Fault diagnosis method for disc slitting machine based on wavelet packet transform and support vector machine

Abstract

ABSTRACT Disc slitting is the most economical and efficient method for shearing wide-width sheet metal into strips of appropriate width. In the high-speed continuous operation, the disc slitting machine is prone to various faults such as circular runout of disc cutter, idling of machine tool, and machine stop. In this study, a measurement system was designed for real-time online detection of three-direction shear forces during the whole process of sheet metal disc slitting, and an experimental platform was constructedfor analyzing the characteristics of the three-direction shearing forces. During the experiments, the detection process was stable, and the shearing force data were reliable. To meet the challenges of detecting shearing force signals with non-distinct features and identifying the fault state online from a remote location, developed a fault diagnosis method based on wavelet packet transform (WPT) and support vector machine (SVM), and verified its feasibility through an experiment. The results show that the proposed method can effectively identify circular runout of disc cutter, idling of machine tool, and machine stop of the disc cutter. Apart from being used for detecting the faults of the disc slitting machine, this method can also be applied in the fault diagnosis of other rotating machines, having good prospects in industrial applications.