Frequency-adaptive Backup-Roll Eccentricity Compensation with Force Feedback

In this paper we present a novel nonlinear adaptive control algorithm aiming at improving the thickness performances of metal strip rolling mills by actively compensating the periodic disturbances due to the eccentricity of the rolls. The main feature of the proposed Roll Eccentricity Compensation method consists in the capability of estimating the frequency of the disturbance without the need to install proximity sensors for synchronization. The proposed compensation scheme relies on a generalized nonlinear phase-locked loop, whose parameters are adapted online to achieve asymptotic cancellation of the unknown perturbation. The local asymptotic stability of the closed-loop system in presence of model uncertainty is proven by a two-time scale averaging analysis.