论文部分内容阅读
针对采集回的变压器局部放电的混合信号,先分别对混合信号和其时移信号进行主成分分析后,进行频谱分析,消除主成分分量的周期性干扰噪声成分,然后重构消除周期性干扰后的数据,对重构的数据进行典型相关分析的盲源分离,针对混有残余的随机脉冲干扰的混合信号再进行典型相关分析的迭代分离,直至提取出变压器局部放电脉冲信号。仿真结果表明,该方法能有效的分离出变压器局部放电脉冲信号。
After the principal component analysis of the mixed signal and its time-shifted signal is carried out, the spectrum analysis is performed on the mixed signal of the partial discharge of the transformer which is collected back to eliminate the periodic interference noise component of the principal component component, and then to reconstruct and eliminate the periodic interference Then the blind source separation of the reconstructed data was analyzed by canonical correlation analysis. The mixed signal with the residual random impulse interference was iteratively separated from the canonical correlation analysis until the partial discharge pulse signal of the transformer was extracted. Simulation results show that this method can effectively separate the partial discharge pulse signal of transformer.