行星齿轮传动系统故障演化机理与稀疏诊断方法研究

Planetary gear transmissions are one type of the heavy-load transmission devices in complex electromechanical systems such as wind turbine and helicopter. The complex structure of planetary gear transmissions always works in changeable environments with variable-speed, variable-load, or strong impact load, which results in the multiple types of gear failures and it is difficult to diagnose the fault of planetary gear transmissions. Moreover, the time-varying characteristics of the vibration transfer path will also increase this difficulty. To solve the problem, this project will research the direct problem and the inverse problem of planetary gear transmissions fault diagnosis in a combined way. The research about the direct problem, fault evolution mechanism, will reveal the developing raw of the gear fault and the mapping mechanism between the vibration response and the fault, thus the fault features can be summarized. Based on the research on the fault evolution mechanism, because of the merit that essential characteristics of complex signals can be described by sparse representation, sparsity-enabled fault diagnosis method will be studied in the project to extract the fault features accurately from the non-stationary and complicated signal and thus to diagnose faults of planetary gear transmissions. The main contents of the project include four-folds as following: the mathematical description for different faults and analytic model of gear time-varying mesh stiffness, the mapping mechanism of the vibration response to the fault in the variable-load planetary gear transmissions, the construction method of the objective function in the algorithm of sparse representation for fault feature extraction, and the solving algorithm for the constructed objective function and performance analysis. The research results of this project will provide theoretical basis and technical support for planetary gear transmissions fault diagnosis, and thus will play a significant role in guaranteeing the safety and reliable operation of complex electromechanical systems.

行星齿轮传动系统是风电装备、直升机等复杂机电系统的重载传动装置，长期运行于变速变载工况并经受强冲击作用，故障频发。加之行星齿轮传动系统结构复杂、失效形式多样、振动传递路径时变等特点，导致其故障诊断困难重重。本项目拟采用正反问题相结合的思想，通过正问题——故障演化机理研究，揭示行星齿轮传动系统不同故障发生、发展的振动响应机制；针对反问题——故障特征提取，利用稀疏表示能“疏而不漏”地刻画信号本质特征的优势，研究非平稳复杂故障特征稀疏表示方法，实现行星齿轮传动系统故障稀疏诊断。研究内容主要包括：行星齿轮传动系统裂纹、点蚀、缺齿等典型故障状态下时变啮合刚度计算模型；变载行星齿轮传动系统不同故障状态的振动响应机理；故障特征稀疏表示超完备字典构造与目标函数构建；目标函数快速求解优化算法与性能分析。本项目将为行星齿轮传动系统故障诊断提供理论支撑与技术支持，对保障复杂机电系统安全可靠运行具有重要意义。

行星齿轮传动系统是风电装备、直升机等复杂机电系统中的重要传动装置，运行工况复杂恶劣，故障频发。本项目针对行星齿轮传动系统故障诊断的难点，正反问题相结合，从正问题方面开展行星齿轮传动系统故障演化机理，揭示行星齿轮传动系统典型故障振动响应机制，反问题方面研究有针对性的系统故障稀疏诊断算法，并开展实验对理论研究及稀疏诊断技术进行验证。经过三年的研究，取得了如下成果：（1）建立了正常状态及故障状态下行星齿轮传动系统时变啮合刚度计算模型，得到了正常状态以及行星齿轮传动系统不同故障类型、不同故障程度下系统时变啮合刚度；（2）建立了行星齿轮传动系统多自由度平移—扭转耦合动力学模型，构建行星轮系运动微分方程，得到正常状态及不同故障状态下系统的振动响应信号，结合振动传递路径分析，研究了不同故障状态下系统的振动响应特征；（3）针对齿轮故障振动信号特点，利用相关滤波法计算最优匹配小波原子，构建基于最优小波基的稀疏表示目标函数，研究了基于分裂增广拉格朗日收缩算法的目标函数快速求解算法，并分析算法性能；结合振动传递路径分析，提出齿轮箱复合故障两步诊断策略；（4）从概率角度出发，依据不同位置原子被选中的概率，对多次基于随机正交匹配追踪算法的稀疏表示结果进行融合，提出基于均化随机正交匹配追踪算法的齿轮箱稀疏诊断方法；（5）设计并搭建行星齿轮传动系统故障模拟实验装置与测试系统，开展行星齿轮传动系统不同故障位置、不同故障类型故障模拟实验，收集行星齿轮传动系统故障振动响应，验证故障机理理论研究及稀疏诊断技术的正确性和有效性。本项目研究成果为行星齿轮传动系统的故障诊断提供了理论基础与技术支持。.在上述研究的基础上，共发表论文10篇，包括期刊论文8篇，会议论文2篇，其中SCI检索4篇，EI检索5篇，完成硕士论文5篇，江苏省大学生创新创业计划训练项目1项，申请发明专利4项，其中3项已授权，申请并授权实用新型专利2项。