Research Seminar- Dr Chuanbin Zhu (Northumbria University)

Talk Overview:

Earth’s near-surface structure (upper few hundred meters of the crust) significantly modifies the earthquake waves passing through it, despite that it rarely represents 1% of the whole path (from an emanating source to an observation site). This modification is referred to as “site effects”. Though the salient physics governing site effects is relatively well understood, its accurate and precise prediction remains elusive using conventional approaches, e.g., physics-based numerical simulations, and classic regression-based empirical models. Machine learning (ML), including deep learning (DL), has been proven to have substantial potential in tackling thorny problems in many scientific disciplines, but its adoption in site effects modelling lags many other subfields. In this talk, I will present the early attempts in the application of ML/DL in site effects characterisation, and its promises and challenges, as well as the potential path forward.

Speaker Bio:

Chuanbin is an Assistant Professor in Geotechnical Engineering in the Faculty of Engineering and Environment, Northumbria University. He earned his PhD degree in 2018 at Queensland University of Technology, Australia, and then held research and/or academic post at the Helmholtz Center Potsdam - GFZ German Research Centre for Geosciences, Germany, Kyoto University, Japan, and the University of Canterbury, New Zealand.

His research aims to improve the hazard and risk evaluations for natural perils, especially earthquakes. He contributed to the Horizon 2020 program “SERA: Seismology and Earthquake Engineering Research Infrastructure Alliance for Europe” which leads to a revised European seismic hazard reference model, and to the New Zealand National Seismic Hazard Model project which re-defines the seismic loading standard in NZ.