Neural Network Model Of Phase Space And Its Application In Hydrologic Mid-And-Long Term Prediction

With the frequent flood occuring and the fast economic developing in China, the flood control departments have much higher demands in watershed for the leading time and forecasting precision of flood and water resources, therefore the study of mid - and - long term runoff prediction is paid more and more attention by researchers, and it is also the most difficult problem which people are trying their best to solve. Because the hydrologic system is a complicated huge system, there exist high non - linear characteristics in the space - time change of hydrologic factors. According to theory of chaotic phase space, the paper established a mid - and - long term runoff prediction model based on the chaotic phase space and neural network. The model is applied in the long term runoff prediction of Baishan reservoir. The results of calculation show that the model has stronger non - linear mapping function and much more information in the time series than traditional ways, it is highly effective and is worthy of being popularized and applied. It is reasonable and superior to use the model in mid - and - long term hydrologic prediction.