Advanced nondestructive evaluation (NDE) and nondestructive flaw-detection technologies (NDT), developed partly through the assistance of a recent Federal Highway Administration (FHWA) applied research program, played a vital role in a successful emergency structural evaluation program undertaken by the Alabama Department of Transportation (DOT) to assess potential wind-induced damage to a major cable-stayed span in Mobile.

The Cochrane Bridge on U.S. Route 90 and state Route 16 in Mobile County is a twin-pylon, cable-stayed bridge with a main span of 238 meters and two side spans of 110 meters. There are two planes of cables at each pylon with a semi-harp arrangement. Forty-eight cables are anchored at each pylon for a total of 96 stay cables on the bridge. Cable sizes range from 26 to 72 strands, each supporting loads between 184.44 and 725.75 metric tons.

On Feb. 26, 1998, a number of stay cables on the Cochrane Bridge unexpectedly began vibrating at high amplitudes during a period of light rain and high winds. According to a report by the Alabama DOT, sustained winds of 46 kilometers per hour (km/h) with gusts up to 64 km/h were recorded at Bates Field in West Mobile that day.

This rain/wind vibration phenomenon has affected a large number of cable-stayed bridges worldwide. In the United States, the Clark Bridge (Illinois), Burlington Bridge (Iowa), Houston Ship Channel Bridge (Texas), Weirton-Steubenville Bridge (West Virginia), and East Huntington Bridge (West Virginia), among others, have experienced the problem. Japanese bridge engineers have reported similar problems, and more recently, the new Erasmus Bridge in Rotterdam, The Netherlands, was affected.

Formation of water rivulets on the cables is believed to be the cause of this aerodynamic instability. As primary load-carrying members of cable-stayed bridges, stay cables are arguably the most important and crucial elements of the entire structure. Therefore, such high-amplitude vibrations can affect the strength and useful life of the cables and the bridge.

Shortly after being approached by Alabama DOT's bridge engineering and inspection consultant A.G. Lichtenstein Inc., the stay-cable assessment team from Construction Technologies Laboratories Inc. (CTL) devised and implemented a comprehensive test program to assess the presence and extent of stay-cable damage. This field assessment program was the first of its kind in terms of scope, comprehensiveness, and speedy response to be conducted on any cable-stayed bridge span in the United States. The following technologies were adapted to resolve the DOT's concerns:

Collaboration Advances State of the Art

The laser-based cable force measurement procedure was conceived at the CTL Structural Engineering Laboratory. Its development was completed in 1997 and supported by the funding of FHWA's NDE research initiative. Dr. Habib Tabatabai and Dr. Armin Mehrabi of CTL devised this technology, which relies on a small laser sensor positioned at bridge deck level. The laser beam is targeted at the mid-length region of the cable from a distance of up to about 150 meters. Minute ambient, cable vibrations created by normal wind speeds are measured with the laser sensor, and the cable's frequency spectra are analyzed. Tabatabai and Mehrabi developed the mathematical formulations that relate a cable's vibration signature to the tension force it supports, accounting for the complex interactions of parameters such as cable mass, length, bending stiffness, external damping characteristics, and other factors. The mathematical algorithm, unlike any other previously available, permits the rigorous incorporation of these cable parameters, enabling much more accurate damage assessments. The method is truly non-contacting because the impacting or the plucking of the cable is not required nor do laser targets need to be placed.

Force changes among cables in an array on a bridge can reveal changes in cable condition attributable to construction activities, corrosion-induced deterioration, long-term bridge foundation settlement, accidental bridge impact, superstructure modifications, and the like. Previously available procedures were either not accurate enough for meaningful structural evaluation, involved the use of the prohibitively costly and complicated lift-off methods (using construction staging, hydraulic jacks, and calibrated pressure gauges), or necessitated installation of expensive load sensors during construction.

This technological advance is of great benefit to assessing the condition of the transportation infrastructure and permitting accurate measurement of cable forces of between 20 to 50 stays per day, employing only two individuals. The available lift-off methods would take a day or more to conduct for each cable. Thus, what may have taken 100 days or more to accomplish using older, less accurate techniques can be performed in less than one week on a typical U.S. cable-stayed bridge!

The nondestructive evaluation concept was successfully verified and affirmed to be repeatable in the laboratory on 1/6th-scale cables. The method was field-verified on the Weirton-Steubenville Bridge over the Ohio River. Good agreement was achieved between the predicted values and the available load cells on the Weirton-Steubenville cables. The sum of predicted forces was within 1 percent of the total of all cable forces calculated by the designer. CTL also achieved excellent agreement between predicted and design forces for cables of the Jindo Bridge in Korea; the Jindo Bridge was instrumented with accelerometers as part of a separate automated structural surveillance project.

The new laser vibrometer system is a product of a collaborative research and implementation effort between FHWA and CTL. Other parties who contributed to this development included the departments of transportation of Florida, Iowa, Louisiana, Texas, Virginia, Washington, and West Virginia. They provided stay-cable database information so that the analytical algorithm relating vibration to cable stay forces could be refined to reflect the stay-cable details for all bridges in the United States. Also, stay-cable suppliers CCS Special Structures Inc., Dywidag Systems International USA Inc., Freyssinet, and VSL Corp. provided specific information about their stay-cable systems and donated materials for the laboratory research.

Cable Force Measurements Support Safety Assessment

Forces in all 96 cables of the Cochrane Bridge were measured in 3½ days. A laser vibrometer was stationed at different positions on the deck (near each pylon), and the laser beam of the vibrometer was targeted at each cable to record vibration histories. To the extent possible, the laser beam was aimed perpendicular to the cable chord. The targeting point along the cable was at approximately the third point to measure ambient vibrations. The velocity-time histories were recorded for all cables. The time histories were analyzed to identify the vibration frequencies of stay cables using a Fast Fourier Transformation (FFT) routine.

The first mode (in-plane) vibration frequencies of stay cables of the Cochrane Bridge were used to calculate the tension force in the cables. The bridge drawings, the cable manufacturer's shop drawings, and field inspection results were used to determine the geometry and mechanical properties of cables as well as the cable-end and neoprene washer conditions. As stated earlier, the new relationship developed by CTL considers all of the important parameters influencing the vibration characteristics of cables, thus providing far better accuracy than would be expected from using a simple string equation.

Live loads were obtained from the Manual for Inspection and Maintenance of the Cochrane Bridge (1991). The Guaranteed Ultimate Tensile Strength (GUTS) for each cable is calculated by multiplying the number of strands times the area of each strand (1.4 cm2) and the nominal strength of the strands (1.9 MPa). The maximum allowable service load in each stay cable is 45 percent of GUTS, according to the 1993 recommendations of the Post-Tensioning Institute (PTI).

The lift-off forces immediately after construction and the anticipated values 4,000 days later (to account for the creep and shrinkage effects) were also reported in the Manual for Inspection and Maintenance of the Cochrane Bridge. Calculated and lift-off forces are, in general, in very good agreement. As a measure of the accuracy of the measurement method, the percentage difference between the total measured forces in all cables and the total anticipated forces (lift-off plus creep and shrinkage) is less than 2 percent. The forces calculated for Cables E1NW and M1NE are lower than the lift-off forces reported. However, the forces in the cables adjacent to these cables do not show any major increase with respect to the original lift-off forces.

On the basis of CTL's measured cable force distribution analysis, it was concluded that seven years of rain/wind-induced cable vibrations had not contributed to degradation in stiffness of the cable to the extent that measurable load redistribution had occurred.

Anchorages Tested for Strand Breakage

As an important complement to the stay-cable structural evaluation protocol, an ultrasonic flaw detection technique was adapted to conduct a complete examination of the stay anchorages. The goal of the nondestructive flaw-detection examinations was to locate possible wire breaks, resulting from cable rain/wind vibration, within the epoxy-grouted portion of 12 selected anchorages. The achorages ranged in size from 32 to 72 strands of 1.52-centimeter diameter.

Each strand is locked at the end of the anchorage by wedges in a wedge plate. The anchorage is grouted with an epoxy grout filled with steel balls. The free ends of the cable strands are enclosed in a steel drum anchor cover, and the wires in the drum are encapsulated in thick grease to prevent corrosion. For this testing program, the anchorage covers were removed, and the ends of each wire strand were ground to a level, smooth surface.

It was observed that most of the anchor covers at deck level contained water, and that the grease around the wire strands had emulsified to a cream-colored paste. In addition, epoxy grout had seeped through the wedge plate at the time of grouting, often encapsulating the free wire strands.

The ultrasonic test method was verified and refined through an initiating development program in CTL's Structural Engineering Laboratory. Similar anchorage components left over from CTL's full-scale fatigue tests of the Burlington, Iowa, cable-stay anchorages were used in the development and calibration of the method. In simplied terms, an ultrasonic pulse is sent into the free end of each strand, and the resulting amplitude-time signal was visualized on a computer screen.

Anomalies such as the anchorage wedges, constricting plates, changes in grout conditions, or wire breaks that reflect ultrasonic energy can be identified on the screen of a flaw detector device. Echoes from the wedge plates were observed in most of the test traces, and echoes from the back plate to the anchors were often observed.

Cut or broken wires in an anchorage sample were tested in the CTL Structural Engineering Laboratory. No anomalies of this type were obtained from the anchors at the Cochrane site. Nearly all test responses identified constrictions at the wedges, and some show echoes from bearing plates at the backs of the anchors. From the amplitude of these pulse echoes at the bearing plates, it was deduced that the signal penetration was at least 1.52 meters along the cables.

Results of the ultrasonic flaw detection corroborated the stay evaluation team's positive findings regarding the structural condition of the stay cables gleaned from stay-force distribution analysis. The only anomalies registered in these tests were from one deck-level stay anchorage at the northeast tower, where three strands exhibited diffuse echoes from 76 to 102 millimeters behind the wedges. These are probably a result of incomplete grouting or voiding immediately around the strands at this depth. The diffuse nature of the echo was not representative of a cut or a break in the wires, which would appear as a much sharper pulse echo on the test response trace.

Rain/Wind Vibration-Control Design

The FHWA NDE research developed while adapting the capabilities of the laser vibrometer to in situ stay-cable structural evaluation, produced other bridge diagnostic methodology benefits. Rapid, accurate field cable-damping measurements are now possible. The utility of the laser sensor for making precision dynamic measurements was taken advantage of on the Cochrane span and integrated with the other resources of the cable evaluation team.

A stay cable's damping characteristics help diffuse rain/wind vibration effects. A bridge cable's intrinsic damping values are a function of a large number of parameters. Design guidelines in the January 1998 draft of the Post-Tensioning Institute's "Recommendations for Stay-Cable Design, Testing and Installation" suggest a wide range in damping ratios - from 0.05 percent to 0.5 percent. The internal vibration diffusion mechanisms within strand wires, between wires and grout, or in the polyethylene sheathing, etc. are highly variable depending on each cable's construction. The magnitude of force in the cable, the cable size, its length, and the force at which the cable was grouted are among the potential factors that could influence the resulting cable damping. The stay-cable evaluation team used the same laser vibrometer sensor adapted for cable integrity evaluation to assess the need for rain/wind vibration corrective measures.

Specific procedures were developed to measure cable-damping ratios for all stays of the Cochrane Bridge. First, a long, nylon strap was wrapped around the cable (near the mid-point of the cable or as high as possible with a 24-meter lift truck). A release mechanism and a rope were attached to the strap. The rope was tensioned with a ratchet system attached to the deck. More than a thousand newtons were applied. Second, while the laser vibrometer was aimed near the cable attachment, the release mechanism was engaged to suddenly free the cable from the rope, and the resulting velocity-time data were recorded. The resulting velocity-time histories were analyzed for calculation of cable-damping ratios. Testing of all 96 stay cables was completed in four days.

The free decay method was employed for calculation of damping ratios. By analyzing the decay patterns, the cable-damping ratios and other damping values, called Scruton numbers, were calculated. Cable-damping ratios varied from 0.9 percent to very negligible. The Scruton numbers for the majority of the cables (86 percent of the cables) were less than 10. Cables with Scruton numbers less than 10 are reported to be vulnerable to rain/wind vibration phenomena.

Based on these test results, it was recommended that mechanical viscous dampers designed to eliminate rain/wind-induced vibration be installed. These measures will be implemented on the Cochrane span's upcoming maintenance cycle.

In brief, the direct and indirect products of the FHWA-sponsored research produced a number of experimental and mathematical procedures that constitute a new and comprehensive stay-cable evaluation protocol for major bridge structures and that helped rapidly diagnose and correct the specific cable vibration problems of the Cochrane Bridge.

Adrian T. Ciolko is CTL's program manager for FHWA contracts, and he directs the operations, technical activities, planning, and management of the CTL Structural Engineering Laboratory in Chicago. Together with Dr. Habib Tabatabai, CTL's principal investigator for the laser vibrometer applied research program, he conceived the idea and approach for applying noncontacting sensor-based NDE technology to bridge stay cables. He has worked for CTL for 22 years. He is a registered professional engineer in Colorado, Illinois, and New York.

Dr. W. Phillip Yen is a research structural engineer in the Structures Division of FHWA's Office of Research, Development, and Technology, and he served as a technical monitor for the laser vibrometer system project. Yen is FHWA's representative in the National Earthquake Loss Reduction Program, and he is a technical committee member of the National Seismic Conference on Highways and Bridges. He received his bachelor's degree in civil engineering from the National Taipei Institute of Technology in Taiwan and his master's degree and doctorate in applied mechanics and civil engineering from the University of Virginia. He is a registered professional engineer in Virginia.