Dynamic Uncertainty Analysis for Random Vibration Signals in Flight Test

Abstract

Random vibration signals with insufficient information can be obtained in flight tests. Traditional uncertainty analysis methods based on statistical theory cannot be employed when the precise probability density and trends are unknown. In this paper, a bootstrap gray method based on gray system theory and bootstrap method is proposed to dynamically analyze random vibration signals. A set of dynamic analysis indices, such as estimated value, variation domain, dynamic uncertainty, reliability, and mean uncertainty, are used to assess the effectiveness of the bootstrap gray method. A group of parameters is discussed to optimize the performance of the proposed method. For time series of random vibration signals, the analysis of variation domain demonstrates the superiority of the bootstrap gray method over the bootstrap method. The instantaneous fluctuation of time-varying signals can be accurately evaluated by dynamic uncertainty, for which prior information on the probability distribution is not taken into account. The statistical characteristic of dynamic uncertainty is quantified with mean uncertainty. The results show that the bootstrap gray method has an advantage in analyzing uncertainty for random vibration signals with insufficient information, and the reliability reaches 100% at the given confidence level.