| 基于改进BP网络对三门峡水库泥沙冲淤量的计算 |
| |
| 引用本文： | 练继建 刘媛媛 胡明罡 张金良. 基于改进BP网络对三门峡水库泥沙冲淤量的计算[J]. 天津大学学报(自然科学与工程技术版), 2004, 37(10): 882-885 |
| |
| 作者姓名： | 练继建 刘媛媛 胡明罡 张金良 |
| |
| 作者单位： | 天津大学建筑工程学院,天津300072 |
| |
| 基金项目： | 国家自然科学基金资助项目(59979020). |
| |
| 摘 要： | 人工神经网络具有很好的分布存储和容错性，适合解决非线性问题．在分析了汛期、非汛期水库泥沙冲淤影响因子的基础上。利用改进的BP网络模型对水库泥沙冲淤量进行计算．网络训练时，非汛期采用动量法和学习律自适应调整策略，拟合误差较小，平均相对误差约为0．10；汛期采用Levenberg-Marquardt优化方法，由于非线性关系复杂及人为因素多。误差相对较大，利用该模型预测不同水库运行条件下泥沙淤积量，计算量小，使用方便．
|
| 关 键 词： | 改进BP网络模型 泥沙冲淤 三门峡水库 |
| 文章编号： | 0493-2137(2004)10-0882-04 |
| 修稿时间： | 2003-04-14 |
Calculation of the Scour and Sediment of Sanmenxia Reservoir Based on Improved BP Network |
| |
| LIAN Ji-jian,LIU Yuan-yuan,HU Ming-gang,ZHANG Jin-liang. Calculation of the Scour and Sediment of Sanmenxia Reservoir Based on Improved BP Network[J]. Journal of Tianjin University(Science and Technology), 2004, 37(10): 882-885 |
| |
| Authors: | LIAN Ji-jian LIU Yuan-yuan HU Ming-gang ZHANG Jin-liang |
| |
| Abstract: | Artificial Neural Network is suitable for solving non-linear problems based on its good performances on distributed storage and error toleration.In this paper,an improved BP network is used to calculate the scour and sediment of reservoir on the basis of analyzing influence factor of flood and dry season.During training the network,for dry season cases,momentum method and auto-adjustive learning rate are used, fitting errors are quite small, and average relative errors are nearly 0.01;as for flood season cases,Levenberg-Marquardt optimizing method is used and errors are relatively large because of complex non-linear relations and artificial factors.Numerical results show that above models can be used to estimate the reservoir sediment and the calculation is more convenient. |
| |
| Keywords: | improved BP network scour and sediment Sanmenxia Reservoir |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
