Project Leader: A.V. Clark
T.L. Anderson

MSEL Program Nondestructive Evaluation	MRD Focus Technology Ultrasonic Characterization
Strategic Thrust Measurement Base and Standards	Character of Research Infrastructure

Technical Description

Acoustoelastic measurements allow the stress in a structural member to be determined in a nondestructive manner with a sensor that can be taken into the field where the member forms part of a major engineering structure such as a building, a bridge, or a crane. As the structure ages or is stressed by unusual events like earthquakes or accidental overloads, acoustoelastic measurements at a few carefully chosen locations can tell whether some parts are now supporting more or less than their design loads. Thus, acoustoelastic techniques applied at a few, judiciously chosen locations may be able to assure the overall or global integrity of a large structure.

Technical Objectives

The objective of this program is to apply acoustoelastic techniques to a specific class of structures where the basic concepts of Global NDE can be demonstrated. Steel bridges have been chosen because the beams are large, and usually accessible, and have reasonably well defined design loads. Furthermore, many of the approximately 500 000 bridges in the US have structural defects but cannot be replaced for economic reasons. As a result, the bridge maintenance community is moving toward inspection techniques that yield a quantitative evaluation of the overall structural integrity of bridges. We expect that properly executed acoustoelastic measurements will be able to satisfy this Global NDE need. To be most effective in the case of steel bridges, the acoustoelastic measurements should be guided by fracture mechanics principles. Therefore, a secondary objective of the program is to develop a rational analysis for the effect of crack-like defects on bridge support beams.

FY95 Accomplishments

Fracture Mechanics of Bridge Girders. In this study, we considerd the effects of three common failure mechanisms in bridges: brittle fracture, fatigue, and plastic deformation. We also performed a comprehensive analysis of how these failure mechanisms affect bridge performance and used probabilistic methods to address questions raised by variability in the bridge material's resistance to fracture.

Brittle fracture is characerized by the stress intensity factor K. When K exceeds K c (the critical stress intensity factor), fracture occurs. We calculated K for the case of a crack in the flange as well as the case of a crack in the web of an I-beam. Thus, we can follow the evolution of K for a crack starting in the flange and propagating into the web of the beam. Such a crack has two possible ill effects: (1) loss of stiffness, which may break the deck and (2) redistribution of the load to stiffer members, which may overload them. We used our calculated values of K to determine the loss of girder stiffness from cracking and found that there was negligible stiffness change until the girder was about 80% fractured. A structural analysis using this change in stiffness revealed negligible load redistribution until the girder was about 90% fractured. For cracks deeper than 90%, load shedding occurs so K finally vanishes and there is no stress to drive the crack. Hence, in principle, total fracture should not occur.

However, such deep cracks may continue to grow due to fatigue. This kind of damage would be caused by heavy truck traffic which induces stresses above the constant amplitude fatigue limit (CAFL). In conventional fatigue-damage monitoring, the effective cyclic applied stress is determined by measuring the various stress levels resulting from bridge vibrations and converting them to an equivalent constant amplitude stress, S. The American Association of State Highway and Traffic Officials (AASHTO) has recommended the use of plots of S versus N to estimate the number of fatigue cycles needed to produce failure. These AASHTO plots are quite conservative, so that when a given bridge reaches its nominal fatigue design life it is likely to have sustained only moderate damage. During FY95, we developed an algorithm to calculate fatigue crack growth for various spectra of expected stress ranges using the case of a plate welded to the flange. This type of welded detail increases the beam stiffness but it is the least resistant to fatigue damage. We found that when the nominal fatigue life was reached, about 50% of the cracks are only slightly larger than their initial size. Even after these cracks extended through the flange, the remainder of the beam could still carry a load. To describe failure when cracks were confined to the web, we included the possibility of plastic deformation and used the failure assessment diagram (FAD) approach employed in the petrochemical industry. This approach allowed us to characterize the probability of failure for steels with varying toughness and yield strengths. We found relatively little improvement with higher yield strength steels and about a 50% increase in fatigue life with higher toughness steels. However, our calculations indicated that the number of fatigue cycles for the crack to completely penetrate the flange was several times larger than for failure due to a web crack. Hence, we expect only moderate improvement in performance by introducing more advanced steels to current bridge design practices.

Acoustoelastic Measurements. In parallel with the analytical studies, we performed field tests to measure actual stresses in highway bridges. Such stresses are conventionally measured with strain gages, which are time-consuming to install and require special precautions when lead- based paint is removed. They also measure only changes in stress and cannot detect residual stresses or the dead-load stress from the weight of the structure itself. In contrast our ultrasonic method employed noncontacting transducers which generate sound directly in the steel and require no paint removal. Furthermore, the transducers employ permanent magnets and are easily attached and removed. One type of these electromagnetic transducers (EMATs) was used to generate and receive polarized shear waves propagating through the thickness dimension of the web. The small difference in sound velocity between the two polarizations caused by the presence of stress was measured. The second type of EMAT used measured the change in stress in the flange of the I-beam when the load on the bridge changed. By using both of these sensors, liveloads (dynamic stresses) and static loads could be measured. The liveloads were measured in a bridge as a test vehicle was driven over it at a range of speeds. Good agreement with strain-gage data was obtained. Static loads (including those due to pre-stressing during fabrication) were also measured on the same bridge and compared with the stresses expected from the design. These field tests were performed in collaboration with personnel from the Virginia Transportation Research Council (VTRC) and the Constructed Facilities Center of West Virginia University.

Additional measurements were made on a second bridge at the request of the Virginia Department of Transportation (VDOT). This bridge was of novel design in that it had no expansion joints and thus eliminated problems of bearing corrosion caused by salting. Thermal expansion causes this bridge to push against backfilled soil at the abutments, with attendant compression of the girders. If this compression is too large, either plastic deformation or buckling can result. VTRC personnel installed strain-gage instrumentation on the bridge at the time of construction and monitored its status for several years. Anomalous readings were recorded at one abutment that indicated potentially dangerous stress. Our ultrasonic measurements made on several girders at opposite sides of the bridge indicated that little stress difference existed. Thus, the bridge was judged to be safe. Subsequent replacement of suspect electronics in the VTRC instrumentation confirmed this result. VTRC personnel returned to this bridge on two further occasions for further monitoring. The ultrasonic data indicated good repeatability and agreement with VTRC instrumentation.

These field tests with EMATs demonstrate that NIST has successfully completed two phases of technology transfer; training of VTRC personnel in use of the ultrasonic equipment, and delivery of transducers and associated electronics to VTRC.

FY95 Outputs

1. "Application of Electromagnetic-Acoustic Transducers for Nondestructive Evaluation of Stresses in Steel Bridge Structures," M.G. Lozev, A.V. Clark, and P.A. Fuchs, to be published as a Virginia Transportation Research Council Report.

2. "New Approaches to Life Assessment in Steel Bridges," T.L. Anderson and A.V. Clark, NISTIR, in review process.

3. "Quantitative Bridge Safety Assessment Utilizing Fracture Mechanics and Ultrasonic Stress Measurements," A.V. Clark and T.L. Anderson, to be published in proceedings of Structural Materials Technology NDE Conference to be held Feb. 20-23, 1996 in San Diego.

4. "Monitoring Bridge Fatigue Loads with Ultrasonic Transducers," P.A. Fuchs, A.V. Clark and S.R. Schaps, SENSORS Magazine, Vol. 12, No. 11, 1995, p.20