Extension of Energetic Macroscopic Representation to Time-Varying Systems - Application to Winder Tension Control

Energetic macroscopic representation (EMR) was introduced to represent energetic systems in a unified way. The result is a structured representation of the system in a succession of layers, containing energy conversion blocks and their corresponding control function counterparts. An extension of EMR to time-varying systems is proposed, considering modeled and unmodeled parameter variation laws. Two layers are added to the EMR, the first to represent parameter variations and the latter to exploit or identify the parametric model. The method is applied to a speed and tension control problem in a paper winder. To obtain high quality paper rolls, high performance and decoupled velocity and tension control must be maintained in presence of parameter variations due to the process and to operating conditions. Predictable parameter variations are compensated by using the parametric model and recursive least squares (RLS) is used to identify unpredictable parameter variations to implement adaptive controllers. The adaptive control law is validated by simulation