Abstract
This paper describes the Mildly Nonstationary Mission Synthesis (MNMS) algorithm which has recently been developed for the purpose of producing short vibration mission signals which are representative of experimentally measured road data. The MNMS method makes use of the Discrete Fourier Transform (DFT), the Orthogonal Wavelet Transform (OWT) and bump (shock) selection and reinsertion techniques. By performing a wavelet grouping procedure, and by setting trigger levels, the user can produce signals which are shortened by up to a factor of 10 compared to the original road data. The resulting missions are representative of the original data in terms of Power Spectral Density (PSD), Probability Density Function (PDF), RMS value, Crest Factor and Kurtosis value. Mission synthesis results vary depending on the level of nonstationarity present in the original data, but obtained mission signal PSD, RMS, and Kurtosis values are typically within +/-10% of the road data targets. The MNMS algorithm has been implemented as a Fortran program for DOS-compatible personal computers