Resources · Solutions Blog · Impact / Brittle Fracture

How Hard Can You Slam a Car Door Before the Window Shatters?

Rand Simulation — Applications Engineering AI
Impact & brittle fracture · Ansys LS-DYNA · 7 min read

AI disclosure: RandSim Labs is an experimental AI-driven engineering simulation platform. Content on this site, including simulations, analyses, figures, and written materials, may be generated or assisted by AI using licensed Ansys tools. AI-generated content may contain errors and is provided for educational, informational, and demonstration purposes only. Users should independently verify all results before relying on them for engineering, design, manufacturing, safety, or other production decisions.

Everyone has done it: closed a car door a little too hard, heard the whole frame boom, and winced — is the window about to go? Someone in the Ideas Lab wanted a real number. So we built the side window in Ansys LS-DYNA, arrested the door frame the way a latch does, and let the glass keep going. Then we turned the slam up until it broke.

The moment of failure — a tempered side window driven far past any real slam (a door-edge closing speed of 10 m/s, roughly 22 mph). It doesn’t crack into daggers; it dices into thousands of small, blunt cubes, exactly the way tempered safety glass is engineered to fail. Color is surface tensile stress; each element vanishes as it reaches the fracture stress. As we’ll show, you cannot actually slam a door anywhere near this hard.

The physics: the glass keeps going when the door stops

A door swings on a vertical hinge, and at the instant the latch catches, the door’s velocity is pointed straight through the window plane. The latch arrests the frame in a millisecond or two — but the pane of glass held inside it still carries all its momentum. For a brief moment the frame has stopped and the middle of the glass has not, so the pane bends outward about its supported edges. Bending a plate puts one face into tension, and glass is brittle: it fails from tensile surface stress. If that stress reaches the glass’s fracture strength, a surface flaw runs and the pane lets go.

That is the whole mechanism, and it is clean to model: give the whole assembly the closing speed, then decelerate the supports to a stop over the latch time while the glass is free to overshoot. The peak tensile stress lands right where you’d expect — along the bottom clamp line and the lower corners.

Inside the model

The window is a tempered pane, 450 × 400 mm and 3.2 mm thick, meshed with about 7,200 fully-integrated shell elements (ELFORM 16, five integration points through the thickness so the bending surface stress is resolved). Glass is a linear-elastic solid (E = 71 GPa, ν = 0.22) with a maximum-principal-stress erosion criterion — when a surface fiber reaches the fracture stress, that element is deleted, which is how the crack and the dicing propagate. The slam itself is a prescribed velocity ramp: everything starts at the closing speed, then the edge supports are brought to rest over a ~1.5 ms latch time. We ran it two ways — a realistic mount (a soft rubber run-channel around three edges, gripped at the bottom by the regulator) and an all-edges-rigidly-clamped bound — so the answer doesn’t hinge on one modeling choice.

The result: a normal-to-hard door slam is a door-edge speed of about 1–3 m/s, which loads the glass to only 40–90 MPa — well under half the tempered fracture stress. To actually shatter the tempered pane you need a door-edge closing speed near 6.5–7 m/s (~15 mph), roughly three times the hardest realistic slam. Your window is safe by a wide margin.

Is it right? Tempered vs. ordinary glass — the number that matters

The whole answer turns on one material number: the glass’s fracture stress. Ordinary annealed glass fails at roughly 40–80 MPa — it’s surface-flaw limited. Automotive side glass is tempered: the surface is locked into about 100 MPa of residual compression, so an applied tension has to first cancel that compression and only then reach the intrinsic strength. That puts tempered fracture near 150 MPa — and it is exactly why side windows dice into harmless cubes instead of shattering into shards. We ran the speed sweep against both strengths:

Peak surface tensile stress in the glass versus door-edge closing speed, for the realistic mount (navy) and the rigid-clamp bound (gray). The two mounting models bracket the same answer. Crossing the tempered strength (green, ~150 MPa) takes ~6.5–7 m/s; crossing the annealed strength (orange, 80 MPa) takes only ~4 m/s. Tempering nearly **doubles** the survivable slam — that is the safety margin engineered into the glass.

So if your side window were ordinary annealed glass, a genuinely violent ~4 m/s slam could crack it. Because it is tempered, you would have to close the door at ~15 mph — which no human arm can do. The control here isn’t a single simulation; it’s the contrast between the two strengths, and the fact that two very different mounting assumptions give the same threshold within a few percent.

The real-world connection

This is the same reason a spring-loaded emergency “window punch” shatters a car window instantly while a hard slam does nothing: the punch is a sharp point load that spikes the local stress past 150 MPa in one tiny spot, while a slam spreads its energy across the whole pane and never gets close. Distributed bending is simply the wrong way to break tempered glass — which, for a side window whose job is to survive years of slams, is precisely the point. And when tempered glass does finally go, it goes the safe way: the stored surface energy releases into thousands of small cubes, the failure mode you saw in the hero clip.

Honest scope. The tempered strength is modeled as an elevated effective fracture stress (~150 MPa, with a 120–200 MPa sensitivity band) that bakes in the residual surface compression, rather than an explicit pre-stress field — the standard demonstration simplification. Single-element erosion visualizes the release and dices at the mesh scale (~5 mm, close to real tempered fragment size) but is not a validated fragment-size prediction. The pane is a flat monolithic idealization of a gently-curved real window, and the load is the one-way inertial-bending mechanism from a latch arrest — not a point impact, a twisting slam, or a pre-cracked edge. What we stand behind: the peak surface stress rises with slam speed as an inertial-bending problem should; a realistic and a rigid mounting bracket the same shatter threshold to within a few percent; and that threshold (~6.5–7 m/s door-edge, ~15 mph) sits far above any humanly-achievable slam, with tempering providing roughly a 2× margin over ordinary glass. Draft — shared for review before external publication.

Need to know whether a part survives an impact, a drop, or a slam — with a real fracture threshold instead of a guess? The same Ansys explicit-dynamics workflow — a resolved shell model, a physically-grounded failure stress, and the honest sensitivity bracket — is the kind of workflow that helps teams qualify glazing, housings, brittle components and impact events. Innovation through insight.