【摘要】 雾霾影响下绝缘子的闪络电压变低,更易发生污闪,威胁电力系统的稳定运行。雾霾中含有不同的离子,故在高压实验室模拟了不同盐分的绝缘子污闪。以闪络电压为因变量,以所加电压、加压时间、泄露电流为自变量,分别建立3种污闪的多元线性回归模型和广义线性回归模型。对比各个模型,共提取出7个故障特征,其中5个可用于判断故障类型。据此设计的判别方法是用所测数据建模,按5条判据逐条判断,以归属次数多的那一种为最终判断结果。经过验证,这种方法的判断结果与实际一致,说明它有较好的故障诊断性能。更多还原

【Abstract】 Under the influence of haze, the flashover voltage of insulators becomes lower, and pollution flashover is more likely to occur, which threatens the stable operation of power system. Fog and haze contain different ions, so the pollution flashover of insulators with different salts has been simulated in high voltage laboratory. Multiple linear regression model and generalized linear regression model of three types of pollution flashovers are established separately by using the flashover voltage as the dependent variable while the applied voltage, the time and the leakage current as independent variables. 7 fault features have been extracted by comparing these models and 5 of them can be criteria for fault diagnosis. The discrimination method is establishing models of the measured data, judging one by one based on five criteria and using the one with more number of attributions as the final result. Its judgment result is consistent with the actual situation after verification, indicating that it has a good fault diagnosis ability.更多还原