Home // DATA ANALYTICS 2019, The Eighth International Conference on Data Analytics // View article
Authors:
Prasad Chetti
Hesham Ali
Keywords: Structural Health Monitoring; Population Analysis; Correlation Networks; Markov Clustering; Sufficiency Rating; National Bridge Inventory database
Abstract:
Traditional Structural Health Monitoring (SHM) methods require bridge inspectors to manually inspect each bridge periodically (usually every two years) and recommend maintenance or rehabilitation services to the bridge if necessary. As limited manpower and budget constraints are the two major shortfalls in traditional SHM methods, in addition to potential human errors and lack of consistency, more rigorous and frequent solutions are needed to assess the health levels of bridges and provide needed recommendations. In this work, we process a new population-based approach that employs the concept of Correlation Networks to evaluate the status of each bridge based on general parameters as well as how it compares to other similar bridges. We propose a Correlation Network Model (CNM) that builds a network of bridges, based on time-series data on Sufficiency Ratings, for a population of 9,546 “steel bridges with stringer/multi-beam or girder design,” taken from the U.S. National Bridge Inventory (NBI) database. We apply Markov Clustering Algorithms to produce clusters of bridges with similar features associated with their fitness ratings over user-defined periods of time. The top five clusters are identified and further analyzed using population analysis algorithms. We were able to identify three clusters with lower fitness ratings and suggest that the bridges in these clusters need to be serviced sooner than those included in the other clusters. Experimental results show that the proposed model provides an efficient approach that allows domain experts to assess the structural health of bridges/civil infrastructures in a robust way that can guide rehabilitation services for all bridges and identify potentially unsafe bridges that need urgent attention
Pages: 12 to 19
Copyright: Copyright (c) IARIA, 2019
Publication date: September 22, 2019
Published in: conference
ISSN: 2308-4464
ISBN: 978-1-61208-741-2
Location: Porto, Portugal
Dates: from September 22, 2019 to September 26, 2019