Accurate prediction of consolidation settlement has been one of the most important geotechnical issues in the land development with reclamation. The soil properties in reclamation ground should be estimated properly in order to predict the consolidation settlement accurately. However, they are varied spatially due to effect of their sedimentary condition. Then, the soil properties obtained through soil investigations have various errors which cannot be eliminated. Also, the number of soil investigations is not enough to estimate properly them in many cases. In this paper, the soil properties are estimated properly from the limited number of soil investigations with an artificial neural network. Also, the estimation error in soil properties and its effect on estimation of consolidation settlement is discussed stochastically. Firstly, an artificial neural network is applied to estimate spatially soil properties, including relationship between void ratio and consolidation pressure of the Holocene clay in Kobe Airport. Secondary, it is elucidated that the estimation accuracy of both void ratio and relationship between void ratio and consolidation pressure is higher than those of compression index and pre-consolidation pressure thought discussions about relative errors. Finally, the Monte Carlo simulations to consolidation settlement is carried out in order to estimate stochastically the consolidation settlement. The consolidation settlement applied practically might be estimated about 1.1 times of that estimated with an artificial neural network judging from the probability.
[1]
Wataru Shiraki,et al.
An Application of Self-Organizing Feature Map to Spatial Estimation Problem for Geotechnical Characterization.
,
2000
.
[2]
Xiping Wu,et al.
Use of neural networks in the analysis and interpretation of site investigation data
,
1994
.
[3]
Spatial Interpolation of consolidation properties of Holocene clays at Kobe Airport using an artificial neural network
,
2013
.
[4]
C. H. Juang,et al.
Three-dimensional site characterisation: neural network approach
,
2001
.
[5]
K. Xiao,et al.
Research on Spatial Estimation of Soil Property Based on Improved RBF Neural Network
,
2013
.
[6]
Kai Hao,et al.
Re search on Spatial Estimation of Soil Property Based on Improved RBF Neural Network
,
2013
.