Periodic Reporting for period 1 - ReStructure 2.0 (A novel physics-based methodology for the seismic analysis of retaining structures leveraging machine learning techniques)
Periodo di rendicontazione: 2021-09-01 al 2023-08-31
The overall cost associated to the construction of geotechnical structures (such as retaining and underground structures) in seismic areas is about 30% of the cost of the entire construction project for a building. This percentage is even higher for complex infrastructure systems such as sport arenas and multi-purpose complexes. According to a recent study coordinated by the United Nations Environment Programme, the construction industry generates roughly 40% of the worldwide energy and process-related carbon dioxide (CO2) emissions. Despite these impressive numbers, for the design of retaining structures, Europe-wide policies, guidelines, and building codes are still based on a century-old theory that unrealistically assumes that the seismic earth pressure increment is proportional to surface acceleration. Approaches based on this theory often lead to over-conservative design of retaining structures, causing an unsustainable consumption of resources without any benefits in terms of performance and safety of the construction. This makes this approach against the principles of the European Green Deal that identified the need of cleaner constructions in the Building and Renovation policy area.
A novel theoretical framework that I have been developing over the years during my research experiences in Italy, the UK, and the USA can solve this issue, making the design of geotechnical structures more technically sound and green. An initial application of this theory, that I developed alongside two USA-based colleagues, was recently added to the US National Earthquake Hazards Reduction Program seismic recommended provision. Such framework is based on robust soil-structure interaction principles. Its assumption is that seismic earth pressure on retaining structure does not have any fundamental relationship with the amplitude of the earthquake shaking. This quantity is instead mainly related to the amount of relative displacement between the structure and the retained soil. This is the theoretical foundation of the ReStructure 2.0 project. Its overall objective is the development of a novel approach to design more sustainable, affordable, and green retaining structures in seismic areas. This transformative method builds upon soil-structure interaction theories and leverages cloud-based high-performance computing capabilities, ad-hoc developed numerical simulations, data science, and artificial intelligence approaches.
A novel theoretical framework that I have been developing over the years during my research experiences in Italy, the UK, and the USA can solve this issue, making the design of geotechnical structures more technically sound and green. An initial application of this theory, that I developed alongside two USA-based colleagues, was recently added to the US National Earthquake Hazards Reduction Program seismic recommended provision. Such framework is based on robust soil-structure interaction principles. Its assumption is that seismic earth pressure on retaining structure does not have any fundamental relationship with the amplitude of the earthquake shaking. This quantity is instead mainly related to the amount of relative displacement between the structure and the retained soil. This is the theoretical foundation of the ReStructure 2.0 project. Its overall objective is the development of a novel approach to design more sustainable, affordable, and green retaining structures in seismic areas. This transformative method builds upon soil-structure interaction theories and leverages cloud-based high-performance computing capabilities, ad-hoc developed numerical simulations, data science, and artificial intelligence approaches.
During the two-year ReStructure 2.0 project I performed a large variety of technical activities, culminating in a novel, innovative, robust, and technically sound model to design retaining structures in seismic areas. This transformative method builds upon soil-structure interaction theories and leverages cloud-based high-performance computing capabilities, ad-hoc developed numerical simulations, data science, and artificial intelligence approaches. This model is based on roughly 115,000 numerical simulations of various retaining structures/soil property combinations (including structure height, soil type, and non-linearity characteristics) subjected to multiple earthquake recordings. These analyses were performed leveraging the computational capabilities of one of the most powerful supercomputer on earth, hosted at the University of Texas at Austin. To perform this large number of analyses, I designed a novel workflow that can be used in the future by other researchers, even in other fields of study. After the definition of the variables to be included in the model, I developed a parallel computing framework. I then performed all of the numerical simulations in the cloud, and automatically populated a relational database that stores all relevant results. This database is one of the most important contributions of this project as it can be leveraged in the future by the international scientific community to develop new models. After populating the database, I explored the importance of all input features leveraging a variety of machine learning models. I then selected the most promising input parameters and developed the final model that is based on a rationale and pragmatic simplified solution that can be used freely by engineers and scientists across the globe. All software tools used and/or developed in this study are open source and will be available to the public.
During the project I also implemented a dissemination plan targeting various audiences. To this end, I organized a total of 10 technical seminars/webinars part of the Restructure 2.0 series (attended by more than 400 people in total). I also organized, hosted, and/or participated in a number of public events, including the European Researchers’ night and meetings with high school students, to raise awareness on gender gaps and attempting at providing younger women with role models in STEM fields. The main event on this subject was organized as part of this project in the main campus auditorium and hosted Senator for Life and Professor Elena Cattaneo, a prominent public figure in the Italian and International research landscape. The event was attended by about 600 people.
The activities of the ReStructure 2.0 project attracted the attention of the nation-wide and international media. The project and my story were featured in a variety of articles and television interviews including an interview published by the national magazine “Il Venerdi di Repubblica” (https://www.repubblica.it/venerdi/2021/09/08/news/studiamo_i_terremoti_in_convento-315526765/(si apre in una nuova finestra)) an interview on the national magazine Prometeo, and an article on the national magazine “D di Repubblica” (https://www.repubblica.it/moda-e-beauty/d/opinioni/2023/07/02/news/elena_cattaneo_progettazione_zona_sismica_stati_uniti_maria_giovanna_durante-405834417/(si apre in una nuova finestra)).
During the project I also implemented a dissemination plan targeting various audiences. To this end, I organized a total of 10 technical seminars/webinars part of the Restructure 2.0 series (attended by more than 400 people in total). I also organized, hosted, and/or participated in a number of public events, including the European Researchers’ night and meetings with high school students, to raise awareness on gender gaps and attempting at providing younger women with role models in STEM fields. The main event on this subject was organized as part of this project in the main campus auditorium and hosted Senator for Life and Professor Elena Cattaneo, a prominent public figure in the Italian and International research landscape. The event was attended by about 600 people.
The activities of the ReStructure 2.0 project attracted the attention of the nation-wide and international media. The project and my story were featured in a variety of articles and television interviews including an interview published by the national magazine “Il Venerdi di Repubblica” (https://www.repubblica.it/venerdi/2021/09/08/news/studiamo_i_terremoti_in_convento-315526765/(si apre in una nuova finestra)) an interview on the national magazine Prometeo, and an article on the national magazine “D di Repubblica” (https://www.repubblica.it/moda-e-beauty/d/opinioni/2023/07/02/news/elena_cattaneo_progettazione_zona_sismica_stati_uniti_maria_giovanna_durante-405834417/(si apre in una nuova finestra)).
During the ReStructure 2.0 project I developed a novel artificial-intelligence-based model to design retaining structures in seismic areas based on 115,000 numerical simulations and soil-structure interaction principles. The model, for the first time, explores the effect of various input parameters, including soil non-linearity, providing a new tool to researchers and practitioners. This new model can help reducing the carbon footprint of new constructions in seismic areas, making the construction industry more green and sustainable. The reduction in emissions also has a positive economic impact, making construction processes more affordable. Thus, the implementation of the new model developed during the ReStructure 2.0 project can result in a win-win strategy that saves money, making constructions more sustainable. The theoretical framework that inspired ReStructure 2.0 was recently adopted to design a major sport arena in California. Using the method I contributed developing, engineers saved 60% of the resources. This is a single, but rather prominent, example of the tangible impact of this project. This method is already implemented in USA-wide design guidelines. The principles and model developed during the ReStructure 2.0 project are suitable to be implemented in the Eurocode (i.e. the Europe-wide building code). Such implementation would constitute a transformational shift from years of past practices.