Certification

Try a material grade (A106), a standard (API 5L) or a term (hazardous area).

↑↓ navigate↵ openAll results & filters →
ProcessMechanicalRotating EquipmentElectricalCivil & StructuralInstrumentation & ControlHSEAutomotivePharmaceuticalLearning pathsGlossaryStandardsNotesCertification

Automotive

Passive safety and crash engineering

Why a crumple zone is a distance and not a cushion, how a soft front and a hard cell work together, what a pretensioner and a load limiter each do, and the millisecond timeline an airbag has to hit.

IntermediateAutomotive

Standards referencedUN R94UN R95UN R137FMVSS 208

Everything up to here has been about avoiding the collision. This topic starts at the moment that has failed, and the problem becomes arithmetic that cannot be argued with.

A sectioned vehicle front structure mounted on a steel test stand in a laboratory, one crash box and longitudinal rail folded into a tight concertina of bright metal while the bumper beam and the rail beside it remain straight and undeformed, cut edges exposing the sheet thickness and the spot-welded flanges, a floodlight stand and a high speed camera on a tripod behind, the hall otherwise empty.
The energy is set by mass and speed. Only the distance is an engineering decision.

The one equation

You cannot reduce the kinetic energy of a moving vehicle. Mass and speed have already decided it. What you can decide is the distance over which it is shed, and deceleration follows directly:

a = v² / 2s

Speed Crush distance Deceleration
50 km/h (13.9 m/s) 50 mm ≈ 1930 m/s², about 197 g
50 km/h 300 mm ≈ 322 m/s², about 33 g
50 km/h 600 mm ≈ 161 m/s², about 16 g

That is the whole of crash engineering in one line. A crumple zone is not a cushion and it does not absorb the impact so you feel nothing. It is 600 mm of purchased distance, and it divides the deceleration by twelve.

Soft at the front, hard in the middle

A diagram of crumple zones and structural load paths. A side view of a vehicle is divided into three regions with a stiffness profile plotted beneath it, soft at the nose, rising through the front structure, and very stiff at the bulkhead and through the passenger cell, with a note that the two regions have opposite jobs and that a front which resists collapse transmits the deceleration straight through while a cell which collapses removes the survival space. A plan view shows the load paths through the front structure, with upper and lower longitudinal rails, a transverse bumper beam, the sills and the A-pillars all drawn as members sharing the impact, and an offset impact shown engaging the upper rail and the sill rather than loading a single rail past its capacity. A materials panel gives mild and formable grades in the crumple zone chosen to fold predictably and absorb energy, and press hardened boron steel at around 1500 megapascals in the cell chosen precisely for not deforming, with a warning that press hardened components generally cannot be heated and straightened because heat destroys the microstructure that gives the strength while the part still looks serviceable. A calculation panel shows deceleration as speed squared divided by twice the crush distance, worked at 50 kilometres per hour for 50 millimetres, 300 millimetres and 600 millimetres of crush, giving roughly 197 g, 33 g and 16 g respectively.
Graded stiffness. Every region is chosen for whether it should fold or refuse to.

The front structure and the passenger cell have opposite requirements.

The front must collapse, in a controlled and predictable order, to provide the distance. A nose that resists collapse passes the deceleration straight through to the occupants.

The cell must keep its shape. Survival space, the position of the steering column, the geometry the belt and airbag were designed around — all of it depends on the cell still being the shape it was. A cell that deforms has removed the thing the crumple zone was buying time for.

So the structure is graded: soft at the nose, progressively stiffer rearward, very stiff at the bulkhead.

Load paths

Real impacts are rarely square and rarely full width. A single front rail can be loaded past its capacity by an offset impact and simply fold sideways, letting the striking object into the cell.

So the load is spread — upper rail, lower rail, a transverse bumper beam tying them together, the sills, the A-pillars, transverse members across the floor. An impact anywhere across the front finds structure to engage.

The material follows the job

Region Material Chosen for
Crumple zone Mild and formable grades Folding predictably, absorbing energy as it goes
Safety cell Press-hardened boron steel, ≈1500 MPa Not deforming

Restraints, in milliseconds

A diagram of the restraint timeline and airbag deployment logic. A millisecond timeline runs from zero, with first contact at zero, the deceleration signature recognised and the fire decision taken by about fifteen milliseconds, the pretensioner firing and removing webbing slack in the same window, the airbag inflating and being fully deployed by around thirty milliseconds, the occupant meeting the bag between forty and sixty milliseconds, the load limiter paying out webbing under controlled force, and the bag venting and deflating so the occupant is not held against a pressurised surface. A note explains that the bag is a cushion the occupant meets rather than a device that pushes them back, so one still inflating when the occupant arrives is more dangerous than none. A deployment logic panel contrasts a kerb strike or pothole, which produces a sharp deceleration spike that stops, with a genuine crash, which spikes and keeps going, and explains that the algorithm judges the shape of the deceleration over time using several sensors rather than a single peak. A restraint panel plots belt force against occupant displacement, showing the pretensioner removing slack at the start, the force rising as the belt engages, the load limiter capping it at a threshold by paying webbing out, and the resulting trade of a little extra forward travel for a large reduction in chest loading. A final panel lists NCAP injury measures, giving the head injury criterion, chest compression, femur load and neck loads, with a caution that designing to the test is a known hazard.
Sensing, decision, pretension, inflation, arrival — all of it inside a tenth of a second.
Time What happens
0 ms First contact
~10–15 ms The algorithm recognises the deceleration signature and commits to firing
~15 ms Pretensioners fire — webbing slack removed
~30 ms Airbag fully inflated
~40–60 ms The occupant arrives at the bag
~60–80 ms Load limiter paying out, bag venting

Why slack is expensive

A pretensioner removes belt slack in the first few milliseconds. Slack sounds trivial — 40 mm of loose webbing, a thick coat, a belt not properly seated.

But slack is time, and at these speeds time is distance. An occupant who begins moving forward before the belt takes up has already spent part of the crush distance bought for them, and will meet the restraint harder when it finally engages.

Why the belt then lets go a little

A load limiter allows webbing to pay out at a controlled force once belt load exceeds a threshold.

That is not a contradiction of the pretensioner. The pretensioner removes slack that serves no purpose. The limiter releases webbing once the force itself would do the injuring — trading a little extra forward movement for a large reduction in chest loading. Engage early, then refuse to crush.

The airbag is a surface, not a push

The bag is something the occupant arrives at. It is not a device that pushes them backwards.

Which is why the timing is absolute. A bag still inflating when the occupant reaches it strikes them with an expanding surface, and that is considerably more dangerous than no bag at all. It is the entire reason an out-of-position occupant — leaning forward, too close to the wheel, a rearward child seat in front of a live airbag — is at such risk.

Telling a crash from a pothole

Both produce a sharp deceleration spike. The algorithm does not look at the peak; it looks at the shape over time. A kerb strike spikes and stops. A crash spikes and keeps going.

Getting that wrong in either direction is a safety failure — firing pyrotechnics during ordinary driving would itself cause a loss of control — which is why the decision is taken from several sensors, including satellite sensors mounted well forward, rather than from one.

How it is measured

Crash protocols instrument a dummy and read injury measures from it:

  • HIC, the head injury criterion — derived from head acceleration over time, not peak alone
  • Chest compression and chest acceleration
  • Femur load
  • Neck tension and shear

What to take away

  • You cannot reduce the energy, only the distance. a = v² / 2s.
  • 600 mm of crush instead of 50 mm divides the deceleration by twelve.
  • Soft front, hard cell. They have opposite jobs and both are deliberate.
  • Multiple load paths exist because real impacts are offset and partial.
  • Boron steel is chosen for not deforming — and generally cannot be heated and straightened.
  • Pretensioner removes slack early; load limiter caps force late. Both, together.
  • The airbag is a surface the occupant arrives at, so 30 ms is not a target but a requirement.
  • The algorithm reads the shape of the deceleration, not its peak.

Next, the discipline that decides how much of any of this is enough — ISO 26262, HARA and ASIL.

Check your understanding

10 questions. Nothing is recorded — this is just for you.

1What does a crumple zone actually achieve?
2A vehicle at 50 km/h stops against a rigid barrier with 600 mm of crush. Roughly what deceleration does the structure see?
3Why must the passenger cell be stiff while the front structure is soft?
4What is the purpose of multiple load paths in the front structure?
5Why is press-hardened boron steel used in the safety cell rather than in the crumple zone?
6What does this imply for collision repair?
7What does a seat belt pretensioner do, and when?
8What does a load limiter do, and why is it not a contradiction?
9Why must an airbag deploy within roughly the first 20 to 30 milliseconds?
10Why does the deployment algorithm need to distinguish a crash from a pothole or a kerb strike?

#vehicle systems#passive safety#crumple zone#load path#boron steel#airbag#pretensioner#load limiter#NCAP#HIC