Featured in Technology Today (Volume 37, Issue 4), a quarterly publication of the Louisiana Transportation Research Center.
South Louisiana is renowned for its unique and picturesque landscape, with numerous bayous, marshes, and other waterways and wetlands. These beautiful features, however, also necessitate the widespread use of piles, deep foundations located well beneath the surface, as the state’s DOTD continues to develop its transportation infrastructure. Testing and analyzing the load capacity of these piles is an extensive and expensive—but also essential—process. As a result, researchers often utilize alternate testing methods, including cone-penetration tests (CPTs), to obtain a faster and more accurate estimate of pile capacity. LTRC researcher Murad Abu-Farsakh, Ph.D., P.E., recently completed a multi-year project evaluating 21 such pile-CPT methods to identify and integrate the best of these methods into Louisiana’s ongoing pile-analysis efforts.
Dr. Abu-Farsakh and his team’s extensive efforts, outlined in a sweeping report entitled “Update the Pile Design by CPT Software to Incorporate Newly Developed Pile-CPT Methods and Other Design Features,” included a comparison of 21 pile-CPT methods with existing data available from 80 traditional pile load tests. The team’s goal in this evaluation was to rank, identify, and select the most accurate and effective methods to implement in Louisiana’s Pile-Design Cone Penetration Test (LPD-CPT) software, updating the software to reflect the latest innovations and incorporate several new testing features. Researchers utilized a combination of mathematical and statistical analyses, Multi-Dimensional Unfolding (MDU) evaluation, and reliability analysis to develop a comprehensive understanding of the best methodologies currently available to DOTD engineers.
The results of these tests were illuminating, enabling Dr. Abu-Farsakh and his team to identify the highest-performing pile-CPT methods and use them to update the LPD-CPT software accordingly. All three evaluation methods—mathematical/statistical, MDU analysis, and reliability analysis—yielded consistent results, elevating researchers’ confidence in their discoveries. Further, the team experimented with several new artificial intelligence (AI) and machine learning (ML) techniques, including artificial neural networks (ANN) and gradient-boosted trees (GBT), to pave the way for even greater testing effectiveness in the future. Based on the positive results of these efforts, Dr. Abu-Farsakh recommends that DOTD compare the AI/ML techniques to direct pile-CPT methods, exploring the potential future benefits of their increased utilization.
South Louisiana’s marshy soils promise to continue challenging the state’s transportation officials to innovate and implement the safest, most durable solutions as they construct and update roads, bridges, and more. Because of the thorough research of Dr. Abu-Farsakh and his team, DOTD engineers are well-equipped to respond to these challenges using the latest and most effective pile-CPT methods, with even further advancement on the horizon as AI/ML technologies continue to mature.
For more information on this project and its findings, please visit LTRC Final Report 682. You can also contact Murad Abu-Farsakh, Ph.D., P.E., at (225) 767-9147 or cefars@lsu.edu.