4. CAE model setup for frontal impact

Frontal impact CAE is rarely a single model. A campaign typically runs two complementary representations: a full vehicle model that captures the complete body-in-white (BIW), powertrain, suspension, closures, and interior, and a sled model that isolates the restraint system on a rigid or simplified buck. The full vehicle answers "does the structure and restraint system work together?" The sled answers "does the restraint system meet its targets, and why or why not?" Getting both right — and keeping them consistent — is the foundation of every downstream result.

Full Vehicle Model: What Must Be Present

A production-grade frontal full vehicle model is assembled from several subsystems, each with its own fidelity requirement:

• BIW and closures: shell elements with thickness from the CAD/CATIA data, spot welds as rigid or deformable beams, structural adhesives as solid or cohesive elements. Mesh size in crush zones is typically 5–8 mm; non-crush areas can relax to 10–15 mm.
• Powertrain and suspension: often represented by concentrated masses and nonlinear springs/dampers, or by a detailed subframe if it participates in load paths.
• Interior: instrument panel, steering column, seats, floor, and toe pan — these define the occupant's immediate environment and must be present for contact.
• Restraint system: airbag modules (folded geometry or *ALE/particle method), seat belts with retractor, pretensioner, and load limiter, plus the steering column collapse.

The model is validated against a rigid barrier test at the component level before it is trusted for occupant response. A common acceptance check is that the vehicle deceleration pulse from the CAE matches the physical test within a defined corridor — for example, ±10% on peak deceleration and ±5 ms on pulse timing.

Sled Model: The Restraint Development Workhorse

The sled model replaces the vehicle structure with a prescribed deceleration pulse applied to a rigid buck. This removes structural variability and lets engineers iterate rapidly on belt routing, airbag venting, and load limiter force. The sled buck carries the seat, steering column, IP, and restraint hardware in their correct relative positions.

The critical input is the sled pulse. It is derived from either a full vehicle CAE run or a physical sled test, and is applied as a prescribed acceleration or velocity boundary. A typical frontal pulse might show a peak of 35–45 g at 60–80 ms, with a total velocity change (delta-V) of 50–64 km/h depending on the test protocol. The pulse must be applied in the correct direction and with the correct sign convention.

Occupant and Dummy Positioning

Dummy positioning is where many campaigns lose correlation. The Hybrid III 50th percentile male is the standard frontal occupant, and its position is defined by a set of joint angles and reference points, not by a single coordinate. The key steps are:

1. Seat reference: establish the H-point (hip point) from the seat CAD or from the physical seat using the SAE J826 H-point machine procedure. The dummy's pelvis must sit at this H-point within ±5 mm.
2. Pelvis and torso: set the pelvis angle to the seat cushion angle, then adjust the torso to the seat back angle. The torso should contact the seat back without pre-load.
3. Head: position the head so the head center of gravity is within the manufacturer's specified tolerance, typically ±10 mm in X and Z. The head must not contact the headrest in the initial state.
4. Arms and legs: set upper arm and forearm angles to the nominal driving posture (e.g., upper arm 20° from vertical, forearm 30° from horizontal). Feet rest on the toe pan or pedals as specified.
5. Belt routing: route the shoulder belt over the clavicle and the lap belt across the pelvis, with the belt in contact but not compressing the dummy. Belt-to-dummy gap should be near zero but not negative.

For the passenger side, the same procedure applies with the seat in its nominal position. For out-of-position (OOP) studies, the dummy is moved to a specified static position — for example, chin on the airbag module — and the airbag deployment is simulated with the dummy present.

Boundary Conditions and Initial State

Boundary conditions define how the model interacts with the outside world. In a full vehicle frontal impact, the ground is represented by a rigid wall or a prescribed velocity field. The vehicle is given an initial velocity (e.g., 56 km/h) and the barrier is fixed. Gravity is applied as a body force. Contact is defined between the dummy and the seat, belt, airbag, IP, and steering wheel.

In the sled model, the buck is constrained to move only in the X direction, and the pulse is applied as a prescribed acceleration history. The seat is fixed to the buck. The dummy is positioned as above, and the same contact definitions apply.

Model Consistency Between Sled and Full Vehicle

The sled and full vehicle models must share the same dummy, the same belt and airbag definitions, and the same contact parameters. If the sled uses a different belt stiffness or airbag vent size than the full vehicle, the two models will diverge and the sled results cannot be used to explain full vehicle behavior. A practical check is to run the sled with the full vehicle's pulse and compare dummy kinematics at 50 ms and 100 ms; the head and chest trajectories should match within a few millimeters.

Practical Setup Checklist

The following table summarizes the key inputs and verification points for each model type. It is intended as a working reference during model assembly, not as a substitute for the detailed procedures above.

Once the model is assembled and the boundary conditions are verified, the campaign can proceed to solver execution and occupant response evaluation. The next chapter covers the solver settings and output requests that turn these models into injury metrics.