|
|||
Date: 28 January 2016
|
Presentation of the Thesis |
|
|
Abstract:The objective of this Thesis is to improve the performance in condition monitoring systems in fault diagnosis using vibration and audio signals in two applications, bearings and pumps, with special emphasis in the feature extraction stage. In bearing vibration fault diagnosis, two nonlinear techniques are proposed. In centrifugal pump application, the feasibility of audio-based fault diagnosis is studied and a comparison with vibration-based fault diagnosis is carried out. A set of 31 new features are proposed for centrifugal pump diagnosis. A success rate of 99.73% is achieved using audio signals and a 100% is achieved in combination with vibration signals. |