MIMO Tension Modelling and Control for Roll-to-roll Converting Machines Chul-Goo Kang* and Bong-Ju Lee** *Dept of Mechanical Engineering, Konkuk University, Gwangjin-gu, Seoul 143-701, Korea (Tel:+82-2-447-2142; e-mail: cgkang@konkuk.ac.kr). **Dept of Mechanical Engineering, Konkuk University, Gwangjin-gu, Seoul 143-701, Korea (e-mail: kk103@konkuk.ac.kr). Abstract: Roll-to-roll converting machines have significant interactions in tensions and speeds among web spans. A reasonable MIMO model for web tension plays an important role for high-performance converting machines. In this paper, we derive a nonlinear MIMO web-tension model of a high-speed gravure printing machine considering span length to be time-varying instead of considering it to be fixed. Then a feedback control system is constructed and web-tension control performance is analyzed at transient and steady-state condition via simulation studies using Simulink software. 1. INTRODUCTION Roll-to-roll converting machines have significant interactions in tensions and speeds among web spans. One main issue in research of roll-to-roll converting machines is to increase web transport velocity as much as possible while controlling tension of the web appropriately in order to obtain high productivity. A reasonable MIMO model for web tension plays an important role for high-performance converting machines. One important roll-to-roll converting machine is a gravure printing machine used for mass printing products. Control of gravure printing machines is basically composed of tension control and register control. Register control is required for high printing resolution in multi-stage printing systems, in which pre-printed marks and a scanning head are generally used for detecting register errors. However, precise register control is impossible without reasonable web tension control of continuous strip. That is, web tension control is a prerequisite for accurate register control to obtain high printing resolution. If web tension in roll-to-roll printing machine is too high, several problems occur such as large register error, rewinder wrinkling, web tearing, and plastic deformation of the web. On the other hand, if web tension is too small, then problems occur such as web oscillations, loose rewinding and web surface damages. Usually highly interactive web tension is modeled under the assumption that the span length is fixed (Shin, 2000; Dwivedula et al., 2003; Mathur, 1998; Weiss, 1985), but at unwinder and rewinder units using turrets, the span length changes during operation, especially during roll change and splicing process. Moreover, web length in one span between rollers varies due to dancer motions. In this paper, we derive a nonlinear MIMO (multi-input multi-output) web-tension model of a high-speed gravure printing machine considering span length to be time-varying instead of considering it to be time-invariant. Time-varying property of the web span length is due to turret motions of unwinding and rewinding units as well as dancer motions. Furthermore, we analyze web-tension control performance at transient and steady-state operations via computer simulation studies using Simulink. The validity of the web-tension model and control system has been shown via simulation using the model of an actual gravure printing machine. 2. GRAVURE PRINTING MACHINE Fig. 1 shows signal flows of the pilot plant of the gravure printing machine with three color printing function installed at Flexible Display Roll-to-Roll Research Center (FDRC), Konkuk University (manufactured by SAM, Inc.), which is presently used for several research purposes. This system is composed of an unwinder unit including turret and splicing mechanism and a passive dancer, an infeeder unit with a passive dancer, three printing units with color register control devices, an outfeeder unit with a passive dancer, and a rewinder unit including an active dancer. Dancers may be bypassed optionally. Eight loadcells are installed at idle rollers in the middle of continuous process for tension pickup. Web tensions can also be estimated using observers instead of using physical sensors (Lynch et al., 2004). When a command is given at HMI in the figure, PLC generates appropriate command signals for motion control of each motor, and then the controller of each motor controls tensions and speeds of the web using motion commands and feedback signals for motion and web tension. Fig. 2 shows a picture of the pilot plant. In this pilot plant, web speed, angular velocity of driving rolls, web tension, angular speed of the turret, radius of the wound roll, arm angles of the dancers can be measured and monitored through control panels and also PCs. Proceedings of the 17th World Congress The International Federation of Automatic Control Seoul, Korea, July 6-11, 2008 978-1-1234-7890-2/08/$20.00 © 2008 IFAC 11877 10.3182/20080706-5-KR-1001.4087