基于改进频谱分布算法的发电机定子槽楔检测  

Generator Stator Slot Wedge Detection Based on Improved Spectral Distribution Algorithm

作　　者：黄艋 周游游 Huang Meng;Zhou Youyou(Shanghai Electrical Automation D&R Institute Co.,Ltd.,Shanghai 200023,China)

机构地区：[1]上海电气自动化设计研究所有限公司,上海200023

出　　处：《电气自动化》2022年第5期92-94,101,共4页Electrical Automation

基　　金：上海市科学技术委员会资助课题(17DZ2283400)。

摘　　要：槽楔松紧度检测是大型发电机组故障诊断的例行项目,基于数字信号处理的槽楔敲击声信号检测是诊断发电机故障的有效手段。然而,人工检查的方法极易受到主观意识的影响,而且不同型号的发电机槽楔的声学特性并不相同。针对大型发电机组定子槽楔基于敲击音频的检测系统,设计了一种基于压缩传感的闭环采样结构,用于分析大型发电机的楔块。通过声学特性对信号进行分析,选择信号的有效部分。此外,还建立了检测精度与抽样误差之间的模型建立数据量以最小化传输能量模式。在系统保持低能量消耗的情况下,实现分类精度达到89.6%。Slot wedge tightness detection is a routine item of fault diagnosis for large generator sets,and slot wedge knocking sound signal detection based on digital signal processing is an effective means to diagnose generator faults.However,the method of manual inspection is easily influenced by subjective consciousness,and the acoustic characteristics of different types of generator slot wedges are not the same.A closed-loop sampling structure based on compression sensing was designed for the detection system of large generator stator slot wedge based on knocking audio frequency;by analyzing the acoustic characteristics of the signal,the effective part of the signal was selected.In addition,the data volume was established according to the relationship model between detection accuracy and sampling error,and the transmission energy was minimized.Thus,the classification accuracy can reach 89.6%while maintaining low energy consumption.

关 键 词：槽楔检测 故障诊断 音频信号 特征提取 压缩感知 

分 类 号：TM31[电气工程—电机]