Strain-based Fatigue Crack Monitoring of Steel Bridges using Wireless Elastomeric Skin Sensor
Fatigue cracks have been a major issue for steel bridges in the nation. State DOTs currently rely on a two-year inspection period to examine steel bridges for detecting fatigue crack activities. However, human inspection is time consuming, labor intensive, cost inefficient, and prone to error. Although Nondestructive Evaluation (NDE) techniques can improve inspection accuracy, the lack of autonomy and continuity in the inspection process still limits its ability to capture critical crack development in a timely fashion. After all, these cracks may occur between scheduled inspection periods and therefore can lead to catastrophic failure of steel bridges.
Measuring strain-induced crack development is an effective way of monitoring fatigue cracks. However, due to the randomness of crack paths, traditional strains sensors, such as metal foil strain gauges, are small and can only be deployed in a small and localized region, limiting their capability of detecting these cracks. Moreover, the limited ductility of these foil gauges leads to breakage under excessive strain, typical during crack formation. Meanwhile, a traditional wired monitoring approach imposes high cost to the overall system due to bulky data acquisition units and expensive cabling work, especially for long-span bridges. The centralized management scheme limits its flexibility and continuous monitoring capability due to potential data inundation.
The main objective of this proposed research is to provide state DOTs a practical and cost-effective long-term fatigue crack monitoring methodology using a wireless elastomeric skin sensor network. This research is intended to demonstrate the value-added of fatigue crack monitoring of steel bridges using wireless skin sensors over the traditional bridge inspection in the following ways:
- The ability to collect objective information regarding fatigue crack activity under in-service loading of bridges in a continuous manner, improving the assessment, safety, and reliability of fractural-critical bridges, and providing early warning regarding evolving internal defects.
- The ability to improve prioritization of bridge repairs (condition-based maintenance) and retrofit for fatigue cracks so as to maximize the effectiveness of limited resources.
- The ability to better assess the effectiveness of various fatigue repair and retrofit techniques for steel bridges through long-term crack monitoring.
Scope of Work:
The proposed research consists of three phases of tasks from the development and validation of the basic crack sensor (Phase I), design, implementation, and small scale validation of the wireless data acquisition system and the autonomous sensor network operation strategies (Phase II), to the robustness, large scale, and long-term full scale validation tests of the integrated wireless crack sensor network system (Phase III).
Phase I tasks: Crack sensor
1. Crack sensor fabrication: Fatigue crack sensors will be developed by fabricating the skin sensors tailored to fatigue crack detection. The thickness of these sensors will be 40 µm to enhance the mechanical robustness under harsh environment.
2. Small scale validation: Tests will be designed and carried out to validate the crack detection capability of the crack sensor. The test setup will consist of a pre-notched steel member subjected to fatigue loading. The fatigue testing facility at the University of Kansas will also be used to perform tensile fatigue tests and bending fatigue tests. An off-the-shelf data acquisition system will be adopted to measure the capacitance change of the sensors under crack development. Metal foil gauges will be used to provide reference measurement for validation.
Phase II tasks: Wireless crack sensor network
3. Wireless data acquisition (DAQ) system: A wireless data acquisition system will be developed for measuring capacitance change of the elastomeric skin sensor. The Imote2 smart sensor platform is selected because of its fully tested software package, which serves as the foundation of this integrated wireless skin sensor network system. The Imote2 sensor can measure voltage inputs between -5 V and +5 V using a stackable data acquisition (DAQ) board. An interface board between the elastomeric skin sensor and the Imote2 DAQ board will be developed to convert capacitance change into voltage signals.
4. Data quality assessment: The quality of data collected by the integrated wireless sensing system will be evaluated. Focus will be placed on the noise level and resolution. The goal is to make sure the integrated system has adequate resolution and low noise level to capture the strain change associated with fatigue crack activity.
5. Crack detection algorithm: Algorithms will be developed to process the measured strain data to reveal the crack activities underneath the skin sensor. Various sensor patterns will be investigated based on typical scenarios of fatigue cracks. Data fusion strategies combining on-board and in-network data processing will be implemented and tested for detecting active fatigue cracks as well as the propagation direction of these active cracks. The goal is to be able to make decisions using the smart sensors by locally combining and processing data measured at different locations without transmitting raw data back to the base station.
6. Autonomous sensor network: Software will be designed to enable autonomous operation of the sensor network. Practical issues such as triggering of the sensor network, sleep-cycling to conserve energy, energy harvesting methods such as solar panels, and other energy efficient operation issues, etc. will be addressed in this task.
7. Small scale system validation: The integrated system will be validated using the small scale pre-notched steel beam from Task 2 of Phase I.
Phase III tasks: Large/full scale validation
8. Robustness test: The robustness of both hardware and software will be tested. Mechanical robustness against external impact on the elastomeric skin sensors, long term stability of the electrical signal generated by the crack sensor, and fault tolerance of the software for autonomous operation will be tested.
9. Large scale validation: A network of crack sensors will be deployed on the existing large scale steel girder test bridge at the University of Kansas, which is a 30-ft long, three-girder bridge that includes realistic fatigue sensitive details.