Residual Stresses due to Butt Welding

Abstract

The accuracy of stress measurement by X-rays has been improved remarkably through recent years. Active researches have been conducted in many laboratories, and as a result the measured stress values have come to the stage of giving a sufficient reliability. However, there still remain some problems which must be further investigated.
In this study, as one of the application problems, the authors took up the problem of welding residual stresses. In connection with this subject, a number of studies have hitherto been made. But there have been few works so far that dealt with the residual stress distributions measured by the X-ray method.
The test piece used was a mild steel plate (SS41) with the dimension of 140×70×4.5mm, and two pieces were butt welded together under the groove angle of 75°by an ordinary A C arc welder, forming a specimen of square type. The welding consisted of two passes on the surface and one pass of reinforcing bead on the back.
The X-ray stress measurement was carried out by the film method using the Glocker, s technique of vertical and 45°oblique incidences. At first, CoKα beams were used, but the diffraction lines showed a considerable broadening. Therefore, CrKα radiation was utilized, and in this case fairly sharp diffraction lines were obtained. The distribution curves of diffraction intensity were drawn by an automatic recording microphotometer. The peaks of these curves were determined, for convenience' sake, by the method of half breadth at the half height. The residual stresses were calculated by the following formula,
σ=(Δ⊥-Δ₊)C⊥₊
The residual stresses were measured in two directions, parallel and perpendicular to the bead, at many points both on the surface and back, and the distribution curves of them were drawn. The results on the surface showed, in contrast to the ordinary results for the plate without a reinforcing bead, compressive residual stresses in the vicinity of the bead and tensile stresses near the edge in both cases of the measurements in the two directions. On the other hand, the results on the back were reverse to the above case. This effect of the reinforcing bead is worthy of note.
For the measurement of welding residual stresses, the Gunnert method or the Mathar method has been widely utilized so far. Consequently, for the purpose of comparing the measured stress values, the Mathar method was also applied to the present specimens. The Mathar method consists of drilling a small hole in the sample and measuring the deformation around the hole. The drill used was of 6mm diameter, and for the measurement of strain four electric resistance wire gages with the gage length of 6mm were pasted on the specimen, each two being attached in the direction parallel and normal to the bead respectively. The calibration test was conducted a priori, and based on this result the residual stresses were calculated according to the formula given by Riparbelli. The Mathar method was applied, as before, to several points on the surface and back, and the results were compared with those of the X-ray method. The stress values obtained by the Mathar method showed a little scattering. On the whole, however, these values were a little smaller than those measured by the X-ray method. This result seems to be natural, since the stress value obtained by the Mathar method gives the average over the distance of the gage length in contrast to the local stress value measured by the X-ray method.
Next, on the supposition of the stress repetition of an extremely low cycle as in the case of a boiler drum, the change in residual stresses due to the repeated stressing of a very low cycle was investigated. Another specimen with the dimension of 140×40×4.5mm, butt welded along the center line parallel to the short side of 40mm was prepared for this experiment.