【摘要】 现有化学检测法检测SF6电气设备放电故障时存在滞后性且耗时耗力,因此提出应用紫外光谱快速在线鉴别设备中的SO2组分,实现放电故障在线监测。利用快速Fourier变换分析、自相关分析、功率谱密度分析对209~219 nm及295~305 nm 2个特征区数据进行处理,选取相关峰数目、相邻峰间距离的期望和标准差、以及快速Fourier变换和功率谱密度的最大峰值处频域分量5个特征值,采用表决原理实现设备内微量SO2组分判别。通过SO2标准气体检测实验,确定特征值阈值,证实提出方法对SO2体积分数的检测限<10-6。通过现场测量实际运行且疑似存在放电故障的SF6电流互感器内气体,证实提出的方法可在线、有效地发现微量SO2特征,且该设备内确已发生放电,判断结果与SF6综合测试仪测试结果一致。更多还原

【Abstract】 Present chemical detection methods by which it takes a long time and is hard to detect discharge failures in SF6-insulated electrical equipment are outdated. Thus, we proposed to identify the presence of SO2 in gas insulated switchgear(GIS) by using ultraviolet spectroscopy(UV) so as to realize online monitoring discharges in SF6 electrical equipment. Moreover, we used the fast Fourier transform(FFT), autocorrelation analysis, and power spectral density(PSD) to process the data of two feature spectroscopy areas, 209~219 nm and 295~305 nm, and chose five characteristic values, namely, the number of correlation peaks, the expectation and standard deviation of the distance between adjacent peaks, the peak frequency components of FFT and PSD, to determine whether SO2 exists based on the voting principle. Through a designed experiment using standard SO2 gas, we defined the thresholds of the characteristic values, and proved that the minimum detectable value was below 10-6. Meanwhile, we tested an operating current transformer with suspect discharge failure in field. The results prove that the method is valid for detecting SO2 trace, and confirm that the suspect discharge failure is in accordance with the result obtained by a comprehensive SF6 test instrument.更多还原