Durability Analysis for Coil Spring Suspension Based on Strain Signal Characterisation

Abstract

This paper aims to predict the durability of an automobile coil spring by characterising the captured strain data. The load histories collected at coil spring are often presented in time domain but time domain cannot provide sufficient information for fatigue life prediction. The objective of this study was to characterise the strain signal in time domain, frequency domain and time-frequency domain for fatigue life prediction. The signal obtained in time domain was used to predict the fatigue life of the coil spring through Rainflow cycle counting technique and models of strain-life relationships. In frequency domain, fast Fourier transform revealed that the frequency components in the strain signal ranged between 0-5 Hz. The frequencies can be further categorised into two ranges: 0-0.3 Hz and 1-2 Hz. Power spectral density confirmed that the frequencies with high energy content were 0-5 Hz and the total energy content in the signal is 4.0872x103 µɛ2. Short time Fourier transform can identify the local time and frequency properties of the signal but it has a limitation in time-frequency resolutions. Wavelet transform can provide a better time-frequency resolutions and it confirmed that the transients in the time domain had frequency range of 1-2 Hz. In summary, this study revealed different possible approaches of signal processing in fatigue life assessment of automotive components as guidance for the selection of suitable approach based on the type of information needed for the analysis.