Integrated smart sensor networks with adaptive real-time modeling capabilities

Abstract

<p>While serviceability, safety, and sustainability of deteriorating infrastructure have received significant attention in civil engineering, accessible approaches are needed to obtain actionable information about a structure over time. In particular, any intelligent infrastructure system's performance is governed by the 1) cost of sensing system required to measure structural states; 2) algorithms used to extract intelligence amongst the enormous quantities of multidimensional data; and 3) different approaches used to link intelligence to decisions. This work presents a theoretical framework for designing integrated SHM systems leveraging smart sensor and material technologies.</p> <p>In this dissertation, two types of sensing techniques for intelligent infrastructure are explored. The first is a bio-inspired sensing skin, termed soft elastomeric capacitor (SEC), that measures surface strain through a chance in capacitance. The SEC is a highly scalable technology and is a low-cost solution for monitoring local states over a global surface. Here, the SEC is used to detect and monitor cracks in a full-scale post-tension concrete component. The second is multifunctional self-sensing Carbon Fiber-Reinforced Polymer (CFRP), capable of detecting changes in its physical state (e.g., strain or damage). The prototyped CFRP capacitor is experimentally characterized through static and dynamic tests.</p> <p>In order to effectively utilize these advanced sensing techniques, methods for sensor network design and implementation are investigated. Of interest are computationally inexpensive real-time adaptive representations using available measurements. This is done through the meta-modeling of the monitored structural components and integrating real-time online parameter estimation algorithms. To explore real-time applicability, the proposed method is applied to the high-rate state estimation problem.</p>