The impact of liquefaction and the need for its accurate prediction cannot be overemphasized. While
traditional physics-based liquefaction prediction models offer valuable insights into soil behavior, they
often face challenges in capturing the full complexity under dynamic loading conditions. Conversely, data-
driven machine learning (ML) techniques excel in pattern recognition and can improve liquefaction
prediction. However, robust ML models with high prediction accuracies usually lack physical
interpretability. This study used cone penetration tests (CPT) data to develop robust ML models to
improve liquefaction prediction accuracy while leveraging the interpretability a simple ML model in
surrogate modeling. The developed models had significantly higher prediction accuracy than existing
empirical methods and can be applied for liquefaction assessment where standard CPT data are
available. More accurate prediction of liquefaction susceptibility can lead to efficient and cost-effective
implementation of liquefaction mitigation practices<\/p>\n\n\n\n
Participation in the upcoming WAIMM webinar series is free and could be accessed from the details below:<\/p>\n\n\n\n
Topic: Machine Learning for Enhanced Liquefaction Prediction<\/strong><\/p>\n\n\n\n Date: Thursday,<\/strong> 13th March 2025<\/strong><\/p>\n\n\n\n Time: 14:30 HRS (GMT)<\/strong><\/p>\n\n\n\n Speaker: Dr Prosper Ayawah<\/strong>, Geotechnical Engineer – STANTEC, Denver – USA<\/strong><\/p>\n\n\n\n Venue: Online – Zoom<\/strong><\/p>\n\n\n\n Register in advance for this Zoom webinar using the link https:\/\/us02web.zoom.us\/meeting\/register\/i8WkM33ORC6QCiX0ESizIQ<\/a> Meeting ID: 898 2103 7791
<\/p>\n\n\n\n
Passcode: 251172<\/p>\n\n\n\n