3-DIMENSIONAL REPAIR WELD SIMULATIONS - BEAD SEQUENCING STUDIES

摘要

In current 3-dimensional multi-pass repair weld simulations each pass is generally modelled as a single continuous weld bead using either a simultaneous bead deposition or a progressive bead deposition moving heat source approach. In reality the length of typical manual metal arc (MMA) repair welds often necessitates the use of multiple weld beads for any individual pass. To investigate the modelling assumptions associated with this approximation a series of results are presented that simulate the deposition of a multi-bead, single MMA weld pass on a stainless steel flat plate. The simulations make use of a fabricated single bead weld test specimen to define the welding conditions. A number of different heat source models and bead sequencing scenarios have been investigated. These include moving heat source simulations, simultaneous bead laying methods (also called "block-dumped" methods) and combinations of the two. The work provides a useful insight into how certain modelling approximations affect weld residual stress distributions. It is concluded that moving heat source end effects are not sensitive to modelling assumptions made in weld beads remote from the stop end. However, residual stress predictions at the mid-length of a multi-bead weld pass on a flat plate are sensitive to the modelling assumptions. In particular, modelling each bead of a multi-bead weld pass can increase the stresses at mid-length relative to a simulation that deposits the whole pass in one go. One further observation made is that modelling the central bead of a multi-bead weld pass using a block-dumped technique (allowing a symmetry boundary condition to be used perpendicular to the weld bead) can result in inaccurate or misleading results.