ERI: A Hybrid Mechanics-Guided Machine Learning-Based Predictive Framework for the Performance of Rocking Foundations During Earthquake Loading

该奖项全部或部分由《2021年美国救援计划法案》（公法117-2）资助。最近的研究结果表明，在地震荷载作用下，摇摆基础是一种有效的岩土隔震机制，有可能消除或减少对其所支撑的建筑物和桥梁结构的破坏。该项目的目标是通过将控制问题物理的力学与使用大数据和机器学习技术发现的知识相结合，为摇摆基础的性能开发一种新颖的混合预测框架。该工程研究启动（ERI）奖是首次尝试将物理学与数据科学结合起来，对摇摆基础支撑结构的抗震性能进行建模，因此该项目直接推进了当前结构系统抗震性能建模的知识和技术水平。在土木工程设计和实践中引入摇摆地基将提高土木基础设施的弹性和可持续性，减少因建筑物和桥梁失效而造成的人员和经济损失，从而直接造福社会。该项目的核心思想是将基于力学的模型与机器学习算法相结合，开发一种混合力学指导的基于机器学习的预测框架，以确保预测的更好的泛化性和准确性，以及结果的一致性。为了实现这一目标，将开展以下研究任务：(i)利用摇摆基础数据库中的离心机和振动台实验数据，开发用于摇摆结构-基础-土壤系统性能预测的机器学习模型；（ii）使用OpenSees有限元框架中可用的基于力学的模型对摇摆系统进行数值模拟；（iii）通过有效地结合任务(i)和（ii）中开发的模型，开发用于预测摇摆系统性能的混合模型。该项目的主要成果将是为摇摆系统建立一个混合建模和预测框架，该框架既是科学驱动的，也是数据驱动的，并且具有在未来不断学习、适应和改进的潜力。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This award is funded in whole or in part under the American Rescue Plan Act of 2021 (Public Law 117-2).Recent research findings reveal that rocking foundations during earthquake loading perform as efficient geotechnical seismic isolation mechanisms and have the potential to eliminate or reduce the damage to the building and bridge structures they support. The objective of this project is to develop a novel, hybrid predictive framework for the performance of rocking foundations by combining the mechanics that governs the physics of the problem with the knowledge discovered from the use of big data and machine learning techniques. This Engineering Research Initiation (ERI) award is the first attempt to combine physics with data science to model the seismic performance of structures supported by rocking foundations, and hence the project directly advances the current knowledge and state of the art in modeling the seismic performance of structural systems. The introduction of rocking foundations in civil engineering design and practice will improve the resiliency and sustainability of civil infrastructure and reduce the human and economic losses resulting from the failures of buildings and bridges, thus directly benefitting society. The core idea of this project is to combine mechanics-based models with machine learning algorithms to develop a hybrid mechanics-guided machine learning-based predictive framework that will ensure better generalizability and accuracy of predictions, as well as consistency of results. In order to achieve this objective, the following research tasks will be carried out: (i) development of machine learning models for performance prediction of rocking structure-foundation-soil systems using centrifuge and shaking table experimental data available in a rocking foundations database; (ii) numerical simulations of rocking systems using mechanics-based models available in the OpenSees finite element framework; (iii) development of hybrid models for prediction of performance of rocking systems by effectively combining the models developed in tasks (i) and (ii). The major outcome of this project will be a hybrid modeling and prediction framework for rocking systems that is both science-driven and data-driven, and has the potential to continuously learn, adapt and improve in the future.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.