抄録
Welding is widely used in the manufacture of steel structures, but weld deformations due to the heat input affects the strength of the structures. Therefore, it’s necessary to be straightened depending on circumstances. However, because straightening is a very costly process, numerical analysis is required to minimize weld deformations. Weld deformations can be simulated by nonlinear thermal elastic-plastic analysis based on the finite element method (FEM). Moreover, the idealized explicit FEM, which is capable of high-speed calculation, has been used to analyze real structures. However, the element sizes relative to the weld length becomes extremely fine when the heat source length of multi-pass welding problem for large steel structures is defined in actual dimensions, resulting in enormous calculation time even if the FEM is used. One way to counter this problem is to increase the heat source length. However, if the heat source length is excessive, analysis accuracy may be reduced due to excessive expansion or shrinkage in the heat input range. Thus, the effect of the heat source length on the analysis accuracy was examined in this study. In addition, the FEM was performed on the welded assembly of a hot dip galvanizing bath, which is one of the large steel structures, and the accuracy of the analysis was verified by comparing the analytical results with measured values. As a result, the maximum achievable temperature decreases with increasing heat source length and that weld deformation is underestimated. Therefore, it’s suggested that optimization of current density according to the length of heat source is necessary to achieve both accuracy of analysis and reduction of calculation time in the analysis of large steel structures.
