Placement of distributed crack sensor on I‐shaped steel girders of medium‐span bridges, using available field data

Abstract

Summary It is critical to detect cracks in steel girders of bridges before they have the potential to compromise the integrity of the structure. Both distributed binary sensors and distributed fiber optic sensors are capable of detecting cracks that are wider than 0.2 mm in steel girders. The objective of this paper is to report the optimum placement of these sensors on the girder to detect smallest possible length of the crack. In this work, the optimized placement of crack sensors was studied using FEM of two typical medium‐span simply supported steel girder bridges (Girder A, 30‐m–long span, and Girder B, 22‐m–long span). Using loads estimated from field monitoring data and FEM, a map of crack opening along the length of the crack was calculated for stable crack lengths. Using these maps and given the detectable crack opening of 0.2 mm, the optimum place to position a distributed crack sensor to detect the smallest crack length was determined. For Girder A, the sensor should be placed at 150 to 250 mm above flange at midspan and at one third from the support, and for the rest of the length of the girder, it should be placed at 200–300 mm above the bottom flange. For Girder B, the optimum placement for installation of binary sensor is estimated to be at 150 to 220 mm above the tension flange. The proposed method of calculation of placement can be used for installation of distributed sensors on other types of bridges.

Monitoring Stress State of H-Shape Steel Using Ceramic Piezoelectric Sensor: A Feasibility Study

“…Tohidi and Sharifi [16] used an artificial neural network system to predict the buckling deformation strength of Ibeams in the inelastic range. Zhou, Glisic, and Raeisi et al used distributed sensors (smart film, optical fiber sensors, and binary crack sensors, respectively) to monitor the health status of the structure [17][18][19]. Although many of these techniques are highly accurate and efficient, most are computationally expensive.…”