Flood and sediment prediction based on BP neural network

Knowledge of the impact factors and variation process of river sediments is the key to solve the increasingly serious river sediment problems. In this paper, flood and sediment prediction model was developed to forecast the sediment load based on artificial neural network, which generated promising results. The model was then applied to the Daling River in the northwest of Liaoning Province. First, the data from 29 historicalflood events from 1984 to 1998 were analyzed using the statistical method to obtain the main impact factors of downstream sediment load. Then, the BP neural network model was developed to character2 ize the relationship between the upstream impact factors and downstream sediment load. Finally, the data from six flood events were used to verify the model. The results showed that the errors between the calculated and measured values are within the reasonable range and meet the accuracy requirement, therefore the model is applicable for downstream sediment prediction in the Daling River.

Project Supported:

Clc Number: