Artificial Intelligence Applications in Post-Earthquake Structural Damage Simulation

الملخص

Earthquakes can pose one of greatest challenge to the designers of buildings and other civil engineering structures. These serious risks threaten human safety and infrastructure, with traditional post-earthquake damage assessments being slow, costly, and sometimes hazardous. This paper summarizes recently developed methods and theories in the developing direction for applications of artificial intelligence in civil engineering, including in simulating and predicting structural damage following seismic events. AI enables rapid and accurate assessment of building damage by collected a lot of data from global earthquake networks and, using methods such as leveraging machine learning (ML), deep learning (DL),  artificial neural networks (ANNs), convolutional neural networks (CNNs) and long short-term memory (LSTM). AI-based approaches offer significant improvements in prediction accuracy, accelerate post-earthquake response, and early warning integration. By these strategies seismic disaster mitigation can be during a pre-earthquake, and be executed at the time of an earthquake and post-earthquake. The research problem is defined in the following main question: Can we, through the use of artificial Intelligence, predict earthquakes, mitigate disaster risks, avoid human losses, and protect buildings and facilities from collapses that cause loss of life and property?  This research aims to study the seismic behavior predicted by artificial intelligence applications for concrete structures as a result of repeated seismic wave loads imposed on them and to know their ability to resist the effects of these waves, represented by shear, torsional, bending and moments forces. This study also aims to evaluate the limits contained in some international codes and compare them with the values predicted by artificial intelligence applications. These codes include the American code ACI-14, the second European code EC-2, and the Egyptian code ECP-2017.