System vulnerability modeling and assessment in seismic-prone areas using machine learning

• Main Author: Chen, Weiyi
• Other Authors: Zhao Ou
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Civil engineering::Construction management
• Online Access: https://hdl.handle.net/10356/163040
• Institution: Nanyang Technological University

Earthquakes have caused great social and economic losses to human societies in recent years. Vulnerability modeling and assessment are important measurements for seismic risk mitigation. Conventional seismic vulnerability analysis frameworks have limitations such as subjectivity, low capacity of index information, or high computation cost. The motivation of this thesis is to overcome these drawbacks by employing machine learning techniques. In this thesis, a machine learning-based seismic vulnerability management framework is proposed. The proposed framework is composed of four objectives: Objective 1 aims to assess the regional seismic vulnerability with the information fusion approach. Objective 2 aims to assess the building vulnerability by XGBoost models learned from earthquake building damage data. Objective 3 aims to predict the casualty rate and economic loss of earthquake disaster areas with AutoML models. Objective 4 aims to explore the effective emergency response plans of hospital networks by Multi-objective optimization.