Eduardo Mendel, Estefhan D. Wandekokem, Flávio M. Varejão, Thomas W. Rauber, and Rodrigo J. Batista
Feature selection, motor pumps, condition monitoring, SVM, classification
The early detection and diagnosis of faults in industrial machinery enables damaged components to be repaired during planned maintenance, which minimizes machinery standstill. In this work we report about fault diagnosis experiments to improve the maintenance quality of motor pumps installed on oil rigs. Vibrational patterns are the basis for describing the condition of the process. As in a real-world domain model-free approaches for creating the fault classification rules are better suited, we rely on the data-driven approach to the learning of the fault classes, i.e. supervised learning in pattern recognition. Our work is motivated by the diversity of the studied defects, the availability of real data from operational oil rigs, and the use of statistical pattern recognition techniques usually not explored sufficiently in similar works. We show the results of automatic methods to define (extract), select and combine features that describe the process and to classify the faults on the provided examples.The support vector machine is chosen as the classification architecture.
Important Links:
Go Back
