صبا علياري شوره دلي

Structural Health Monitoring (SHM) plays a crucial role in assessing the safety of steel railway bridges by optimizing instrumentation while maximizing information and minimizing costs. This thesis focuses on damage detection methods using artificial intelligence and examines optimization techniques for sensor placement tailored to steel railway bridges. Additionally, a new framework for reinforcement learning-based optimization is introduced. Two influential sensor placement optimization methods, Effective Independence (EFI) and Genetic Algorithm are eva‎luated using modal analysis results from a finite element model of two steel structures: a steel beam and the Kw51 steel railway bridge in Belgium. These methods are compared with the reinforcement learning-based approach presented in this research. The criteria for optimal sensor placement include the condition number of the mode shape matrix and the root mean square of the off-diagonal elements of the modal assurance criterion matrix, which were used to identify the optimal sensor configuration. The reinforcement learning-based method demonstrated suitable accuracy compared to the other two methods but required more computational time. Additionally, this thesis explores the application of artificial intelligence in damage detection. By introducing a stacked neural network based on a random forest algorithm, an accuracy of over 98% was achieved in classifying different structural states, enhancing the structural health monitoring capabilities of the bridge.

پايش پل , بهينه سازي , جايگذاري حسگر , تشخيص خرابي , هوش مصنوعي

Bridge Monitoring , Sensor placement , damage detection , Artificial Inteligence , Optimization

Saba Aliyari shouredeli

Dr. Javad Mirmohammad sadeghi