Prepared by I.N. Komsky, J.D. Achenbach, Center for Quality Engineering and Failure Prevention
and P.J. Stolarski, California Department of Transportation
STRUCTURAL MATERIALS TECHNOLOGY
-- An NDT Conference
Edited by Robert J. Scancella, Mary Ellen Callahan
Sponsored by the New Jersey Department of Transportation, the Federal Highway Administration
Held February 23-25, 1994 in Atlantic City, New Jersey
Copyright © 1994 Technomic Publishing Co., Inc.
ABSTRACT
Bridge inspection statistics nationwide show that more than a third of the United States' half million highway bridges are either "structurally deficient" or "functionally obsolete". As for steel bridges, corrosion, fatigue damage, and brittle fracture are among the major problems. Inadequate design and wear of a bridge structure can produce high concentrations of stress which may result in development and growth of cracks. There is a high probability that crack growth in a critical bridge component such as a pin connection will lead to catastrophic failure. A number of pin failures have happened over the last several years, some with catastrophic consequences. These failures prompted the Federal Highway Administration to require evaluation of all pinned connections in bridges throughout the country. The final report to the Missouri Highway and Transportation Commission on "Investigation of Bridge Pin Failures" [2]. However, immediate replacement of a pin after the detection of a crack of any size may not be desirable due to high cost of the replacement. Cost effective maintenance of bridges requires implementation of the so called "retirement-for-cause" concept, which means that the component (pin) will be replaced only if the flaw is larger than an allowable size. Application of this concept requires periodical inspection of a pin connection for quantitative measurement of crack sizes.
Currently available ultrasonic techniques do not have the capability of accurately measuring crack size and orientation because of unpredictable coupling between the ultrasonic contact transducer and the pin surface. To remove this unpredictability it has been suggested to keep the ultrasonic transducers permanently attached to the pins, which is however a highly expensive approach since a large number of pins must be inspected.
In this paper the self-compensating ultrasonic technique, previously developed by two of the authors for the evaluation of aircraft structures [3] requires application of two transducers positioned on different sides of the inspected area. The transducers are used in both the pulse-echo and the through-transmission mode, which means that ultrasonic waves are generated by both transducers, partially reflected from and transmitted past the crack, and finally received by the transducers. However, this transducer configuration is not effective for field inspection of bridges, particularly for application to pin connections. Hence a modified one-sided technique for a self-compensating measurement has been developed. A schematic of the modified self-compensating technique, is shown in FIGURE 2.
An ultrasonic transducer, placed on the front face of the pin, is fired. This transducer used in the pulse-echo mode produces a beam of ultrasound which is partially reflected by the crack and partially transmitted. The signal reflected from the crack generates a voltage which in the frequency domain may be expressed as
The signal, transmitted past the crack and reflected from the back face of the pin, is also received by the transducer with voltage
The lines with arrows in FIGURES 2 and 3 are not actual rays. Rather they indicate the general paths traveled by the beams of ultrasound.
Next we consider the ratio VR / VT. It then easily follows that
The purpose of measuring 
is to obtain quantitative
information on the depth of the crack. Specific values of the
reflection coefficient, Rc, and the transmission coefficients,
TlC and T2C, are indeed representative of a specific crack depth,
and hence a measured value of the ratio
would also
determine the depth of a crack. Unfortunately Eq. (4) also contains
the constant K which is unknown and can not be measured independently.
To retain the usefulness of Eq. (4), the voltage ratio 
must therefore be compared either with synthetic data obtained
by analytical and numerical methods for the crack-pin configuration
or with calibration data obtained with separate but identical
pins containing cracks of known lengths. The calibration experiments
would be carried out in exactly the same manner as described above
and would yield for cracks (possibly EDM notches)
of known depths. A comparison of a field measurement of
with a calibration measurement would give the depth of the
crack, independently of the coupling between transducer and pin
in either the field test or the calibration.
EXPERIMENTAL CONFIGURATION
The efficiency of ultrasonic bridge inspection depends to a high degree on the portability of testing equipment. To meet this requirement the selfcompensating technique has been adjusted to utilize pulse signals from commercially available flaw detectors instead of previously used tonebursts [3]. A portable ultrasonic flaw detector EPOCH II (Panametrics, Inc.) was used to generate high voltage pulses which were converted into ultrasonic longitudinal or shear waves by a contact transducer (AEROTECH MSW-QC 2.25MHz and 5MHz, Krautkramer-Branson). Longitudinal or shear wave techniques have been selected for each particular application based on pin size and configuration. Ultrasonic waves were received by the same transducer after being reflected from the crack and back face of the pin. The output signals were then digitized by the ultrasonic flaw detector and the data was subsequently acquired by a personal computer. For automatic discrimination and measurement of the signals the SQL software module from PC-based TestPro system (Infometrics, Inc.) has been used. The ratio of the signals reflected from the crack and the back face was calculated using multiple time domain gates.
APPLICATION: PINS WITH WEAR GROOVES AND ARTIFICIAL FLAWS
The modified self-compensating technique has been tested to detect and characterize cracks in the form of saw-cuts in middle parts of pins. Two saw-cuts of maximum depths 0.5" and 0.25" have been made on opposite sides of a pin. The width of the saw-cut was about 0.02". An ultrasonic transducer was placed on one of the pin faces with a layer of a coupling fluid. The position of the angle-beam 2.25MHz transducer was selected so that the signals reflected from the saw-cut and from the back face could be detected and measured. FIGURE 3 shows rectified waveforms acquired by the ultrasonic flaw detector and the data processing system. The time domain gates used for the signals discrimination and analysis are also shown in FIGURE 3. Despite differences in the responses obtained from the saw-cuts they cannot be directly used for the quantitative determination of the flaw sizes. The calculation of the ratio of the voltages defined by Eq. (4) yields the values of 1.16 and 0.77 for the 0.5" and 0.25" saw-cuts respectively.
Reflections of ultrasonic waves from wear grooves result in false crack indications and unnecessary replacement of the pin. To eliminate false indications from wear grooves it was proposed by the authors to apply ultrasonic waves of a higher frequency. Application of straight-beam 5MHz transducer made it possible to detect a 0.125" saw-cut made in the 3" diameter pin. On the other hand even the deeper wear grooves did not produce detectable responses.
Classification of the cracks can be performed by comparing experimental results from actual structures with the signals measured on calibration specimens or with signals generated by theoretical modeling. Development of a set of calibration specimens with different sizes of flaws can be expensive and time-consuming due to the large variety of pin configurations. Therefore, physical calibration specimens should be machined for only a few crack sizes and pin configurations. On the other hand a numerical approach suggested in [4] makes it possible to calculate forward-scattered and back-scattered fields for the incidence of body waves on a surface-breaking crack. The numerical approach has been experimentally tested to determine the angle of inclination and the depth of the surface-breaking crack from ultrasonic measurements . A neural network has been used to classify cracks using data acquired by an immersion ultrasonic system. Although the backscattering data have only been utilized in the calculations the same approach can be applied using R/T measurements as shown in [5] for an application with plate waves.
CONCLUSION
A self-compensating ultrasonic technique has been applied to bridge components. The technique and transducer configuration were modified for one-sided inspection. The electronic configuration and the data acquisition system have been adjusted to provide a portable automated system for field applications. The ultrasonic system was tested on a specimen with saw-cuts of different sizes and locations which model cracks in the side wall of the pin. The feasibility of computerized data acquisition and processing for discrimination and measurement of ultrasonic signals has been discussed. An application of theoretical modeling and physical specimens combined in one calibration set is recommended to obtain information on crack sizes.
ACKNOWLEDGMENT
This work was supported by the Infrastructure Technology Institute of Northwestern University. Helpful discussions with Phil Fish are gratefully acknowledged. The authors would like to thank the Wisconsin Department of Transportation and the Illinois Department of Transportation for supplying the pins used in this work.
REFERENCES