Robust Output Only Health Monitoring of Steel Railway Bridges: Analysis of Applicability of Different Sensors

Introduction

Monitoring aging infrastructure performance under various loading and environmental conditions continuously via an automated monitoring system that evaluates their structural health quantitatively is very significant to ensure safety and avoid progressive collapses. One of the important elements in any transportation system are bridges which currently subjected to increased traffic load and frequency. In the United States, bridges structural health is usually assessed via visual inspection which is costly, might be unsafe and subjected to human interpretation as examined by Phares et al. (Phares et al., 2001). Structural Health Monitoring (SHM) often involves detecting structural damages automatically which explored previously by Farrar and Worden (Farrar & Worden, 2012). SHM involves extracting structural health information via signal processing layer from responses collected with a set of sensors deployed on a structural system. Significant information features usually extracted from the data using damage identification methods and integrated into analyses or probabilistic models to evaluate the structure health and expected service life (Achenbach, 2009).

SHM systems usually incorporate one or more of strains, accelerations and displacement where acceleration measurements provide insights on the structural global behavior while strain measurements provide unique understanding of structural local behavior. Research work was conducted to investigate the effectiveness of certain types of strain sensors in damage detection within SHM applications (Glisic & Inaudi, 2008; Glisic & Inaudi, 2012; Harmanci et al., 2016). Fiber Optic (FO) sensors application in SHM damage detection was investigated using two main FO techniques, where the first technique was based on fiber Bragg-gratings and the other was based on Brillouin optical time-domain analysis. The first technique allows the use of long gage FO sensors and the second one allows for distributed FO sensors. Glisic et al. (2013) applied the framework to a full-scale reinforced concrete structure and it was found that both sensing techniques are suitable for SHM damage detection (Glisic et al., 2013). To locate structural damage from dynamic strain measurements using local modal filters, a framework was proposed and validated experimentally in the laboratory by Tondreau and Deraemaeker (2014) on a steel beam with a dense array of strain sensors (Tondreau & Deraemaeker, 2014).