To solve the problem that the mode mixing phenomenon of empirical mode decomposition (EMD) in processing the non-linear and non-stationary signal and the anti-noise performance of traditional Prony method is not insufficient,on this basis,this paper presented a new method of harmonic detection based on complete ensemble empirical mode decomposition (CEEMD) and the improved Prony method.Firstly,the harmonic signal with noise is decomposed into intrinsic mode function (IMF) by CEEMD.And then,the obtained IMF component carries on threshold de-noising based on energy threshold method.Given the original Prony method,the solving process of the actual order and the linear prediction parameters are optimized.At last,the harmonic components after de-noising are detected and analyzed by improved Prony method.The MATLAB simulation experiment and the measurement data of three-phase uncontrollable load verify that this method can not only accurately identify the characteristic parameters of harmonics,but also has good anti-noise performance,which provides a new way for the harmonic analysis.%为了解决经验模态分解(EMD)处理非线性非平稳信号的模态混叠现象以及传统Prony方法抗噪性能差的不足.提出基于完全经验模态分解(CEEMD)和改进Prony相结合的谐波分析方法.先将含噪的谐波信号进行CEEMD分解得到固有模态函数,再通过能量门限法对分解的固有模态函数进行阈值去噪.针对原始Prony方法优化了实际阶数和线性预测参数的求解过程.最后将去噪后的谐波分量进行改进Prony变换检测分析.并通过MATLAB仿真实验以及依托三相不可控负载的实测数据验证了该方法能够精确辨识出谐波分量的特征参数且具有良好的抗噪性能,从而为谐波分析提供了一种新思路.
展开▼
