Application of Artificial Neural Network in Runoff Prediction:A Case Study of Lanzhou Hydrological Station in Yellow River Basin
Runoff prediction is crucial for hydropower generation,reservoir scheduling,flood forecasting and other fields,so the establishment of an efficient and high-precision model is the focus of research in the field.Aiming at the nonlinear function characteristics of runoff,this paper proposes a BP neural network model based on PCA to reduce the dimensionality of input variables and genetic algorithm to optimise the initial parameters.Firstly,PCA principal component analysis is applied to reduce the dimensionality of the input variables to ensure the maximum response to the characteristics of the variables and improve the network speed to achieve the goal of high efficiency.Secondly,genetic algorithm is used to optimise the initial weight threshold of the network by taking the absolute value of the error between the predicted value and the real value as the fitness function,so as to achieve the goal of accuracy optimisation.Finally,the data are divided for training and testing.The model prediction effect is obtained according to the evaluation index.This paper takes the monthly runoff of Lanzhou station in the Yellow River Basin as the research object for data training and prediction,and optimises the model establishment by dividing the runoff size into flat water year and abundant water year.The test results show that compared with the traditional BP model and the models proposed by other researchers such as Support Vector Machine,Random Forest,Adaboost Regressor,etc.,the PCA-GA-BP neural network model proposed in this paper has higher prediction accuracy on the two types of runoff datasets,faster training speed,and is a powerful tool for river runoff prediction.
万方数据
维普
