To solve the problem of land resources,the rapid development of Shenzhen was accompanied by reclamation projects.The marine silt layer in the reclamation area is a highly saturated soft soil with high compressibility.The effect of upper loads and the self-weight consolidation of backfill can easily lead to the occurrence of geological hazards of ground settlement,and the prediction study of ground settlement problems in reclamation areas has become crucial.Taking the monitoring data from May 2019 to April 2022 in the Airport New City area as an example,various mathematical models were selected to predict the ground subsidence,including the NAR neural network model,the GM (1,1 )model,the polynomial regression model,and the ARIMA model.The prediction shows that all four models can better reflect the future development trend of the cumulative settlement,which can provide a basic basis for the prevention and control program of ground settlement and foundation treatment in the later stage,and effectively reduce the risk of geological disasters of ground settlement.Combining the stability and accuracy of their prediction results and other factors,the advantages and disadvantages of the four models can be ranked in order of the ARIMA model,NAR neural network model,polynomial regression model,and GM (1,1)model.
关键词
地面沉降/数据驱动模型/预测/填海区
Key words
land subsidence/data-driven model/prediction/reclamation
维普
万方数据