Researchers have developed and validated a ConvLSTM framework designed to predict retaining wall deformation during staged excavation. This framework, trained on simulated data augmented with Gaussian noise, integrates multiple temporal resolutions via a stacking ensemble. Field data from 34 inclinometers across 11 South Korean excavation sites demonstrated the framework's effectiveness, achieving an average mean absolute error of 1.4 mm and a coefficient of determination of 0.93 for predicting deformations up to 5.0 m of excavation. AI
IMPACT This framework offers a novel approach to geotechnical engineering, potentially improving safety and efficiency in construction projects by accurately predicting structural deformations.
RANK_REASON Academic paper detailing a new framework and its field validation. [lever_c_demoted from research: ic=1 ai=0.7]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →