| 神经网络在洪水实时预报中的应用研究 |
| |
| 引用本文： | 熊立华,郭生练,王元.神经网络在洪水实时预报中的应用研究[J].水电能源科学,2002,20(3):28-31. |
| |
| 作者姓名： | 熊立华 郭生练 王元 |
| |
| 作者单位： | 1. 武汉大学,水利水电学院,湖北,武汉,430072
2. 郑州黄河工程公司,河南,郑州,450008 |
| |
| 基金项目： | 国家重点基础研究发展规划“973”项目 ( G1 990 4 36) |
| |
| 摘 要： | 建立了一种基于神经网络的洪水实时预报模型。运用向后演算法，该模型的权重系数可以在每一时间步长上进行自动更新，较好地反映了实际水文过程和参数的时变性；由于该模型不再需要单独的误差序列实时校正模型，因而更加简洁。最后利用淮河鲇鱼山水库1975－1999年的小时降雨和入库洪水资料对模型参数进行了率定和校核。结果表明，洪水实时预报的效率系数超过96％，洪峰值合格率为92．5％，峰现时间误差都在1h以内。
|
| 关 键 词： | 神经网络 洪水实时预报 实时校正 向后演算法 神经网络 |
| 文章编号： | 1000-7709(2002)03-0028-04 |
| 修稿时间： | 2002年1月7日 |
Study and Application of Artificial Neural Network in Real-time Flood Forecasting |
| |
| XIONG Li-hua,GUO Sheng-lian,WANG Yuan.Study and Application of Artificial Neural Network in Real-time Flood Forecasting[J].International Journal Hydroelectric Energy,2002,20(3):28-31. |
| |
| Authors: | XIONG Li-hua GUO Sheng-lian WANG Yuan |
| |
| Affiliation: | XIONG Li-hua 1 GUO Sheng-lian 1 WANG Yuan 2 |
| |
| Abstract: | Based on the artificial neural network(ANN), a real-time flood forecasting model is proposed. According to the back-propagation method,the weighted coefficients of the ANN are updated at each time step,reflecting the time-varying characteristics of the hydrological processes in nature.Also,without an independent time-series model for forcasting the simulation errors,the model is very flexible and simple. The model was tested by using the hourly data of rainfall and floods at Nianyushan reservoir. The model efficiency exceeded 96% during both calibration and verification periods. The correctness flood peaks value is 92.5%. The errors between the forecasted and observed flood peak time were always within one hour. |
| |
| Keywords: | hydrological model flood forecasting real-time correction back-propagation method neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
