摘要
针对船舶制冷压缩机运行故障自动检测中的误检、错检问题,提出基于支持向量机的自动检测方法。该方法利用小波变换对压缩机振动信号进行多尺度分解,提取频域信号熵作为故障特征参量,以表征非平稳信号中的故障信息。在此基础上,构建支持向量机分类模型,基于结构风险最小化原则实现故障状态的自适应识别。实验结果表明,该方法在多种工况下均能获得较高的故障识别精度,误检率与漏检率均低于1%,优于传统智能诊断方法,具有较高的工程应用价值。
To address the issues of false and missed detections in automatic fault diagnosis of marine refrigeration compressors,an automatic detection method based on support vector machines is proposed.The compressor vibration signals are decomposed into multiple scales using wavelet transform,and the frequency-domain signal entropy is extracted as a fault characteristic parameter to effectively represent fault information in non-stationary signals.On this basis,an support vector machines classification model is constructed to achieve adaptive fault state identification based on the principle of structural risk minimization.Experimental results demonstrate that this method achieves high fault recognition accuracy under various operating conditions,with false detection and missed detection rates both below 1%,outperforming traditional intelligent diagnostic methods and demonstrating high engineering application value.
作者
徐海青
李伟
XU Haiqing;LI Wei(School of Marine Engineering,Jiangxi Vocational Technical University,Jiujiang 332000)
出处
《现代制造技术与装备》
2026年第1期185-187,共3页
Modern Manufacturing Technology and Equipment
基金
江西省教育厅科学技术研究项目“船舶制冷装置干燥系统的应用与研究”(GJJ2204807)。
关键词
支持向量机
制冷压缩机
小波变换
support vector machine
refrigeration compressor
wavelet transform
