Authors: Hector, D.D., Udom, G.J., Giadom, F.D. and Tse, A.C.
Abstract
The integration of artificial intelligence (AI) techniques with Geoscientific data has revolutionized the creation of 3D Engineering Geological Models for subsurface soil profiles. ChatGPT-4 is a large multimodal model AI tool, it accepts both text and image inputs, that allows its users to specify the type of language task, by combining natural language understanding with visual information GPT-4 generates accurate and nuanced 3D models, aiding in proper foundation design, groundwater flow modeling, and environmental management. Its reliability, creativity and ability
to handle complex design makes it a valuable tool for geotechnical applications. In this paper, ChatGPT-4 was used to create 3D Engineering Geological Models of subsurface soil profiles in parts of Port Harcourt Metropolis. Data, including coordinates and static water levels from five geotechnical boreholes drilled to depth of 30 meters were used as input parameters. The elevation of the area ranged from 10 m to 7 m above sea level. The soil profile consists of 3 meters thick medium grained sand succeeded by 10 meters thick clay of low plasticity which in turn, is underlain by 17 meters of thick poorly graded sand (SP). A 3D soil profile of the study area was created showing the elevations, topography and weak stratigraphic layers. This model enables engineers and researchers to optimize infrastructural
design, reduce costs and mitigate risk.