基于模糊神经网络的绝缘子表面污秽在线监测

在恶劣的自然环境下,积污绝缘子随时可能发生闪络。在实验室模拟试验和现场实测数据基础上,分析了环境因素对不同污秽程度绝缘子外部电气特性的影响。通过选择泄漏电流有效值、泄漏电流峰值及泄漏电流脉冲频次、环境温湿度等参量作为输入参数,提出采用模糊神经网络方法实现对输电线路绝缘子污秽状况在线监测结果的综合评定。介绍了基于模糊神经网络的污秽评定模型的构建过程,最后列举部分试验数据验证了该方法的可行性。

On-line Detecting of Insulator Surface Contamination Based on Fuzzy Neural Network

Polluted insulators may flashover at any time under disadvantage weather conditions. Based on laboratory simulation experiments and on-site data, the influence of environmental factors on the exterior electrical characteristics of insulators with different contamination conditions are analyzed. This paper introduces the principle of the integrative evaluation on the on-line detection results about the surface contamination states of the transmission line insulators by using fuzzy neural network method. The r.m.s value, the peak value, the amplitude and times of the pulses of leakage current, the environmeutal temperature and humidity are chosen as input variables. The model establishment of the contamination evaluation based on fuzzy neural network method is recommended emphatically, and the feasibility of method is proved by experiments in laboratory.