变转速下行星齿轮箱时变非平稳复合故障诊断方法研究

Planetary gearbox is widely used in wind turbine, aviation, lifting and transportation and other complex machinery transmission system. Due to the complex working environment (corrosion, pollution) and operating state (variable speed, impact load), the key components of the planetary gearbox often occur serious wear and fatigue cracks and other faults. In addition, multi-fault prone to occur at the same time. Compared with the traditional single-fault under constant rotational speed, the amplitude and waveform of the fault impulse under variable rotational speed are time-varying nonstationary. Meanwhile, complex nonlinear modulation coupling make the diagnosis more difficult. Traditional methods cannot explain the nonlinear time-varying non-stationary modulation coupling phenomenon of multi-fault, cannot eliminate the gear mesh noise modulation and fault characteristic frequencies resulting from nonlinear coupling, cannot solve the problems of low frequency weak impact separation and quantitative characterization. Hence, this project will be fully research the new theory and new method of the planetary gearbox time-varying non-stationary multi-fault feature extraction. The main contents include: (a) dynamic modeling and vibration signal modulation coupling characteristics of typical complex fault combination under complex conditions; (b) automatic order decomposition of target characteristic components based on time-frequency analysis technology; and (c) establish generalized characteristic index model based on adaptive generalized demodulation transform and fault type matching identification.

行星齿轮箱广泛应用于风力发电、航空、起重运输等复杂机械的传动系统，由于复杂的工作环境（腐蚀、污染）和运行状态（变转速、冲击负载），行星齿轮箱中的关键部件常发生严重磨损和疲劳裂纹等故障，而且，多个部件故障容易同时产生。相对于传统恒速下的单一故障，变转速下行星齿轮箱复合故障冲击的幅度和波形表现为时变性，同时产生复杂的非线性调制耦合，诊断更为困难。针对传统方法不能分析解释行星齿轮箱时变非平稳复合故障特征的非线性调制耦合现象，不能消除齿轮啮合噪声与非线性调制耦合带来的伪故障特征频率，不能解决低频弱冲击现象中复合故障特征分离与量化表征问题，本课题将全面深入研究行星齿轮箱时变非平稳复合故障特征提取的新理论、新方法。主要研究内容包括：典型复合故障组合复杂条件下的动力学建模及其振动信号调制耦合特性、基于时频分析技术的目标特征成分自动阶比分解和基于自适应广义解调的广义特征指标模型建立和故障类型匹配识别。

实现变转速条件下行星轮系复合故障特征提取对于重大装备传动系统的安全、经济和有效运行具有重要的工程意义。本项目致力于探明时变工况下故障冲击时变特性、行星轮系故障特征频率分布规律，提出时变故障特征频率高聚集性时频表征、高精度模式分解和量化表征理论。取得的突破性进展概述如下：(1) 揭示了时变工况下故障冲击幅值和故障冲击间隔的双时变特性，明确了双时变特性导致的故障特征提取难题，提出了自适应广义解调变换理论；(2) 探明了行星齿轮箱振动信号中瞬时频率分布规律，提出了行星轮系时变且密集故障特征频率高聚集性时频表征和高精度模式分解技术；(3) 基于行星轮系故障特征频率分布规律，提出了行星轮系时变且密集故障特征频率高精度量化表征理论。.基于本项目研究成果公开发表学术论文16篇，其中以第一作者/通讯作者发表SCI论文13篇（JCR1区11篇）、EI论文14篇，以第一作者发表在《机械工程学报》上的关于自适应广义解调变换的论文入选 2020 年度 Top10 高被引。基于本项目的资助与相关研究，项目负责人成功申请到了博士后科学基金面上项目和博士后科学基金特别资助项目；清华大学博士后出站考核优秀；入选北京工业大学高层次人才（校聘教授、博士生导师和硕士生导师）；担任期刊《Journal of Dynamics, Monitoring and Diagnostics》和《兵器装备工程学报》青年编委；参加国内外学术会议4次，其中报告2次、分会场主持人1次；培养博士研究生1人、硕士研究生5人、本科生1人。