基于竞争神经网络的变电站集中监控全面监视周期分类方法

Comprehensive Monitoring Cycle Classification Method of Centralized Monitoring in Substation Based on Competitive Neural Network

摘要:

对变电站集中监控全面监视周期的影响因素进行了分析,包括变电站电压等级、集中监控缺陷发生频次、设备运行情况、变电站重要程度、变电站运行环境和设备平均负载6个方面。在此基础上提出了基于竞争神经网络的变电站全面监视周期分类方法,利用机器学习手段实现变电站全面监视周期的科学合理分类。仿真分析结果表明该方法可行有效。

Abstract:

An experimental research platform for the impact of mountain fire on safe operation of 10 k V and 35 k V transmission lines is designed and established. Pine,straw and shrubs are burnt to simulate the mountain fire fault of 10 kV and 35 kV transmission lines,and the variation characteristics of three-phase voltage,conductor to ground leakage current and zero-sequence voltage of transmission lines under mountain fire are studied. The effects of flame height and flame width on the insulation of 10 k V transmission line are studied and analyzed too. The test results show that mountain fire can cause single-phase high resistance grounding and phase to phase short circuit of transmission line,which are the results of the joint action of voltage level,distance to ground,flame height,flame width and other factors,and whether the flame bridges the air gap is an important factor affecting air insulation. The phenomena that the transmission line discharges to the flame and the voltage of one phase affected by the flame is lower than that of other phases during the test are studied,and the causes and laws of mountain fire on 10 k V and 35 k V transmission lines are obtained.