| 联想记忆神经网络在高压输电线路故障诊断的应用(英) |
| |
| 引用本文： | 姜惠兰,孙雅明.联想记忆神经网络在高压输电线路故障诊断的应用(英)[J].天津大学学报(英文版),1999(1). |
| |
| 作者姓名： | 姜惠兰 孙雅明 |
| |
| 作者单位： | 天津大学电气自动化与能源工程学院 |
| |
| 摘 要： | 建造了实现高压输电线路故障诊断的两种联想记忆神经网络的模型结构并提出了提高其容错能力的有效方法．对于FNN，提出“伪吸引域”的概念以人为扩大样本吸收域，在不降低网络效率的前提下大大提高了网络的容错性能；对于BAM－NN，提出增加冗余神经元来提高网络的记忆能力和容错性能，但该措施增加了网络的复杂度，降低了网络的效率．仿真结果表明，本文构造的两种联想记忆NN都能正确识别故障，且对干扰输入信息序列有相当的容错能力，而FNN则是解决电力系统故障诊断问题的一个更加简便而且有效的方法．
|
| 关 键 词： | 神经网络 电力系统 故障诊断 容错性 |
APPLICATION OF ASSOCIATIVE MEMORY NEURAL NETWORK IN HIGH VOLTAGE TRANSMISSIONLINE FAULT DIAGNOSIS |
| |
| JIANG Huilan , SUN Yaming.APPLICATION OF ASSOCIATIVE MEMORY NEURAL NETWORK IN HIGH VOLTAGE TRANSMISSIONLINE FAULT DIAGNOSIS[J].Transactions of Tianjin University,1999(1). |
| |
| Authors: | JIANG Huilan SUN Yaming |
| |
| Abstract: | Effective methods of enhancing the fault-tolerance property are proposed for two kinds of associative memory (AM) neural network (NN) used in high voltage transmission line fault diagnosis. For feedforward NN (FNN),the conception of "fake attaction region" is presented to expand the attraction region artificially,and for the feedback Hopfield bidirectional AM NN (BAM-NN),the measure to add redundant neurons is taken to enhance NN's memory capacity and fault-tolerance property. Study results show that the NNs built not only can complete fault diagnosis correctly but also have fairly high fault-tolerance ability for disturbed input information sequence. Moreover FNN is a more convenient and effective method of solving the problem of power system fault diagnosis. |
| |
| Keywords: | neural network power system fault diagnosis fault-tolerance property |
| 本文献已被 CNKI 等数据库收录！ |
