基于高阶累积量的齿轮系统辨识方法研究

高阶累积量分析技术是近年来迅速发展起来的新技术,作为处理非高斯信号、非线性信号以及盲信号的重要工具而受到日益重视.它有许多优点,特别是对于外加高斯噪声不敏感以及可以检测系统的非线性特性,故理论上可以完全抑制高斯噪声的影响,提高分析精度.本文在对其研究的基础上,将它应用于齿轮旋翼传动系统动态特性分析之中,对传动系统和旋翼桨叶的动态固有频率等进行了识别,并与频域和时域识别法进行了比较,表明该方法是一种可行的有发展前途的方法,为齿轮系统的动力学研究提供一种新的有效分析方法.

A Gear System Identification Method Based on Higher Order Cumulants

A gear transmission system exhibits a coupling nonlinear vibration due to the nonlinearity of gear meshing,bearing and gearbox supporting system stiffnesses incorporated with the modulating of gear and bearing as well as the nonlinear coupling of rotary wings.The dynamic signals of the system appear of complicated non-Gauss and nonlinear characteristics.A new method of modern signal processing-the higher order cumulant was adopted.The nonlinear modulating characteristics of the system were monitored by determining the dynamic response of the transmission system and helicopter rotary wings.The dynamic natural frequencies of the transmission system and the rotary wings were obtained by analyzing in the cumulant domain.Its comparison with the power spectrum method and the time-series analysis (ARMA) identification Method shows that the higher order cumulant method is supperior in noise inhibiting,higher identification accuracy and nonlinear characteristic extracting of the system.