3. THOR-M injury criteria

THOR-M (Test Device for Human Occupant Restraint, Modified) is a 50th-percentile male frontal impact dummy whose value in restraint optimization comes not from its anatomy alone but from the injury criteria it reports. Each body region carries a set of channels, filters, and threshold definitions that convert raw transducer signals into pass/fail judgments. Understanding these definitions is essential before any sled or vehicle test, because the choice of restraint parameters — pretensioner timing, load limiter force, airbag venting, knee bolster stiffness — is ultimately judged against these numbers.

Two conventions must be fixed before reading any metric. First, the coordinate system: +X is forward (the direction of travel of the vehicle before impact), +Y is to the occupant's left, +Z is upward. Accelerations are expressed in g (1 g = 9.81 m/s²), and the sign follows the axis direction, so a chest acceleration of −45 g during a frontal pulse means 45 g of deceleration along −X. Second, all channels are filtered per SAE J211/1: head and chest accelerations at CFC 1000, neck forces and moments at CFC 600 (or CFC 1000 for moments in some protocols), and chest deflection at CFC 180. Applying the wrong filter class changes peak values by several percent and can flip a marginal result, so the filter class is part of the criterion definition, not an afterthought.

Head

The head criterion is the Head Injury Criterion (HIC), computed from the resultant translational acceleration at the head center of gravity:

HIC = max over t1,t2 of [ (t2 - t1) * ( (1/(t2-t1)) * ∫ a_R(t) dt )^2.5 ]
subject to t2 - t1 ≤ 15 ms

where a_R is the resultant acceleration in g. The 15 ms window constraint prevents a very short, very high spike from dominating. The commonly cited threshold is HIC₁₅ = 700, with HIC₃₆ = 1000 used in some regulatory contexts. In practice, HIC is dominated by the deceleration phase between roughly 30 and 80 ms in a 50 km/h frontal pulse; a head that contacts the airbag early and rides it down typically produces a broader, lower peak than one that bottoms out the bag and strikes the steering wheel hub. A related but distinct metric, the Brain Injury Criterion (BrIC), uses rotational velocity about the three axes and is increasingly reported alongside HIC for THOR-M tests, with a threshold around 0.68–1.0 depending on the protocol.

Neck

Neck injury is assessed through four families of metrics: axial force (tension and compression), shear force, bending moment (flexion and extension), and the Neck Injury Criterion (Nij). The Nij formulation couples axial force and bending moment:

Nij = (F_z / F_int) + (M_y / M_int)

where F_int and M_int are the intercept values for the relevant load direction (tension/extension, tension/flexion, compression/extension, compression/flexion). The threshold is Nij ≤ 1.0. Typical intercept values used with THOR-M are on the order of F_int = 6.8 kN (tension) or 6.2 kN (compression) and M_int = 200–300 N·m depending on the direction pair; these must be taken from the specific test protocol rather than assumed. Separately, peak tension is often limited to about 4.0 kN and peak compression to about 4.0 kN, with flexion and extension moments limited to roughly 190 N·m and 57 N·m respectively in many frontal protocols. The neck is where restraint tuning is most sensitive: an over-aggressive pretensioner that pulls the occupant into the bag too early can raise extension moment, while a bag that deploys too late lets the head rotate forward and loads flexion.

Chest

THOR-M measures chest injury with three independent channels. Chest deflection is measured by an IR-TRACC (Infra-Red Telescoping Rod for Assessment of Chest Compression) or a similar displacement sensor, reported as absolute deflection in millimeters and as a normalized deflection ratio. The classic threshold is 63 mm absolute, or a normalized deflection of 0.33–0.35 when scaled to chest depth. Chest acceleration is the resultant of the three accelerometers at the chest center, with a 3 ms clip threshold typically at 60 g. The Viscous Criterion (VC) combines deflection rate and compression:

VC = max over time of [ (dD(t)/dt) * (D(t)/D_max) ]

with a threshold of 1.0 m/s. VC captures the rate-dependent soft-tissue injury that a pure deflection limit misses, and it is the metric most directly influenced by load limiter force level: a higher limiter force increases deflection rate even if peak deflection stays constant. THOR-M's four-point chest deflection measurement (upper left, upper right, lower left, lower right) also allows asymmetry to be reported, which is useful when evaluating oblique or offset frontal configurations.

Abdomen

The abdomen is assessed with a pressure-based or deflection-based criterion depending on the build. THOR-M uses abdominal pressure sensors and, in some configurations, a deflection measurement at the abdomen. The commonly applied threshold is a peak abdominal pressure of about 1.0 bar (100 kPa) summed across the sensors, or an abdominal deflection limit around 50 mm. Abdominal loading is strongly coupled to belt geometry: a lap belt that sits too high, above the iliac crest, concentrates force on the abdomen and drives pressure up even when chest metrics look acceptable. This is why the abdomen channel is a useful diagnostic when a restraint package passes chest but fails overall.

Lower Extremity

Lower extremity metrics in THOR-M cover the femur, knee, and tibia. Femoral force is measured along the femur axis, with a threshold typically around 10 kN for the 50th-percentile male. Knee loading is captured through knee slider force and knee moment, with thresholds on the order of 6 kN and 200 N·m respectively in many protocols. Tibia axial force is limited to roughly 8 kN, and the Tibia Index combines axial force and bending moment:

TI = (F_z / F_c) + (M / M_c)

with F_c and M_c being critical values (approximately 35.9 kN and 225 N·m in the standard formulation) and a threshold of TI ≤ 1.0. Lower extremity results are dominated by the knee bolster, the footwell geometry, and the lap belt path; a stiff bolster that stops the knee early reduces femur force but can raise tibia index if the load path is misaligned.

THOR-M Injury Criteria by Body Region Coordinate system: +X forward, +Y left, +Z up · Filters per SAE J211/1 Head HIC₁₅ ≤ 700 (HIC₃₆ ≤ 1000) BrIC ≈ 0.68–1.0 a_R at head CG, CFC 1000 Window: t₂ − t₁ ≤ 15 ms Neck Nij ≤ 1.0 Tension / compression ≈ 4.0 kN Flexion ≈ 190 N·m · Extension ≈ 57 N·m F_int ≈ 6.2–6.8 kN · M_int ≈ 200–300 N·m Chest Deflection ≤ 63 mm (norm. ≈ 0.33) 3 ms clip ≤ 60 g VC ≤ 1.0 m/s 4-point IR-TRACC deflection Abdomen Peak pressure ≈ 1.0 bar (100 kPa) Deflection ≈ 50 mm Sensitive to lap-belt height Lower Extremity Femur axial force ≈ 10 kN Knee slider force ≈ 6 kN · knee moment ≈ 200 N·m Tibia axial force ≈ 8 kN · Tibia Index ≤ 1.0 (F_c ≈ 35.9 kN, M_c ≈ 225 N·m) Restraint Parameters That Drive Each Region Head: airbag venting, bag timing, steering column stroke Neck: pretensioner timing, load limiter force, bag position Chest: load limiter force, belt geometry, bag coverage Abdomen: lap-belt height relative to iliac crest Lower extremity: knee bolster stiffness, footwell geometry All channels filtered per SAE J211/1

Reading the Metrics Together

The criteria are not independent, and a restraint package that optimizes one region often degrades another. A useful engineering workflow is to treat the channels as a coupled system. Consider a representative 50 km/h frontal sled case with a THOR-M in the driver position, a force-limiting retractor set to 4 kN, a pretensioner firing at 12 ms, and a passenger airbag with two vent sizes under evaluation. If the 4 kN limiter produces chest deflection of 58 mm and VC of 0.92 — both passing but close — while HIC₁₅ sits at 620, the package is chest-limited. Reducing the limiter to 3.5 kN typically lowers deflection and VC but allows more forward excursion, which raises head travel and can increase HIC if the bag is already bottoming out. The correct move is usually to adjust bag venting or bag stand-off distance rather than to keep lowering limiter force, because the head channel has more margin than the chest channel in this example.

A second common pattern appears in the neck. If Nij is passing but extension moment is near its limit, the pretensioner is likely firing too early relative to bag deployment, pulling the occupant forward into a bag that has not yet reached full inflation. Delaying the pretensioner by 4–6 ms, or increasing bag inflator output, usually reduces extension moment without materially changing chest metrics. Conversely, if flexion moment is the limiting channel, the bag is deploying too late or is positioned too high, and the head rotates forward under the bag rather than being caught by it.

When a test fails a single channel, first check the filter class and the coordinate sign convention before changing hardware. A CFC 600 versus CFC 1000 mismatch on neck moment, or a sign error on chest acceleration, produces failures that no restraint tuning can fix.
Threshold values quoted here are representative of common frontal protocols and are not a substitute for the governing regulation or OEM specification. Intercept values for Nij and Tibia Index, in particular, vary between protocols and must be taken from the applicable document before any pass/fail judgment is issued.

Finally, the lower extremity channels deserve attention in any package that uses a knee bolster as a load path. Femur force and tibia index respond in opposite directions to bolster stiffness: a stiffer bolster reduces femur force by stopping the knee earlier but can raise tibia index if the resultant load vector is not aligned with the tibia axis. The practical check is to compare the femur force and tibia index trends across bolster variants; if both improve, the geometry is sound, and if they diverge, the load path needs realignment rather than a stiffness change. Used this way, the THOR-M criteria set is not a scorecard applied after the fact but a diagnostic map that tells the engineer which restraint parameter to move next.

← Previous 📖 Contents Next →