Windshield 3-point-bending test with /FAIL/ALTER
The provided model showcases a typical 3-point bending test simulation for a windshield glass material. The material is modeled with a sandwich layout of shell-solid-shell elements with coincident nodes. Both model layout and glass material formulation can be shown in Figure 1. For this model /FAIL/ALTER have been used, because of its good precision for windshield glass applications.
Figure 1: Model layout
Model explanation
The given model contains the following parts:
- 2 stable supports for the structure
- 1 support that impose the bending load and is under a constant imposed velocity load (/IMPVEL)
- 2 parts with shell elements for the inner and outer glass
- 1 part with solid elements for resin material (here polyvinyl butyral – PVB)
All three supports contain a rigid body (/RBODY) defined to set the proper boundary and kinematic conditions they have. For the moving support a mass equal to 1.5 tons is introduced as an external load. Supports material is steel modeled with a linear load (/MAT/LAW1 – ELAST) and a shell property is defined for the element formulation (/PROP/TYPE1 – SHELL).
The glass material (inner and outer layers) is modeled with an elasto-plastic material law with a user defined plastic area (/MAT/LAW36 – PLAS_TAB). The property used is for the modeling of laminate stack of plies (/PROP/TYPE51 – STACK). Here just one ply is introduced for each of the inner and outer layers with a specific thickness. Ipos = 4 defines that the bottom of the ply layout is coincident with the element mid-surface, in this way we can adjust the bottom glass layers that is not represented at mid-surface initially.
The PVB material is modeled with the use of a Hyper-elastic material law (/MAT/LAW42 – OGDEN). A properly defined solid element property is introduced (/PROP/TYPE14) with Ismstr =10 to enable Lagrange type total strain formulation.
For the formulation of contacts 4 different /INTER/TYPE7 interfaces were introduced. They can be categorized into 2 pairs of 2 symmetrical contacts. One pair of interfaces is used to model the contact between fixed supports and the material specimen, while the other for the contact between the moving support and the specimen.
Finally, an advanced nonlinear stress-based failure criteria for glass applications (/FAIL/ALTER) is defined in corresponding with the Elasto-plastic material law defined for the glass parts. With the use of this formulation parameters of micro-cracks and crack propagation speed are used to define failure stress. Additionally, with mode = 1 parameter X-FEM approach can be used to set stress equal to zero perpendicular to the crack direction. The elements layout that enables the use of this failure load is explained in Figure 2.
Figure 2: Elements layout for /FAIL/ALTER formulation
A time history file request (/TH/RBODY) is added for capturing the forces applied in supports rigid bodies during the test.
Two identical models with just a change in the mode of the failure law are solved to showcase the differences between the different formulations.
Results
In Figure 3, the failure imposed by the different formulations is displayed. As we can see the standard approach imposes the erosion of the whole element, while the X-FEM approach can provide results which identify precisely the crack created.
Figure 3: Element erosion for different formulation of /FAIL/ALTER
In Figure 4, the total resultant force applied in the moving support is displayed. Initially, the different fluctuations are caused by the load case imposed and the stiffness of the material. At the final time-steps of the simulation great noise can be seen and is caused by glass element fracture.
Figure 4: Moving Support Force