ON-LINE IDENTIFICATION OF THE TIME PREDITION MODEL IN HOT ROLLING

摘要

Mill pacing control is the key to success operation in the hot rolling mill. Accurate rolling time models are essential to improve manufacturing productivity and energy efficiency, but the models include error due to uncertain time-variant factors. This paper presents a novel identification scheme for the rolling time model. The identification method is based on recursive least-square method. Applied to actual plant data, the effectiveness of this method is certified. This scheme has been used for the mill pacing control system in one of the hot rolling mills at West Japan Works.