Session: 03-06-01 Advances in Materials and Manufacturing Technology

Paper Number: 87473

87473 - Benchmark Study of Welding Deformations in Stiffened Plate 

The fabrication of ship hulls involves welding operations, one typical operation is fillet welding of stiffeners to plates (panels) a so-called T-joint. When constructing the panels, the GMAW (Gas Metal Arc Welding) process is often used. Molten material is then deposited at a joint thus heating up and melting the edges of the plates to be joined. This local heating followed by a rapid cooling of the supplied metal and the plate edges will lead to local yielding and development of plastic strains which, in turn, will lead to residual deformations and residual stresses in the welded panel. The residual stresses will influence the fatigue life of the structure, and residual deformations will increase the risk for buckling of the panel and give poor fabrication tolerances which may complicate the assembly of a built-up structure.

 

Fillet welding of a stiffener to a plate (panel) has been frequently studied in the literature both  experimentally and numerically, using non-linear FE-analyses. However, most references focus on the residual stress field and if residual deformations are studied, the angular distortion, and the typical V-shape seen in the welded plate (panel) are reported. It was therefore decided to carry out a benchmark exercise with a two-pass fillet welding of a stiffener to a plate (T-joint) where experimental results were available and focus on residual deformations, determined using nonlinear FEA and including the full shape of the welded plate (and stiffener) and the influence of the constraints used during welding. With the focus on welding residual deformations, information useful for the fabrication of larger complex welded structures may be obtained.

 

Five universities participated in the benchmark. The welding experiments were carried out within the Australian research program for DMTC Ltd., where residual deformations were recorded for a fillet welded T-joint using a weld pass on each side of the stiffener (but moving in opposite directions).

 

The benchmark clearly shows that careful consideration is needed when defining a reference plane for comparison of resulting deformations from FE-simulations and from experiments. It is also important to have a good documentation of the experimentation, besides welding process parameters and geometrical dimensions of the material also the actual boundary conditions and steps carried after the completion of the welding and before measuring residual deformations and stresses. The FE-simulations have shown that it is difficult to obtain the full antisymmetric, twisted shape of the plate as measured in the experiments, although a trend similar to the experiments can be seen with a slight influence of the weld pass direction, as the plate tends to move upwards at the end of each weld pass. Both experiments and FE-simulations show that the deformations resulting from welding one stiffener to the plate may cause further fabrication issues, if the welded plate is released before subsequent fabrication operations.

Presenting Author: Lennart Josefson Chalmers University of Technology

Authors:

Lennart Josefson Chalmers University of Technology
Bai-Qiao Chen Instituto Superior Técnico
Koji Gotoh Kyushu University
Fang Wang Shanghai Ocean University
Kun Liu Jiangsu University of Science and Technology
Stephen Van Duin University of Wollongong
Pingsha Dong University of Michigan