Automatic Construction of Three-Dimensional Ground Model by Data Processing

Abstract

There have been attempts to improve the operational efficiency of construction projects and plan response countermeasures for estimated damage following disasters through the utilization of accumulated electronic data, which constructs a digital twin that can reproduce a physical space in cyberspace and feedback the cyberspace simulation results to physical space. However, the application of such simulations is limited, unless numerical models can be automatically constructed from the data. In this study, we develop a program that utilizes a data processing platform to read, transform, and integrate data to create mediated data with a common data structure. This mediated data can be used to construct analytical models for various numerical analyses. Using the program developed, a grid model of 3-D ground surface as the mediated data was constructed based on borehole data obtained through a ground survey. Each grid point has basic material parameters of soil, and these parameters are estimated from borehole data and other investigation reports. Each grid point has general geotechnical parameters and can be easily converted to a 3-D finite-element model. When the borehole data is added or changed, the analytical model can be updated with almost no cost, whereas it would be very costly to create the model manually.