基于传递函数特征的机械结构故障诊断方法新进展

故障诊断对估计机械结构的健康状态非常重要，诊断方法一直以来也是研究的热点问题。与其他故障诊断方法相比，基于传递函数的故障诊断方法操作简单，提供的故障特征和指示因子也更加灵敏有效。根据所利用的线性传递函数或非线性传递函数的类型不同、基于传递函数的故障诊断方法目前大概有4类，即一般线性方法、一般非线性方法、基于广义频响函数（GFRF）方法和基于二阶输出频谱（SOOS）方法。本文旨在对各种基于传递函数的故障诊断方法研究进展进行梳理，通过带阻尼弹簧单元的多自由度模型解释各种方法的基本原理和操作流程，通过具体的仿卫星结构中螺栓松动故障实验比较各种方法的优缺点，探讨现有故障指示因子灵敏度的提高方式及现有方法在其他复杂机械结构中的运用前景。

Recent advances of fault diagnosis methods based on transmissibility function for mechanical structures

The importance of fault diagnosis to health monitoring of mechanical structures draws extensive attention from researchers to diagnosis methods. Transmissibility function based methods, compared with other diagnosis methods, are much simpler and provide more sensitive and effective damage features and indexes. According to different transmissibility functions adopted, either linear or nonlinear, transmissibility function based methods can be classified into four categories:general linear methods, general nonlinear methods, methods based on Generalized Frequency Response Function (GFRF) and methods based on the Second Order Output Spectrum (SOOS). The main purpose of this article is to provide a comprehensive review on recent advances of fault diagnosis methods based on linear and nonlinear transmissibility functions, explain their basic principles and procedures through nonlinear multiple degrees of freedom models, demonstrate their merits and demerits with experimental results on a bolted satellite-like structure, and discuss their improvement in sensitivity of related damage indicators and potential applications to some complex structures.