RD-E: 5400 Cut Methodology

The target of the cut methodology is to study one area of the model by applying the deformation of the full model to a smaller model.



Figure 1. Full Model and Sub-Model

Options and Keywords Used

Input Files

The input files used in this example include:
Frame Modified
<install_directory>/hwsolvers/demos/radioss/example/54_cut_model/*

Model Description

The full model contains a rigid plate with an imposed displacement that impacts the bumper. The crash box behind the bumper is fixed in all directions. Contact is defined between the parts using /INTER/TYPE7. Advanced mass scaling is used to reduce the solution time of the model.


Figure 2. Problem Description

Units: mm, ms, kg, N, GPa

The result of the full bumper model shows that most of the deformation is in the crash box area, which is just behind the rigid plate. The crash box on the other side shows very little deformation. The term sub-model refers to a smaller model that is created by cutting or trimming the full model. This is different than the //SUBMODEL model organization method that can be used in Radioss.


Figure 3. Full Model Results with Suggested Cut Location to Create the Sub-Model

In order to focus on the crash box area just behind rigid plate, the sub-model (blue highlighted area in Figure 3) is cut from the full model. To have the same behavior as the sub-model and the full model, the force and displacement on the boundary of the selected are needed. Using /SECT, it is possible to save and read section force or displacement.

In this example, two sections (section ID 10 and ID 9) are defined in the full model. The HyperCrash, Process, Submodeling option was used to define the sections to the full model and create the sub-model. This process will create the /SECT in the full model with ISAVE=2 to save the section displacement or section force data during the computation with frequency Δ t . The file name of section data (SC file) is defined in /SECT, flag_name. Next, use the SC files created from the full model as input to the smaller sub-model. Using the same /SECT definitions as the full model, change ISAVE=101 to read the section data from the SC file created from the full model.
Note: The /SECT, Δ t input how often the section displacement and force results are written to the SC file. If Δ t is too small, it can lead to very large SC output files. However, if Δ t is too big, then the results may be wrong due to aliasing. A recommended value for Δ t is 10 times the model timestep.

The node groups and element groups in the section should be the same in the full model and sub-model to ensure consistency between the forces. It is also possible to use HyperMesh to create the /SECT in the full model and create the sub-model by deleting elements from the full model.



Figure 4. Cut Methodology in Radioss

Results

For this example, the reduced sub-model is about 58% of the full model. With the same CPU, about 54% of the elapsed time is saved.
Table 1. Total Calculation Time Comparison
  Full Model Sub-Model Ratio

Sub-Model / Full Model

CPU 1 1  
Number of Elements 16722 9730 58.19%
Time step [ms] 0.1500E-02 0.1500E-02 1
Elapsed time [s] 498.62 272.16 54.58%

The computation time is reduced because there are fewer elements and contacts in the simulation.

The correlation of deformation in the crash box in the full model and sub-model is good.


Figure 5. Deformation Overlay of the Full Model and Sub-Model
Figure 6. Section Force in Deformed Crash Box in Full Model and Sub-Model

The cross-section force in the crush box is very similar between the two simulations. If the sub-model results are very different than the full model, more section output results can be saved in the SC file by reducing the /SECT Δ t . Alternatively, the location of the sections could be modified to include a larger area from the full model that influence the deformation of the sub-model.

Conclusion

The cut methodology allows you to reduce the computation time by running smaller models. Its main limitation lies from the fact that the interactions between the selected parts in the cut model and the unselected parts must be small enough so that the behavior of the selected parts is not significantly affected by unselected parts. It is important to compare the section force (or moment) between the full model and sub-model to make sure they are similar.

In industrial applications, such as a car simulation with crash dummies, it is advised to keep only the dummy and its direct environment as sub-model. Then about 90% of the CPU cost could be saved compared to the full car model. In low speed impacts or the beginning of impact with high speed, it is advised to keep only the deformed area as sub-model. Then about 30% of the CPU cost could be saved compared to the full car model.