RBF neural network medium-and-long-term runoff prediction method based on runoff production mechanism

Abstract

A RBF neural network medium-and-long-term runoff prediction method based on a runoff production mechanism comprises the following steps: exploring a main physical factor set influencing monthly runoffformation from the physical process of runoff formation; taking seasonal difference of runoff formation into consideration, and respectively screening the physical factor sets primarily selected in each month by adopting a stepwise regression analysis method to obtain a remarkable physical factor set influencing monthly runoff; establishing a monthly runoff forecasting RBF neural network model byadopting a Matlab language; optimizing the optimal combination parameters of the monthly runoff forecasting model; and evaluating the forecasting effect of the monthly runoff forecasting model. A significant factor set is screened from a runoff generation physical mechanism, so that the physical significance of the prediction model is effectively ensured, and the calculation stability of the prediction model is improved. According to the method, the runoff prediction effect of the model is quantitatively evaluated by adopting the relative error absolute average value and the relative error qualification rate.

Description

technical field [0001] The invention relates to hydrological forecasting technology, in particular to a medium and long-term runoff forecasting method for establishing a forecast model month by month in consideration of the physical cause of runoff and seasonal differences. Background technique [0002] Reliable medium- and long-term runoff forecasting is of great significance for the macro-control and micro-scheduling of water resources, ensuring the safety of water conservancy projects and the benefits of flood control, power generation, and shipping. In water resource system engineering, medium and long-term runoff forecasting has always been one of the hot spots and difficulties in research. From the current research, there is no forecast method that can be applied to all situations and watersheds, and it is still in continuous development and exploration. research stage. [0003] With the continuous development of computer technology, more and more advanced theories an...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem of medium and long-term runoff forecasting in the existing water resource system engineering only for the time series of runoff itself, while less consideration is given to the physical process affecting the formation of runoff, resulting in inaccurate forecasting models that consider the physical causes of runoff technical issues

Images