Study on parameters of cone penetration test considering data pre-processing and feature selection
Determination of the parameters of soil is of great significance for the design and construction of geotechnical engineering.The cone penetration test (CPT)is one of the most widely used in-situ testing methods for marine soil.The cone tip resistance is measured by CPT and can be transformed to the various parameters of marine soil.In this paper,the cone tip resistance obtained from CPT survey is taken as the research object.By analyzing the characteristics of spatial location,overburden pressure and porosity,a machine learning model is established using the XGBoost algorithm.By considering data preprocessing and feature selection in the modeling process,the prediction performance of the model is improved.The data preprocessing methods include box plot analysis and noise smoothing.A marine soil parameter prediction model based on feature engineering and XGBoost is proposed to predict the cone tip resistance and this method is applied to a practical engineering project in Zhejiang Province.The research results show that the proposed model has a correlation coefficient of 0.951 between predicted values and measured values,indicating high prediction accuracy.Boxplot analysis and noise smoothing can effectively improve the prediction accuracy.When the soil porosity is used as an input,the prediction accuracy can further be improved.
cone penetration testcone tip resistancedata preprocessingXGBoostfeature selectionsoil porosity
万方数据
维普
