自适应精简经验Ramanujan分解及其在复合故障诊断中的应用

Adaptive Concise Empirical Ramanujan Decomposition and Its Application in Composite Fault Diagnosis

Ramanujan傅里叶模态分解采用低频向高频扫描的方式获取分量信号,易出现过量分解和信息分散的现象,致使分解分量不具有单一完整的状态信息.为了解决上述问题,论文提出了一种自适应精简经验Ramanujan分解(Adaptive Concise Empirical Ramanujan Decomposition,ACERD)方法.在ACERD方法中,采用功率谱密度获取分割频带,旨在进行准确的频带划分.同时,利用Ramanujan傅里叶变换提取每个分割频带所对应的模式分量,提高周期分量的识别能力,并获得具有单一周期特征信息的模式分量.通过复合故障仿真信号和实测信号分析,结果表明:ACERD方法具有优异的频带分割和周期脉冲特征提取能力,适用于复合故障诊断.

Ramanujan Fourier mode decomposition uses scanning from low frequency to high frequency to obtain component signals,which is prone to excessive decomposition and information dispersion,resulting in decomposed components not having a single and complete mode information.To address the above issues,this paper proposes an adaptive concise empirical Ramanujan decomposition(ACERD)method.In the ACERD method,the power spectral density is used to obtain the split frequency band for accurate frequency band division.Meanwhile,the Ramanujan Fourier transform is used to extract the mode components corresponding to each segmented frequency band,improve the recognition ability of periodic components,and obtain mode components with a single periodic feature information.The analysis results of composite fault simulation signals and measured signals indicate that the ACERD method has excellent capability of frequency band segmentation and periodic pulse feature extraction,which is suitable for compound fault diagnosis.