February 1996
Integrity of bridge supports, namely abutments, columns and piers, and foundations, can be diminished by scour, abutment and foundation movement, and seismic excitation, among other processes. Time domain reflectometry (TDR) analyses of signals pulsed through cables incorporated within the bridge structure can be employed to monitor suspected degradation of these supports. Two general approaches are described in this paper to locate and quantify damage of critical bridge support members using TDR. First, the overall stability of the structure can be assessed by monitoring movement of abutments and foundations with respect to the foundation materials. Second, columns and piers can be monitored for internal cracking. Both approaches utilize the same method of measuring deformation of metallic, coaxial cables caused by displacement or deformation of the surrounding medium.
These sensing cables require an electronic monitoring system to collect both the global measurements of external displacement and local measurements of internal deformation. It is important that this system function during critical events, such as earthquakes and floods, when major damage is most likely to occur. The necessary miniturized, low power consumption, puler for the monitoring system has been developed in conjunction with the U.S. Army Corps of Engineers and HYPERLABS.
Scour has been linked to nearly 95% of all severely damaged and failed highway bridges constructed over waterways in the United States (Lefter, 1993). The greatest loss of sediment to scour occurs at high water velocities, particularly during floods. Bridge pier movement can occur as a result of material loss beside and beneath the base of its footing, which produces undesired displacement of the bridge supports and may ultimately result in structural collapse. For these reasons, development of technologies to monitor scour and resulting structural movements is needed.
Recent earthquakes have demonstrated that many bridge columns and piers designed to resist earthquake shaking can be severely damaged yet remain standing. While physical damage can be obvious, often it is internal or obscured by exterior reinforcement. Thus it is important to quickly evaluate the integrity of many standing bridge piers immediately following an earthquake.
Time domain reflectometry (TDR) technology may offer a simple and rapid measurement technique for identifying external and internal changes of bridge piers and other supports with embedded coaxial cables and exterior electronics. This paper presents on-going investigations of installation techniques required for monitoring of bridge pier displacement and deformation. Cable placement in newly constructed columns and as part of a retrofit reinforcement program for existing columns is described. Early laboratory test results of detecting internal column damage during simulated earthquake loading are also discussed.
Time Domain Reflectometry Background
TDR is an electromagnetic testing technique originally developed to detect faults along power transmission lines by pulsing voltage along the line and searching for voltage reflections caused by cable discontinuities. Reflection travel times are measured from the reflection site to the pulsing electronics attached at one end of the cable. Knowing the pulse propagation velocity of the transmission line, the accurate location of cable faults can be accomplished by converting measurements from the time domain to the lineal distance domain.
The amplitude of reflected voltage, called the reflection coefficient, at a cable shearing fault can
be directly related to the magnitude of shear deformation (Dowding, Su, and O'Connor, 1989).
TDR signals shown in FIGURE 1 demonstrate that the reflection coefficient, measured as the percentage
of reflected voltage and expressed in 
, increases as the cable is sheared incrementally. These initial
findings have been extended by others to include extension and combined deformation modes. To
accurately determine the extent of cable deformation, a correlation between reflection
coefficient and deformation magnitude must be known for the selected cable(s).
A TDR measurement system provides several advantages for long-term monitoring of bridges. Previous deployment of TDR-based instrumentation suggests that it is robust enough to withstand emplacement during heavy construction and should perform during potentially damaging conditions such as floods and earthquakes. A miniaturized, low power, intelligent TDR pulser is currently being developed for remote monitoring. This signal pulsing and recording instrument is designed to reduce the size, complexity, and cost of current cable testing electronics. Remote access to multiplexed cables has been demonstrated successfully in the field with current electronics and existing telecommunication lines (Dowding and Huang, 1994). These advantages show that a survey of all bridge supports can be performed remotely through one central unit, saving time and reducing costs.
External Displacement: Bridge Abutment and Pier Movement
Measurements of scour at bridges founded on shallow footings indicate maximum scour occurs around the upstream side of footings during floods. As scour progresses and soil deposits are eroded, footings may be exposed and eventually undermined, leading to intolerable pier movement. Optimal placement of TDR cables would therefore be through the footing section on the upstream side. Most importantly, the cable monitoring system must operate at high water flows, when footing and pier displacement is most likely to occur as a result of scour.
Although bridge abutments may be less affected by scour, under certain circumstances they may need to be monitored for detrimental movements. Abutments and piers can be inspected using the same system designed to measure very small structural movements (on the order of 2 mm) relative to their foundation materials. Detection of such small displacements should provide an early warning of progressive movements which may reduce the stability of the bridge.
Several cable installation schemes have been investigated for monitoring pier movement (Dowding and Pierce, 1994b), but the most practical orientation of TDR cables is illustrated in FIGURE 2. A single cable is shown extending from an accessible location on or below the bridge deck, down along the pier, and into a hole drilled through the footing and foundation materials. The cable must be encased in grout through the hole to ensure direct transfer of soil-structure displacements. In addition, the cable should be enclosed by a protective pipe along the entire length of the pier to screen the cable from debris and weather. A more durable transmission line can be connected to the monitoring cable at the top of the pier and extended to the electronics.
The number and arrangement of TDR cables installed on an individual bridge pier will depend on many factors, including the scour-critical classification, risk associated with structural collapse, and installation and monitoring costs, among others. It is expected that a TDR system can be installed on new bridges, since cables can be easily incorporated into new construction. However, the complexity and cost of cable installation on existing bridges are significantly greater, and installation techniques to minimize these problems are currently being researched.
Movement of the pier footing can be detected from voltage reflections generated by local cable deformation. For instance, lateral translation or rotation of the footing would locally shear the cable-grout composite at the interface of the footing and foundation soils, as illustrated in FIGURE 2. A minimum cable deformation of 2 mm produces a distinguishable voltage reflection in the TDR signal. Further footing displacement progressively shears the cable and produces increasingly larger voltage reflections at that point along the cable, creating similar TDR signals to those shown in FIGURE 1.
To accurately measure these structural displacements, the cable and grout must be sufficiently weak to deform under the applied loads. A special cable-grout system is currently being developed to measure localized soil deformation in earth structures (Dowding and Pierce, 1994a). This project involves the design and production of an integrated cable and grout composite which exhibits low shear strength and compliance to deform uniformly with soft soils. In addition to embankments and excavations, this system can also be applied to monitor movements of footings or abutments founded on soft soils.
Internal Deformation: Bridge Column Cracking
Field observations of seismic failure of bridge columns in Northridge, California after the 1994 earthquake indicate shearing of the upper and lower connections occurred as a result of excessive lateral movements. In a recent field study, lateral loading of two-column bridge piers in East St. Louis, Illinois near the New Madrid fault revealed that tensile deformation occurs at the interface of the column and its base, and that vertical slippage or shearing occurs between overlapped, or spliced, steel reinforcement within the column (Lin, Gamble, and Hawkins, 1994). Splicing is a common construction practice which weakens the column against large moments that may develop during an earthquake. For this reason, remaining bridge columns are being retrofitted with external reinforcement along the spliced lengths to provide additional resistance.
To continue studies on the behavior of reinforced concrete bridge columns subjected to lateral loading, researchers at the University of Illinois constructed three model bridge piers, each with two columns per base beam (Lin, Gamble, and Hawkins, 1994). Each steel reinforced concrete column had a unique reinforcement geometry, but all exterior dimensions were the same. Columns were 12 ft high and 2 ft in diameter, with 6 ft center-to-center spacing on the beam. Steel reinforcing bars were spliced through the bottom 3 ft of each column. To investigate the response of the overlapping steel, columns were loaded under simulated earthquake conditions to failure.
To monitor internal deformation in the spliced region, four cables were attached to the inside of each reinforcement cage before columns were cast with concrete. After the columns were fully constructed, cables remained in the positions shown in FIGURE 3. Two cables were aligned vertically to primarily monitor tensile deformations at the column-base interface. The other two cables were installed transversely through the column diameters and perpendicular to the vertical cables to monitor shearing between the spliced reinforcement. The dashed lines in FIGURE 3 mark the cable segments encased in concrete; solid lines represent the exposed lengths of cable. Each vertical cable passes through the beam and exits on one side. Both ends of each transverse cable are exposed to allow pulsing from either end. Cable selection, preparation, and details of installation for these laboratory columns are described by Pierce and Dowding (1995).
Cables have also been installed on existing bridge columns during a seismic retrofitting project in East St. Louis, Illinois (Pierce, 1995). This work was performed solely to assess the feasibility of cable installation under rigorous field construction conditions, and as such no measurements are expected at this site, although cables are accessible if monitoring becomes warranted. The field demonstration was effective, showing that cables can be easily affixed to the exterior steel reinforcement and remain intact after concrete placement. Further studies in this area are anticipated, including optimization of cable attachment and arrangement, and monitoring cable response to lateral loading.
Laboratory columns were individually tested under reversed cyclic lateral loading to simulate earthquake conditions. Essentially, the free end (top) of the column was displaced laterally in controlled cycles, creating large bending moments at the column-base interface. Displacements were increased incrementally over time to failure, defined as a post-peak loading equal to 80% of the maximum measured load. A more thorough discussion of the testing procedure is given by Lin, Gamble, and Hawkins (1994).
To measure deformation, cables were pulsed when the column reached the maximum positive and negative displacements of each cycle. TDR signals were immediately collected and analyzed for signs of cable deformation, which only occurred during post-peak loading cycles. Not all of the cables deformed significantly to indicate internal damage, but several cables provided excellent results. Analyses showed that, in general, the vertical cables measured extension at the column-base interface, and the transverse cables sheared between the spliced steel as the bars slipped relative to each other, causing the interior concrete to shear and eventually crack.
In one column test, shear deformation was detected on the lower transverse cable at points A and B, shown in FIGURE 3. Tensile deformation occurred along one of the vertical cables at the column-base interface in another test, designated as point C. Reflection coefficients measured from these cable deformations are compared to simultaneous strain gage measurements of the nearest reinforcing bars, as illustrated in FIGURE 4. This plot shows that reflection coefficient, measured from both shearing and extension deformation modes, increases as the steel reinforcing bars were strained. The greater slope for cable shearing suggests that this cable was more sensitive to the strain environment than the vertical cable for these particular two tests. Tensile strain measured in the reinforcing steel does not necessarily correlate with cable shearing or extension, so a direct comparison cannot be made here. However, further studies can be conducted to quantify the reflection coefficient-cable deformation relationships for these cables.
The TDR response may be more meaningful if it were plotted against the load-deformation response of the column. FIGURE 5 shows the maximum load and displacement points for column D for each cycle of loading, in the positive and negative directions. Displacement measurements were taken at the top of the column, meaning these displacements are the maximum sustained by the column. The hatched band represents the area between global maximum load in the two directions.
Also shown in FIGURE 5 are the responses of two transverse cables, on located 1 m (3 ft) above the base (upper) and one located 0.15 m (0.5 ft) above the base (lower). One would expect damage to be concentrated in the region closest to the column connection to the base. These responses sugest that more damage occurs in the lower part of the column near the base. The lower cable was deformed at two locations, and the reflection coefficent are 4 to 12 times geater than the reflection measured on the upper cable. Lower cable response was initiated earlier, since first signs of measurable damage on the lower cable occurred one cycle before the upper cable detected damage.
Both cables did not respond until after global peak loading was achieved. This suggests that the column was behaving elastically up to peak, such that no permanent (plastic) damage or cracking occurred during pre-peak loading, as one would expect. When the cables were deformed, the points on each cable corresponded to locations within the lap splice. This suggests that cracking may have initiated in the concrete between the spliced rebars as the steel strained significantly to break the bond between them.
Response of vertically oriented cables is shown in relation to the load-deformation response of column F in FIGURE 6. Column F was similar to column D except there was more concrete cover over the exterior reinforcing bars. One of the vertical cables measured small reflections during the post-peak loading. These measurements correspond to a point on the cable located at the interface between the column and base. Reflections only occurred when the column was displaced in the positive direction, which placed this cable under tensile loading. Thus these results suggest that the vertical cable deformed in extension at the column-base interface, as expected.
TDR technology may provide a means to inspect bridge piers and abutments by measuring internal deformation and external movement. Installation of metallic, coaxial cables in newly constructed laboratory columns and retrofitted field columns verified the cable robustness needed for monitoring long-term performance. Early results from laboratory tests showed that cables deform in high strain environments inside reinforced concrete columns. These observations point to the need to investigate placing cables in existing piers for monitoring scour-induced movement.
This project has been supported by the Infrastructure Technology Institute (ITI) at Northwestern University. The assistance and cooperation of Dr. William Gamble and Yongqian Lin of the University of Illinois, and their support by the Illinois Department of Transportation (IDOT), are gratefully acknowledged.
Charles Dowding
Charlie Pierce
