Multiscale modeling of 3D printed steel

Additive manufacturing of steel is a novel manufacturing/construction technique for which several processes have been developed. The most promising for civil infrastructures is called wire and arc additive manufacturing because it allows higher deposition rates than the other processes. The benefits of 3D printing, in general, are due to its unmatched levels of flexibility for designers. With 3D printing, materials and structures can be made with virtually any shape. Accordingly, with this manufacturing technique, structural elements with optimized geometries can, perhaps for the first time, actually be realized. 
One of the main challenges here is the lack of predictive modeling frameworks for these materials due to the additional complexities raised due to their layer-by-layer structure. These interfaces not only create an anisotropic behavior (i.e. different stiffness coefficients in different loading directions), but it also complicates the failure response of the material. For example, when loaded in the weakest direction, i.e. the direction perpendicular to the interface direction, the material will fail at a lower load. Further, for more general loading directions, the existence of interfaces can result in crack paths different than those observed in isotropic materials, i.e. conventional steel.  
In this project, we will use a multi-scale framework that systematically studies material behavior in different directions. Effectively, we will derive the macroscopic response of the material as the average of the responses in all directions, making use of the “Microplane model” that will be adapted to the 3D printed steel. With this approach, the directional effects of interfaces, the change in stiffness and strength based on the direction of the printing interfaces, will be automatically incorporated into the macroscopic behavior of the material, without significant additional computational cost. This framework can then be implemented into Finite Element software to analyze failure conditions and crack paths. Also, the framework can be used for finding optimized topologies.

Figure 1. Microplane-predicted vs experimental crack path in steel specimen. (From: Ožbolt, J., Tonković, Z., & Lacković, L. (2016). Microplane model for steel and application on static and dynamic fracture. Journal of engineering mechanics, 142(2), 04015086.)