Groove formed along the welding direction in FSP simulation in DEFORM software

[preetham_alluri]

Dear members,

1. I am trying to perform friction stir processing simulation for Aluminum alloy (AA6xxx, thickness 1.5 mm, 800 RPM, 500 mm/min) in DEFORM 3D (v10,v12 and v13), but there is a groove or wormhole like defect formed along the welding direction.
2. I am using Lagrangian incremental method. I tried to change many parameters, such as friction co-efficient (shear 0.1 to 1), mesh strategy (fine mesh, moving mesh with tool), material properties (DEFORM database, other constitutive models), heat transfer co-efficients, solvers (Conj Grad ==> Sparse).
3. Depending on the friction co-efficient (low to high) and plunge depth, the temperature during welding phase is from 260 C to 500 C.
4. If I consider dwelling phase then there is a step fall in the temperature to 150 C or less before starting the welding. So I also tried to avoid this stage and continued welding.
5. But these things did not help in solving the groove defect formation.
6. After the plunging phase, when the tool moves in the processing direction the cavity does not fill and a wormhole, tunnel like defect is formed

I attached sample photos from the archives for better visualization. Any help, suggestions and insights from the members is highly appreciated.

image 1: initial stage where the mesh moves along the tool synchronously
image 2: top view of the groove formed along the weld direction.
image 3: side view of the groove formed along the weld direction (groove highlighted in red ellipse).
Thank you