Data-driven strategies for register error mitigation in roll-to-roll manufacturing system

초록

Roll-to-roll gravure printing is widely used for the large-scale production of flexible electronics, such as displays, solar panels, and electronic paper. This process involves the continuous transport of web materials via rollers to achieve precise patterns. However, challenges (tension fluctuations) resulting from roll eccentricity or bearing defects can cause register errors and reduce printing accuracy. In this study, a predictive compensation strategy was developed to address the aforementioned issues utilizing tension data to predict eccentric roll conditions in gravure printing systems. The impact of the tension-data segmentation on the performance of the predictive model was systematically investigated. To optimize data preparation for eccentricity compensation, a novel metric termed the “segment score” was introduced. The segmentation approach resulted in a 74.0% improvement in the eccentricity prediction accuracy. The compensation outcomes from the optimally and suboptimally segmented data were compared and experimentally validated. The results showed that the proposed methodology improved tension control by 42.5% and register accuracy by 32.4%. These findings suggest that the proposed compensation strategy effectively mitigates the impact of eccentric roll conditions, thereby providing a reliable solution for minimizing register errors in practical roll-to-roll gravure printing applications.

키워드