基于BP人工神经网络的绝缘子泄漏电流预测
发布时间：2018-12-16 18:51
【摘要】：外绝缘泄漏电流和绝缘子的污闪过程存在密切的对应关系,相对湿度是影响泄漏电流产生和发展的关键因素之一。首先对两种常用悬式绝缘子进行人工污秽试验,利用泄漏电流测量系统记录其在运行电压作用下、不同相对湿度时的泄漏电流波形并对此进行分析;采用BP人工神经网络的方法,建立了不同湿度下的泄漏电流最大值之间的对应关系;选择Levenberg-Marquardt快速学习算法对建立的BP神经网络进行训练。利用部分试验数据进行的验证以及采用不同方法获得的泄漏电流随相对湿度变化曲线均表明,使用该神经网络来建立不同表面受潮状态时泄漏电流最大值之间的关系是准确有效的。
[Abstract]:There is a close relationship between the leakage current of external insulation and the flashover process of insulator, and the relative humidity is one of the key factors affecting the generation and development of leakage current. Firstly, two kinds of commonly used suspended insulators were tested by artificial pollution test. The leakage current waveform of two kinds of commonly used insulators under different relative humidity was recorded and analyzed by the leakage current measurement system. The corresponding relationship between the maximum leakage current under different humidity is established by using BP artificial neural network, and the Levenberg-Marquardt fast learning algorithm is selected to train the established BP neural network. The verification of partial test data and the curve of leakage current with relative humidity obtained by different methods show that, Using the neural network to establish the relationship between the maximum leakage current in different surface damp states is accurate and effective.
【作者单位】： 清华大学电机系 清华大学深圳研究生院 清华大学深圳研究生院 清华大学深圳研究生院
【基金】：国家自然科学基金项目(50377020)。~~
【分类号】：TP183;TM853
[Abstract]:There is a close relationship between the leakage current of external insulation and the flashover process of insulator, and the relative humidity is one of the key factors affecting the generation and development of leakage current. Firstly, two kinds of commonly used suspended insulators were tested by artificial pollution test. The leakage current waveform of two kinds of commonly used insulators under different relative humidity was recorded and analyzed by the leakage current measurement system. The corresponding relationship between the maximum leakage current under different humidity is established by using BP artificial neural network, and the Levenberg-Marquardt fast learning algorithm is selected to train the established BP neural network. The verification of partial test data and the curve of leakage current with relative humidity obtained by different methods show that, Using the neural network to establish the relationship between the maximum leakage current in different surface damp states is accurate and effective.
【作者单位】： 清华大学电机系 清华大学深圳研究生院 清华大学深圳研究生院 清华大学深圳研究生院
【基金】：国家自然科学基金项目(50377020)。~~
【分类号】：TP183;TM853
【参考文献】
相关期刊论文 前10条
1 关志成;张仁豫;薛家麒;马霜;李楠;;自然污秽可溶盐构成及其对污闪电压值的影响[J];电瓷避雷器;1989年06期
2 李武峰;范建斌;李鹏;宿志一;;直流绝缘子串污秽闪络特性研究[J];电网技术;2006年15期
3 关志成,王绍武,梁曦东,王黎明,范炬;我国电力系统绝缘子污闪事故及其对策[J];高电压技术;2000年06期
4 吴光亚,蔡炜,卢燕龙,肖勇,赵全江,薛春林;直流输电线路绝缘子串片数的防污设计[J];高电压技术;2001年06期
5 李治,王永江,王黎明,关志成;绝缘子泄漏电流和放电现象的初步研究[J];高电压技术;2002年S1期
6 张寒,文习山,丁辉;用神经网络预测基于气象因素的绝缘子等值附盐密度[J];高压电器;2003年06期
7 聂一雄,尹项根,刘春,文远芳;用模糊逻辑方法对绝缘子串在线检测结果的评定[J];中国电机工程学报;2003年03期
8 乐波,谢恒X;基于模糊输出BP神经网络的电机主绝缘老化状态评估方法[J];中国电机工程学报;2005年02期
9 杨庆,司马文霞,蒋兴良,白困利,梅冰笑;复杂环境条件下绝缘子闪络电压预测神经网络模型的建立及应用[J];中国电机工程学报;2005年13期
10 焦尚彬;刘丁;郑岗;张青;;基于最小二乘支持向量机的绝缘子等值附盐密度预测[J];中国电机工程学报;2006年02期
【共引文献】
相关期刊论文 前10条
1 张之f，
本文编号：2382873
本文链接：https://www.wllwen.com/kejilunwen/rengongzhinen/2382873.html
