抑制周期性窄带干扰的时域重构优化算法性能分析

Analysis of the Optimization Reconstruction Method in Time Domain for Periodic Narrow-band Interference Eliminating

为更好地检测电力设备早期绝缘放电信号,结合遗传优化算法和傅立叶级数提出了一种改进的时域重构窄带干扰抑制方法。它通过遗传算法优化选择窄带干扰的重构参数,然后基于傅立叶级数对其进行时域重构,并通过与实测信号相减,获得待检测放电信号。此方法的干扰抑制效果与放电信号的频率分布及其幅值大小无关,仅与窄带干扰的重构精度有关。理论分析和仿真计算结果表明,该方法在窄带干扰频率分量个数及其频率值存在估计误差时,依然可以获得较高的窄带干扰重构精度,并具有较强的抗随机干扰能力。实测数据处理结果也验证了此方法的可行性。

A new improved method of narrowband interference suppression in time-domain based on genetic algorithm（GA） and the Fourier series was proposed.In this method,the genetic algorithm was used to optimize the reconstruction parameters of narrow-band interference,so that the narrow-band interference could be reconstructed in time-domain based on the Fourier series.So the discharge signal can be detected with the measured signal subtracting the reconstructed narrow-band interference.The effect of interference suppression by this method does not depend on the spectrum and amplitude of the discharge signal,but relate to the accuracy of the reconstructed narrow-band interference.The results of theoretical analysis and simulation show that the method can get a good interference suppression effect although there are estimation errors in the narrow-band interference frequency components and its value,and has a good ability of anti-random interference in the same time.Measured data processing results also verify the feasibility of this method.