Modeling and identification of the source of oscillations in web tension

Although there has been much work in dynamic modeling of different web handling elements and web longitudinal behavior, efforts to systematically validate models by experimentation on a web platform are non-existent. Existing literature has extensively used dynamic models for numerical analysis and/or design of control systems without adequate experimental validation of the models. One of the well-known modeling techniques for creating a model for the entire web line is based on the concept of primitive elements. In this approach every primitive element of the web line is modeled separately using first principles approach, and then the entire web line model is obtained by appropriately combining the primitive element models. The goal of this paper is to present results from recent investigations on validation of key primitive element models. Model refinement and modifications are also considered when sufficient level of agreement between model and experimental data was not obtained.Since the dynamic model for web tension in a span is nonlinear, designing experiments for web line model validation is a formidable task. There are a few known model validation techniques for nonlinear systems but these do not provide any clear procedures that can be applied to the web line. Therefore, the approach taken in this study was to consider test cases of experimentation that mimic typical web line operations in the industry such as acceleration/deceleration of the line and running the line at a constant speed. A number of test cases for model simulations and experimentation were considered. A representative sample of the results is shown and discussed.Data from the simulations of existing models did not contain the oscillations found in measured tension signals. This study also considered refinement and modifications of dynamic models that would lead to better agreement between the model and experimental data. It was found that span length variations introduced by out-of-round and/or eccentric rollers are the direct cause of oscillations in the tension signal. A refined model for web tension that includes span length variations is given. Comparison of the data from simulations of the refined model and experimental data shows a high level of agreement. These results are shown and discussed.Mechanical and Aerospace Engineerin