油液光谱数据诊断综合传动装置异常磨损定位方法  

Study on Abnormal Wear Location of Integrated Transmission Components Based on Oil Spectral Data

作　　者：徐峰[1] 张倩倩 季文龙[1] 贾然[3] 张鹏[4] 郑长松[4] XU Feng;ZHANG Qian-qian;JI Wen-long;JIA Ran;ZHANG Peng;ZHENG Chang-song(No.32184 Unit of PLA,Beijing 100072,China;No.32381 Unit of PLA,Beijing 100072,China;Key Laboratory of Modern Measurement and Control Technology of Ministry of Education,Beijing Information Science and Technology University,Beijing 100192,China;School of Mechanical Engineering,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]中国人民解放军32184部队,北京100072 [2]中国人民解放军32381部队,北京100072 [3]北京信息科技大学现代测控技术教育部重点实验室,北京100192 [4]北京理工大学机械与车辆学院,北京100081

出　　处：《光谱学与光谱分析》2024年第5期1398-1404,共7页Spectroscopy and Spectral Analysis

基　　金：装备预研重点实验室基金项目(141009AT03091H)资助。

摘　　要：磨损是影响综合传动装置工作可靠性及使用寿命的重要因素之一,当前相关研究中常用的聚类、主成分分析、加权融合等油液光谱数据分析方法,缺乏对特定元素浓度指标异常磨损情况下随时间增长的考虑。为分析综合传动装置不同零部件的磨损状态,提出一种基于油液光谱数据的零部件异常磨损定位分析方法。针对综合传动装置异常磨损过程中部分元素在特定阶段会出现快速增长的情况,提出基于时间窗相关距离的聚类方法,分离表征不同零部件磨损状态的元素;提出磨损元素的磨损趋势分级方法,以高磨损趋势元素为聚类中心,使聚类结果具备可解释性;通过分级系数确定零部件磨损元素权重,融合各零部件磨损元素,获取不同零部件磨损状态表征;通过异常磨损界限值识别异常磨损,实现零部件异常磨损定位。以综合传动装置润滑油液光谱数据为例,检测判断该装置异常磨损的零部件及时间段。检测结果表明:Fe、Cu、Pb三种元素的磨损趋势分级系数最高,携带大量磨损信息,能够有效表征装置的磨损状态;基于时间窗相关距离的有中心聚类方法,成功将油液光谱数据分为Fe、Cu、Pb三类,可用于有效表征整体、摩擦片、齿轮组的磨损状态;基于分级系数的加权融合方法可以有效对该装置的异常磨损部位和时间周期进行检测和判断,为后续的故障预防和维护提供技术指导。Wear is one of the important factors affecting the working reliability and service life of the integrated transmission device.Abnormal wear of the components of the integrated transmission device will reduce its operating efficiency and even cause its random failure,resulting in significant economic and military losses.Therefore,it has become an important method to improve the reliability of integrated transmission devices to quickly and accurately detect the characteristics of wear elements and locate the parts with abnormal wear by using oil spectral data.However,oil spectral data samples generally contain many interfering and additive elements,etc.Clustering,principal component analysis,weighted fusion and other methods commonly used in current studies lack consideration of the increase of abnormal wear of specific element concentration indexesover time.In order to analyze the wear state of different parts of the integrated transmission,a method of abnormal wear location analysis of parts based on oil spectral data was proposed.A clustering method based on the correlation distance of the time window was proposed to separate the elements representing the wear states of different parts.The wear trend classification method of wear elements was proposed,with high wear trend elements as the cluster center,so the clustering results could be interpreted.The weight of component wear elements was determined by classification coefficient,and the wear elements of each component were fused to obtain the representation of the wear state of different components.Abnormal wear can be identified by abnormal wear threshold value to locate abnormal wear of parts.The parts and period of abnormal wear were detected and judged.The test results show that Fe,Cu and Pb have the highest wear trend classification coefficient and carry a lot of wear information,which can effectively characterize the wear state of the device.The centralized clustering method based on the correlation distance of the time window successfully divides the oil

关 键 词：机械磨损 油液光谱数据 磨损趋势分级 异常磨损定位 

分 类 号：TH212[机械工程—机械制造及自动化]