HIC calculator: head injury criterion (HIC15 and HIC36) from your acceleration curve
HIC is the largest value of (t2 − t1) × (mean acceleration)^2.5 over any window up to 15 ms (HIC15) or 36 ms (HIC36), with the acceleration in g and the time in seconds. Paste your head acceleration curve, pick the units and an SAE J211 filter class. You get HIC15, HIC36, the window, the peak, the Euro NCAP and UN R127 ratings and the injury risk. Free, runs in your browser.
The second tab sizes the pulse and the space under a bonnet for a target HIC. The third checks the HIC in a head impact report from Skill-Lync's LS-DYNA course for unit, window and HIC(d) slips. Skill-Lync read all 155 public Week 7 head impact reports to build it.
Euro NCAP VRU protocol v1.2 · UN R127 · FMVSS 201 and 208 · SAE J211-1 Appendix C · NHTSA risk curves · checked against a separate solution (4,234 checks at build) · Updated 22 Nov 2026
The search runs over every pair of samples, with the integral by the trapezoid rule between them, as SAE J211-1 9.4.3 asks ("use all sampled data points"). With four columns each axis is filtered first, then the resultant is formed. Times must increase; repeated or unsorted times are refused.
The pulses bring the head to rest (no rebound); a symmetric pulse of length T then needs a stopping distance of vT/2. The spring is elastic: the head leaves at the speed it came in, so the speed change is 2v. Euro NCAP tests at 40 km/h (11.1 m/s); UN R127 and the NHTSA proposal at 35 km/h (9.7 m/s).
The default is one public report's numbers: a mean of 49.1 g from 2.5 to 30 ms, called HIC and compared with the 650 Euro NCAP band. Type yours. Leave the mean blank if you only have the plot.
HIC15 windowHIC36 windowbefore filteringresultant used
- Euro NCAP pedestrian headform (VRU protocol v1.2): Brown, 0.25 of 1 point for this grid point at 40 km/h.
- UN R127 / GTR 9: above 1,000, so only acceptable inside the manufacturer's HIC 1,700 zone (at most a third of the bonnet top).
- For comparison only (dummies in a crash test, not a headform): FMVSS 208 limits a Hybrid III 50th male to HIC15 700. FMVSS 201 interior free-motion headform; not a pedestrian test: HIC(d) = 0.75446 × HIC36 + 166.4 = 1,201 against 1,000.
- One acceleration column was filtered. SAE J211-1 9.4.1 filters before forming a resultant; with four columns the tool filters each axis first, as LS-PrePost's pre-filter does.
| Risk curve (NHTSA, lognormal) | At your HIC15 |
|---|---|
| AIS 2+ (skull fracture) | 62% |
| AIS 3+ (serious) | 38% |
| AIS 4+ (severe) | 24% |
| Pulse that stops the head | Duration | Peak | Stopping distance |
|---|---|---|---|
| Half-sine | 14.5 ms | 122 g | 80.7 mm |
| Haversine | 17.6 ms | 129 g | 98 mm |
| Triangle | 15.4 ms | 147 g | 85.4 mm |
| Square | 12.3 ms | 92 g | 68.4 mm |
- The 27.5 ms window is fine for HIC36, but compare it with the Euro NCAP or R127 limits only after searching again with a 15 ms window: HIC15 can only be the same or lower.
- Recomputed: (27.5 ms = 0.0275 s) × (49.1 g)^2.5 = 465.
- Your 464 matches it (within 2%).
- A 98 g peak at 40 km/h (11.11 m/s) is a contact of 18.16 to 36.32 ms if the pulse is a half-sine (head stopped to full rebound). The head travels 101 to 128 mm into the surface, and HIC15 is 716 to 1,194.
- Case 3 is the test that matters: the hood bends and gives the head stopping distance. Compare its HIC15 with the Euro NCAP bands and the R127 limits.
What is the HIC formula?
HIC is a sliding-window average of the head's resultant acceleration, raised to the power 2.5 and multiplied by the window length. FMVSS 208 and UN R127 define it with a in g and t in seconds:
The 2.5 power makes HIC grow faster than the acceleration: double the mean and HIC goes up 22.5 = 5.7 times, while doubling the time it lasts only doubles it. So the units matter. An explicit solver run in kg-mm-ms writes acceleration in mm/ms², and
Read mm/ms² as g and HIC comes out 101.972.5 = 105,000 times too small. Put t in ms instead of seconds and it is 1,000 times too big. This calculator converts both for you. It also searches t1 and t2 only at your samples, with the trapezoid rule between them, and filters before it forms the resultant, as SAE J211-1 9.4.1 asks.
What HIC is acceptable in Euro NCAP and UN R127?
HIC15 below 650 scores full marks in Euro NCAP, and HIC15 1,000 is the UN R127 limit over two-thirds of the bonnet. Both use the 15 ms window and a free-flying headform. Euro NCAP's Vulnerable Road User protocol (v1.2, July 2026) colours each grid point by its HIC15:
| Euro NCAP colour | HIC15 | Points |
|---|---|---|
| Green | below 650 | 1.00 |
| Yellow | 650 to below 1,000 | 0.75 |
| Orange | 1,000 to below 1,350 | 0.50 |
| Brown | 1,350 to below 1,700 | 0.25 |
| Red | 1,700 or more | 0.00 |
The points over all grid points scale to 10 for the head zone, after a correction factor from Euro NCAP's own verification tests. Those tests allow 10% on the HIC before a colour counts as wrong. The test conditions differ between the two:
| Euro NCAP VRU v1.2 | UN R127 (04 series) | |
|---|---|---|
| Speed at contact | 11.1 ± 0.2 m/s (40 km/h) | 9.7 ± 0.2 m/s (35 km/h) |
| Child headform | 3.5 kg, 165 mm, 50° ± 2° to the ground, WAD 1000 to 1500 | 3.5 ± 0.07 kg, 165 mm, 50° ± 2° |
| Adult headform | 4.5 kg, 65° ± 2°, rearward of WAD 1700 | 4.5 ± 0.1 kg, 165 mm, 65° ± 2° |
| Limit | colour bands above | HIC 1,000 over two-thirds of the bonnet top (and over half the child area), 1,700 elsewhere |
| Acceleration channel | CFC 1000, recorded at 20 kHz or more | CFC 1000, CAC 500 g (ISO 6487) |
Euro NCAP's headforms are those of UN R127 Annex 4 (Euro NCAP CP 401). WAD is the wrap-around distance from the ground over the car's front. NHTSA proposed the same GTR 9 test as FMVSS 228 in September 2024 (35 km/h, HIC15 1,000 and 1,700); it is not yet in force.
Two more limits often turn up in reports, but neither is a pedestrian test. FMVSS 208 limits HIC15 to 700 for the Hybrid III 50th percentile male in a crash test (390, 570 and 700 for the 12-month, 3-year and 6-year-old dummies). FMVSS 201 uses HIC(d), below.
Should you use HIC(d) for a pedestrian head impact?
No. HIC(d) belongs to FMVSS 201, the US standard for heads hitting the inside of a car:
There a free-motion headform is fired at up to 24 km/h into pillars, headers and roof rails, and its HIC (36 ms window) is turned into the value a full dummy would see. NHTSA calls it a transfer function. A pedestrian headform is judged on its own HIC15 by Euro NCAP and R127, with no conversion. Note too that HIC(d) can never fall below 166.4: one report gives HIC36 = 0.2468 and HIC(d) = 166.6 for the hood case (report), which means the HIC under it was nearly zero, a sign of a unit setting, not of a safe hood. An HIC(d) of 1,000 is a headform HIC of 1,105.
48 of the 155 reports quote the HIC(d) formula and 34 print an HIC(d) value for their run; only 7 name FMVSS at all.
How do you filter head acceleration with CFC 1000?
With the 4-pole phaseless Butterworth filter of SAE J211-1 Appendix C, at channel frequency class 1000 for the head. Euro NCAP, UN R127 and the NHTSA proposal all ask for CFC 1000. The filter is a 2-pole recursion run forward and then backward, so it does not shift the peak in time:
T is the sample period in seconds. The coefficients are a 2-pole Butterworth at 2.0775 × CFC; run twice, the response is 3 dB down at 5/3 × CFC (1,667 Hz for CFC 1000). Skill-Lync's verifier checks this against scipy's own Butterworth design. To start the filter cleanly, J211 copies at least 10 ms at each end and turns it about the end point, filters, then drops the copies. This calculator does exactly that and starts each pass with its first two outputs equal to its first two inputs (J211 asks you to report the start-up you used).
Three sampling rules come with it. J211-1 9.2 asks for at least ten times the CFC: 10 kHz, a 0.1 ms output interval, for CFC 1000. Euro NCAP records at 20 kHz or more. And the formula itself breaks down below 4.155 × CFC, where the tangent term blows up, so the calculator refuses to filter there.
None of the 155 Week 7 reports filters the head acceleration. On a long hood pulse it changes little. On the very short case 1 pulse it changes everything: the synthetic case 1 curve in the first tab (a half-sine of 31,000 g lasting 0.115 ms) gives HIC15 8.054e6 unfiltered and 1.332e6 after CFC 1000, and its peak drops to 10,011 g. A pulse that short is mostly above the filter's band.
How often should LS-DYNA write NODOUT for HIC?
Every 0.05 ms (20 kHz) to match Euro NCAP, and never coarser than 0.1 ms. HIC is only as good as the samples it is searched on. Of the 26 reports that state their ASCII output interval, 14 write 0.5 ms or 1 ms, that is 2,000 or 1,000 samples a second. One sets 1 ms for both D3PLOT and ASCII output (report). The same synthetic pulses show what that does:
| Synthetic half-sine at 40 km/h | Pulse | Fine samples | Every 0.1 ms | Every 0.5 ms |
|---|---|---|---|---|
| Case 1, 31,000 g | 0.115 ms | 8.054e6 | 5,84,805 | 0 |
| Case 2, 4,700 g | 0.757 ms | 4,75,496 | 4,50,487 | 1,92,273 |
| Case 3, 98 g | 18.2 ms | 716 | 716 | 714 |
HIC15, unfiltered. Fine samples: 0.001 ms, 0.005 ms, 0.05 ms. The pulse starts at 2 ms, so the 0.5 ms samples fall at 2.0 and 2.5 ms. These are illustrative curves, not LS-DYNA output.
At 0.5 ms the case 1 pulse falls between two samples and HIC is 0. The hood case, 18.2 ms long, barely moves (716 against 714). If your HIC changes when you halve the output interval, the interval was too coarse.
What injury risk does a HIC value mean?
NHTSA fits the probability of a head injury of at least a given AIS level (Abbreviated Injury Scale, 1 minor to 6 maximum) with lognormal curves:
| HIC | AIS 2+ (skull fracture) | AIS 3+ (serious) | AIS 4+ (severe) |
|---|---|---|---|
| 650 | 28% | 9.4% | 2.6% |
| 700 | 31% | 11% | 3.4% |
| 1,000 | 47% | 23% | 11% |
| 1,350 | 61% | 37% | 23% |
| 1,700 | 71% | 49% | 36% |
AIS 2+: μ 6.96352, σ 0.84664 (Eppinger et al. 1999, 2.3: skull fracture, fitted on 54 head impacts of 0.9 to 10.1 ms, so it applies to HIC15 and HIC36). AIS 3+: μ 7.45231, σ 0.73998; AIS 4+: μ 7.65605, σ 0.6058 (Kuppa 2004, written for HIC36).
The numbers check out against what NHTSA publishes: 31% skull fracture risk at HIC 700 (Eppinger: 31%), and 11% and 36% AIS 4+ risk at 1,000 and 1,700 (NHTSA's 2024 proposal: 11% and 36%). One mismatch: Kuppa's own table puts the AIS 3+ 25% and 50% points at HIC 950 and 1,680, while the μ and σ printed beside it give 1,046 and 1,724. The AIS 2+ and AIS 4+ rows match their tables. The calculator uses the curves as printed.
35 reports copy "HIC 650 = 90% level 1, 55% level 2, 20% level 3, 5% level 4" and 29 copy "at a HIC of 1000, an 18% probability of a severe, 55% of a serious and 90% of a moderate head injury"; 52 quote one or the other, and some quote both, though they cannot both be right. Skill-Lync could not find the original coefficients of either in an NHTSA or primary document, so this page does not use them. On NHTSA's curves HIC 650 is 28% AIS 2+, 9.4% AIS 3+ and 2.6% AIS 4+.
How much space under the bonnet does a head need?
About 80.7 mm of travel for HIC15 1,000 at 40 km/h, if the bonnet could make a smooth half-sine pulse, and about 108 mm for 650. The second tab works it out for any speed and target. The cleanest case is a square pulse, constant deceleration for a time T:
For the other shapes HIC is a fixed fraction of T × peak2.5 while the pulse fits in the window. Skill-Lync's verifier finds the constants by optimising t1 and t2 on the closed-form integrals:
At 40 km/h, to stay at HIC15 1,000:
| Pulse | Duration | Peak | Stopping distance | For HIC15 650 |
|---|---|---|---|---|
| Half-sine | 14.5 ms | 122 g | 80.7 mm | 108 mm, 91.9 g |
| Haversine | 17.6 ms | 129 g | 98 mm | 131 mm, 96.4 g |
| Triangle | 15.4 ms | 147 g | 85.4 mm | 114 mm, 111 g |
| Square | 12.3 ms | 92 g | 68.4 mm | 87.9 mm, 71.6 g |
The square pulse needs the least room (68.4 mm): a perfectly flat deceleration uses the space best. Any real bonnet does worse, and the head must not reach the engine or hinges inside that distance. A bonnet acting as a simple spring does worse still, because an elastic spring throws the head back and doubles the speed change:
For the 3.5 kg child headform at 40 km/h, HIC15 1,000 needs k = 22.2 N/mm, and the spring then deflects 139 mm over a 39.4 ms contact. That is why bonnets are designed to dent and absorb, not to spring back.
Why is the Week 7 case 1 HIC in the millions?
Because a shell sphere hitting a rigid wall stops in a fraction of a millimetre, so the number says "rigid against rigid", not "fatal". One report gives a peak of about 31,000 g for case 1. If that is a half-sine and the sphere bounces off at the 11.11 m/s it arrived with, the contact lasts 0.115 ms, and HIC15 is 8.056e6. The spring that gives that pulse, for a 3.5 kg head, is k = m(π/T)² = 2.620e6 N/mm, and it compresses 0.406 mm.
No head does that. What decides it is how stiff the contact and the sphere's own shell are, not anything about injury. The reports bear this out: 61 of the 155 print a HIC of 10,000 or more somewhere, 36 of them a million or more, for example 2.435e8 (report). Case 2, the headform on a rigid wall, is a reference too: only its 14 mm headskin (R127 Annex 4) cushions it. Case 3, the headform on the hood, is the one to rate.
What did 155 Week 7 head impact reports do?
Skill-Lync read every public report of the challenge on 22 Nov 2026. The counts below come from text searches checked by reading the matches; text in screenshots is not counted, so each is a floor.
- Window: 75 mention HIC36 and 47 HIC15; 45 only ever mention HIC36, though Euro NCAP and R127 rate HIC15. One averages from 2.5 to 30 ms, a 27.5 ms window, and compares the result with the 650 band (report); it is the default in the third tab.
- Units: 52 state kg-mm-ms and 93 type 11.11 mm/ms for 40 km/h. Only 1 says it set LS-PrePost's gravity to 0.00981 and the time unit to msec (report). 17 describe a as "a multiple of 10 m/s² or about 1 g"; 1 g is 9.80665 m/s².
- Angle: 68 mention 45° and 50 mention 50° (9 both). Euro NCAP and R127 fire the child headform at 50°, the adult at 65°.
- Accelerometer: 80 mention an accelerometer; 4 take HIC from a random or any node.
- Standards: 39 cite Euro NCAP, 23 GTR 9 or R127. None states the 3.5 kg of the current child headform; 2 quote 2.5 kg, the older EEVC value.
What this calculator does not cover
- Rotational head injury (BrIC and similar): HIC uses linear acceleration only.
- The current SAE J211/1 (2022) text. It is not public; the filter follows the public 1995 Appendix C.
- How LS-PrePost searches t1 and t2, and its filter defaults. The tutorial we read doesn't say, so compare the window you get here with the one LS-PrePost shows.
- A published figure for the space a bonnet needs. None we could check; the second tab gives the physics only.
Running the head impact is the easy part; explaining why case 1 reads 10⁸ and what the hood's HIC15 means for a Euro NCAP score is what safety teams hire for. The pedestrian head impact is a project in Skill-Lync's LS-DYNA course.
Frequently asked questions
How do you calculate HIC?
Take the resultant head acceleration in g and the time in seconds. For every pair of times t1 and t2 up to 15 ms apart, multiply the window length by the mean acceleration raised to the power 2.5. HIC15 is the largest value. A 100 g mean held for 10 ms gives 0.010 × 100^2.5 = 1,000.
What is the difference between HIC15 and HIC36?
Only the longest window allowed: 15 ms or 36 ms. HIC36 searches more windows, so it is never lower than HIC15. For a short hard impact, such as a headform on a bonnet, they come out equal. Euro NCAP and UN R127 pedestrian tests use HIC15; FMVSS 201 interior tests use 36 ms.
What HIC value is acceptable for pedestrian protection?
Euro NCAP awards a headform grid point full marks below HIC15 650, then 0.75, 0.50 and 0.25 points up to 1,000, 1,350 and 1,700, and nothing above 1,700. UN R127 requires HIC 1,000 or less over two-thirds of the bonnet top and 1,700 or less everywhere tested.
What is HIC(d) in FMVSS 201?
HIC(d) = 0.75446 × HIC + 166.4, with a 36 ms window, must stay at or below 1,000. It converts the HIC of a free-motion headform fired at up to 24 km/h into pillars, headers and roof rails into a dummy-equivalent value. It belongs to that interior test, not to a pedestrian bonnet test.
How do I set up HIC in LS-PrePost for a kg-mm-ms model?
Ansys's LS-PrePost tutorial sets Settings, Post Settings, Hic/Csi constants to time unit msec and gravity 0.00981, then plots HIC from the NODOUT file of the node you requested. Write NODOUT often enough: Euro NCAP records at 20 kHz, which is a 0.05 ms output interval.
What does a HIC of 1000 mean?
On NHTSA's lognormal curves, HIC 1,000 carries about 11% risk of an AIS 4+ head injury, 23% of AIS 3+ and 47% of a skull fracture (AIS 2+). In Euro NCAP it is the boundary between yellow and orange; in UN R127 it is the limit for two-thirds of the bonnet.
Related on Skill-Lync
- Crash box calculator: mean crushing force and energy of crush tubes
- Explicit time step calculator: the LS-DYNA time step, and why NODOUT can be written far more often than every 0.5 ms
- Consistent units calculator: mm/ms², kg-mm-ms and 11 more systems
- Bharat NCAP ratings in 2026 and how to read car safety ratings
- Designing a car bonnet's frontal BIW enclosure
- Student project: Week 7 head impact
- LS-DYNA courses compared
- All free engineering tools
Sources
- Euro NCAP, Crash Protection: Vulnerable Road User Impacts Protocol, v1.2 (July 2026): 1.3.1 (20 kHz), 3.4.2.12-21 (headforms, 50° and 65°, 11.1 m/s), 4.2.1 colour bands, 4.4.2 points and HIC tolerance. Technical Bulletin CP 401 v1.0 (March 2025): headforms per UN R127 Annex 4, CFC 1000, CAC 500 g.
- UN Regulation No. 127, pedestrian safety, as published in the EU Official Journal L 2025/460 (incorporating Supplement 2 to the 04 series): 2.24 HIC definition, 5.2 head limits, Annex 4 headforms (3.5 and 4.5 kg, 165 mm, 14 mm skin, CFC 1000), Annex 5 4.6-5.7 (9.7 m/s, 50°, 65°).
- 49 CFR 571.201 (FMVSS 201), S6.1, S7 and S8.9; NHTSA final rule, 60 FR 43031 (18 Aug 1995), on HIC(d) as a dummy-equivalent transfer function. 49 CFR 571.208 (FMVSS 208), S6.2, S6.3, S15.3.2, S19.4.2, S21.5.2, S23.5.2.
- NHTSA, proposed FMVSS No. 228 Pedestrian Head Protection (GTR 9), 89 FR 76922 (19 Sept 2024): headforms, 35 km/h, HIC1000 and HIC1700 areas, 11% and 36% AIS 4+ risk, CFC 1000.
- SAE J211-1 (Rev. MAR95), Instrumentation for Impact Test, Part 1 (public copy incorporated by reference): Appendix C filter, 3.7, 9.2, 9.4.1, 9.4.3, Table 1.
- R. Eppinger et al., Development of Improved Injury Criteria for the Assessment of Advanced Automotive Restraint Systems II, NHTSA (Nov 1999), 2.3. S. Kuppa, Injury Criteria for Side Impact Dummies, NHTSA (May 2004): head injury risk curves.
- Ansys LS-PrePost tutorials, Post tutorials, chapter 1 (Calculate occupant injury criterion): 1.2 unit settings (msec, 0.00981), 1.3 HIC plot from NODOUT, 1.5 pre-filtering. NIST: gn = 9.80665 m/s² (exact).
- Skill-Lync, all 155 public Week 7 head impact reports, read 22 Nov 2026. scripts/verify/hic.py recomputes every case by brute force over all sample pairs, optimises the pulses on closed-form integrals, re-implements the J211 filter and checks it against scipy, and simulates the spring; the page checks itself against it at build time.