Research on Karst Groundwater Level Prediction Based on SSA-RBF Neural Network Model
In this paper,an improved RBF neural network time series prediction model based on the Singular Spectrum Analysis( SSA) method was established,and the real-time monitoring data of groundwater level was denoised by SSA to extract the water level change trend,and then the radial basis( RBF) neural network time series model was used to predict the denoised data. The karst water level data from 2020 to 2021 in Wuhan were selected for processing and analysis,and the results showed that:①the prediction accuracy of the improved RBF neu-ral network model based on SSA was 50％ higher than that of the RBF neural network model alone;②The SSA-RBF neural network prediction model performed well in the prediction of karst groundwater level,and the maximum prediction error of karst groundwater level in the next 24 hours was not more than 0.2m,and the prediction accuracy reached more than 99％.
NETL
NSTL
万方数据
