重载行星齿轮系统的多状态参数在线监测与故障诊断新方法研究

Planetary gearboxes are commonly employed in heavy-duty power trains for transport, energy and military applications. However, their operation and maintenance face challenges due to the heavy-duty loads and harsh environments. Any failures with those planetary gear systems have potential to cause expensive overhaul and safety concerns. The existing condition monitoring techniques have constrains in particular when dealing with the diagnostic problems in planetary gearboxes because of the complex signal transfer path and the fluctuating operational conditions. It also lacks the summary of diagnostic features for heavy-duty planetary gear system, which contain rich fault information. To address these issues, the proposed project aims to study the dynamic behaviors of heavy-duty planetary gear system for fault diagnosis and develop a novel data fusion monitoring system that uses multiple fiber Bragg grating (FBG) based sensors of no interference with the normal operation of gearboxes. By using the FBG excellent wavelength-encoded property, various FBG-typed sensors can realize the precise measurements of multiple variables, such as strain (or stress), acceleration and temperature. They also have attractive characteristics such as small size, anti-corrosion and multiple measuring points on one fiber. The main research topics of this project include: the study of the influence of failure on dynamics of rigid-flexible coupled helical planetary gear systems; the research of the sensing principles of FBG sensors and the optimization of measuring layout for the condition monitoring of planetary gearboxes; the development of a new hybrid diagnostic technique by employing the data fusion of various information from multi-variables and different measured locations, which improves the effectiveness and reliability under non-stationary operational conditions. The achieved results can assist in the condition monitoring of advanced heavy-duty machines.

行星齿轮箱是运输、能源和军事等现代工业体系的关键部件，但恶劣的工况使其容易出现故障，严重威胁系统安全经济的运行。本项目针对故障下重载行星轮系动态特性研究缺乏，以及困扰行星齿轮箱在线诊断的信号传递路径复杂和外部非平稳激励干扰抑制困难等难题，探索重载行星轮系在故障与外部激励下的响应特性，利用光纤光栅可嵌入箱体内部、一线多测点及对应变、振动和温度多参数敏感的特点，提出融合多状态参数和多测点信息的重载行星轮系在线监测与故障诊断新原理、新方法。主要研究内容包括：研究刚柔耦合斜齿行星轮系故障作用与各状态参数时频响应特性的关系；研究在线监测行星轮系应变、振动和温度动态响应的光纤光栅分布式传感系统；研究多状态参数分布动态测量信号的时空配准和信息融合的故障诊断算法，削弱非稳态工况干扰，提升故障辨识精度。本研究成果对丰富机械诊断理论具有重要意义，可为保障高端复杂装备可靠性的在线监测提供新的有效途径和科学依据。

以行星齿轮箱为代表的传动设备是机械装备中的关键基础部件之一，广泛应用于风力发电、海事船舶、航空航天以及国防武器装备等重大领域。随着我国高端机械装备朝高功率密度和重型化方向发展，传动设备故障概率日益增高，制约了我国高端机械装备的安全、高效运行。而及时发现、定位传动设备故障可有效为高端机械装备视情维护提供基础依据，对我国高端机械装备安全服役与降低运行成本具有十分重要的意义。本项目针对振源耦合、传递路径复杂、变运行工况造成的传动设备故障特征微弱、波形形貌不断变化等难题,围绕系统动态响应特性、先进传感监测手段以及非稳态、变工况下故障特征提取算法等方面展开了研究。.项目主要研究内容工作与成果包括：.(1) 定量分析了齿轮副跳动、分度误差对其静态传动误差的影响，探究了分度误差和跳动误差对齿轮副动态特性的影响规律；.(2) 探索并分析验证了将光纤光栅嵌入潜在故障点附近以排除传递路径干扰来诊断传动设备故障的可行性；.(3) 首创了一种在非稳态工况无转速计情况下基于动态时间规整的齿轮副动态信号重采样与故障诊断算法；.(4) 利用卷积神经网络和稀疏自编码器等深度学习框架结合迁移学习算法进一步提升了变工况情况下齿轮箱故障诊断的精度和效率；.(5) 项目还将有关理论和方法应用于板壳结构状态监测与故障诊断以及非稳态工况条件下机械装备的振动抑制等研究方向。.相关成果对我国高端机械装备的健康管理、动态测试与分析提供重要的理论和技术支撑。项目在Mechanical Systems and Signal Processing等权威期刊和知名国际会议共发表高水平论文8篇，其中SCI收录期刊论文4篇，EI收录论文4篇。