| 基于EKF训练的归一化RBF神经网络在旋转机械故障诊断中的应用 |
| |
| 引用本文： | 张雷,胡彦红,陈巍巍,刘秋鞍,林建中,张丽芳.基于EKF训练的归一化RBF神经网络在旋转机械故障诊断中的应用[J].中国工程机械学报,2010,8(1):86-90. |
| |
| 作者姓名： | 张雷 胡彦红 陈巍巍 刘秋鞍 林建中 张丽芳 |
| |
| 作者单位： | 上海理工大学,机械工程学院,上海,200093 |
| |
| 基金项目： | 上海市研究生创新基金资助项目 |
| |
| 摘 要： | 在径向基函数(Radial Basis Function,RBF)神经网络成熟的基础上,对旋转机械的转子系统进行故障诊断,针对梯度下降法容易产生梯度消失的问题,提出用扩展卡尔曼滤波器(Extended Kalman Filter,EKF)对权重进行调节训练,并将结果与反向传播(Back Propagation,BP)算法和梯度下降调节进行比较,用EKF训练的RBF神经网络不仅在性能上有优势,在精度和迭代速度上亦优于其他方法.相信在今后的实际应用中尤其在旋转机械故障诊断中可以更大地发挥其优势.
|
| 关 键 词： | 径向基函数神经网络 旋转机械 梯度下降法 扩展卡尔曼滤波器 |
Applying EKF-trained normalized RBF neural network for fault diagnosis on rotary machinery |
| |
| ZHANG Lei,HU Yan-hong,CHEN Wei-wei,LIU Qiu-ling,LIN Jian-zhong,ZHANG Li-fang.Applying EKF-trained normalized RBF neural network for fault diagnosis on rotary machinery[J].Chinese Journal of Construction Machinery,2010,8(1):86-90. |
| |
| Authors: | ZHANG Lei HU Yan-hong CHEN Wei-wei LIU Qiu-ling LIN Jian-zhong ZHANG Li-fang |
| |
| Affiliation: | 1. Department of Mechanical Engineering,University of Shanghai for Science and Technology, Shanghai 200093, China) |
| |
| Abstract: | By applying the radial basis function （RBF） neural network, a fault diagnosis is proposed for the rotor system of rotary machinery. Based on the understanding that the potential gradient disappearance via descending gradient method, the Extended Kalman Filter （EKF） is employed to adjust the trained weights. In comparison with the backpropagation and descending gradient techniques, the EKF-trained RBF network possesses such advantages as high performance, precision and iterative speed. Therefore, this approach could secure a crucial position in industrial applications,particularly for rotary machinery. |
| |
| Keywords: | radial basis function neural network rotary machinery descending gradient Extended Kalman Filter |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
