用傅立叶级数法消除局部放电检测中窄带干扰的研究  

RESEARCH ON ELIMINATING NARROW FREQUENCY NOISE BY FOURIER SERIES IN PARTIAL DISCHARGE DETECTION

作　　者：程养春[1] 李成榕 王伟[3] 

机构地区：[1]华北电力大学电气工程学院,北京市昌平区102206 [2]高电压与电磁兼容北京市重点实验室,北京市昌平区102206 [3]电力系统保护与动态安全监控教育部重点实验室(华北电力大学),北京市昌平区102206

出　　处：《中国电机工程学报》2005年第20期106-111,共6页Proceedings of the CSEE

摘　　要：提出了一种新的抗窄带干扰方法——傅立叶级数法,利用原始数据通过傅立叶级数估计原始数据中的窄带干扰波形,然后与原始数据相减,以得到局部放电脉冲信号;通过仿真数据和现场实际数据验证了该方法的抗干扰效果,发现该方法能够在完全消除窄带干扰的同时,基本保留局部放电脉冲的波形、幅值和极性,尤其在脉冲信号的中心频谱与干扰的频谱非常接近时;讨论了随机干扰、频率误差和模数转换量化误差对抗干扰效果的影响,认为这些因素对抗干扰效果的影响不大;与常见的时频转换法和自适应滤波法进行了对比分析,结果表明傅立叶级数法在保持局部放电原始波形、幅值和极性信息方面具有明显优点。A new method of eliminating periodical narrow frequency noises in partial discharge detection of power apparatus, Fourier series method, is presented. It uses Fourier series of the original data to evaluate the noise, and subtracts the original data with the evaluated noise. So, the noise is eliminated but partial discharge pulses are retained. The anti-interference effect of this method is proved by simulation data and field test data. At the same time of eliminating noises, the shape, amplitude and polarity of pulse signals can be completely remained, even if the main frequency of partial discharge pulses are close to that of narrow frequency noises. The influence of random noise, frequency error and quantization error are also discussed. It is found that this influence is very limited. The effect of this method is also contrasted with time-frequency domain transform method and ~adaptive filter method. The results show that the new method has obvious predominance of keeping the shape, amplitude and polarity of partial discharges.

关 键 词：局部放电 窄带干扰 傅立叶级数 电力设备 

分 类 号：TM835[电气工程—高电压与绝缘技术]