A mission synthesis algorithm for fatigue damage analysis

Abstract

This paper presents a signal processing based algorithm, the Mildly Nonstationary Mission Synthesis (MNMS), which produces a short mission signal from long records of experimental data. The algorithm uses the Discrete Fourier Transform, Orthogonal Wavelet Transform and bump reinsertion procedures. In order to observe the algorithm effectiveness a fatigue damage case study was performed for a vehicle lower suspension arm using signals containing tensile and compressive preloading. The mission synthesis results were compared to the original road data in terms of both the global signal statistics and the fatigue damage variation as a function of compression ratio. Three bump reinsertion methods were used and evaluated. The methods differed in the manner in which bumps (shock events) from different wavelet groups (frequency bands) were synchronised during the reinsertion process. One method, based on time synchronised section reinsertion, produced the best results in terms of mission signal kurtosis, crest factor, root-mean-square level and power spectral density. For improved algorithm performance, bump selection was identified as the main control parameter requiring optimisation.