Structural Health Monitoring Can Improve Bridge Monitoring Capabilities

Although not initially designed for earthquake monitoring, structural health monitoring applications are proving their worth during seismic events. Image shows detail of sensor on New Jersey’s Burlington Bristol Bridge.

Civil + Structural Engineer – Seismic Zone
January 2012

Nathaniel Dubbs, EIT and Matthew Yarnold, EIT

New and unpredictable situations are boosting the use of structural health monitoring applications. The signature bridges of the Burlington County Bridge Commission (BCBC) in New Jersey were first targeted for structural health monitoring (SHM) applications after the BCBC tasked their engineer of record, Pennoni Associates, with incorporating the latest available technology in assessing the safety of their structures following the collapse of I-35 in Minneapolis. The bridges selected for SHM applications include the Burlington Bristol Bridge (BBB), an 80-year-old steel truss structure with a main 540-foot vertical lift span, and the Tacony Palmyra Bridge (TPB), an 82-year-old structure consisting of a 550-foot steel arch span, 280-foot rolling bascule span and steel through truss approach spans.

As with any SHM system, it is important to identify specific and relevant hazards and vulnerabilities to monitor due to the costs associated with sensors and data acquisition systems. To this end, a set of unique SHM systems were designed and implemented on the BBB and TPB. The first SHM system was designed with the intention of characterizing the live load demands placed on critical structural members of the BBB. The system was designed to automatically capture strain response to heavier traffic events, defined within the data acquisition system as approximately one-half of a legal truck load for the span, with electrical resistance strain gages. These sensors are well-suited for this application — which provides accurate measurements as vehicles travel along the span — due to the high sampling speeds allowed. A second type of sensor was also employed to characterize the long-term demands of the same structural member due to daily and seasonal temperature variations experienced by the structure. These sensors consisted of vibrating wire strain sensors and are sampled at very slow speeds. A similar SHM system was installed on the TPB. It consisted of many vibrating wire strain sensors installed along a cross section of the main arch span, coupled with vibrating wire displacement sensors, with the intention of characterizing the behavior of the structure due to daily and seasonal temperature variations. Both of these systems are equipped with cellular modems, allowing for remote data collection and connection to data acquisition systems.

Following the earthquake in the northeastern region of the United States on Aug. 23, 2011, the BCBC immediately tasked engineers with inspecting the structures for any related damage. To assist the engineers with the field inspections of structural components along the bridges, the data acquisition systems on both bridges were accessed and the collected data was inspected for any information about the performance of the structure during and after the earthquake. Although the monitoring systems were not originally intended to monitor for structural response to earthquakes — both in terms of sensors utilized and measurement triggering processes — the available measurements provided information on the state of dead load distribution before and after the earthquake event.

The triggering mechanism for the electrical resistance strain gages on the BBB constantly scans the average stress within a critical member of the lift span truss. If the stress within this member exceeds a threshold consistent with approximately a half legal truck load, the data acquisition system will record data from a period of time before the trigger and will remain recording all sensors until a period of time after the average stress has fallen below the predefined threshold. This type of recording typically allows for the full capture of the truck entering the span, traveling across and exiting the span. After inspecting the data corresponding to these electrical resistance strain gages, engineers noticed that no event was triggered at the time of the earthquake. To ensure that the triggering system was working properly, the recorded events before and after the earthquake were examined and both found to be consistent with all data recorded since the system was installed. These events happened within minutes of the earthquake, suggesting that the monitoring system was operating as expected. By not triggering the threshold corresponding to the half legal truck load, it was concluded that the earthquake did not create a significant enough stress within the monitored members to trigger the system to begin recording.

In addition to inspecting the data from the electrical resistance strain gages, engineers also inspected the stress levels measured by the vibrating wire strain gages for both the BBB and TPB before and after the earthquake. To determine if any significant changes in the dead load distribution were experienced by either structure, the baseline strain measurements from the sensors were carefully examined for periods of time before and after the earthquake. While the sensors do not measure absolute dead load stress, they will measure any change in dead load stress that would be caused by any significant structural damage. In the case of both bridges, the long term stresses within the monitored members were found to be both stable before and after the earthquake event and consistent with measured data from weeks leading up to the event. Since the SHM system on the TPB also monitors bearing displacements of the main arch span, it was confirmed that the span did not experience any major movements.

While the SHM systems installed on the Burlington Bristol and Tacony Palmyra Bridges were not originally designed for capturing the structural response due to earthquakes, the information available from the systems lent itself to a prompt analysis of the demands placed on the structures during the event. Before emergency field inspections were complete, Pennoni notified the BCBC that preliminary analysis from the SHM systems gave engineers confidence that the monitored spans did not experience any abnormal changes in overall stress distribution and that the structural responses did not exceed what is seen many times during the course of a single day. The inspection team then confirmed these findings by examining critical elements and details throughout both structures. SHM systems can be applied to a wide variety of situations and with the ability for incorporation of video and networking technologies, this approach could be useful for a wide variety of structures and a broad set of applications.

Nathaniel Dubbs is a staff engineer with Intelligent Infrastructure Systems, A Pennoni Company. Matthew YarnoldEIT, also contributed to this article. He is a project engineer with Intelligent Infrastructure Systems.