The evolution of functional recovery methods in earthquakes, from historical foundations to artificial intelligence opportunities
The evolution of functional recovery methods in earthquakes, from historical foundations to artificial intelligence opportunities
Majid Rostami Bahnamiri1 Hamidreza Tavakoli2
1) PhD student in Civil Engineering, Earthquake Orientation, Babol Noshirvani University of Technology
2) PhD in Civil Engineering, Earthquake Engineering - Professor, Department of Structures and Earthquake Engineering, Babol Noshirvani University of Technology
IOI Code
:
ABZE-ABZGCK
Publication :
5th International Congress on Civil Engineering, Architecture, Building Materials and Environment(caucong.com)
Abstract :
The occurrence of devastating earthquakes around the world, massive human and financial losses, and long recovery periods have made the importance of developing accurate and comprehensive methods for evaluating the performance of structures more evident than ever before. Traditional performance assessment methods, which were mainly based on two-state approaches of efficient and inefficient , were not able to provide a complete and probabilistic picture of the extent of damage and consequences caused by earthquakes. The evolution of seismic performance assessment from traditional force-based approaches to functional and ultimately intelligent frameworks has created a new paradigm in earthquake engineering. This research, by reviewing the historical and analytical methods of performance assessment and recovery, explains their evolution path to the integration of artificial intelligence. Initially, the theoretical foundation of hierarchical methods such as fault tree analysis (FTA) was transferred from safety-oriented industries to structural engineering to explain the relationships between component failure and system consequences in the form of probabilistic models. Then, the development of generalized limit state functions was proposed for various structures and materials to enable a more comprehensive analysis of the probability of failure and composite consequences (such as fatigue, corrosion, and simultaneous fire). In the next step, the application of machine learning (ML) and deep learning algorithms in the analysis Extensive data and functional recovery time prediction were investigated. The results show that the integration of graphical component dependency models with AI computational algorithms can reveal hidden relationships between structural and non-structural component failures and enable real-time assessment after earthquakes from building sensor data. Finally, the research provides a data-driven framework for resilient design and functional optimization that, while aligning with FEMA P58 standards, provides a platform for integrating AI into the post-earthquake decision-making process. The proposed approach, by combining structural engineering insights, probabilistic analysis, and machine learning, is a step towards realizing smart, resilient, and recoverable structures in the age of AI.
Keywords : functional recovery fema p58 machine learning algorithms smart structures