滚动轴承磨损区域静电监测及性能退化评估方法研究

Condition monitoring and performance degradation assessment of rolling bearing are important tools for condition based maintenance and safety protection of equipment. In view of the traditional monitoring techniques for the inadequacy of ability of early failure found, in this project, based on electrostatic induction, a new wear site monitoring method is investigated and a multiple point charge sensing measurement model with random distribution of spatial position is modeled for the analysis of the influence of probe and insulating layer materials, structural parameters, etc. on sensor capability, for the guidance of the optimization design of sensors. Electrostatic induction signal denoising method is presented and multi-dimension features of sensitive performance degradation trends in the whole life cycle of rolling bearing are extracted in the following. On such basis, local-global dimension reduction algorithm is adopted for the extraction of the internal structure of multidimensional characteristics to construct a performance degradation assessment model. An improved local outlier factor of weighting distance method is created to automatically wipe off outlier in real time, improving the accuracy and robustness of algorithm. Verification and application research of sensors as well as proposed theories, methods and models are performed through the transducer calibration test, pin-plate wear monitoring test, rolling bearing monitoring test and wind power gear box test. The results have important theoretical significance and engineering value for enriching the methods of rolling bearing condition monitoring, improving the security, reliability and economical efficiency of equipment service.

滚动轴承状态监测和性能退化评估方法是实现视情维修管理和保障设备安全服役的重要手段。针对常规在线状态监测技术对于滚动轴承早期退化发现能力不足的问题，本项目研究一种基于静电感应的磨损区域监测新方法，建立空间位置随机分布多点电荷传感测量模型，分析探极和绝缘层材料、结构参数等对传感器性能影响的规律，优化设计传感器；研究磨损区域静电感应信号预处理方法，从多角度提取对滚动轴承全寿命周期性能退化趋势敏感反应的特征；在此基础上，运用局部-全局降维算法提取多维特征内在结构，构建改进加权距离局部离群因子性能退化评估模型，实时自动去除野点，增强算法的准确性和鲁棒性；通过传感器标定试验，销盘磨损监测试验，滚动轴承监测试验以及风电齿轮箱搭载试验对传感器性能以及所提理论、方法、模型进行验证与应用研究。本项目的研究成果对于丰富滚动轴承在线状态监测手段，提高设备服役安全性、可靠性和经济性，具有重要的理论意义和工程价值。

滚动轴承状态监测和性能退化评估方法是实现视情维修管理和保障设备安全服役的重要手段。针对常规在线状态监测技术对于滚动轴承早期退化发现能力不足的问题，本项目研究了一种基于静电感应的磨损区域监测新方法，在传感器设计、信号去噪和特征提取以及性能退化评估方面进行了深入研究。分析静电测量原理并设计研制了磨损区域静电传感器，建立带电电荷经过探极表面的感应测量模型，探讨静电感应信号的主要影响因素；研究磨损区域静电感应信号去噪方法，从时频域和复杂度等多角度提取轴承退化的敏感特征；采用谱回归完成特征集的降维和冗余消除，进而建立健康状态的高斯混合模型，以基于贝叶斯推断的距离作为轴承性能退化程度的定量评估指标。通过销盘磨损监测试验，滚动轴承监测试验对传感器性能以及所提理论、方法、模型进行验证与应用研究。本项目的研究成果对于提高机械设备关键部件的状态监测和性能退化定量评估能力具有重要的参考价值和指导意义。