Pattern recognition of mechanical failure based on Kohonen network
Article Text (Baidu Translation)
-
摘要: 在自动聚类分析的基础上, 根据某汽轮机减速箱运行状态特征数据, 采用Kohonen网络方法, 确定该机械的运行状态, 结果与选用其它方法所确定的结果一致。该方法能实现对汽轮机减速箱运行状态的监控Abstract: The application of Kohonen network to the diagnosing of mechanical failure was studied. The results show that the Kohonen network can discribe the machine's running through testing the characteristic data of decelerating box. Compared with other methods, it is pointed out that this method is reasonable, and this method can monitor and control the decelerating box running.
-
Key words:
- Kohonen network /
- pattern recognition /
- decelerating box of steamship /
- running
-
表 1 减速箱运行状态特征数据
Table 1. Character data of decelerating box in running state
样本号 样本输入特征 故障模式 1 -1.7817 -0.2786 -0.2954 -0.2394 -0.1842 -0.1572 -0.1584 -0.1998 1.0 0.5 0 2 -1.8710 -0.2957 -0.3494 -0.2904 -0.1460 -0.1387 -0.1492 -0.2228 1.0 0.5 0 3 -1.8347 -0.2817 -0.3566 0.3476 -0.1820 -0.1435 -0.1778 -0.1849 1.0 0.5 0 4 -1.4151 -0.2282 -0.2124 -0.2147 -0.1271 -0.0680 -0.0872 -0.1684 0.5 1.0 0.5 5 -1.8809 -0.2467 -0.2316 -0.2419 -0.1938 -0.2103 -0.2010 -0.2533 1.0 0.5 0 6 -1.2879 -0.2252 -0.2012 -0.1298 -0.0245 -0.0390 -0.0762 -0.1672 0.5 1.0 0.5 7 -1.5239 -0.1979 -0.1094 -0.1402 -0.0994 -0.1394 -0.1673 -0.2810 0.5 1.0 0.5 8 -1.4087 -0.2773 -0.2759 -0.2181 -0.0575 -0.0829 -0.0592 -0.1240 0.5 1.0 0.5 9 -0.5147 -0.1839 -0.1432 -0.0694 0.0285 0.0991 0.1326 0.0592 0 0.5 1.0 10 0.2741 0.1442 0.1916 0.1662 0.2120 0.1631 0.0318 0.0337 0 0.5 1.0 11 0.2045 0.1078 0.2246 0.2031 0.2428 0.2050 0.0704 0.0403 0 0.5 1.0 12 0.1605 -0.0920 -0.0160 0.1246 0.1802 0.2087 0.2234 0.1003 0 0.5 1.0 13 -0.7915 -0.1018 -0.0737 -0.0945 -0.0955 0.0044 0.0467 0.0719 0 0.5 1.0 14 -1.0242 -0.1461 -0.1018 -0.0778 -0.0363 -0.0476 0.0160 -0.0253 0 0.5 1.0 
下载: 导出CSV
表 2 训练好后网络的连接权
Table 2. Network's connecting power after training
输入层节点 竞争层节点 1 2 3 1 -1.771 1 -1.354 2 -0.880 4 2 -0.2838 -0.2317 -0.1612 3 -0.3262 -0.2374 -0.1290 4 -0.2842 -0.1994 -0.1011 5 -0.1565 -0.0960 -0.0410 6 -0.1380 -0.0736 0.0002 7 -0.1488 -0.0572 0.0439 8 -0.1907 -0.0980 0.0113
下载: 导出CSV
表 3 第二组样本的网络识别结果
Table 3. Outcome of network recognice of the secondly stylebook
样本号 样本的网络识别输出结果 期望输出 4 0.5000 1.0000 0.5000 0.5 1.0 0.5 5 1.0000 0.5000 0 1.0 0.5 0 6 0.5000 1.0000 0.5000 0.5 1.0 0.5 7 0.5000 1.0000 0.5000 0.5 1.0 0.5 10 0 0.5000 1.0000 0 0.5 1.0 11 0 0.5000 1.0000 0 0.5 1.0 12 0 0.5000 1.0000 0 0.5 1.0
下载: 导出CSV
-
[1] 闻新, 周露, 王丹力, 等. MATLAB神经网络应用设计[M]. 北京: 科学出版社, 2000. [2] HANSEN L K, SALAMON P. Neural network ensembles[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990, 12(10): 993-1001. [3] 温熙森, 胡茑庆, 邱静. 模式识别与状态监控[M]. 北京: 国防科技大学出版社, 1997. [4] PANDYA AS, MACY R B. Pattern recognition with neural network in C++. IEEE Press(中译本)[M]. 北京: 电子工业出版社, 1999. [5] 张立明. 人工神经网络的模型及其应用[M]. 上海: 复旦大学出版社, 1992. [6] 王文成. 神经网络及其在汽车工程中的应用[M]. 北京: 北京理工大学出版社, 1998. [7] CARMAN C S, MERICKEL M B. Supervising isodata with an information theoretic stopping rule[J]. Pattern Recognition, 1990, 23(1/2): 185-197. [8] BRYANT J. A fast classifier for image data[J]. Pattern Recognition, 1989, 22: 45-48. doi: 10.1016/0031-3203(89)90037-X -
点击查看大图
图(2) / 表(3)
计量
- 文章访问数: 320
- HTML全文浏览量: 103
- PDF下载量: 276
- 被引次数: 0
