Crash structures: why composites absorb energy differently from steel

Two completely different mechanisms

A steel crash box absorbs energy by yielding. It buckles into a concertina of plastic hinges, and every fold consumes energy as the metal deforms permanently. The material stays in one piece throughout, which makes the behaviour intuitive, repeatable and straightforward to simulate.

A composite crash tube does something that looks like failure: it destroys itself progressively from one end, in a continuous crush front of fibre fracture, matrix cracking, delamination between plies and friction as the fragments slide past each other. Nothing yields, because these materials barely yield at all — they are brittle. The energy goes into creating an enormous amount of new surface area.

The counter-intuitive part is that the second mechanism is the more efficient one per unit mass.

The numbers

The relevant metric is specific energy absorption, SEA — energy absorbed per kilogram of crushed material. Published values for CFRP tubes commonly fall in the 50–80 kJ/kg range, with individual studies reporting around 78.7 kJ/kg at higher crush velocity, up to about 99.7 kJ/kg quasi-statically, and foam-filled composite tubes above 100 kJ/kg.

Typical metallic crash structures sit substantially lower. That advantage is precisely why composite crash structures are mandatory in top-level motorsport and why they appear on the nose and side of a Formula 1 car: when the crash structure itself is part of the mass you are trying to minimise, energy per kilogram is the figure of merit.

A second metric matters as much: crush force efficiency, the ratio of mean crush force to peak force. A structure that spikes to a huge initial load and then collapses is worse for an occupant than one that holds a steady force through its stroke, even at the same total energy. Reported optimised composite designs quote efficiencies around 0.77 — and much of crash-structure design is chasing that number rather than the SEA headline.

The catch: composites do not fail gracefully by default

Progressive crushing is not what a composite tube does naturally. Hit one squarely without preparation and it is entirely capable of failing catastrophically instead — splitting, buckling globally, or shattering — and absorbing a fraction of the energy it should.

The fix is to force the failure to start where you want and then propagate steadily. In practice that means a trigger: a chamfered end, a tulip profile, a ply drop-off or a deliberate thickness reduction that concentrates stress at one end so the crush front initiates there and marches along the tube. Published work is consistent on this point — the amount of energy absorbed correlates directly with how much genuinely progressive crushing you achieve.

This is the single most important practical difference from metal. A steel crash box degrades gracefully if your design is imperfect. A composite one can fall off a cliff, and the gap between the SEA in a paper and the SEA in your part is usually the trigger geometry.

Why this makes design harder, not easier

Crush behaviour is sensitive to things that barely matter in a metal structure: the fibre architecture (braided, filament-wound and laminated tubes behave differently), the ply stacking sequence, the crush velocity, the trigger, and even off-axis loading. A tube optimised for an axial hit can perform poorly in an oblique one, which is a real-world load case, not an academic one.

Simulation is correspondingly harder. Metal plasticity is a mature, well-validated modelling problem. Composite crush involves several interacting failure modes at once, and models generally need calibrating against physical crush tests rather than being predictive from first principles.

And the material is not repairable in the way a metal crumple zone is. After any significant impact the structure is replaced, and — as with the battery enclosure — internal delamination from a sub-critical impact may leave the surface looking undamaged while the crush performance is compromised.

Where it connects to the rest of this site

Failure criteria are the entry point. Our failure envelope simulation plots the Tsai–Wu and maximum-stress surfaces for a lamina, which is how you establish the load state at which a ply first fails — the beginning of the process described here, and the point at which conventional laminate analysis stops being valid.

That boundary is worth stating plainly. Classical laminate theory and first-ply failure tell you when damage starts. Everything in this article happens afterwards, in the progressive-damage regime, and it needs explicit damage modelling and test correlation. Knowing which side of that line you are on is most of the skill.

A note on the numbers here

The SEA and crush force efficiency values above come from published academic literature on composite crashworthiness. They vary widely with tube geometry, fibre architecture, trigger design and test speed — the ranges quoted are representative of the literature, not values you should design to.

Any real crash structure is validated by physical testing to a regulated protocol. Nothing here substitutes for that.