Tension Control of Digital Web Winding System Based on Dynamic Bayesian Network Model

In the industry of fiber, paper, plastic and metal plating, one of the key techniques to guarantee the product quality is web tension control of linear, striping and planar elastic material. This paper introduces a dynamic Bayesian network training algorithm of multi-agent technology which can overcome the shortcomings of traditional tension control method, and effectively weaken the coupling effect of the velocity and tension in the tension control system. The simulation result shows the excellent self-learning and adaptive performance of this algorithm.