Time-Frequency-Based Fatigue Data Editing For Automotive Applications

This paper presents fatigue data editing (FDE) techniques using time-frequency domain. The time-frequency methods identified and extracted higher amplitude segments and produced shorter edited signals. Based on the comparison of the edited signals resulted, it was found that the wavelet transform gave the shortest signals. It was able to summarize strain signals up to 77% and maintain more than 90% of the statistical parameters and the fatigue damage. Meanwhile the short-time Fourier transform and the S-transform summarized the signals only of 23% and 22%, respectively. It proved that the wavelet transform was the best technique for FDE, especially for the automotive applications.