Prediction for soil design properties based on a multivariate database for Shanghai soft clay

This study compiles a database for Shanghai soft clay labeled as SH-CLAY/12/4191 based on data collected from 51 investigation reports of geotechnical engineering in Shanghai. The database contains 4191 data points, each of which is composed of multivariate parameters measured simultaneously in close proximity at the same depth. There are five physical parameters and seven mechanical parameters. Then, a 12-dimensional multivariate probability distribution is constructed based on the compiled database. To make it more useful for engineering practice, the transformation models between the design parameters (i.e., K0, Es, c, and φ) and the index parameters as well as the in-situ test parameters are derived. The transformation models are explicit nonlinear formulas of multiple variables without cross-term. The transformation models are validated by the data from the new site, showing the generally reliable performance in predicting the design parameters.