| Abstract |
The application of intelligent algorithms, in electro-mechanical diagnosis systems, is increasing inorder to reach high reliability and performance ratios in critical and complex scenarios. In this context,different multidimensional intelligent diagnosis systems, based on different machine learningtechniques, are presented and evaluated in an electro-mechanical actuator diagnosis scheme. The useddiagnosis methodology includes the acquisition of different physical magnitudes from the system,such as machine vibrations and stator currents, to enhance the monitoring capabilities. The featurescalculation process is based on statistical time and frequency domains features, as well as timefrequencyfault indicators. A features reduction stage is, additionally, included to compress thedescriptive fault information in a reduced feature set. After, different classification algorithms such asSupport Vector Machines, Neural Network, k-Nearest Neighbors and Classification Trees areimplemented. Classification ratios over inputs corresponding to previously learnt classes, andgeneralization capabilities with inputs corresponding to learnt classes slightly modified are evaluatedin an experimental test bench to analyze the suitability of each algorithm for this kind of application. |
