2. Linear impact test fundamentals

Before any CAE model can be trusted, the physical test it reconstructs must be understood at the level of hardware, instrumentation, and physics. A linear impact test is not a single apparatus but a family of configurations that share one defining trait: the occupant or a surrogate structure is accelerated along a nominally straight path, and the resulting crash pulse is controlled by the interaction between a moving mass and a deceleration or acceleration device. The two dominant configurations in restraint-system development are the sled test and the linear impactor. They differ in what moves, what is instrumented, and which part of the occupant-loading chain they isolate.

Sled versus linear impactor: what each configuration isolates

In a sled test, a buck — a partial vehicle body carrying seats, restraints, and instrumented dummies — is mounted on a carriage that runs on rails. The carriage is accelerated or decelerated to reproduce a target vehicle pulse, and the occupant loads the restraint system exactly as in a full-vehicle crash, but without the cost and variability of destroying a complete car. The sled isolates the restraint–occupant interaction while the pulse is prescribed by the test engineer rather than emerging from a real vehicle structure.

In a linear impactor test, a moving mass strikes a stationary target — a door trim panel, a side airbag module, a steering column, or a body section. The impactor carries the instrumentation and the energy; the target is fixed. This configuration isolates the component-level response: how a specific airbag, trim, or structure behaves when loaded at a defined velocity and contact geometry. For airbag CAE, the linear impactor is the primary tool for validating deployment kinematics, fabric stress, and venting behavior before those submodels are embedded in a full sled or vehicle model.

Crash pulse generation

The crash pulse is the acceleration–time history imposed on the sled carriage or the impactor. It is generated by a controllable energy-absorbing device, most commonly:

• Hydraulic servo sleds — a programmable hydraulic actuator drives the carriage, allowing arbitrary pulse shapes (e.g., a trapezoidal or half-sine pulse) to be commanded and repeated with high fidelity.
• Deceleration sleds with crushable tubes or honeycomb — the carriage is launched to a target velocity and then decelerated by plastically deforming metal tubes or crushing aluminum honeycomb. The pulse shape is set by the material and geometry of the absorber.
• Bungee or pneumatic launchers — used for linear impactors, where a known mass is accelerated to a target velocity and released; the pulse is essentially a velocity step at contact.

For CAE reconstruction, the pulse is usually supplied as a measured acceleration–time curve from the physical test, or as a target pulse defined by a vehicle program. The two quantities that must be matched are the velocity change (Δv, the integral of acceleration over time) and the pulse shape (rise time, peak, and duration). A common error is to match peak acceleration while ignoring Δv; the occupant's forward excursion is governed primarily by Δv, so a model that matches peak but not integral will predict the wrong belt loading and airbag interaction timing.

Target pulse (example, side impact sled):
  t = 0 ms      a = 0 g
  t = 5 ms      a = 18 g   (rise)
  t = 15 ms     a = 22 g   (plateau)
  t = 30 ms     a = 0 g    (decay)
  Δv = ∫ a dt ≈ 6.5 m/s

These values are illustrative. In practice the pulse is digitized at 10 kHz or higher and filtered per SAE J211 (channel class 60 or 180, depending on the measurement). The filter class matters: an unfiltered pulse contains high-frequency content that will not be reproduced by the sled hardware and should not be fed into a CAE model as a boundary condition.

Occupant loading mechanisms

Once the pulse is defined, the occupant loading path determines what the airbag must do. Three mechanisms dominate:

1. Belt loading. The seatbelt couples the occupant to the decelerating structure. Belt force is a function of pretensioner timing, load limiter threshold, and belt geometry. In CAE, the belt is modeled with seatbelt elements and a slipring; the pretensioner and load limiter are represented by force–displacement or force–time characteristics.
2. Airbag loading. The airbag deploys into the gap between the occupant and the interior surface. Its pressure–time history depends on inflator mass flow, vent area, fabric permeability, and the volume it fills. The occupant feels this as a distributed pressure over the contact area.
3. Structural contact. When the airbag bottoms out or the occupant moves beyond the airbag's coverage, direct contact with trim, door, or steering column occurs. This is the mechanism that produces the highest injury metrics and is the primary target of airbag optimization.

The timing relationship between these mechanisms is the core of reconstruction. The airbag must be fully deployed before the occupant arrives, but not so early that it has already vented. The belt must engage before the airbag is fully loaded, or the occupant will "submarine" under the belt. These timing constraints are what the CAE model must reproduce, and they are why the pulse shape — not just Δv — matters.

From test configuration to CAE boundary conditions

The practical consequence of this chapter is a mapping exercise. When you build a reconstruction model, you must decide which configuration you are reproducing and set boundary conditions accordingly:

For a side-impact sled reconstruction, for example, the model's buck is given the measured carriage acceleration as a prescribed motion, the door trim and airbag module are positioned per the test setup, and the dummy is seated per the pre-test photograph. The airbag inflator is represented by a tank-test-derived mass flow and temperature curve. The model is then run and the dummy's head, chest, and pelvis accelerations are compared to the test data. If the model's head acceleration peaks earlier than the test, the likely cause is an airbag that deploys too fast or a pulse that rises too steeply; if it peaks later, the airbag may be under-inflated or the pulse may be too soft.

With the test configuration and pulse generation understood, the next step is to translate these physical quantities into a CAE model that reproduces them. That translation — geometry, materials, inflator characteristics, and contact definitions — is where the reconstruction either succeeds or fails, and it is the subject of the chapters that follow.