珠江河口网河区水位变化过程的神经网络模型

Artificial Neural Networks in the Study for the Stage Fluctuation Process in Pearl River Estuary

珠江三角洲河口地区的水位变化同时受到洪水与潮水的影响，河网各水道的相互作用使得水位变化更加复杂和具有非线性的特点。人工神经网络模型具有生物神经网络的某些特征，具有学习能力，易于应用于不同类型的水文系统，尤其是在河网与水道不断变化的情况下，模型具有不断学习和自动调整的能力。初步探讨表明该模型能较有效地模拟本质为非线性的河口网河地区实际水文过程。

In the Pearl River Delta, the fluctuation in stage of the estuarine areas is connected with both flood and tide. The interaction of channels in the waterway network causes the fluctuation more complicated and nonlinear. The artificial neural network model has human-like performance (e. g. it can ‘learn’), hence it can be easily adapted to different types of river systems. The model possesses the function of self-studying and self-adjusting especially in the changing of level of waterway network and channels.The model is a highly nonlinear dynamic system which can efficiently simulate the nonlinear process of the real hydrologic variables (tide, discharge) in river or estuarine environment.