Geotechnical Meetings@PoliTo

On Friday 21 April 2023, the meeting entitled Geotechnical Meetings@PoliTo, organised by the Department of Structural, Geotechnical and Building Engineering - DISEG of the Politecnico, will be held in the Aula Albenga. The speakers at this meeting will be Maria Giovanna Durante (graduated cum laude in Civil Engineering at the University of Sannio) and Paolo Zimmaro (Assistant Professor in Geotechnical Engineering at the University of Calabria).

Programme and Abstract

Maria Giovanna Durante - When Geotechnical Earthquake Engineering Meets Artificial Intelligence: The ReStructure 2.0 Project:

• "Standard seismic design of retaining structures in Europe is based on a century-old theory, that does not account for the actual physical behavior of soil-structure systems. Methods based on this theory, commonly referred as Mononobe-Okabe methods, often lead to conservative design of retaining structures that causes an unsustainable consumption of resources without any benefits on the performance and safety of the construction. Such design approach is against the principles of the European Green Deal that identified the need of cleaner constructions in the Building and Renovation policy area. The main goal of ReStructure 2.0 is to develop a novel physics-based framework based on soil-structure interaction principles, recognizing the relative displacement between wall and retained soil as the driving factor in the seismic response of wall-soil system. This novel approach is based on the combination of computational simulations, experimental and field data, relational databases, and artificial intelligence techniques. This more adequate design approach can lead to a significant reduction of the resources used during the construction, making the process more sustainable, affordable, and green. This presentation illustrates two important parts of ReStructure 2.0: 1) A simplified solution for the evaluation of the seismic increment of the earth pressure acting on a retaining structure; 2) Some preliminary results of the AI methods applied on the results of an extensive parametric analysis carried out in OpenSees. The simplified solution presented here has been recently adopted in the US National Earthquake Hazards Reduction Program (NEHRP) Recommended Seismic Provisions for New Buildings and Other Structures".

Paolo Zimmaro - Towards a new paradigm for the evaluation of liquefaction risk: the Next Generation Liquefaction Project:

• "In the last 50 years, the cost associated with damage produced by natural disasters increased by an order of magnitude, going from $14 billion in the 1976-1985 period to $140 billion in the 2005-2014 decade. A substantial portion of such damage is caused by earthquakes. Notably, earthquake-induced liquefaction is one of the dominant earthquake-related effect. Despite the relevance of liquefaction-related damage, current models to predict earthquake-induced liquefaction susceptibility, triggering, and related consequences are largely based on relatively data-scarce pre-2000 datasets and lack transparency in modeling choices. As a result, they are not ideal in accurately predicting liquefaction-induced effects. Since 2000 many earthquakes around the world generated hundreds of new liquefaction case histories. Furthermore, many of them were documented with modern reconnaissance tools and represent high-quality data points. To fill the existing gap in the liquefaction modeling state-of-the-art, in 2016 the Next-Generation Liquefaction (NGL) project was launched. This project has two main scopes: 1) Create a global, open-source, and transparent database of liquefaction case histories; 2) Use this data to generate novel semi-empirical, data-informed liquefaction susceptibility, triggering, and consequences models. This presentation will present the NGL database (Zimmaro et al., 2019 and Ulmer, Zimmaro et al., 2023), main features of recent liquefaction case histories, new techniques to analyze liquefaction data leveraging big data and advanced data analytics tools, and present, for the first time, the on-going development of a novel semi- empirical liquefaction prediction model. This model is being developed to distinguish explicitly between liquefaction triggering and surface manifestation of liquefaction within a Bayesian framework and will represent a transformational shift from years of past practices".

For more information contact andrea.ciancimino@polito.it.