature regulation of a 4-litre crystallizer are shown to validate the effect and advantage of the proposed ILC scheme.