论文部分内容阅读
利用金属磁记忆(MMM)技术进行故障检测时,较弱的故障信号提取成为检测准确度高低的关键。采用小波分析和奇异值分解相结合的方法,对金属磁记忆信号经行故障特征提取。通过小波分析将故障信号分解为不同尺度的分量,以形成初始向量特征矩阵,并对该矩阵进行奇异值分解,选择代表特征信号的奇异值分量重构,从而实现对故障信号的特征提取。经过实验证明该方法有效。
When using metal magnetic memory (MMM) for fault detection, the weak fault signal extraction becomes the key to the detection accuracy. Using the combination of wavelet analysis and singular value decomposition, the metal magnetic memory signal is extracted through the fault feature. Wavelet analysis of the fault signal is decomposed into different scales of components to form the initial vector feature matrix, and the matrix singular value decomposition, select the representative of the singular value of the signal reconstruction component, in order to achieve fault signal feature extraction. After experiments show that the method is effective.