Adaptive Control with Self-Tuning for Center-Driven Web Winders
Authors
Jensen, Peter ; Larsen, Jeppe
Term
10. term
Publication year
2007
Pages
90
Abstract
Denne afhandling undersøger adaptiv styring af et centerdrevet web‑oprullingssystem med feedback fra en dancer‑mekanisme, der hjælper med at holde banespændingen stabil. Vi udviklede en matematisk model og implementerede den i Simulink. Flere adaptive styringsmetoder blev vurderet, og vi anvendte en modificeret udgave af Model Identification Adaptive Control (MIAC), som justerer sin styring i takt med opdateringer af modellen og de estimerede systemparametre. Selve controlleren er udformet som optimal styring med en integral omkostningsfunktion baseret på den udviklede model og de estimerede anlægsparametre. Den adaptive controller blev testet på en skalamodel leveret af Danfoss og gav acceptabel ydeevne, når vægtningsparameteren for styreindsats var tilstrækkelig høj. Vi udviklede og testede også to metoder til at bestemme den aktuelle rulle‑radius og én metode til at bestemme systemets samlede træghed. Begge radiusestimeringer fungerede tilfredsstillende, men inerti‑estimatoren var kun pålidelig, når der var nok forstyrrelser til at aktivere systemets dynamik.
This thesis examines adaptive control of a center‑driven web winder with feedback from a dancer mechanism that helps keep tension stable. We built a mathematical model and implemented it in Simulink. Several adaptive control methods were evaluated, and we used a modified Model Identification Adaptive Control (MIAC) approach that adjusts its behavior as the model and estimated system parameters are updated. The controller design applies optimal control with an integral cost function, based on the developed model and estimated plant parameters. The adaptive controller was tested on a Danfoss scale model and delivered acceptable performance when the control‑effort weighting parameter was sufficiently high. We also developed and tested two methods to determine the current winder roll radius and one method to determine total system inertia. Both radius estimators performed well, but the inertia estimator was reliable only when the system experienced enough disturbances to excite its dynamics.
[This abstract was generated with the help of AI]
Keywords