APPLICATION OF SIGNAL PROCESSING TECHNIQUES FOR HEALTH MONITORING AND DAMAGE DETECTION OF MEMORIAL BRIDGE
Date of Award
Program or Major
Doctor of Philosophy
Yashar Eftekhar Azam
Aging and deterioration and damage to infrastructure elements such as bridges, roads, and tunnels, are some of the biggest concerns to the uninterrupted operation of the US transportation network.Structural health monitoring (SHM) is one of the most powerful tools proposed by the civil engineering research community as an economical and objective method for civil infrastructure management and condition assessment. A typical SHM process consists of continuous or periodic estimations of structural parameters based on a sensor network, feature extraction, and statistical modeling for feature classification to detect any damage that occurred to the structure. Due to the evolution of building codes, progress in design methods, advancing development in sensing technology, improvement in computing and information technology, and the arrival of machine learning methods, SHM has evolved over the decades. This development can provide a massive amount of valuable information to monitor and manage the serviceability and safety of in-service structures. Civil engineering structures, specifically bridges, can employ methodologies and techniques where SHM and machine learning intersect to mitigate operational challenges to identify and locate damages in early states. The research completed in this dissertation investigates damage detection and localization through structural health monitoring methods and machine learning algorithms. The newly reconstructed Memorial Bridge, a vertical lift steel truss bridge located in Portsmouth, NH is considered as a case study. The work includes proposing novel condition assessment techniques by implementing Artificial Neural Networks for damage recognition using structural response of the bridge; Development and application of classification models (support vector machine and multivariate logistic regression) for model evaluation in classifying data based on the bridge condition to find the optimal solution; Developing a tool to create a stochastic traffic simulation as a database based on different vehicular traffic characteristics; examining the effects of noise on the performance and accuracy of machine learning models prediction.
Mazaherimeybodi, Forouzandeh, "APPLICATION OF SIGNAL PROCESSING TECHNIQUES FOR HEALTH MONITORING AND DAMAGE DETECTION OF MEMORIAL BRIDGE" (2022). Doctoral Dissertations. 2706.
Available for download on Thursday, October 26, 2023