Acceptance Criteria of Complete Joint Penetration Steel Bridge Welds Evaluated Using Enhanced Ultrasonic Methods (2019)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4 Background 1.1 Problem Statement and Research Objective Reliable detection of internal weld flaws using any non­ destructive testing (NDT) technique is essential to ensuring the desired performance of a structure. Presently, two NDT methods are used for evaluation of complete joint penetration (CJP) groove butt welds in steel bridges: radiographic testing (RT) and ultrasonic testing (UT). Using RT, discontinuities are dis­ tinguished from sound weld or base metal based on contrast variations that appear on a radiographic film. UT, on the other hand, utilizes reflections from high­frequency sound waves to inspect for internal discontinuities within the weld and base metal. While RT can reliably identify volumetric discontinui­ ties, this method is typically not as effective for thin planar discontinuities such as cracks and lack of fusion. UT typically launches the sound waves at an angle into the material in such a way that planar discontinuities can be readily detected. Advances in ultrasonic methods, including the develop­ ment of phased array ultrasonic testing (PAUT), provide enhanced ability to detect and characterize weld flaws, per­ form automated data collection, and generate images of ultrasonic results. Although improvements have been made to the ultrasonic equipment, the current acceptance criteria for PAUT in the 2015 edition of American Welding Society (AWS) D1.5 Bridge Welding Code [1] provided in Annex K are not based on the criticality of a weld discontinuity on bridge performance measures such as the resistance to fatigue and fracture. Rather this acceptance criteria is a workman­ ship criteria meant to provide an arbitrary control on the level of quality. In other words, the acceptance and rejection criteria are not related to the structural performance of the weld in any way. Because of the apparent “good experience” with RT, thresholds for flaw rejection using conventional UT were “calibrated,” though not systematically, to criteria used traditionally for RT that also were not based on structural performance [3, 4]. While this approach may seem reason­ able at first glance, it is absolutely critical to recognize that RT and UT are totally different approaches for flaw detec­ tion simply due to the physics associated with the technolo­ gies. One technique responds to changes in density, which is recorded on a 2­D film, while the other measures reflection of sound in both the amplitude and time domain. Hence, the common statement that “UT or PAUT can be used to see the same flaws as RT” is incorrect. Described herein is a summary of the proposed research for the development of updated acceptance criteria for evalu­ ation of CJP welds in steel bridges using enhanced ultrasonic testing methods. The proposed research included both ana­ lytical and experimental programs. The analytical testing program included parametric studies of various plate geom­ etries, welds, residual stress fields, flaw types, and locations. The analytical results were intended for establishing accep­ tance criteria and identifying the minimum flaw size that must be detected reliably. The experimental testing program had two components. The first consisted of developing pro­ cedures for characterizing and sizing flaws in CJP welds using enhanced ultrasonic technologies. This included establishing practical acceptance criteria that could be incorporated into the AASHTO/AWS D1.5 Bridge Welding Code and guidelines on their use. The second component consisted of a “round robin” exercise in which plates with known flaws would be circulated to a number of testing firms in order to evaluate the proposed procedures and criteria. The primary objectives of NCHRP Project 14­35 were to use data collected through analytical studies and experimen­ tal testing to (1) develop guidelines to evaluate complete joint penetration welds in steel bridges based on updated accep­ tance criteria and (2) develop proposed modifications to the existing AASHTO/AWS D1.5 Bridge Welding Code. As a minimum, the guidelines were to cover shop and field fabrication and in­service evaluation and include procedures for using enhanced ultrasonic testing methods to evaluate CJP welds in steel bridges and pertinent acceptance criteria. C H A P T E R 1

5 1.2 Scope of Study The research conducted for NCHRP Project 14­35, “Accep­ tance Criteria of Complete Joint Penetration Steel Bridge Welds Evaluated Using Enhanced Ultrasonic Methods,” focused on the application of PAUT to inspect steel bridge welds. Accep­ tance criteria, scanning procedure requirements, calibration requirements, and technician qualification requirements were developed using analytical and experimental methods. The analytical methods included finite element models, fitness­for­ service (FFS) models, and ultrasonic models. The experimental methods included measurement of acoustic material properties and ultrasonic testing of weld flaw specimens representing typical bridge weld geometries. These specimens included embedded and surface discontinuities representing both planar and volumetric discontinuities. The ultrasonic testing of weld flaw specimens included blind round robin testing by outside conventional UT, PAUT, and TOFD technicians along with inspection of the weld flaw specimens using RT. Weld flaw specimens will also be used to perform final verification tests of the proposed modifications to the PAUT inspection procedure. 1.3 Phased Array Ultrasonic Testing (PAUT) Since phased array relies on the same basic physics as con­ ventional UT to generate and receive ultrasound, many of the details of PAUT inspection do not change from conventional UT. However, unlike the single element transducer used in conventional UT, PAUT uses multiple element transducers and electronic time delays to generate and receive ultrasound. The electronic time delays use constructive and destruc­ tive interference that allow the ultrasonic beam to be steered, scanned, swept, and focused electronically. Figure 1 shows the electronic time delays for 16 active elements of a 64­element transducer (i.e., elements 1–16 active) in order to produce a 40° incidence angle (left) and 70° incidence angle (right). The array in the transducer can be constructed from a linear array, a two­dimensional matrix array, or a circular array. Linear arrays are used for most applications since they are cheaper than more complex arrays and easier to program [5]. Phased array probes commonly have between 16 to 128 elements. Focal flaws are calculated by the software, which controls the time delays and firing sequence of the transducer. The frequency of PAUT is very similar to conven­ tional UT, typically between 2–5 MHz for bridge weld testing. Two types of scans are typically used for PAUT: • Electronic scans (E­scans) are performed by multiplexing the same focal flaw along a linear array. This will produce a scan which is similar to manual scanned conventional UT. • Sectorial scans (S­scans) are performed by altering the time delays as the elements are fired, which creates a beam which sweeps through a range of incidence angles. PAUT can utilize encoded scanners to capture a continuous stream of data from different transducer positions, either auto­ matically or semi­automatically. Semi­automatic scanning— using a wheel or string encoder attached to the transducer—is typically utilized for bridge welds due to the variation in geometry associated with bridge fabrication. PAUTs using Electronic time delay for 40° incidence angle Electronic time delay for 70° incidence angle Figure 1. PAUT time delays.

6 encoded scans have multiple views that can be displayed to the technician, including the following: • A­scan (x–y plot of amplitude vs. time for a single beam; top left of Figure 2) • B­scan (end view when volume corrected) • C­scan (top view when volume corrected; bottom of Figure 2) • D­scan (side view) • E­scan (end view of all A­scans when multiplexing same focal law) • S­scan (end view of all A­scans for a range of incidence angles, top right of Figure 2) PAUT calibration involves correction of the wedge delay and sensitivity calibration. For conventional UT, sensitivity calibration involves measuring the reference amplitude of a standard 1.5 mm diameter (0.06″) side­drilled hole (SDH) reflector on an IIW­type calibration block; material attenu­ ation at other sound paths is accounted for by implement­ ing a correction through the attenuation factor equation. For PAUT, the reference amplitude is calculated across the full range of angles that will be used during the scanning. The standard SDH reflector on the IIW­type block is still used, but the beam is swept through all of the angles by moving the transducer along the IIW­type block surface. After calibra­ tion, the reference reflector will have the same amplitude at each angle (e.g., 70 degrees and 45 degrees). Time corrected gain (TCG) is used to account for material attenuation by sweeping the ultrasonic beam through SDH reflectors at varying depths. After performing TCG calibration, identi­ cal reflectors will have the same amplitude regardless of the depth or beam angle. PAUT has many advantages over conventional UT, one of which is the increased sound coverage. Compared to con­ ventional UT, PAUT can provide the UT technician with the ability to scan a material using multiple beam angles simul­ taneously. The UT technician also has additional views such as the S­scan and E­scan, which are two­dimensional repre­ sentations of all of the A­scans plotted simultaneously. This can aid the technician in distinguishing false call signals due to geometric indications. It can also help in flaw characteriza­ tion, through the use of tip diffraction signals or signals at the surface. Weld overlays showing the geometry of the weld preparation can also be drawn on the S­scan or E­scan views, which can help UT technicians inspect locations where dis­ continuities are more likely, such as the fusion face or weld root. If PAUT is used as a direct replacement of conventional UT in manual raster scanning, these advantages are likely to improve flaw detection and rejection if the same amplitude­ based acceptance criteria were implemented. Encoded PAUT scanning offers the ability to collect the raw scan data and save it for future reference or viewing. Conven­ tional UT indications, on the other hand, are typically reported Figure 2. Sample PAUT image (top left) A-scan, (bottom) C-scan, and (top right) S-scan.

7 as tabulated values of indication amplitude, length, and loca­ tion. Operator error is introduced into the reporting process since these values often are manually transferred from physical measurements or instrument results. Conventional UT A­scan data is also not typically saved for future reference. Although PAUT can provide more coverage than conven­ tional UT, full coverage of the weld does not ensure that all dis­ continuities within the covered region will be detected. When line scanning is performed with a single transducer, each point in the volume of the weld will only be primarily covered by sound with a single angle of incidence (it is recognized that due to beam spread, a given location will be “hit” by more than one angle of incidence but not with significant amplitude). If the flaw is not oriented in a manner to reflect adequate ultra­ sound back to the transducer based on the specific angle of incidence, the discontinuity may not be detected (or very little sound reflected) even though sound is covering that region. For this reason, it is often recommended to scan with angles that are normal to “expected” discontinuities, such as fusion faces of welds. When line scanning is performed, the probe is typically kept normal to the weld axis to inspect for discontinuities which are primarily oriented parallel to the weld axis. Con­ ventional UT, on the other hand, is typically performed by raster scanning where the probe is moved with rotation, transverse, and longitudinal movements. This movement helps to maximize the amplitude response from discontinu­ ities that are not oriented perfectly parallel to the weld axis. Prior PAUT research found that a skew angle of only 10° from the alignment of the discontinuity caused the signal amplitude to drop considerably and flaw detection become marginal [6]. A change in skew angle of 20° from perpen­ dicular to the discontinuity resulted in total loss of disconti­ nuity response. Therefore, lack of raster scanning when line scanning with PAUT is likely to result in decreased ampli­ tude for some weld flaws.