Computational weld mechanics

所提出的研究是为了解决焊接结构的设计问题，即不再假定设计是给定的。取而代之的是，我们假定我们得到了使用中的载荷和没有指定焊接接头的初步设计。分析应预测所有焊接接头的最佳位置和尺寸、焊接程序、焊接规程、夹具设计和夹具规程。对于一类重要的焊接结构，主要的失效模式是疲劳。在给定大型拖拉机等结构的零件几何形状、每个零件的材料类型和相关在役载荷的CAD文件后，我们提出了以下策略：自动对所有零件进行网格划分。然后，用户应用服务中的负载和约束。然后，在相邻部件之间施加约束后，自动执行线性静态(或动态)三维有限元分析。此外，通过部件之间的邻接关系自动识别潜在的焊接路径。从每个潜在焊接路径的单位长度约束力出发，计算试焊焊缝尺寸和焊接程序。然后，对结构进行重新划分，以包括建议的带有填充金属的焊接接头。然后，对焊接结构中的焊接接头进行热应力分析，以计算焊接时的瞬时温度、组织演变、残余应力和焊接变形。此外，将使用Ohji-Sudnik熔池模型计算熔池大小和形状。微观结构的演变将使用蒙特卡罗算法和计算热力学进行建模。应力分析将包括对组织演变的单晶热粘塑性分析。然后，对新的焊接接头重新进行线性或非线性的带约束的塑性有限元分析。在这些焊接接头中，计算由于在役载荷和焊接接头中的残余应力而产生的三维应力分布，特别是在诸如焊趾等高危特征中。根据应力分布和在役载荷循环，估计焊缝的疲劳寿命。如果疲劳寿命过长，可以减小焊缝尺寸。如果疲劳寿命太短，则必须对焊缝进行强化或重新设计。此过程将自动完成。

The proposed research is to solve for the design of welded structures, i.e., we no longer assume the design is given. Instead we assume that we are given in-service loads and a preliminary design with no weld joints specified. The analysis should predict the optimum position and size of all weld joints, weld procedures, weld schedules, fixture design and fixture schedule.  For an important class of welded structures, the primary failure mode is fatigue.  Given CAD files for the geometry of the parts for a structure such as a large tractor, the material type for each part and the associated in-service loads, we propose the following strategy:  Mesh all parts automatically. Then the user applies in-service loads and constraints. Then with constraints between adjacent parts applied automatically perform a linear static (or dynamic) 3D FEM analysis.   Also potential weld paths are identified automatically by adjacency between parts.  From the constraint force per unit length along each potential weld path, a trial weld size and weld procedure will be computed.  Next the structure is remeshed to include the proposed weld joints with filler metal.  Then a thermal stress analysis of the weld joints in the welded structure is done to compute the transient temperatures, microstructure evolution, residual stress and distortion due to welding.  The weld pool size and shape will be computed using an Ohji-Sudnik weld pool model. Microstructure evolution will be modelled using Monte Carlo algorithms and computational thermodynamics. The stress analysis will include single crystal thermo-visco-plastic analysis of the evolving microstructure.  Then either a linear or a non-linear plastic FEM analysis with constraints is redone with the new weld joints.  In these weld joints, the 3D stress distribution due to in-service loads and residual stress in the weld joints and particularly in high risk features such as the weld toe are computed.  From the stress distribution and in-service load cycles, the fatigue life of the weld is estimated. If the fatigue life is too long, the weld size can be reduced. If the fatigue life is too short, the weld must be strengthened or redesigned. This process will be done automatically.