Neural Network Based Adaptive Control of Web Transport Systems

EasyChair Preprint 6152

Abstract

A tension control problem of a roll-to-roll web handling system is considered in the paper. It is shown that the dynamical response of the web handing system heavily depends on roll inertia. Dissimilar to other researches that based on the assumption of rolls with perfect cylindrical form and the web material having homogenous thickness, the paper takes imperfect roll-shape and nonhomogeneous web material into account. The two factors directly affect roll inertias during operating process. The novel contribution of the paper is a presentation of a neural network to estimate inertia momentums of unwinding and rewinding rolls that vary according to web material movement. The neural network is designed based on RBF network, estimating error is proven to be asymptotically stable. The information on estimated inertia fed into a backstepping-sliding mode controller that ensure tension and velocity tracking of the system. The control design is presented in a systematical approach in the paper. The closed loop system stability is proven mathematically. The tracking performance is shown through several simulation scenarios.

Keyphrases: Web Transport System, backstepping sliding mode control, neural network, tension control

@booklet{EasyChair:6152,
  author    = {Manh Cuong Nguyen and Duc Dinh Nguyen and Tien Dung Pham and Van Manh Tran and Tung Lam Nguyen and Thi Ly Tong},
  title     = {Neural Network Based Adaptive Control of Web Transport Systems},
  howpublished = {EasyChair Preprint 6152},
  year      = {EasyChair, 2021}}