Modeling and simulation of non-stationary vehicle vibration signals based on Hilbert spectrum

Abstract

Vehicle vibrations are random and highly non-stationary in many occasions, however, so far there has been no suitable model to characterize and simulate these non-stationary random vibrations. Under the circumstances, some primary research was made in this paper on modeling and simulating non-stationary random vibration signals. After reviewed various tools available to analyze non-stationary data and compared their relative merits, this paper has established a new model to represent non-stationary random data based on the Hilbert spectrum, and has proposed a novel five-step method to simulate non-stationary vehicle vibration signals, which can mostly retain statistical characteristics of the sample signal. This model and this method have been verified the effectiveness and feasibility by numerical simulations and practical experiments based on various records of random vibration signals collected from transport vehicles, further they could be applied in many fields such as vehicle dynamic analysis, and vibration environmental testing.