Hurst parameter and fractal dimension concept for fault detection in electric motors

In this paper we aim to define the degradation of the bearings in electric motors related to parametric values. Aging tests of an electric motor are performed according to IEEE test procedures. The vibration signals of the bearings are acquired via accelerometers. This step is followed by the analysis of the data in the spectral frequency domain. Finally, the Hurst parameter is calculated as a measure of the degradation. Since there is a simple relationship between the Hurst parameter and the Fractal Dimension concept, the latter can be considered as another parametric approach to define the degradation. We recognize that the positive Hurst parameter in the healthy case changes to a negative value for the faulty case, which indicates that the degradation is produced by the aging process. The difference between the parameters for the healthy and faulty cases is a measure of the degradation.