摘要
为解决电力集中监控设备缺陷识别滞后、预测精度不足等问题,提出一种基于声音信号分析的典型缺陷预测技术。该技术通过构建设备运行声音信号分析模型,将声音频谱特征与缺陷发展规律深度关联,建立了集声音采集、特征提取、模式识别以及预测决策于一体的智能预测框架。实用性能测试结果表明,该技术在缺陷预测准确率、故障提前发现时间、误报率控制等方面均明显优于传统定期巡检方法。
In order to solve the problems of lagging defect identification and insufficient prediction accuracy of centralized power monitoring equipment,a typical defect prediction technology based on sound signal analysis is proposed.This technology establishes an intelligent prediction framework integrating sound collection,feature extraction,pattern recognition and prediction decision-making by constructing an analysis model of sound signal during equipment operation,and deeply associating sound spectrum characteristics with defect development law.The practical performance test results show that this technology is obviously superior to the traditional regular inspection method in the aspects of defect prediction accuracy,fault early detection time and false alarm rate control.
作者
杨亚星
刘扬
刘欢
YANG Yaxing;LIU Yang;LIU Huan(Inner Mongolia Electric Power(Group)Co.,Ltd.,Ordos Power Supply Branch,Ordos O17000,China)
出处
《电声技术》
2025年第10期66-68,共3页
Audio Engineering
关键词
电力集中监控设备
声音信号分析
典型缺陷预测技术
power centralized monitoring equipment
sound signal analysis
typical defect prediction technology
