智能水文预报模型的研究与应用

Research and application of intelligent hydrological forecasting model

洪水过程具有高度非线性、复杂性和非平稳性特征。将自适应步长的布谷鸟搜索(ASCS)算法应用于神经网络水文模型参数优化中,构建ASCS-LSTM洪水预报模型,并采用注意力机制进一步提高输入输出的相关性,实现高精度的智能洪水预测。在秦淮河流域的水位预测实验表明,ASCS-LSTM预报模型的预报结果要优于传统机器学习模型,稳定性和精确度得到提升,可为水文预报提供新思路。

The flood process is highly nonlinear,complex and non-stationary.The ASCS-LSTM flood forecasting model is built,which adopts adaptive step size cuckoo search(ASCS)algorithm to optimize parameter of LSTM neural network hydrological model,and which applies the attention mechanism to further improve the relevance of input and output to achieve high-precision intelligent flood prediction.The water level prediction experiments in Qinhuai River basin show that ASCS-LSTM forecasting model can achieve more stable and accurate forecasting results than those of traditional machine learning model,and thus providing a new idea for hydrological forecasting.