欢迎访问《计测技术》杂志网站

基于熵变匹配追踪的叶端定时数据缺失识别方法研究

Research on identification method of blade tip timing data loss based on Correntropy Matching Pursuit

中文摘要:

为解决叶端定时系统在实际应用中存在的数据缺失问题，提出基于熵变匹配追踪的叶端定时数据缺失识别方法。该方法利用相关熵诱导度量基于高斯核函数度量样本的权重。不同于正交匹配追踪对所有观测数据赋予相同权重，熵变匹配追踪基于相关熵诱导度量变化，对观测数据赋予不同范数类型的权重，使得其对异常值具有较好的鲁棒性。通过仿真分析与实验对该方法的性能进行验证，结果显示所采用的熵变权重因子为数据缺失位置分配了接近于零的权重，有效降低了数据缺失对特征提取结果的影响，证明了该方法的鲁棒性。基于熵变匹配追踪的叶端定时数据缺失识别方法为叶端定时系统的装机应用提供了理论支撑，具有技术借鉴价值。

英文摘要:

To address the issue of data loss commonly faced by tip timing systems in practical applications, a method for identifying missing tip timing data based on Correntropy Matching Pursuit is proposed. This method uses Correntropy Induced Metric based on Gaussian kernel functions to measure sample weights. Unlike orthogonal matching pursuit, which assigns the same weight to all observed data, Correntropy Matching Pursuit assigns weights of different norm types to the observed data based on changes in correlated Correntropy Induced Metric, making it more robust to outliers. The performance of this method was verified through simulations and experiments. The results showed that the correntropy weight factor assigned nearly zero weights to the missing data locations, effectively reducing the impact of data loss on fea? ture extraction results, thus demonstrating the robustness of the method. The tip timing data loss identification method based on Correntropy Matching Pursuit provides a theoretical basis for the implementation of tip timing systems in practi? cal applications.

作者	单位
杨志勃	（西安交通大学，陕西西安 710049）
吴淑明*	（西安交通大学，陕西西安 710049）
乔百杰	（西安交通大学，陕西西安 710049）
王亚南	（西安交通大学，陕西西安 710049）
陈雪峰	（西安交通大学，陕西西安 710049）

中文关键词: 叶端定时数据缺失特征识别熵变匹配追踪

英文关键词:blade tip timing data loss feature identification Correntropy Matching Pursuit

基金项目:

DOI：10.11823/j.issn.1674-5795.2024.02.04

引用本文：杨志勃,吴淑明*,乔百杰,王亚南,陈雪峰.基于熵变匹配追踪的叶端定时数据缺失识别方法研究[J].计测技术,2024,(2):.

关闭