surrounding materials. From this viewpoint, it is less harmful to the welded work than an elastic residual stress. However, if the plastic deformation due to the residual stress is sever, that particular point can be the weakness that leads to a failure of the work. It is important to distinguish elastic residual stresses from plastic residual stresses. Since acousticelasticity is sensitive to only force, there is no way to distinguish a given reduction in the elastic constant is due to elastic or plastic deformation. Independent measurement is necessary to determine if the deformation is elastic or plastic. Application of ESPI solves the above issues. By applying an external tensile load to a welded specimen and measuring the resultant displacement with an in-plane sensitive ESPI setup, one can evaluate the resultant strain as a whole-field image of the specimen. The observed strain reflect the elastic constant at all the point on the specimen; if the elastic constant is low, the strain is large. This solves the first issue of point-to-point variation. By keeping the external load at a level substantially lower than the residual stress, it is possible to make the external load not cause additional plastic deformation to the specimen. By applying the same external tensile load after releasing the specimen from the initial load and analyzing the reproducibility, it is possible to determine whether the observed strain is plastic or elastic at each point on the specimen. If the strain is not reproduced in the second loading at a certain area of the specimen, it is likely that the specimen has been plastically deformed by the welding. With this idea in mind, we conducted the following experiment on butt-welded specimens. We measured residual stresses near the weld line without applying an external load to the specimen using a typical SAM setup (Fig. 14.1). We applied a tensile load to another butt-welded specimen and a non-welded specimen of the same material and dimensions at a minimum stress level (much lower than typical residual stress) and measured the resultant strain as a whole image. Figure 14.2 shows the phase velocity measured with the SAM near the weld line parallel to it. The dashed line indicates the SAW velocity measured at a certain distance away from the weld where the effect of residual stress is negligible. Figure 14.2 indicates that the phase velocity is altered as much as 10 %. Since the elastic constant is proportional to the square root of the phase velocity, this indicates that the residual stress alters the elastic constant of this specimen accordingly, and the change must be reflected on the strain data obtained with the ESPI. We conducted an ESPI experiment on a welded specimen shown in Fig. 14.2 (100 mm long 20mmwide 0.4mm thick butt-welded at the middle point along the 100 mm length or 50 mm away from the end). The specimen was attached to a tensile machine for a tensile load applied perpendicular to the weld (the specimen was gripped by the tensile machine at the two 20 mm-wide ends). A conventional, dual-beam ESPI setup sensitive to in-plane displacement along the tensile axis was arranged in front of the specimen. Some initial tensile load was applied until the specimen stopped slipping inside the tensile machine’s grips. The first interferometric image was taken at that point with a digital camera. After a load increment of 10 N, or an average tensile stress of 1.25 MPa, the second interferometric image was taken. A fringe pattern was formed by subtracting the second image from the first image. (In the actual experiment, a carrier fringes of known phase was added to the second image to make the fringe analysis easier.) For comparison, the same procedure was taken for a non-welded Fig. 14.2 (a)Welded specimen and (b) SAW velocity near weld line. The weld line is center of the specimen running vertically in above figure. Nominal phase velocity was measured 7 mm away from weld line. Other SAW velocity data were taken along a line parallel to and 3 mm away from weld, 2 mm and 3 mm from center 138 I.-K. Park et al.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==