基于傅里叶变换的GIS局部放电信号处理研究

Signal processing of the GIS' partial discharge based on Fourier transform

当气体绝缘组合开关设备(GIS)内部发生局部放电(PD)时,检测到的信号可能存在周期窄带干扰,对传统的FFT阈值置零滤波方法进行了模拟,对它在抑制窄带干扰上的不足,可能的原因进行了分析,并在此基础上对算法提出了改进。选取合适的压缩倍数,对窄带干扰所在的频带进行分块阈值压缩处理,并对周围的频带利用插值补偿或最小二乘法进行了加宽处理,利用Matlab仿真,通过对不同的压缩倍数、窗函数和消噪效果进行对比分析,得到最合适的压缩处理方法,并验证了其有效性,结果显示这种方法可以快速有效地抑制窄带干扰。

When partial discharge （PD） exsit in gas insulated switchgear （GIS）, the narrowband interference may be detected. The traditional FFT threshold denoising method is simulated. The possible reasons for its deficiencies in the suppression of narrowband interference are analyzed, and improvement is made about the algorithm. By selecting the suitable compression ratios, the frequency band where narrowband interference exist is block threshold compressed, and the frequency band around is widening by interpolation compensation or the least square algorithm. By using Matlab, the most suitable compression method is obtained after comparing different compression ratioes, window functions and denoising effects. Test results show the effectiveness of the method on the suppression on the narrowband interference.