径流长期预报的人工神经网络方法

Long-Range Runoff Forecast by Artificial Neural Network

提出径流长期预报的人工神经网络方法．运用神经网络的一典型模型──“反向传播”模型，以大伙房水库在补水期的径流状况作为研究对象，尝试了神经网络方法的效果。结果表明，该方法预报成功率较高，容错能力较强，可望成为径流长期预报的有效的辅助手段。

The Artificial Neural Network(ANN )method for long-range runoff forecast was dresented in this paper.With the application of one of the typical artificial neural network models'back-propagation'model,we selected the runoff status of Da Huofang Reservoir as an object of study to test the performance of the ANN mehtod approach.As a result,the successful rate is higher and the fault-tolerant ability is stronger.Therefore the ANN method might be referred as an effective assistant technique for long-range runoff forecast.