Prediction of Salinity in Qiantang Estuary Based on Wavelet Neural Network Optimized by Particle Swarm Optimization

Abstract

Saltwater intrusion is a serious problem for most of macro-tidal estuaries, especially when it happened at the downstream region of drinking water sources. Therefore, accurate prediction of salinity variation becomes absolutely necessary. A three-layer salinity prediction model is established in this paper, based on wavelet neural network (WNN) optimized by particle swarm optimization (PSO). The optimized model overcomes the local minimum problem and accelerates the convergence speed of WNN method. Hourly observed salinity data at Ganpu station in Qiantang Estuary was used for case study. Additionally, we compare this model with another two models, WNN and BP, by predicting the salinity at Ganpu station with the same sample. The results show that PSO-WNN model has higher convergence rate and prediction accuracy.