Abaqus 焊缝模拟分析实例

1. Introduction
Full 3-dimensional simulation of multi-pass repair welds is now feasible and practical given the development of improved analysis tools (i.e. ABAQUS, (2001)) and significantly greater computer power. A single weld bead of limited length, deposited on a flat plate (bead on plate), introduces a strongly varying 3-dimensional residual stress field that has similar characteristics to a repair weld. This simple welded component is therefore a good vehicle to demonstrate new numerical methods and validate thermal and mechanical predictions against experimental data. A series of results are presented in this paper that simulate the deposition of a single bead on a flat plate. For the purposes of this paper, macrographs of a bead on plate specimen have been used to validate the penetration of the weld fusion boundary. The work has also been used to define the locations at which future residual stress measurements (neutron diffraction) should be taken as planned in the near future under the European funded ENPOWER research programme. The work reported in this paper compares the results obtained using simultaneous bead laying and moving heat source approaches and also investigates the effects of plastic strain annealing
2003 ABAQUS Users’ Conference
1
HOME
BLeabharlann BaiduOKMARKS
SEARCH
(introduced in ABAQUS 6.2, (2001)) and the use of alternative elements types. The work is presented in more detail, including the comparison of two further block-dumped heat source models, in Elcoate (2003).
a
b
Figure 1. Schematic diagram (a) and photograph (b) of the bead on plate specimen
2
2003 ABAQUS Users’ Conference
HOME
BOOKMARKS
SEARCH
3. Bead on plate finite element model
An ABAQUS, (2001) 3-dimensional model was constructed based on the measured dimensions of the welded test specimen 1. The model developed for the moving heat source analysis was a half model assuming that the geometry is symmetrical about the weld centreline (that is ignoring the small offset actually observed in the experimental weld). For the purposes of the block-dumped analysis advantage of the additional symmetry plane was exercised by reducing the half model to a quarter. The half model contains 5655 quadratic elements, within which 595 elements represent the bead. The bead geometry is an idealisation consistent with the measured width (7mm), length (60mm) and volume of weld deposit (mass = 5.5g). However, the top of the idealised bead is flatter than the observed profile (Figure 2) and the detailed profiles at the start and stop ends (Figure 3) are ignored. The physical geometry of the actual bead is such that the start end is rounded, while the stop end is flattened. The start and stop ends of the bead were idealised using a uniform cross-section along the entire length.
2. Welded plate test specimens
In support of this study, and other ENPOWER investigations, two test specimens have been prepared. Both of the specimens are similar in specification and the main features of one of the specimens (specimen 1) are provided in this section. Figure 1 is a diagrammatic picture and photograph of specimen 1 with the key dimensions. A piece of AISI Type 316L austenitic stainless steel plate, 180mm x 300mm x 17mm, was supplied by Mitsui Babcock for the validation studies. The as-received material test certificate composition (Heat 55841) was: C=0.02; Mn=1.404; P=0.027; S=0.0011; Si=0.582; Ni=11.11; Cr=17.834; Mo=2.06; N=0.0132 (% weight). A test specimen, 120mm x 180mm, was machined from the 17mm thick plate, and datum edges marked. Both the test specimen and the remaining plate were then solution heat treated in a vacuum furnace for 1 hour at 1050oC and slowly cooled, to eliminate fabrication residual stresses. The absence of residual stress in the test plate was confirmed by X-ray diffraction measurements. A single manual metal arc weld bead, with a low heat input procedure, was deposited along the centre-line of the plate by Mitsui Babcock, as shown in Figure 1. A Babcock type S-316, 2.4mm diameter electrode (approx. 19Cr12Ni21/2Mo, classification BS2926 19.12.3.L BR) was used. The following welding conditions were measured: current = 85A, closed circuit voltage = 25.4V, deposition time = 19.4s, bead length = 60mm, giving an average electrical heat input for this weld of 0.7kJ/mm. The increase in weight of the plate was measured to be 5.5g. After welding, a mould of the weld bead shape was made using a dental impression compound. The average width of the bead at the plate interface was about 7mm, and the bead height approximately 1.8mm.
HOME
BOOKMARKS
SEARCH
3-Dimensional Repair Weld Simulations – Bead on Plate Comparisons
Dr CD Elcoate1, PJ Bouchard2, and Dr MC Smith2
1) Frazer-Nash Consultancy Limited, 1 Trinity Street, College Green, Bristol, BS1 5TE, UK. 2) British Energy Generation Limited, Barnett Way, Barnwood, Gloucester, GL4 3RS, UK. Full 3-dimensional simulation of multi-pass repair welds is now feasible and practical given the development of improved analysis tools (i.e. ABAQUS) and significantly greater computer power. A single weld bead of limited length, deposited on a flat plate, introduces a strongly varying 3dimensional residual stress field that has similar characteristics to a typical repair weld. This simple welded component is therefore a good vehicle to demonstrate new methods and validate thermal and mechanical predictions. This paper presents a series of results that simulate the deposition of a single weld bead on a flat plate. The work exploits newly developed features within ABAQUS (i.e. plastic strain annealing) which have been demonstrated to be robust and efficient and readily extendible to larger and more complex models. Good comparisons are achieved between simultaneous bead laying and moving heat source approaches whilst it is also noted that the simultaneous bead laying methods offer significant modelling advantages in terms of model creation and execution time.