Chapter29 Spatiotemporal Sensing for Pipeline Leak Detection Using Thermal Video Ganesh Sundaresan, Seung-Yeon Kim, Jong-Jae Lee, Ki-Tae Park, and Hae-Bum Yun Abstract Auto Modulating Pattern Detection Algorithm (AMP) is a novel data processing technique to detect adverse hazards for various monitoring applications. This ongoing experimental study seeks to expand the one-dimensional “pointsensing” AMP to the two-dimensional “plane-sensing” case. AMP will be validated as an effective spatiotemporal technique for detecting leakage in fluid distribution pipeline networks. Pipelines fail due to various causes, including external interference (mechanical damage) and corrosion caused by external environmental factors. These damage signatures are usually very tiny and are masked by other more dominant trends such as the ambient air temperature, making these patterns difficult to detect. An experimental pipeline setup has been fabricated to simulate a real pipeline network. Damage is simulated through the use of holes that can be closed using watertight bolts. A stationary 360 240 resolution infrared camera will be used to measure spatiotemporal temperature signatures on the pipeline surface over time. AMP can be implemented on each pixel’s time history to detect abnormal changes in temperature that are associated with hazards. Preliminary results show that AMP successfully provides spatiotemporal information related to adverse hazards. Keywords Auto-Modulating pattern detection algorithm • Spatiotemporal analysis • Hilbert-Huang transform • Damage detection • Civil infrastructure monitoring 29.1 Introduction As of 2009, the USA has about 600,000 bridges, 85,000 dams, 160,000 miles of national highway network, 140,000 miles of rail transportation network, 100,000 miles of levees, and 12,000 miles of navigable inland waterways [1]. The above mega infrastructure networks are continuously monitored for various operation and maintenance purposes using numerous monitoring technologies, such as structural health monitoring (SHM) systems [2–4], intelligent transportation systems (ITS) [5], road weather information systems (RWIS) [6], waste monitoring systems (WMS), and so on. With recent advancements in sensing and networking technologies, quantitative data collected from various civil systems are becoming more available than ever. However, obtaining sensor data is necessary, but not sufficient, to understand the complicated spatiotemporal behavior of civil systems in infrastructure networks. The research community is overwhelmed with the extensive nature of data that is required to be collected from infrastructure networks comprised of a large number of heterogeneous civil systems. Future construction, operation and maintenance of infrastructure are envisioned to be G. Sundaresan • H.-B. Yun ( ) Department of Civil, Environmental and Construction Engineering, University of Central Florida, Orlando, FL, USA e-mail: g.sundares@gmail.com; Hae-Bum.Yun@ucf.edu S.-Y. Kim • J.-J. Lee Civil and Environmental Engineering Department, Sejong University, Seoul, South Korea e-mail: eswaiii@hanmail.net; jongjae@sejong.ac.kr K.-T. Park Korea Institute of Construction Technology, GoYang-Si, GyeongGi-Do, South Korea e-mail: ktpark@kict.re.kr A. Wicks (ed.), Structural Health Monitoring, Volume 5: Proceedings of the 32nd IMAC, A Conference and Exposition on Structural Dynamics, 2014, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-04570-2__29, © The Society for Experimental Mechanics, Inc. 2014 263
RkJQdWJsaXNoZXIy MTMzNzEzMQ==