测量误差对接地网故障诊断影响的分析

为分析测量误差对接地网故障诊断结果的影响,将导体的腐蚀程度分为4个等级。对给定的测试方案,分别对接地网设置各种故障,并采用均匀分布的随机数反映测量误差的分布。在此基础上反复进行故障诊断计算,对诊断结果分析和统计,得出各个支路故障的诊断可信度、误判及漏判可能性和支路分辨率等指标。提出了根据提出的指标改进测试方案和算法以提高接地网故障诊断结果准确性的方法,并给出一实例说明方法的可行性,分析结果表明,提出的指标能反映测量误差对接地网故障诊断结果的影响,综合利用诊断指标提高了接地网故障诊断精度。

Influence of Measurement Error on Grounding Grids Diagnosis & its Applications

In order to conveniently analyze the influence of measurement error on grounding grids diagnosis results, the corrosion of a branch of a grounding grid is described by four degrees according to the extend of corrosion of the conductor. As for a certain test management, various damages are set on the grounding grid. Random numbers with evenly distributed probabilistic density are used to simulate the measuring error. On above bases, corrosion diagnosis is carried out repeatedly, The results of diagnosis are treated by a statistic method, based on which, the indexes of diagnosis belief level, probability of mis-finding, probability of dis-finding and branches identification level are obtained. The diagnosis belief level is defined as the ratio of correct detection times to the summation of correct detection times and mis-finding times. The probability of mis-finding is defined as the ratio of mis-finding times to the summation of correct detection times and mis-finding times. The probability of dis-finding is defined as the ratio of dis-finding times to the summation of correct detection times and dis-finding times. The branches identification level is defined as the ratio of the diagnosis belief level of one branch to the probability of mis-finding of the other branch. It is concluded that the diagnosis belief level and the branches identification level increase while the probability of mis-finding and the probability of dis-finding decrease with the measurement error goes down. The approach to improve the test management and the algorithm to obtain more accurate corrosion diagnosis results in accordance with the proposed four indexes are discussed. An example is given showing the feasibility of the proposed methods. It is concluded that the proposed indexes are helpful to evaluate the influence of measurement error on diagnosis results. The diagnosis results can be improved obviously by the proposed methodology.