© 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

This paper presents a fault detection system for rotative machinery. A permanent-magnet DC motor is used as case of study. The main idea is to estimate on-line the non-load torque (To) in order to monitor the bearing health condition. The fault detection system is based on the design of a generalized Takagi-Sugeno PI (proportional-integral) observer. The main advantage of this approach is that it can be easily implemented because the observer gains are obtained by solving a set of LMIs (linear matrix inequalities). Moreover, the method can be extended to more complicated nonlinear systems by using the Takagi-Sugeno approach. A simulation is performed to show that this fault detection scheme can be applied to detect abrupt faults on rotative machinery which can lead the system to undesirable performance caused by vibrations or breakdown.

_