Identifying an Appropriate Daubechies Wavelet in Analyzing the Fatigue Failure Utilizing Fuzzy C-Means

This work aims to determine the best Daubechies wavelet for analyzing fatigue failure. This study involved 180 strain data ranging from 200 to 2000 με. The wavelet coefficients of each data were determined based on db4, db12, db20, and db30. Furthermore, they were further correlated with fatigue life predictions obtained from strain-life models utilizing the Fuzzy C-Means. Simulation results revealed that wavelet coefficients were negatively correlated with fatigue life such that an increase in the wavelet coefficients led to a reduction in fatigue life. Moreover, discrete db12 produced the strongest correlation with the coefficient of determination up to 0.477, indicating that this wavelet is the most suitable for fatigue failure analyses.