Data-Driven Reliability-Based Design Methods for Piles in Glacial Deposits

点击翻译按钮获取中文摘要

This proposed research will conduct first of its kind and comprehensive investigation on a large amount of full-scale pile load tests in Ontario soils using innovative artificial intelligence (AI) techniques and develop reliability-based design methods for piles in glacial deposits. Although piles have been used for centuries to support structures in Canada and around the world, challenges still exist today to accurately predict their capacities and settlements due to many uncertainties, in particular ground conditions. Glacial deposits covering most of Canada are well known for their heterogeneous material properties. Thus, it is difficult to accurately characterize the soil properties for optimum designs. Due to these uncertainties, expensive and time-consuming pile load tests are commonly used in practice to verify designs in local ground conditions. Despite plentiful research on pile load tests around the world, a limited number of research has been conducted on piles in glacial deposits, particularly in Canadian soils. For this research, a database of over 200 pile load tests was established over the last few years with accompanying geotechnical reports. These load tests will be back-analyzed using innovative and efficient AI techniques to correlate pile behavior with soil measurements. Based on these AI-driven back-analyses, innovative and reliable reliability-based design methods will be developed to assist engineers to mitigate the risks associated with inconsistent soil conditions. The research findings will result in cost-effective infrastructure developments in Canada. The proposed research is also vital in training HQP for their career development in Canada.