摘要
针对传统变电设备检修模式识别精度低、响应速度慢、人工成本高的问题,采用自适应噪声抑制算法、小波变换特征提取、支持向量机分类识别以及模糊逻辑预警决策等关键技术,提出基于声音识别技术的变电设备异常预警应用方案。通过在500 kV变电站的应用验证,所提方案在异常识别准确率、预警响应时间等关键指标上均显著优于传统方案,有效提升了变电设备异常预警的智能化水平。
Aiming at the problems of low recognition accuracy,slow response speed and high labor cost of the traditional substation equipment maintenance mode,this paper proposes an scheme of abnormal warning of substation equipment based on sound recognition technology,which uses the key technologies such as adaptive noise suppression algorithm,wavelet transformation feature extraction,support vector machine classification and fuzzy logic early warning decision.Validation at a 500 kV substation shows that the method significantly improves anomaly recognition accuracy and early-warning response speed compared with conventional approaches,enhancing the intelligence of substation equipment anomaly management.
作者
夏梦超
姬改改
XIA Mengchao;JI Gaigai(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144-146,共3页
Audio Engineering
关键词
声音识别技术
变电设备异常预警
支持向量机分类
sound recognition technology
substation equipment anomaly early warning
support vector machine classification
