Using Wireless Data Acquisition for Bridge Collapse Testing

Challenge:
Lifting road fill via a crane boom can cause a bridge to collapse when it reaches a breaking point, so a method was needed to quickly and safely measure the weight of the fill at this point.
Solution:
By attaching load cell sensors to the load-bearing cables that suspend the bucket of a truck-mounted crane and connecting them to the NI WLS-9237 Wi-Fi data acquisition (DAQ) module, the weight of the bridge subgrade can be easily read and recorded in the field.
"Due to the danger of bridge collapse, the most important concern during testing is safety. Therefore, the WLS-9237 wireless data acquisition module is used to collect the weight data applied to the subgrade on the bridge measured by the load cell sensors. The wireless system provides the measurement accuracy required to quickly calculate the collapse load while eliminating the safety issues that may be caused by cables."

Developing Wireless Measurement Systems Using NI Hardware and SoftwareThe
Ferguson Structural Engineering Laboratory has a suite of structural testing tools and various loading devices that enable extensive research in the field of structural behavior. In 2009, researchers at the laboratory successfully collapsed a 120-foot bridge with an artificial crack face to study its classification as a critical fracture as defined by the American Association of State Highway and Transportation Officials. After three rounds of testing, the bridge finally collapsed due to being unable to bear more than 360,000 pounds. The researchers tested it to determine the damage characteristics of the bridge after the crack face was generated and the vulnerability of the girder after the fracture. The result was that the bridge withstood a weight of about 4.5 times the maximum theoretical load.
The purpose of the test was to observe the sequence of failure mechanisms and determine the maximum load required to cause the initial collapse of the entire bridge. The system behavior of the double box girder structure bridge required three rounds of testing, two dynamic and one static. In the first round of testing, we used explosive fill to artificially damage the bottom edge of one girder. If the bridge was at a critical fracture at this time, this fracture would cause the bridge to collapse; however, the bridge had no obvious damage after the first round of testing.
We conducted another dynamic test to induce collapse. The bridge was lifted from its original position. The cracks in the broken beam expanded in a network and were quickly eliminated by the lifting part. Once again, the bridge remained intact after the test.
In the final test, we continued to load the bridge until it collapsed. Initially, we loaded it with 1 times the theoretical load, and then continued to increase the load by 1,500 to 3,000 pounds each time by dumping filler from a crane bucket (Figure 1). After more than 100 times, the bridge was loaded with more than 360,000 pounds and finally collapsed.
Because of the danger of bridge collapse, the most important concern during the test was safety. Therefore, we used the NI WLS-9237 wireless data acquisition module to collect data from load cell sensors on the weight applied to the roadbed on the bridge. The wireless system provides the measurement accuracy required to quickly calculate the collapse load while eliminating the safety issues that may be caused by cables. The system allows for direct wireless connection to the sensors using the simple and secure IEEE 802.11g protocol.
We used LabVIEW software to remotely monitor the load signals. Built-in signal conditioning and the highest level of commercial network security allow us to transmit data in real time to a remote monitoring location about 50 feet away (Figure 2).
Testing Procedure
We simulated vehicle loads by placing box-type concrete beams on the bridge deck and added weight by loading the bridge with dumped materials (Figure 3). Road base was chosen as the filler material due to its easy availability, low cost and high density.
To quickly and safely measure, we attached the load cell sensor to the load-bearing cable that hangs the bucket of the vehicle-mounted crane. We also built a wooden box to store the wireless transmitter and power supply (Figure 4) and hung it on the side of the sensor. We used a simple laptop battery to power the WLS-9237. Finally, we connected the load cell sensor to the bucket with a steel cable to provide enough space so that the dumping of filler would not damage the load cell sensor or related equipment. The WLS-9237 Wi-Fi DAQ device is connected to the load cell sensor, so the load data can be easily read and recorded directly from the site without being affected by the operation of the crane.
Bridge Data Acquisition System
In addition to safely planning and executing the full test, it was important to collect data during the experiment for future analysis of the bridge’s behavior. We designed and implemented an instrumentation scheme to measure deformation and material stress. We connected strain gauges directly to the bridge members to measure material deformation.
All hardware came from National Instruments (NI), including a 244-channel DAQ device (Figure 5). Two NI SCXI-1001 12-slot chassis were populated with 24 NI SCXI-1520 8-channel universal strain modules. We also used two NI SCXI-1000 4-slot chassis, which were populated with five NI SCXI-1520 8-channel universal strain modules and three NI SCXI-1121 4-channel isolation amplifiers. NI SCXI-1314 8-channel terminal blocks were connected to all 29 NI SCXI-1520 modules, and three NI SCXI-1321 4-channel terminal blocks were connected to NI SCXI-1121 isolation amplifiers. All four chassis were connected to the PC via the NI PCI-6250 DAQ board, which was configured using LabVIEW.
Testing to Collapse
After three days of adding fill, the majority of the concrete at the mid-span extension where it connects to the exterior railings began to crumble as the total bridge load increased to approximately 360,000 pounds. As the bulk of the material began to slide off, an additional three buckets of fill were added to the bridge before the bridge collapsed to the bottom concrete bed (Figure 6). As the bridge deck load exceeded the weight of the test process, a series of bridge member collapses occurred. The load redistributed on the bridge after the mid-section collapsed, indicating the contribution of redundant force paths to maintaining the bridge balance as the bridge damage progressed. These findings will then be used to develop strength models to evaluate double box girder bridges.