Numerical residual strength prediction of stationary shoulder friction stir welding structures. (1st May 2020)

• Record Type: Journal Article
• Authors: Scheider, I.; Barbini, A.; dos Santos, J.F.
• Is Part Of: Engineering fracture mechanics. Volume 230(2020)
• Publication Date: 2020-05-01
• Subjects: Residual strength -- Cohesive zone model -- Welded structures -- Parameter identification -- Stress-strain curve; Fracture mechanics -- Periodicals; Rupture, Mécanique de la -- Périodiques; Fracture mechanics; Periodicals; 620.112605
• DOI: 10.1016/j.engfracmech.2020.107010
• Languages: English
• ISSNs: 0013-7944

Abstract

Highlights: Residual strength of a welded structure was characterized numerically for different crack propagation directions. Spatially varying stress strain curves were identified from single specimens using DIC. Spatially varying crack propagation parameters were identified by multi-specimen method. The residual strength of welded specimens could be predicted with reasonable accuracy. Procedure for numerical residual strength assessment of welded structures was developed. Abstract: The residual strength of a structure made of dissimilar Aluminium panels joined by a stationary shoulder friction stir welding (SSFSW) process is predicted by numerical simulation using finite elements including cohesive elements for crack propagation. The yield strength and strain hardening parameters within the stirred zone and adjacent thermo-mechanically affected and heat affected zones are derived from a tensile specimen cut out from the panel perpendicular to the weld seam. The identification is conducted by a hybrid numerical/experimental procedure with an inverse search by help of digital image correlation in order to obtain the strain field at the welded surface and to compare them to the numerical calculation. The crack propagation parameters are retrieved from specimens with crack in the stir zone and heat affected zone on either side. After this identification procedure, the fracture behaviour of a coupon specimen with a crack crossing the weld is predicted.