Mass Unbalance Default with Daubechies Wavelet

Abstract

In this study we used a fault simulator and MATLAB to compare Daubechies db4 wavelet analysis and FFT spectrum analysis for diagnosing unbalance faults under stationary and non-stationary speed conditions. FFT-based spectrum analysis excelled at identifying mass imbalance and misalignment faults during stationary operation, but struggled with resonance frequency detection. Conversely, the db4 wavelet analysis proved superior at detecting resonance frequencies and was more effective under non-stationary speeds. The complementary strengths suggest a combined approach could provide a more comprehensive mechanical fault diagnostic tool.