A data-driven approach for forecasting embankment settlement accounting for multi-stage construction. (December 2022)

• Record Type: Journal Article
• Authors: Wan, Xiao; Doherty, James
• Is Part Of: Computers and geotechnics. Volume 152(2022)
• Publication Date: 2022-12
• Subjects: Embankment -- Soft soil -- Consolidation -- Data-driven approach; Engineering geology -- Data processing -- Periodicals; Soil mechanics -- Data processing -- Periodicals; Rock mechanics -- Data processing -- Periodicals; 624.1510285
• DOI: 10.1016/j.compgeo.2022.105001
• Languages: English
• ISSNs: 0266-352X

Abstract

Abstract: Reliable forecasts of the behaviour of embankments on soft soil are important for safe and efficient infrastructure design. Despite the development of advanced models in the past few decades, recent soft soil embankment prediction exercises revealed that forecasts from these models can be poor due to the subjective judgement required in the parameter selection process. Data-driven approaches, which connect measured soil data directly to a forecasting model, aim to overcome this by eliminating or reducing the subjectivity in assigning parameters. This paper builds on an existing data-driven approach for forecasting the behaviour of embankments on soft soil by including the ability to account for staged embankment construction. The multi-stage method was found to provide a significantly improved match to measured embankment settlements during and shortly after construction. The long-term settlements are similar to the single stage approach and both methods eventually converge to the same settlement-time response.