A Soft Technique for Fault Detection and Identification in Mechanical Systems

Abstract

There are two approaches to diagnose a fault for mechanical systems: add hardware sensors, design software algorithms. This paper will focus on software approach to detect and identify a class of faults. The proposed approach uses an estimator built to enable the detection of fault occurrences through comparison of observed states with available sensor information, while other nonlinear estimator is used to identify the fault extracting its signature. Advantages of the proposed fault detection and identification methods are that they are based on the established dynamic model, they do not require acceleration measurements. Experimental study is given to illustrate the effects of the proposed approach.

Keywords

• Fault detection
• Software
• Estimator
• Industrial machine