CEE Ph.D. Defense Announcement: Advancing the Design and Analysis of Drilled Displacement Piles Through Field Data and Machine Learning

Abstract: Drilled Displacement Piles (DDPs) are relatively new deep foundation elements increasingly used to support infrastructure projects. Their primary advantages include reduced noise, vibration and soil spoils during installation. Additionally, the installation process densifies the surrounding soil, enhancing its resistance and improving pile performance compared to other traditional systems. A review of more than 150 tested piles, and the collection of  high-quality pile-soil data, which included in-situ testing (CPT tests) and axial load results, showed that current empirical methods consistently underestimate DDP axial load capacity. To address this, correction factors derived through a Bayesian approach are proposed to improve the accuracy of current conventional models. More importantly, a new analytical method is introduced to estimate axial load capacity using CPT data, leveraging direct measurements of sleeve friction to reflect the shaft resistance mechanism more accurately. Lastly, pile installation parameters are utilized to develop a real-time, in-situ capacity prediction method. Together, these advancements offer a data-driven framework for performance-based DDP design, enhancing reliability and supporting modern geotechnical practice.