| 神经网络在变压器超高频局部放电模式识别中的应用 |
| |
| 引用本文： | 殷录民,张慧芬. 神经网络在变压器超高频局部放电模式识别中的应用[J]. 山东大学学报(工学版), 2004, 34(3): 51-54 |
| |
| 作者姓名： | 殷录民 张慧芬 |
| |
| 作者单位： | 1. 山东大学,后勤处,山东,济南,250061
2. 济南大学,信息科学与工程学院,山东,济南,250022 |
| |
| 摘 要： | 结合自适应遗传算法 (AGA)和BP算法各自的优点 ,构造了AGA BP混合算法作为神经网络的学习算法 .实验结果表明 ,AGA BP神经网络既解决了BP神经网络对初始权值敏感和容易局部收敛的问题 ,又提高了AGA神经网络的收敛速度、稳定性和求解质量 ,具有较高的识别率和较强的推广能力
|
| 关 键 词： | 超高频局部放电 模式识别 神经网络 |
| 文章编号： | 1672-3961(2004)02-0051-04 |
| 修稿时间： | 2003-11-24 |
Application of neural network in UHF PD pattern recognition of transformers |
| |
| YIN Lu-min ,ZHANG Hui-fen. Application of neural network in UHF PD pattern recognition of transformers[J]. Journal of Shandong University of Technology, 2004, 34(3): 51-54 |
| |
| Authors: | YIN Lu-min ZHANG Hui-fen |
| |
| Affiliation: | YIN Lu-min 1,ZHANG Hui-fen 2 |
| |
| Abstract: | By combining Adaptive Genetic Algorithm(AGA) with Back-Propagation(BP) algorithm, an AGA-BP hybrid algorithm is presented to train network. Test results that the proposed AGA-BP-NN algorithm overcomes the sensitivity of BP-NN to initial weight and local convergence, and improves the convergence rate, stability and solution quality of AGA-NN. As a result, AGA-BP-NN algorithm is possessed of higher recognition rates and generalization ability. |
| |
| Keywords: | ultra-high-frequency PD detection pattern recognition neural network(NN) |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
