灾害天气条件下电力系统故障诊断特征匹配方法

A Feature Matching Based Method for Fault Diagnosis in Power Systems Under Disaster Weather

灾害天气会导致电力通信系统的可靠性降低,警报信息在通信过程中发生畸变的可能性上升,进而增加对停电区域内元件准确识别故障的困难。现有研究表明,通过对线路发生故障期间的外部环境数据进行分析可得到该线路的故障风险程度,但此信息尚很少被用于故障诊断之中。在此背景下,提出一种利用外部天气环境信息进行故障特征匹配的电力系统故障诊断方法。首先,利用故障期间采集得到的外部环境信息根据灰色模糊理论得到线路综合故障概率。然后,在现有的电力系统故障诊断解析模型(优化模型)的目标函数中增加外部环境故障特征匹配指标,对与灾害天气相关故障特征不匹配的故障假说进行惩罚。接着,利用模拟退火遗传算法对改进的解析模型进行求解。最后,针对广州市局部电网的故障案例进行仿真测试,结果表明所提的改进故障诊断方法对通信故障的容错性强;此外,还比较分析了考虑与不考虑外部天气环境影响时的诊断结果及差异原因。

Reliability of the communication system in a power system decreases under disaster weather,and then the possibility of communication distortions of alarm messages rises. As a result,it will become more difficult to accurately identify faults in the outage area（ s）. The existing research has demonstrated that the fault risk degree of a transmission line can be obtained by analyzing the external environmental data during the line fault. How ever,this information source is rarely used in existing fault diagnosis methods. Given this background,a power system fault diagnosis method based on fault feature matching using external weather environment information is proposed. First,the fault probability of a transmission line is attained by employing the grey fuzzy theory on the basis of the external weather information collected during a fault.Then,the matching index of the fault feature related to the external weather condition is added to the objective function of the existing analytical model（ optimization model） for fault diagnosis,and the fault hypotheses which do not match the corresponding fault features will be punished. Next,the simulated annealing genetic algorithm is used to solve the improved analytical model. Finally,the proposed method is demonstrated by employing fault scenarios in a regional power system in Guangzhou. Simulation results show that the improved fault diagnosis method is more fault-tolerant to communication failures. In addition,the fault diagnosis outcomes with and without employing external weather information are compared and the reasons for their differences analyzed.