Development, Testing, and Evaluation of Bridge Bearing Movement Sensor System

Bridge bearings and expansion joints are critical features of bridges as they prevent the buildup of thermal stresses caused by changing temperatures. Current assessment techniques include visual inspections which are subject to inaccuracies based on the inspectors' judgement. Therefore, innovation in the way bearings and expansion joints are monitored and maintained is required. Structural health monitoring systems have been developed to measure important features of bridges though one limiting factor of their use in practice is their high cost. With the advent of low-cost microelectromechanical systems (MEMS) and internet of things (IoT) platforms a bearing movement sensing system has been developed. The proposed sensing system uses an Arduino platform to measure acceleration from a multi-axis accelerometer and converts changes in gravitational acceleration values into bearing displacements. The sensing nodes were calibrated to measure displacement with changing temperature. The sensor nodes were wired together and connected to a single battery that was charged using a solar panel. Data from the sensing system was transmitted over the internet to the cloud. A newly constructed 1.2 km bridge in Kingston, Ontario, Canada was used as the testbed for the developed sensing system. One bridge pier located at an expansion joint was equipped with 10 sensing nodes, one at each bearing, to measure the movement of the joint. Additionally, sensors on bearings atop 2 piers at opposite ends of an 86 m bridge section measured the expansion/contraction of the section. The proposed monitoring system had a cost of approximately $100 per measurement node. The results suggested that the coefficient of thermal expansion of the bridge was approximately 7.45×10-6 /°C, within the range of 6.1 to 13.1×10-6 /°C used in the design of concrete structures. Additionally, the rate of expansion and contraction of the bridge section was measured as 0.61 mm/°C and 0.67 mm/°C, respectively. The design value assumed for the bridge section was 0.86 mm/°C. Using the collected data, together with environmental conditions, machine learning models capable of predicting bridge movements were developed. The models were effective at predicting movements from limited datasets, though they require further development to increase accuracy.

Department dc:contributor.department

Civil Engineering

Author dc:creator

• Heykoop, Isabel

Advisor dc:contributor.supervisor

• Hoult, Neil

• Internet of Things
• Structural Health Monitoring
• Machine Learning
• Sensor Systems
• Microelectromechanical Systems

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/1974/32546

OAI identifier oai:identifier

oai:queensu.scholaris.ca:1974/32546