摘要
针对同步发电机故障预测与健康管理(Prognostics and Health Management,PHM)系统故障特征提取困难,信号容易受到噪声干扰,诊断结果可靠性低的缺点,本文以故障率较高的轴承故障为例,提出以小波包熵值作为故障特征,提取轴承典型故障的振动信号。通过小波包分析,计算出不同故障、不同故障程度的小波包Shannon熵值。与正常轴承对比进行故障程度预测及故障定位。仿真结果表明小波包Shannon熵值能够清楚地反映出轴承故障程度及故障位置,该方法简单可靠,进行故障预测及诊断效果显著,克服了传统故障特征提取方法的不足。
Based on many deficiencies that exist in extracting fault characteristics in the synchronous motor PHM system, such as the signals are difficult to extract, are easily influenced by the variation of the loads, the paper introduces a method to extract the wavelet packet Shannon entropy as the broken character, using wavelet packet to analyze the oscillatory signals, and geting the wavelet Shannon entropy, through comparing to prognosticate the failure and aclfieve failure orientation. Experiment results in extracting the bearing fault of wavelet Shannon entropy shows this method useful, efficient, and easy to be used in engineering.
出处
《大电机技术》
北大核心
2014年第1期31-34,共4页
Large Electric Machine and Hydraulic Turbine