首页|基于叶尖定时数据奇异值分解的振动事件识别

基于叶尖定时数据奇异值分解的振动事件识别

扫码查看
原文链接
  • 万方数据
  • 维普
叶尖定时数据自动化测量及处理是旋转机械在线监测和智能运维的必要环节,快速、准确判断叶片振动类型,实现振动事件识别是数据自动化测量及处理的关键.提出了一种基于加窗叶尖定时数据奇异值分解的振动事件识别方法,仅需单只传感器准确识别叶片同步、异步振动事件.基于不同叶片叶尖定时数据时延特性,加窗构造了"类重构吸引子矩阵",根据矩阵奇异值特征实现振动事件识别.开展了方法仿真及实验验证,仿真与实验结果一致性良好,压气机试验件测试数据表明,叶片发生振动事件时第1奇异值增大为7倍以上,其中发生异步振动事件时第2奇异值增大为14倍以上,提出方法能够准确识别叶片同步、异步振动事件.
Vibration Events Identification Method Based on SVD of Blade Tip Timing Data
The automatic measurement and processing of blade tip timing data is an essential part of online monitoring and intelligent ma-intenance of rotating machinery. The key is to identify vibration events quickly and accurately. A vibration events identification method based on singular value decomposition( SVD) of windowed blade tip timing data is proposed. The synchronous and asynchronous vibration events can be accurately identified by a single probe. Based on the time-delay characteristics of different blades,a"Hankel-like matrix"is constructed. The vibration events are identified by the singular value characteristics of the matrix. The simulation and experiment are car-ried out and show good consistency. The experiment data of the engine compressor show that when the vibration events occur,the first sin-gular value is increased by more than 7 times. When the asynchronous vibration events occur,the second singular value is increased by more than 14 times. It proves that the proposed method can accurately identify blade synchronization and asynchronous vibration events.

vibration measurementblade tip timingvibration eventssingular value decomposition( SVD)hankel matrix

支烽耀、牛广越、段发阶、邓震宇、钟国舜

展开 >

天津大学精密测试技术及仪器全国重点实验室,天津 300072

中国电子科技集团公司第十一研究所,北京 100020

振动测量 叶尖定时 振动事件 奇异值分解 重构吸引子矩阵

国家自然科学基金项目国家自然科学基金项目中国博士后科学基金项目天津大学科技创新领军人才培育"启明计划"项目精密测试技术及仪器全国重点实验室(天津大学)青年教师科研启动项目国家科技重大专项项目

52205573U22412652022M7201062024XQM-0012Pilq2304J2022-V-0005-0031

2024

10.3969/j.issn.1004-1699.2024.05.009

传感技术学报
东南大学 中国微米纳米技术学会

传感技术学报

CSTPCD北大核心
影响因子：1.276
ISSN：1004-1699
年,卷(期)：2024.37(5)