2. Restraint system fundamentals

Occupant restraint correlation engineering begins with a precise mental model of what each device does, when it acts, and how its output couples into the occupant's kinematics. A restraint system is not a collection of independent safety features; it is a tuned chain in which the seat belt carries the majority of the energy management duty, the pretensioner removes slack before the occupant loads the belt, the load limiter caps the force transmitted to the chest, and the airbag distributes head and torso loading over a larger area. Correlation work fails most often when engineers treat these as separate subsystems and miss the timing and force interactions between them.

Seat Belt: The Primary Load Path

The three-point belt is the backbone of occupant restraint. Its geometry defines how crash energy is transferred from the occupant into the vehicle structure. The lap portion loads the pelvis through the iliac crests, while the shoulder portion loads the upper torso through the clavicle and rib cage. The diagonal routing across the chest creates a force vector that both restrains forward motion and induces a controlled rotation of the torso.

In correlation testing, the belt's effective stiffness and routing are the first variables to verify. A belt that is routed 10 mm differently on the dummy compared to the sled setup can shift chest deflection by several millimeters. Key parameters to document include:

• Belt anchor positions (D-ring, buckle, outboard and inboard anchors) in vehicle coordinates
• Belt webbing width, material, and elongation characteristics
• Belt routing path over the dummy's shoulder and pelvis
• Initial belt slack and the gap between webbing and occupant surface

During a frontal impact, the belt begins loading the occupant after the pretensioner has removed slack — typically within 10–20 ms of the crash pulse onset. The belt then stretches elastically, and the occupant's forward excursion is governed by the combined stiffness of the webbing, the retractor, and the anchor points.

Pretensioners: Removing Slack Before Loading

A pretensioner is a pyrotechnic or mechanical device that retracts the belt webbing in the first milliseconds of a crash. Its purpose is to eliminate the slack between the belt and the occupant's body so that the belt engages the occupant earlier and more uniformly. Without pretensioning, the occupant travels forward several centimeters before the belt becomes taut, which delays restraint and increases peak excursion.

There are two common pretensioner types: buckle pretensioners, which pull the buckle downward, and retractor pretensioners, which wind the webbing onto the spool. Some systems use both. The pretensioner's firing time is controlled by the airbag control unit (ACU) and is typically triggered within 5–15 ms after the crash pulse is detected, depending on the severity and the algorithm's threshold.

For correlation, the critical parameters are pretensioner stroke, firing time, and the resulting belt force buildup. A pretensioner that fires too late allows the occupant to move forward before the belt is tight; one that fires too early may cause the belt to dig into the occupant before the crash pulse has fully developed. The correlation engineer must verify that the pretensioner timing in the simulation matches the physical test, often by comparing belt force traces and occupant displacement.

Load Limiters: Capping the Force on the Chest

Once the belt is taut, the load limiter controls how much force the belt transmits to the occupant. A load limiter is typically a torsion bar in the retractor or a deformable element in the buckle that allows the belt to pay out webbing at a controlled force level. This reduces chest deflection and rib fracture risk, but it also increases forward excursion, which must be managed by the airbag and the occupant compartment geometry.

The load limiter's force level is a design trade-off. A typical front-row load limiter might be set between 2.5 kN and 4.5 kN, depending on the vehicle platform and the occupant size. In correlation, the load limiter's force-displacement curve is a primary input. If the simulation uses a linear force cap while the physical part has a progressive or stepped characteristic, the chest deflection correlation will degrade.

Airbags: Distributing Load and Managing Head Kinematics

The frontal airbag deploys from the steering wheel or instrument panel and inflates within 20–40 ms. Its primary functions are to cushion the head and chest, reduce peak acceleration, and prevent the head from striking hard interior surfaces. The airbag's interaction with the belt is critical: the belt controls the torso, while the airbag controls the head and supplements torso restraint.

Airbag correlation involves several parameters that directly affect occupant response:

• Deployment timing: The time from ACU trigger to full inflation. A late deployment may allow the occupant to contact the airbag while it is still inflating, causing high neck forces.
• Vent size and pressure: Controls the airbag's stiffness and energy absorption. A larger vent reduces pressure but may allow the head to bottom out.
• Bag geometry and tethering: Determines the contact surface shape and how the bag interacts with the belt and steering wheel.
• Inflator output: Mass flow rate and gas temperature affect the inflation profile.

In a typical 56 km/h frontal rigid barrier test, the airbag begins deploying at approximately 15–25 ms, reaches full inflation by 40–60 ms, and the occupant contacts the bag at 50–70 ms. The belt and airbag together manage the occupant's forward excursion, with the belt carrying the torso load and the airbag controlling head rotation and chest compression.

Interaction and Timing: The Coupled System

The restraint system's performance is determined by the sequence and overlap of these devices. The pretensioner fires first, removing slack. The belt then loads the occupant, and the load limiter begins paying out webbing once the force threshold is reached. The airbag deploys in parallel, and the occupant contacts it after the belt has already begun restraining the torso. If any of these events is mis-timed, the occupant's kinematics change, and the correlation between simulation and test diverges.

The following SVG illustrates the timing and force interaction of the restraint chain during a frontal crash event.

Practical Correlation Example: Belt Force Trace Matching

Consider a correlation task where the physical sled test shows a belt force peak of 3.8 kN at 55 ms, while the simulation predicts 4.5 kN at 48 ms. The discrepancy suggests the simulation's load limiter is engaging too early or the belt stiffness is too high. To diagnose:

1. Extract the belt force–time trace from both test and simulation.
2. Compare the pretensioner phase: if the simulation's force rises earlier, the pretensioner firing time or stroke may be too aggressive.
3. Check the load limiter force level: if the simulation's plateau is higher than the test, the force cap is set too high.
4. Verify the belt routing: a shorter effective belt length in the simulation increases stiffness and shifts the peak earlier.

After adjusting the load limiter force from 4.0 kN to 3.5 kN and delaying the pretensioner firing by 3 ms, the simulation's belt force peak moved to 3.9 kN at 53 ms — within 5% of the test. This kind of iterative tuning is the core of restraint correlation.

Key Takeaways for Correlation Engineers

• The seat belt is the primary load path; its routing and stiffness dominate torso kinematics.
• Pretensioners remove slack and must be timed to match the crash pulse; early or late firing changes occupant excursion.
• Load limiters cap chest force but increase excursion; their force–displacement curve must be modeled accurately.
• Airbags manage head and chest loading; their deployment timing and venting must align with belt engagement.
• Correlation is a coupled problem: adjust the belt first, then the airbag, and verify the full timing chain against test data.

With this foundation, subsequent chapters will address how to build and validate these components in simulation, how to design physical tests that isolate each device, and how to quantify correlation quality across the full restraint system.