Structural analysis and fault detection on manufacturing systems: A complex network approach

Abstract

This paper presents a structural analysis and proposes an online fault detection and location scheme for Flexible Manufacturing Systems (FMS) based on Complex Network theory. The results are applied to a canning system and they indicate that the methodology of design and built of these systems, gives scale free networks. In consequence, these systems present an acceptable tolerance to random failures and are extremely vulnerable to attacks in elements with a high degree (Hubs). In terms of this analysis, an online monitoring system and signals acquisition is designed to implement two fault detection schemas: 1) Fault detection for abrupt changes in the derivatives of the ith order of linear time invariant (LTI) systems, and 2) Model of bands and components of bands of homogeneous complex networks. The experimental results of the implementation of the monitoring system in the detection and location of faults in a canning system were satisfactory.