| 多重遗传神经网络在微量特征气体临界值识别中的应用 |
| |
| 引用本文： | 李昕,张勇,刘君华,吴浩扬.多重遗传神经网络在微量特征气体临界值识别中的应用[J].西安交通大学学报,2001,34(4):335-337,342. |
| |
| 作者姓名： | 李昕 张勇 刘君华 吴浩扬 |
| |
| 作者单位： | 西安交通大学电气工程学院, |
| |
| 基金项目： | 国家自然科学基金资助项目(69676004)；国家教委博士点基金资助项目(980699828)． |
| |
| 摘 要： | 将气敏元件阵列技术和遗传神经网络相结合，检测了电力变压器油中的4种微量故障特征气体(1×10
|
| 关 键 词： | 电力变压器 故障特征气体 临界值识别 故障诊断 遗传神经网络 气敏元件阵列技术 |
| 文章编号： | 0253－987X(2001)04-0335-04 |
Multiple Neural Network
with Genetic Algorithm for Critical Pattern Recognition of Trace Quantity Gas |
| |
| Li Xin,Zhang Yong,Liu Junhua,Wu Haoyang.Multiple Neural Network
with Genetic Algorithm for Critical Pattern Recognition of Trace Quantity Gas[J].Journal of Xi'an Jiaotong University,2001,34(4):335-337,342. |
| |
| Authors: | Li Xin Zhang Yong Liu Junhua Wu Haoyang |
| |
| Abstract: | Critical pattern recognition of trace gas by multiple neural network with genetic algorithm is presented. The trace gas concentration was measured by the above method， for example，1～70 parts per million by volume of hydrogen, or acetylene or ethene, or 50～550 parts per million of carbon monoxide. The single network can recognize gas species in large range, but cannot acquire the precise output at critical value. Since the concentration threshold of failure characteristic gas in transformer oil is very important to early failure diagnosis, the measurement precision of it should be improved. A new method of multiple neural network with genetic algorithm is presented, that can keep the recognition range and also improve the precision. |
| |
| Keywords: | neural network genetic algorithm transformer characteristic gas |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
