ASCE has honored Chao Shi, Ph.D., C.Eng, and Yu-Hsing Wang, Ph.D., P.E., F.ASCE, with the 2023 Thomas A. Middlebrooks Award for the paper “Development of Subsurface Geological Cross-Section from Limited Site-Specific Boreholes and Prior Geological Knowledge Using Iterative Convolution XGBoost,” Journal of Geotechnical and Geoenvironmental Engineering, September 2021.
Shi and Wang tackle a long-standing geotechnical challenge of subsurface stratigraphy from limited boreholes and develop an innovative image-based machine learning algorithm, called Iterative Convolution eXtreme Gradient (IC-XGBoost), which can automatically develop and update subsurface geological cross sections from sparse site-specific data and a single training image reflecting prior geological knowledge at the site of interest.
IC-XGBoost can automatically produce high-resolution subsurface stratigraphy using limited site-specific boreholes and prior geological knowledge as inputs. It is a game changer that revolutionizes the current geotechnical practice of manually developing subsurface stratigraphy, which is often highly simplified with a loss of many valuable details from the available measurements, with automatic development of digital underground models, based on which geotechnical analysis models (e.g., finite element models) can also be developed automatically. It will bring about a paradigm shift in subsurface stratigraphy and greatly improve both the accuracy and efficiency of subsequent geotechnical analysis and design that use the subsurface stratigraphy as inputs. It also paves the way to digital twins of geotechnical structures in the coming era of digital intelligence and industry 4.0.
The Thomas A. Middlebrooks Award is given to the author or authors of a paper published by the Society that is judged worthy of special commendation for its merit as a contribution to geotechnical engineering. Papers written by young engineers are given preference.