루퍼시스템 응답의 자동진단기술

Abstract

This paper proposes an algorithm to automatically diagnose the looper angular response characteristics of a looper system installed in multi-stand type rolling mills in real time. The looper angular response pattern is clearly distinguished three operating areas such as the front part, the controlled part, and the back part. There are two key techniques that compose diagnosis algorithm. First, physical indicators including information on the physical state for each part were designed. Second, the performance indexes, which means the current state of each part numerically, were suggested in the form of a judgment formula using physical indicators. According to the simulation results, the proposed diagnosis algorithm computed each performance index for given the looper angle response data and gave information on the presence or absence of abnormality and the abnormal part of the looper angle response.