摘要
电力系统输电线路绝缘子运行中受到空气环境的污染,绝缘子表面污秽主要有大气中的盐密和灰密成分,不同地区不同环境下盐密和灰密对对绝缘子闪络电压影响具有非线性关系,为了较好辨识此非线性关系,运用基于GNBR算法的BP神经网络诊断该非线性关系,BP网络中采用GNBR算法,使网络具有很好的收敛性和很高的辨识精度,解决了BP网络局部最优、训练速度慢和辨识精度低的问题。实验结果表明设计的基于GNBR算法的BP神经网络能够很好地辨识绝缘子污秽对闪络电压的影响关系。
The insulator of the power system transmission line can be easily polluted by the air environment in operation and the contamination of the insulator surface mainly includes salt density and ash density of the atmosphere. The impact of the salt density and ash density on the insulator flashover voltage is nonlinear in different regions and different envir- onments. In order to have a better identification effect, the BP neural network based on the GNBR algorithm is used to diagnose the nonlinear relationship. The GNBR algorithm is used in the training of BP network so that the network has good convergence and high recognition accuracy. The designed method has better solved the problems with the BP network such as the local optimum, slow speed and low recognition accuracy. The experimental result shows that the designed BP neural network based on the GNBR algorithm has good identification ability to identify the relationship between the insulator contamination and the flashover voltage.
出处
《电网与清洁能源》
2013年第6期6-9,15,共5页
Power System and Clean Energy
基金
湖北省自然基金项目(2010CDB02503)
湖北省教育厅项目(Q20091406)~~
关键词
绝缘子污秽
诊断
闪络电压
BP网络
GNBR算法
insulator contamination
diagnosis
flashover voltage
BP neural networks
GNBR algorithm
