Stress, strain and stiffness: what actually fails

A beam does not fail because a large force was applied to it. It fails because the force was large relative to the area carrying it, and because the material it is made of had already run out of the ability to stretch elastically. Those are two different ideas, and keeping them apart is most of what stress analysis is.

Stress is force spread over area

Stress is load divided by the cross-sectional area resisting it. That division is the whole point: a 20 kN pull is meaningless on its own, because the same load is trivial through a girder and fatal through a wire.

Take a steel bar 10 mm in diameter. Its area is π × 0.005² = 78.54 mm², or 7.854 × 10⁻⁵ m². Pull it with 15 kN and the stress is 15,000 ÷ 0.00007854 = 191 MPa. Halve the diameter and the area falls to a quarter, so the same load produces four times the stress — area scales with the square of the dimension, which is the trap that makes thin sections fail unexpectedly.

Megapascals and newtons per square millimetre are the same number, which is convenient once you notice it: 191 MPa is 191 N/mm². If your answer comes out a million times too large or too small, this is almost always where it happened.

Strain is deformation as a proportion

Strain is how much something stretched divided by how long it was to start with. It has no units — it is a ratio, often quoted as a percentage or in microstrain.

Because it is a proportion, the same strain means very different absolute movements. Our bar at 191 MPa has a strain of about 0.0955%, which over a 2 m length is an extension of 1.91 mm. Over a 200 mm length the identical strain is 0.19 mm. Neither number tells you whether the part is safe; only the stress compared against the material's limits does that.

Young's modulus is stiffness, and stiffness is not strength

Divide stress by strain, in the region where the material still springs back, and you get Young's modulus — the material's stiffness. For our bar, 191 MPa ÷ 0.000955 gives 200 GPa, the textbook figure for steel.

Here is the distinction that catches people out: stiffness and strength are unrelated properties. Steel is roughly three times as stiff as aluminium, so it deflects a third as much under the same load — but a high-strength aluminium alloy can carry more stress before it yields than mild steel does. Choosing steel because it is "stronger" is often really a decision about deflection. Ask which one the design is actually limited by, because they lead to different materials.

Glass is stiffer than aluminium and shatters; rubber is enormously extensible and very weak. Stiffness tells you how much it moves. Strength tells you when it stops coming back.

Where the elastic formulas stop applying

Everything above assumes the material is still in its elastic region, where stress and strain are proportional and the part returns to its original shape. Push past the yield point and the relationship becomes non-linear, the deformation becomes permanent, and Young's modulus no longer predicts anything useful.

Mild steel yields at around 250 MPa. Our bar at 15 kN sat at 191 MPa, comfortably elastic. Raise the load to 20 kN and the stress becomes 254.6 MPa — past yield, and the extension the elastic calculation predicts is simply wrong. A calculator will still return a tidy figure. The number being plausible is not the same as the model being valid, and this is the failure mode that matters: not arithmetic, but using an elastic formula outside the elastic range.

Beams: the load is not the problem, the moment is

A beam in bending is not uniformly stressed. The load creates a bending moment that varies along the span, and the stress it produces is highest at the surfaces furthest from the neutral axis — which is why an I-beam puts its material in the flanges and hollows out the middle.

For a simply supported beam with a point load at mid-span, the maximum moment is the load times the span divided by four. A 5 kN load on a 3 m span gives 5 × 3 ÷ 4 = 3.75 kN·m, at the centre. Move that same load towards a support and the maximum moment falls, without the load changing at all. This is why "how heavy" is never a sufficient question about a beam — where, and over what span, decide the answer.

The order to work in

Find the load path, then the area or section resisting it, then the stress. Compare that stress with the material's yield strength and apply whatever factor of safety the application demands. Only then ask about deflection, using the modulus. Doing it in that order means you find out the model is invalid before you have built a chain of calculations on top of it.