Learning Forum

[KARTHIK S]

I am a final-year Mechanical Engineering student currently working on an academic project involving the replication of a published research paper titled "Optimization of Bi-Hexagonal Crash Box Design Using Metal 3D Printing Technology for Enhanced Crashworthiness." (https://doi.org/10.1016/j.nxmate.2026.101841)

For our validation study, we replicated the M2 crash box model from the research article using ANSYS Workbench 2026 R1 student version (Explicit Dynamics). We carefully followed the geometry, material properties, boundary conditions, meshing strategy, and loading conditions described in the paper.

However, after completing the simulation, we observed that our crashworthiness results differ significantly from those reported in the paper. The values of:

Energy Absorption (EA)

Specific Energy Absorption (SEA)

Crush Force Efficiency (CFE)

Peak Crushing Force (PCF)

show approximately 40% deviation from the published results.

We have repeatedly reviewed our model, including the material assignment, contact definitions, mesh quality, boundary conditions, and analysis settings, but we have not been able to identify the reason for such a large difference.

Since this project is part of our academic research, we would greatly appreciate your guidance. We would be grateful if you could review our simulation setup or suggest possible reasons that could lead to this level of deviation. Any recommendations regarding solver settings, contact formulations, element types, material modelling, or other simulation parameters would be extremely valuable.

[aabedini]

Hello Karthik, 
Ansys employees cannot download or review models posted in the learning forum. However, please check if the following general guidelines are helpful to you.

1. Material model, hardening curve, and plasticity output

• Confirm the hardening law and parameters (e.g., multilinear isotropic, piecewise plasticity, etc.) and that the curve extends to the maximum plastic strain reached in the crash; if the plastic strain in the simulation exceeds the last tabulated point, extend the curve appropriately or results can be nonphysical. 
• Check that the plasticity parameters (yield stress, hardening slope, etc.) are calibrated against the same test data as in the paper; discrepancies in the stress–strain curve, incorrect hardening law, or poor calibration are a common cause of mismatch.

•  

2. Contact formulation and boundary conditions

• Verify that the contact type and parameters match the study; explicit and implicit solvers can give different stress levels for the same geometry if contact is handled differently, especially in large sliding/self-contact situations typical of crash boxes. 
• Review all boundary conditions carefully; explicit dynamics does not always warn when BCs are inconsistent, and in models with joints or rigid parts you should enforce kinematic conditions via joint loads when appropriate to avoid unintended constraints. 

3. Loading rate, time step, and dynamic effects

• Compare the loading rate to the experiment/paper: for quasi-static crashes, you should keep kinetic energy below about 1% of internal energy; if kinetic energy is higher, results (EA, SEA, CFE, PCF) can be significantly off due to inertial effects. 
• Be aware that explicit results can show sharp force oscillations after impact from contact and wave propagation; ensure that any force–displacement data you use for EA/SEA/CFE/PCF is appropriately filtered or interpreted consistently with the paper. 

4. Post‑processing 

• Check that energy absorption and force metrics are computed exactly as in the study: same integration domain, same reaction-force location, same displacement reference, and same filtering or averaging on the force–displacement curves.

• Check whether local and global responses, such as local and global buckling patterns, are predicted correctly in your model compared to the experiment.