On bending stresses at longitudinal weld joints of cylindrical shells due to peaking

Junctions in shell structures: A review

“…Stress concentration factor was derived, and the measurement procedure proposed. Bock and Zeman [345] proposed a powerseries approximation for the local type of peaking in welded longitudinal joints of cylindrical shells with angular misalignment of the cylinders. Good agreement with the FE analysis was noted.…”

Discussion: “Bending Stresses at Longitudinal Weld Joints of Pressurized Cylindrical Shells due to Angular Distortion” (Ong, L. S., and Hoon, K. H., 1996, ASME J. Pressure Vessel Technol., 118, pp. 369–373)

“…In the derivation of the analytical second-order results by Bock and Zeman (1994), there appeared the problem that an additive constant could not be determined directly. This constant was determined such that the results of the second-order theory, i.e., the geometrically nonlinear theory, agreed with the results published in the literature, e.g., in BS PD 6493 or in the report by Maddox (1985), as results of the first-order theory, i.e., the linear theory.…”

“…This constant was determined such that the results of the second-order theory, i.e., the geometrically nonlinear theory, agreed with the results published in the literature, e.g., in BS PD 6493 or in the report by Maddox (1985), as results of the first-order theory, i.e., the linear theory. VKPr=v / Pr 3 /D comment is just meant to clarify a minor flaw in the text, a flaw that goes through the whole literature on this subject and which has also crept into the results in the investigation by Bock and Zeman (1994); and it is intended to shed light onto the whole problem of local peaking.…”

“…Having not been satisfied with this situation, despite the fact that it already gave a pronounced improvement over those used so far, we studied the problem further, and found that the same problem-a nondeterminable additive term, which corresponds to a rigid body displacement-also occurs actually in the firstorder theory investigation of the problem, if one uses the same approximation approach suggested by Timoshenko and Gere (1961), and also used by Ong and Hoon (1996), Bock and Zeman (1994), and Zeman (1994). It turned out that the obvious choice of the coordinate system-in the center of the circle-is also a consistent one, and that no rigid body displacement correction is required.…”

“…Using this result (1) to ' 'calibrate'' the nonlinear result given by Bock and Zeman (1994), one obtains for the maximum moment…”