Early-stage monitoring on faults of rolling bearings based on fractal feature extraction

摘要

Vibratory signals from a machine, under different faulty conditions, often show themselves different geometric structures, which is the basis of applying principle of fractal geometry to its health condition monitoring and fault diagnosis. An approach for fractal feature extraction from vibratory signals of rolling bearings was studied in this paper. We judged the existence of an early-stage inner race defect by analyzing the trend-changing characteristics on all fractal feature parameters synthetically, determined the window time when the early-stage inner race defect occurs by means of the change points of maximum linearity-mean values on fractal characteristic curves, and made a final diagnosis decision by quantitatively evaluating symptom likelihood on the fractal feature parameters. The results from a case study based on the test-to-failure data showed that the propose approach in this paper is effective and of great application potential to early-stage health condition monitoring and fault diagnosis of rolling bearings.