Keynote Lecture: Assessing the Potential of CFA Piling Rig Telemetry Data for Stratigraphic Profiling and Optimised Pile Installation-PPT Presentation

PPT Presentation

Pdf Summary

The presentation evaluates whether continuous-flight-auger (CFA) piling rig telemetry can improve ground profiling and pile installation. Rig measurements provide high-resolution time-series data that may complement conventional site investigation, which can be limited and lead to conservative designs.<br /><br />A workflow was developed to collect and clean telemetry, label data, select features, train supervised and unsupervised models, detect material boundaries, handle anomalies, smooth predictions spatially, and validate results with rig operators and engineers. The study used data from two UK projects: 1,977 piles and 537,723 records. Four geotechnical designers labelled pile data using a web app, taking about 20 seconds per pile.<br /><br />Models classified ground materials by depth and helped identify boundaries such as the top of rock. Reported classification accuracy reached 98.2% for a supervised model; an unsupervised approach achieved 92.6%. Spatial smoothing and statistical methods were used to improve boundary estimates and flag anomalous predictions.<br /><br />Applied retrospectively to a historical project of around 2,000 piles, the approach indicated potential savings of 11.2 drilling hours (7.1%), 442.6 m³ of concrete (5.5%), and 154 tonnes of CO₂e. However, accuracy fell by 10–15% when models were applied across projects, highlighting the need for site-specific calibration and further testing.<br /><br />Wider implementation must account for differences in rig capacity, operators, ground conditions, and drilling tools. Future work aims to combine telemetry with pile load tests, live monitoring, geotechnical design parameters, and site-investigation data to improve ground and pile–soil understanding and support more efficient foundation designs.

Keywords

continuous-flight-auger

CFA piling

rig telemetry

ground profiling

pile installation

geotechnical data

machine learning

material boundaries

supervised classification

ground investigation