【摘要】 提出一种基于核密度估计独立分量分析(KDICA)的间谐波分层提取方法。对信号时间序列的Hankel矩阵进行奇异值分解(SVD),利用奇异值的加权能量贡献率确定谐波个数。采用快速傅里叶变换(FFT)、最小二乘法(LS)并结合Teager能量算子,确定基波、谐波与间谐波的相关参数,构造合适的虚拟观测信号,采用KDICA算法,分层提取基波、谐波及间谐波信号。在实时性要求不高和缺少先验知识(如频率和相位)的情况下,能够较高精度地检测出小幅值间谐波成分。仿真结果验证了方法的有效性。更多还原

【Abstract】 An inter-harmonic multi-extracting method based on kernel density independent component analysis(KDICA) is proposed in this paper.The Hankel matrix is reconstructed based on time series,then according to the singular values’ percent of contribution to total energy(PCTE),the number of harmonics are known.The method combines FFT,least square method and Teager energy operator to determine the fundamental and harmonic and inter-harmonic parameters.And it constructs the appropriate virtual observed signals and hierarchically extracts the fundamental and harmonic and inter-harmonic with the help of KDICA.In the case of less real-time demand,the method can accurately detect the small amplitude inter-harmonic without prior knowledge.The proceeding result of simulation model proves its effectiveness.更多还原