Numerical Investigation of the Effect of Rolling on the Localized Stress and Strain Induction for Wire + Arc Additive Manufactured Structures

Abstract

Cold rolling can be used in-process or post-process to improve microstructure, mechanical properties and residual stress in directed-energy-deposition techniques, such as the high deposition rate wire + arc additive manufacturing (WAAM) process. Finite element simulations of the rolling process are employed to investigate the effect of rolling parameters, in particular rolling load and roller profile radius on the residual stress field as well as plastic strain distribution for the profiled roller. The results show the response to rolling of commonly used structural metals in WAAM, i.e., AA2319, S335JR steel and Ti-6Al-4V, taking into account the presence of residual stresses. The rolling load leads to changes in the location and the maximum value of the compressive residual stresses, as well as the depth of the compressive residual stresses. However, the roller profile radius only changes the maximum value of these compressive residual stresses. Changing the rolling load influences the equivalent plastic strain close to the top surface of the wall as well as in deeper areas, whereas the influence of the roller profile radius is negligible. The plastic strain distribution is virtually unaffected by the initial residual stresses prior to rolling. Finally, design curves were generated from the simulations for different materials, suggesting ideal rolling load and roller profile combinations for microstructural improvement requiring certain plastic strains at a specific depth of the additive structure.

Computationally Efficient Models of High Pressure Rolling for Wire Arc Additively Manufactured Components

“…The creation of non-uniform mesh density does not require special techniques or algorithms and can be done using standard finite element analysis (FEA) software tools. Such a strategy to generate efficient models has been widely adopted in 2D [40,49,50] and 3D [33,34,38,[41][42][43][45][46][47] simulations of rolling and burnishing.…”

“…Cozzolino et al [33] used an FEM model with significantly shorter length (465 mm) compared to the actual workpiece (750 mm) that was rolled during experiments, so as to reduce computational time, and the short model predicted similar results to the full scale model. For rolling in a WAAM component, Abbaszadeh et al [34] developed a model with a length of 100 mm for a deformable component, which was rolled along the minimal length under steady state, and they found that a further increase of the rolling distance in analysis did not make any marked difference in the result. Short models have also been used in other rolling scenarios.…”

“…However, Hassani-Gangaraj et al [41] found that their deep rolling model underestimated compressive RS in a depth of 0.2 mm below the rolled surface, and they attributed the error to the use of a rigid roller in their model. To eliminate potential errors, Abbaszadeh et al [34] considered elastic deformation of the roller in their rolling model.…”

“…Isotropic friction (µ = 0.3) was specified for the contact interaction. Coules et al [47] and Abbaszadeh et al [34] suggested that a friction coefficient of 0.3 is representative for non-lubricated contact between the metallic roller and components. It was also found that the predicted RS distribution was not sensitive to the friction coefficient between the roller and component [33,34,47].…”

“…In current literature only two publications were found to address rolling simulation for WAAM built components [34,35]. For similar problems, the modelling research relevant to WAAM rolling includes the simulations of burnishing [36][37][38][39], deep rolling processes [40][41][42][43], hot and cold manufacture rolling processes [44,45], and rolling for reducing RS in quenched block [46] and weld seams [33,47].…”