Automatic classification of landslide kinematics using acoustic emission measurements and machine learning

Founded on understanding of a slope's likely failure mechanism, an early warning system for instability should alert users of accelerating slope deformation behaviour to enable safety-critical decisions to be made. Acoustic emission (AE) monitoring of active waveguides (i.e. a steel tube with granular internal/external backfill installed through a slope) is becoming an accepted monitoring technology for soil slope stability applications; however, challenges still exist to develop widely applicable AE interpretation strategies. The objective of this study was to develop and demonstrate the use of machine learning (ML) approaches to automatically classify landslide kinematics using AE measurements, based on the standard landslide velocity scale. Datasets from large-scale slope failure simulation experiments were used to train and test the ML models. In addition, an example field application using data from a reactivated landslide at Hollin Hill, North Yorkshire, UK, is presented. The results show that ML can automatically classify landslide kinematics using AE measurements with the accuracy of more than 90%. The combination of two AE features, AE rate and AE rate gradient, enable both velocity and acceleration classifications. A conceptual framework is presented for how this automatic approach would be used for landslide early warning in the field, with considerations given to potentially limited site-specific training data.

Automated summary

This study successfully developed and demonstrated the feasibility of automatically classifying landslide kinematics using acoustic emission (AE) measurements with machine learning (ML) methods. ML models were trained and tested using data, achieving an accuracy of over 90%, and a potential field application framework was proposed.