State of The Practice and Art for Structural Health Monitoring of Bridge Substructures

CHAPTER 2. STATE OF THE PRACTICE

From an investigation into the state of the practice of SHM, it is evident that there are a number of different monitoring systems and techniques. All of them have their pros and cons, but each can be useful to a certain degree. Most of the advances in SHM have been made in the monitoring of the superstructure elements of bridges and other structures; however, the importance of SSHM cannot be underestimated.

Since a large amount of the modern SHM technology is already widely used and documented as it pertains to superstructure monitoring, this review of the state of the practice will primarily focus on common technology and its practicality for use in a SSHM system.

GENERAL MONITORING SYSTEMS

Monitoring systems range in their functionality, cost, applied technology, and monitoring approach. A system generally contains three components: (1) a measuring device, (2) a method of reading that device, and (3) a method of storing the measurements. Depending on the complexity of the measurement being taken, the measuring device and readout component may be one and the same, such as dial gauges or pressure gauges (see figure 1).

Figure 1. Photo. Standard rotary dial gauges.

These devices convert a measurement parameter into mechanical gauge movement and can be considered the most basic of transducers as they transfer one physical aspect into another. Virtually all types of measurements have specialized devices to read that particular occurrence (i.e., time, displacement, velocity, acceleration, load, pressure, frequency, electromagnetic field (EMF), light intensity, strain, sound intensity, X-rays, voltage, inductance, capacitance, etc.). For most measurement types, there are many ways to take those measurements, which in turn, dictate the capabilities and/or limitations of a monitoring system.

The most basic systems use fully manual devices and readouts (e.g., dial gauges, proving rings, pressure gauges, etc.) coupled with manual record keeping. The limitations imposed on this method by requiring physical onsite personnel (i.e., recording/storage rate, man hours, and travel) are in some ways offset by the unforeseen observations and the ability to react to and record unplanned secondary happenings. The most exotic systems use complex measurement devices requiring sophisticated readout units coupled with a multifunctional data acquisition system (DAS) capable of sending the recorded data via cellular or satellite communications. These systems are often enabled to accept remote configuration/scheme changes, are self powered or self contained, and require little to no site visits. The most extreme cases of this type of system would likely be used by the National Aeronautics and Space Administration for space exploration because it is impossible to access the unit during use. Aside from the obvious cost, these systems are rarely adaptable to unforeseen occurrences. For SHM and SSHM applications, some midrange systems can be selected to provide a balance between equipment costs and required onsite man hours, which will allow most projects to be affordable.

CASE STUDY

One sample study performed by FHWA, the Washington State Department of Transportation, the city of Seattle, WA, and the bridge design team on the West Seattle freeway bridge incorporated a SSHM protocol.⁽¹⁾ This study is one of few that focused on substructural elements of a bridge pier during the construction of the bridge, as well as data collection over time. The West Seattle freeway bridge was built between 1981 and 1984. The original bridge was struck by a freighter in 1978 and was deemed inoperable as a result of the incident. The goal was to advance the state of the art of pile group design and analysis, and the information collected would be used in increasing pile group efficiency.

Authorities in Seattle, WA, authorized the use of instrumentation on pier EA-31, which is a single-column pier that supports the eastbound approach ramp from Spokane Street near the East Waterway and the Duwamish River (see figure 2).⁽¹⁾

Figure 2. Illustration. Pier EA-31 site map.⁽¹⁾

As stated previously, the purpose of this project was to improve the state of the art of pile group design and analysis. This was done by collecting information regarding the load distribution among the pile group, the load transfer from the piles to the soil, the portion of the load transferred from the pier footing to the piles, and the settlement of the pier footing. Furthermore, the results gathered from this data were compared with theoretical predictions that would either validate the theoretical models or allow for the modification of those models.

In order to provide measurements for the data collection criteria, measurements were selected by first measuring pile tip load as well as the load at six elevations along the pile to determine the individual pile load distribution. A load cell placed at the pile tip permitted direct measurement of the load. Next, six telltale rods were installed on each pile to determine the pile tip displacement. The pile deformation measured by the rods was converted to strain and used as a check. Strain gauges were then installed at the top of the piles, which provided information about the load transferred from the pier footing to the individual piles. Settlement of the pier footing was then measured by using a precise surveying measurement at the four corners of the footing. Last, soil settlement below the pier footing and within the pile group was measured to determine the soil’s reaction to the loading and subsequent deformation of the piles.

In total, 3 of the 12 piles were instrumented with a load cell at the pile tip, along with 6 elevations of strain gauge pairs and a 5-position telltale extensometer (see 1 ft = 0.305 m figure 3). Data from the instrumentation were collected in the field using portable manual readout units and recorded on field sheets. During construction, the measurements were made at irregular intervals dependent on accessibility and other constraints due to the construction progress. The instruments were monitored as each significant phase of construction was completed to provide realistic data from the construction process. Instrumentation monitoring was conducted throughout construction and continued through 1987, 5 years after the start of construction.⁽¹⁾ Data were again collected in September 1988, September 1989, and October 1993. Two additional sets of data were taken in 1999 and 2002, which extended the period of monitoring to 20 years. The report presents a summary of the existing working gauges as well as the date at which failed gauges no longer worked (see table 1).⁽¹⁾

1 ft = 0.305 m
Figure 3. Illustration. Pier EA-31 pile instrumentation layout.

Table 1. Summary of gauge failures.

As reported, all pile tip load cells were functioning after 20 years of service, with the exception of one transducer from pile 7, which was damaged during pile driving. From the data collected in 2002, the average load for all three piles was 94 tons (85.26 Mg) with a maximum deviation of approximately 11 percent (see 1 ton = 0.907 mg figure 4). This data suggest that some eccentricity was present wherein slightly more load was being taken by the eastern-most piles (represented by pile 1).

1 ton = 0.907 Mg
Figure 4. Graph. Pier EA-31 tip load in 3 of the 12 piles.

During the instrumentation phase, pairs of strain gauges were installed into the three monitored piles at six different levels along the pile. This provided 12 gauges in each pile for a total of 36 strain gauges. All of these gauges were located beneath the groundwater level, and 17 of these gauges were no longer functioning after 20 years of service. However, all the gauges were reported to have worked until at least October 1987, which provided 4 years of data collection. Since all of the gauges were installed below the groundwater level, it is possible that their failure was due to the water resistance of the system. The data from the strain gauges that were still in commission were plotted over time (figure 5 through figure 7).⁽¹⁾ For piles 1 and 10, the average strain change in the pile was between -300 and -500 µe(microstrain), with pile 1 being on the higher end of that range. However for pile 7, the average strain change in the pile was approximately -225 µe. This suggests that the piles further away from the center of the pile cap where the column is sitting experienced more strain change likely due to bending. The gauges installed at the top of the other piles and the strain gauges in the column were all still functioning after 20 years.

Figure 5. Graph. Pier EA-31 average strain change pile 1.

Figure 6. Graph. Pier EA-31 average strain change pile 7.

Figure 7. Graph. Pier EA-31 average strain change pile 10.

This study showed that SSHM using wired gauges is extremely useful. With the advances in the durability of data collection and monitoring systems, it is likely that this same system, if installed today, would not have the number of failed gauges. While this study required a worker to be onsite to record the data, the usefulness of the instrumentation provided insight into the design of foundations and instrumentation. While the technology record-keeping used in this study is somewhat outdated, the types of instrumentation are readily applicable and available (in a more robust form) for future studies. Automated data acquisition, monitoring, and remote data recovery are also available for this type of instrumentation and could be easily retrofit to the existing gauges.

WIRELESS SENSORS FOR HEALTH MONITORING

Wireless instrumentation has two connotations: (1) truly wireless gauges that minimize or even eliminate wiring attached to instrumentation, which is the topic of this section and (2) wireless communication (cellular or satellite) with instrumentation that may or may not employ onsite gauge wiring between the transducers and the data logger. Remote monitoring in itself is not automatically wireless, but rather, it may make use of landline communication between the data logger and querying parties.

Wireless systems use basically the same measurement devices (or transducers) as wired systems, but they use a transmitter and receiver system instead of lead wires. Wire costs range between $0.40 and $1.00 per 1 ft (0.305 m) per gauge and may require even greater expenses depending on the complexity of the installation site. Transmitters, similar to data logging equipment, are limited by their sampling and transmission rates-higher sampling rates come at higher costs with an upper rate limit of $5,000 to $10,000 samples/channel. The cost comparison of wireless to wired systems is generally site specific, but it leans toward wired systems. However, in the case of moveable structures or mechanical devices, slip rings or other features which allow the movements of the wires are required and tend to tip the scales in favor of wireless systems.

Wireless sensors for SHM systems are used more frequently as the technology becomes more available. Since no wires are required between the gauges and the DAS, installation time and associated costs are reduced as compared to traditional wired systems. Typically, wireless sensors are installed over an entire structure to get a full mapping of the desired measurement (i.e., stress, strain, displacement, temperature, velocity, etc.) across the entire structure. A wireless DAS collects the data sent back from these sensors and stores the collected data to an onsite data logger. As with most health monitoring programs, almost all wireless instrumentation used to-date involve superstructure and not substructural elements.

A study by Arms et al. introduced the idea of a SHM system in which the data acquisition software could be reprogrammed remotely.⁽²⁾ The goal was that the operating parameters of a monitoring system, such as sampling rate, triggering parameters, downloading intervals, etc., should be alterable from a remote location. As a result, operators should never have to go back to the site after initial installation. This provides a fully remote monitoring system in which all the parameters of the data logging and collection can be altered from a separate location.⁽²⁾

The wireless transmittable gauges were installed on the existing structure at main points of interest. Wireless sensors received transmitted data, and the data were uploaded to an onsite laptop (see figure 8). The laptop transmitted the data through a cellular uplink to the base station. From this base station, the software on the laptop could be altered to change the data collection parameters. The software could also be altered with trigger parameters so that the system could be sleeping but would wake up when an event occurred that increased the change in strain levels, such as a train crossing a bridge (see figure 9).

Figure 8. Photo. Wireless data collection and transmit setup.⁽²⁾

Figure 9. Photo. Train crossing bridge causing a strain event.⁽²⁾

While the software allows for a completely wireless system, its use as a SSHM system is not as probable. For installation in the deep foundation system, wireless sensors would have to be extremely powerful to transmit data wirelessly through surrounding soil, sometimes at depths of 100 ft (30.5 m). Even if they were available, sensors capable of this would most likely be too expensive to negate the cost savings from not using wired sensors. Furthermore, sensors used for reinforced concrete structural elements can provide much better data when installed within the concrete member where the reinforcement is located. Once again, a typical wireless sensor would not have the capability to transmit signals through hardened concrete. However, the wireless DAS could still be used with no obstructions.

Systems that overcome deep concrete embedment that are presently used are quasi-wireless where gauges are installed deep within the structure tethered to a transmitter at the concrete surface. These systems still suffer from power draw, and the useful unattended lifespan is limited, especially at high-sampling and transmission rates.

Susoy et al. researched the development of a standardized SHM system for the movable bridges in Florida.⁽³⁾ The assumption was that due to the multitude of elements, movable bridges were more prone to damage and deterioration and that the typical visual inspection as required by FHWA was not adequate. The study detailed the SHM system that was installed on the SR-401N Bascule Bridge over the Barge Canal in Port Canaveral, FL (see figure 10 and figure 11). A detailed finite element analysis was run to determine the probable locations for stress concentrations on the bridge. Once this was complete, wireless transmitting strain gauges were mounted on the bridge in these locations (see figure 11). The strain sensors transmitted their data wirelessly to the installed DAS, and the data were logged on a field computer also installed onsite.⁽³⁾

Figure 10. Photo. Bascule Bridge on SR-401N in Port Canaveral, FL.⁽³⁾

Figure 11. Illustration. Locations and types of sensors on Bascule Bridge.⁽³⁾

For this study, the wireless sensors were almost a necessity due to the type of project. Installing wired sensors on a movable bridge could prove to be difficult and cause damage to the wires. No mention was made concerning the accessibility of the data once they were collected, so it is assumed that the data were downloaded by a worker sent to the site. However, this study was based on the idea of wireless sensors for the monitoring system and therefore would have the same difficulty translating to SSHM as the Arms study.⁽²⁾

A study by Watters et al. introduced the idea of a special design for a wireless sensor capable of detecting threshold levels.⁽⁴⁾ The sensor was coupled with a radio frequency identification (RFID) chip. The sensor is read by scanning the system with a radio frequency (RF) transceiver. The RF transceiver alerts the RFID chip to power the sensor to collect data. Once the data are collected, the RFID chip transmits the data back to the transceiver to be read.⁽⁴⁾

The study focused on the use of the sensor to determine whether certain data may have crossed a threshold, namely chloride ingress into reinforced concrete structures. A particular threshold was set, and the system read the data and determined if the threshold had been met. The system was extremely useful for data that did not need to be streamed. For chloride intrusion into reinforced concrete structures, the critical point at which the chloride concentration is reached could take years to be met. Therefore, a DAS capable of collecting and logging data at a high rate was not needed. In typical concrete inspection, a core sample of the concrete deck must be taken and analyzed in a lab. With this technology, a sensor can be embedded into a structure and then routinely checked at a predetermined interval. Furthermore, the trends can be plotted over time to help owners and engineers predict when the chlorine intrusion will reach a critical level. The capability to send an alert when a certain threshold level is reached would be extremely useful in bridge monitoring. If an alert is programmed into the transducer that reacts when a certain level is met, it will allow authorities to react and make a decision about keeping a bridge open or closing it down depending on the severity of the event, which may save lives.

While this is a useful system for data that need to be monitored over long intervals, from a SHM point of view, the system is not beneficial for structures loaded with highly irregular or dynamic loading, such as a bridge. The sensors for a bridge SHM system would need to be read and have the data collected and stored at a relatively high rate in order for the owner or engineer to determine what is happening to the structure during its service life.

FIBER OPTIC SENSORS FOR HEALTH MONITORING

With the recent advances in the telecommunications field with fiber optics, the interest in fiber optic sensors (FOSs) has increased and has made way for powerful new sensors to be used for SHM. FOSs send light beams through the fiber optic cable at regular intervals and measure the return time. When the cross sectional area of the cable changes, the return time changes. This change in return time can be related to engineering parameters (i.e., strain, displacement, etc.) of the structural member to which they are attached. They are considered to be beneficial because they are relatively immune to interference from radio frequencies, electric or magnetic fields, and temperature differences.

A study by Udd et al. introduced the use of FOSs in existing structures.⁽⁵⁾ The study introduced single-axis fiber grating strain gauges for the use of nondestructive evaluation of existing structures. The benefits of these include a long service life and the fact that they can be installed in long gauge lengths, providing more accurate results. There was nothing in the study that related to remote or wireless monitoring; it focused on the sensitivity of the gauges as well as the installation requirements of working on an existing structure.

In this case, the bridge required structural strengthening in order to accommodate increased loads on the structures that were not expected at the time of construction. The bridge was strengthened using fiber-reinforced polymer (FRP) composites that did not alter the appearance of the bridge while still providing increased strength (see figure 12).⁽⁵⁾ The fiber grating strain gauges were installed by embedding them into saw cuts in the bottom of the bridge girders and on the outside of the adhered FRP coating (see figure 13 and figure 14).⁽⁵⁾

Figure 12. Photo. FRP wrap installation on bridge superstructure.⁽⁵⁾

Figure 13. Photo. FOS installation on bridge superstructure.⁽⁵⁾

Figure 14. Photo. FOS installation over FRP wrap on bridge superstructure.⁽⁵⁾

The Udd et al. study focused primarily on the monitoring of the bridge superstructure, but the FOSs could have been installed just as easily to the pile foundation of the bridge.⁽⁵⁾ This would have provided data showing how the bridge foundation reacted to the same loads that were visible in the data from the superstructure. The sensors proved to be sensitive. The gauges detected not only small cars crossing the bridge, but also, on one occasion, the effect of a single person running out to the center of the bridge, jumping up and down five times, and walking back off the bridge (see figure 15). Furthermore, gauges were easily installed by embedding them within the structure and applying them to the exterior of the structure with adequate results from each installation.

1 mi = 1.61 km
Figure 15. Graph. Measurement of strain induced on bridge from varying events.⁽⁵⁾

FOSs are helpful in a SSHM system because of their relative immunity to temperature effects. Typically, bridge foundations are designed with mass concrete elements such as drilled shafts or piles for the subsurface foundation, a shaft or pile cap, and large concrete columns. The temperature changes that can take place inside these mass concrete elements are large. Typical vibrating wire gauges show large frequency changes due to temperature that must be corrected when analyzing. Fiber optics results showed only the strain that is truly induced by temperature change in the structure and not that of the gauges.

A study by Hemphill examined the combining of wireless technology with FOSs.⁽⁶⁾ The study proposed and tested the idea of a fully integrated continuous wireless SHM system for the East 12th Street bridge in Des Moines, IA (see figure 16). Fiber bragg grating (FBG) strain sensors were installed at 40 different locations on the bridge. The data collector scanned the FBG sensors and transmitted the data wirelessly to a computer in a secure facility close to the site (see figure 17). The data were stored as a data file and automatically uploaded to a file transfer protocol (FTP) site. When this site was accessed, the data file was downloaded and deleted from the FTP site to make room for the next data file. These data were compiled, processed, and posted to a Web site that allowed users to view real-time strain data along with real-time streaming video of the bridge.⁽⁶⁾

Figure 16. Photo. East 12th Street bridge in Des Moines, IA.⁽⁶⁾

Figure 17. Photo. Host computer near East 12th Street bridge site.⁽⁶⁾

This system is useful because it provides the end user with simple, easy-to-follow data viewing that is easily monitored. With the addition of the real-time streaming video, a data monitor can review the data, compare them with the live traffic on the bridge, and make the necessary correlations to the loading on the structure. The wireless transmitting of the data is also useful because it reduces the man hours that are normally required to go to the site and download the data from the collection system, which can be time consuming and expensive. This system is efficient and has few drawbacks, if any. The FOS gauges can be installed in the substructure and on the superstructure, and there are no limiting factors to the system.

CURRENT AND FUTURE POSSIBILITIES FOR HEALTH MONITORING

Weyl summarized the proposal for a full-scale SHM system for the Indian River Inlet Bridge in Delaware.⁽⁷⁾ The design of the SHM system was fully integrated throughout the design phase of the project so that it fit seamlessly with the construction phase. The following types of gauges were installed throughout the bridge: vibrating wire strain gauges, weldable foil strain gauges, accelerometers, global positioning system sensors, load cells, linear potentiometers, corrosion monitors, etc. This creates a total of 240 sensors, 11 DASs, and 39 data loggers.⁽⁷⁾

The project will be carried out in three phases. Phase I took place during construction to determine live construction loads. Phase II will take place immediately after bridge construction to determine the initial response of the bridge to traffic, thermal, and wind loading. Phase III will take place during the intended service life of the bridge to compare against the data collected during phase II.⁽⁷⁾

Finally, a Web-based user interface was developed to present data in an easy-to-read and understandable format for the University of Delaware, the Delaware Department of Transportation, and those that worked on the project. At the time of this report, there were no data to report from this project because it was still in the preliminary construction phase.

This project provides an example of the future possibilities that SHM holds for the sustainability of the Nation’s infrastructure. Fully integrating the monitoring system into the design phase of the project does not delay construction or hold back the monitoring system. The data collected from this type system can be archived as data that are useful for the history of the bridge and that help with the determination of any possible problems that might occur.

This particular study involved a high number of sensors, gauges, and DASs for the full SHM system, but it is still similar to the proposed monitoring for the I-35W bridge monitoring system that is studied in this report. The use of everyday technology, such as a Web site that provides interested users with real-time data from the bridge, coupled with the advanced technology of resistance and vibrating wire strain gauges, will propel SHM systems into practice.