Damage identification tools for cable-stayed bridges: monitoring the key structural elements
Date
2022
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
In this, the era of the 4th Industrial Revolution (Industry 4.0), Structural Health Monitoring (SHM) has enabled the evolution of the maintenance of our infrastructure, and in particular of bridges. In the past, maintenance procedures were periodic or depended on in-person visual inspections that rely on the experience and subjectivity of the inspector. SHM enables a pivot from this outmoded process to a data-driven procedure, that is objective, higher accurate and can capture structural changes at early stages, well before they become visually obvious. Through SHM, infrastructure can be maintained based on its needs, increasing the lifespan of individual structural members. Potential damage can be addressed early on, when the repair cost and time are lower. In this way, the limited available resources can be more effectively utilized. Additionally, real time monitoring of bridges can enhance the serviceability of our transportation system and increase the safety for the commuters who use it. ☐ In order for SHM to be used to comprehensively monitor the condition of bridges, and to capture behavioral changes, it is crucial that appropriate tools that can utilize the collected data, be developed. Sensors used in SHM systems capture the local and global response of the structure to the combination of loads applied to it. These include static loads (e.g., self-weight), semi-static or slowly changing loads (e.g., thermal loads), and dynamic loads (e.g., traffic or wind loads). An unexpected response to a load combination can indicate a behavioral change of the structure. This in turn can signal the need for further investigation. By quantifying how damage to different elements of the bridge alters bridge response, the cause of the potential damage can be identified. ☐ With the focus being on cable-stayed bridges, this research presents two methods that can utilize the response of the structure to identify potential damage to key structural components of the bridge (i.e., the beam, stay cable, and bearing). Through two simplified models, the potential of these two methods to identify behavioral changes, distinguish between the various types of behavioral change, and decipher the type of damage that has caused the change, was assessed. ☐ The first method uses the strains developed in the edge girders of the bridge due to traffic loading, and in particular to trucks passing across the bridge. This method relies on calculating and monitoring the position of the Neutral Axis (NA) in the edge girder. With the NA position depending on (1) the cross-sectional properties of the girder and (2) the internal forces in the girder, its position changes if damage causes a change to either of these parameters. ☐ A simplified analytical model, consisting of the key structural elements of a cable-stayed bridge (i.e., a cable, a beam and a bearing) and loaded with a moving point load (to simulate traffic loads) was developed. Through this model, it was shown that NA position is expected to move upward: i) in the damaged section due to damage on the beam, ii) due to cable damage with the change been more intense close to the cable, and iii) due to bearing damage affecting the whole length of the beam; and downward in the undamaged section due to beam damage. Thus, NA position can be used to capture changes in any of the key elements of a cable-stayed bridge, and equally importantly, it can distinguish among the effects of different types of damage and locate the damage. ☐ In an effort to understand the behavior of the NA in a real structure, data collected from Indian River Inlet Bridge (IRIB), a cable stayed bridge, were used. In the controlled environment of non-destructive load tests, it was seen that the NA position would depend on the load characteristics (magnitude and position). To increase the sensitivity of monitoring NA under ambient traffic, a data mining model was developed to predict NA position based on estimations of the load characteristics. This model is shown to be capable of capturing changes in NA position as small as ¼ of inch. ☐ Like the first method, the second method used the strains developed in the edge girders of the bridge due to the static and the slowly changing loads (i.e., the dead and thermal loads). Due to the direct connection between the thermal induced strains and the temperature, this method suggests the monitoring of these two parameters using a linear regression model, and in particular, focusing on the slope and the intercept of the strain-temperature curve. A change in either of the two parameters indicates a change in the way the structure absorbs and distributes the loads. Once again, a simplified model was used to understand the expected behavior of these parameters due to different types of damage, considering strains measured in the top and the bottom of the girder. Using this model, it was shown that this method can also distinguish among different types of damage. ☐ Based on the research conducted, these two methods, show potential for capturing, distinguishing, and locating different types of damage on a cable-stayed bridge structure, and have great promise for use as powerful damage identification tools. Due to the similarities of the data needed for each method, it is suggested that they be implemented together, resulting in an even more robust monitoring procedure.
Description
Keywords
Damage identification, Cable-stayed bridges