2. Child occupant biomechanics and injury criteria

Evaluating a child restraint system (CRS) in simulation requires more than a seat model and a crash pulse. The virtual occupant must reproduce the mass distribution, joint stiffness, and geometric proportions of a real child, and the results must be expressed in metrics that correlate with actual injury risk. This section covers the anthropomorphic test devices (ATDs) used for child occupant protection, the Q-series family that dominates current regulatory and consumer testing, and the injury criteria applied to interpret their response.

Why Child ATDs Differ From Adult ATDs

A 50th-percentile adult Hybrid III is a mature, ossified body with a single dominant load path through the spine and pelvis. A child is none of these things. The head is proportionally far larger and heavier relative to the body, the neck is slender and incompletely ossified, the rib cage is compliant, and the pelvis is not yet a rigid load-bearing structure. These differences change both kinematics and injury tolerance.

Three consequences matter directly for simulation:

• Head mass fraction. For a 3-year-old, the head accounts for roughly 15–18% of total body mass, versus about 6–8% for an adult. During frontal deceleration, the neck must restrain a much larger inertial load, which is why neck tension and head excursion dominate CRS performance.
• Compliant thorax. A child's chest deflects more under belt loading, so chest deflection criteria are scaled differently and belt routing is more sensitive to submarining.
• Developing pelvis. The iliac crests are not fully formed, so a lap belt can slip above the pelvis (submarining) more easily, allowing the torso to jackknife.

These are not cosmetic differences. A model calibrated only against adult corridors will produce non-physical child kinematics, particularly in neck force and chest deflection.

The Q-Series ATD Family

The Q-series was developed under the EEVC/ISO framework to provide a biofidelic child ATD across the full range of child sizes. The family is defined by nominal age and mass, and each size has its own instrumentation set and calibration corridors.

Each Q dummy carries a head accelerometer triad, upper and lower neck load cells, chest deflection sensor, and pelvic/seat-belt load cells. The Q3, for example, uses a six-axis upper neck load cell that reports Fx, Fy, Fz, Mx, My, Mz — the same channel set used in the injury criteria below. In a simulation, these correspond to output requests on the neck element or a dedicated joint.

Injury Criteria for Child Occupants

Injury criteria convert measured ATD response into a normalized risk value. The most relevant for CRS evaluation are:

• HIC (Head Injury Criterion). HIC = max over the time window of [(1/(t2−t1)) ∫ a dt]^2.5 · (t2−t1), with a in g and the window capped at 15 ms (HIC15) or 36 ms (HIC36). For child occupants, HIC15 is standard. A common regulatory limit is 1000 for HIC36, but child seats are typically evaluated against lower targets.
• Neck injury criteria. Nij combines axial force and bending moment: Nij = (Fz/Fzc) + (My/Myc), where Fzc and Myc are critical values scaled by age. Nij > 1.0 indicates elevated risk. For a 3-year-old, Fzc is on the order of 2.5–3.0 kN in tension and Myc around 40–50 N·m.
• Chest deflection. Peak sternal deflection, typically limited to about 40 mm for a 3-year-old (roughly 20% of chest depth).
• Head excursion. Forward displacement of the head center of gravity relative to a reference point, used in FMVSS 213 and ECE R44/R129. Limits are typically 550–720 mm depending on the standard and CRS type.

These criteria are not independent. A CRS that reduces head excursion by over-tensioning the harness may increase neck tension. The design task is to find a configuration that keeps all criteria below threshold simultaneously.

From ATD Response to Risk: A Worked Example

Consider a forward-facing Q3 in a 50 km/h frontal sled test, simulated with a validated FE model. The output channels are head acceleration (resultant), upper neck forces and moments, and chest deflection. Suppose the simulation returns:

Head resultant accel (peak): 62 g at t = 78 ms
Neck Fz (tension, peak):      2.1 kN at t = 82 ms
Neck My (flexion, peak):      38 N·m at t = 82 ms
Chest deflection (peak):      32 mm at t = 85 ms
Head excursion (max):         610 mm

Applying the criteria:

• HIC15: compute over the 15 ms window around the peak. With a 62 g peak and a typical pulse shape, HIC15 lands near 700 — below the 1000 limit but above a 700 target, so it is marginal.
• Nij: (2.1/2.8) + (38/45) = 0.75 + 0.84 = 1.59. This exceeds 1.0 and indicates a neck injury risk that must be addressed.
• Chest deflection: 32 mm < 40 mm — acceptable.
• Head excursion: 610 mm — within the 720 mm ECE R129 limit but above a 550 mm target.

The dominant problem is neck tension and flexion moment. In practice, this points to excessive head rotation relative to the torso, often caused by a harness that is too low or a CRS that allows the torso to pitch forward before the head is restrained. A common fix is to raise the top tether anchor point or adjust the harness routing to engage the shoulder earlier, which reduces both Fz and My at the cost of slightly higher chest deflection.

Visualizing the ATD-to-Criteria Chain

The diagram below shows how a Q-series ATD's instrumentation maps to the injury criteria and how a design change propagates through the chain. It is not a process flow; it is a signal-and-criteria relationship map, which is the structure that matters when interpreting simulation output.

Practical Guidance for Simulation Setup

When setting up a CRS evaluation with a Q-series ATD, three configuration details have an outsized effect on the validity of the injury criteria:

1. Output sampling rate. HIC and Nij are sensitive to peak values. Sample at 10 kHz or higher; a 1 kHz output will miss the acceleration spike that drives HIC.
2. Filtering. Apply the correct CFC (Channel Frequency Class) filter before computing criteria. HIC uses CFC 1000; neck forces use CFC 1000; chest deflection uses CFC 180. Using the wrong filter changes the peak by 10–20%.
3. Coordinate consistency. Define the ATD local axes once and document them. The neck load cell's Fz is positive in tension; the chest deflection sensor is positive in compression. Mixing these signs is the most common source of erroneous pass/fail results.

With these controls in place, the simulation output becomes a reliable proxy for sled test measurements, and the injury criteria provide a defensible basis for comparing CRS designs before any physical prototype is built.