Title

COMPENSATION OF IRREGULAR BACK-UP ROLL ECCENTRICITIES IN 4-HIGH AND 6-HIGH APPLICATIONS

COMPENSATION OF IRREGULAR BACK-UP ROLL ECCENTRICITIES IN 4-HIGH AND 6-HIGH APPLICATIONS

Authorship

Zipf, Mark E.; Elwell, Thomas G.; Godwin, Craig K.

DOI

10.5151/2594-5297-0096

Downloads

heading.download_article_detail 54 Downloads

Abstract

The compensation of back-up roll eccentricity is an established, mature and well understood concept that has seen many implementations with varying degrees of performance and success. A typical assumption in the compensator’s internal model’s structure and dynamics is that the offending and disruptive eccentricity signals are based on combinations of periodic roll ovality and rotation axis offset. In real-world cases, the disruptive eccentricities may have far more complex behavior stemming from improper roll grinding practices and / or poor metallurgical quality rolls having irregular / inconsistent patterned variations in surface hardness. These additional components tend to exceed the capabilities of contemporary eccentricity methods, leading to unacceptable exit gauge control and potentially unstable compensation reactions. This paper will examine and illustrate efforts to improve existing eccentricity compensation practices by considering disturbance estimation and feedforward cancellation coupled with adaptive parameter estimation and self- tuning regulation. Periodic classical and higher-order eccentric disturbances are estimated as a function of roll rotation from spatial decomposition of exit gauge temporal signals. Correlating these decomposed components and relating them to the tracked periodic rotation of the rolls within the stack, allows these un-modeled (pseudo-random periodic) components to be directly mapped to the surfaces of the individual rolls. These mapping functions can be applied to directly compensate for the overall roll stack irregularities by feedforward disturbance cancellation. Controller adaptation is provided by explicit descriptions within the internal models and through parameter identification coupled with self-tuning regulation to accommodate mill / material related variants.

 

The compensation of back-up roll eccentricity is an established, mature and well understood concept that has seen many implementations with varying degrees of performance and success. A typical assumption in the compensator’s internal model’s structure and dynamics is that the offending and disruptive eccentricity signals are based on combinations of periodic roll ovality and rotation axis offset. In real-world cases, the disruptive eccentricities may have far more complex behavior stemming from improper roll grinding practices and / or poor metallurgical quality rolls having irregular / inconsistent patterned variations in surface hardness. These additional components tend to exceed the capabilities of contemporary eccentricity methods, leading to unacceptable exit gauge control and potentially unstable compensation reactions. This paper will examine and illustrate efforts to improve existing eccentricity compensation practices by considering disturbance estimation and feedforward cancellation coupled with adaptive parameter estimation and self- tuning regulation. Periodic classical and higher-order eccentric disturbances are estimated as a function of roll rotation from spatial decomposition of exit gauge temporal signals. Correlating these decomposed components and relating them to the tracked periodic rotation of the rolls within the stack, allows these un-modeled (pseudo-random periodic) components to be directly mapped to the surfaces of the individual rolls. These mapping functions can be applied to directly compensate for the overall roll stack irregularities by feedforward disturbance cancellation. Controller adaptation is provided by explicit descriptions within the internal models and through parameter identification coupled with self-tuning regulation to accommodate mill / material related variants.

Keywords

Eccentricity disturbance estimation

Eccentricity disturbance estimation