Mechanics of materials / Sheet 04
Bending and deflection
Formula reference for mechanical engineering interviews. Bending stress in a determinate beam is independent of material; deflection is not. Check both, and say which one governed.
The flexure equation
Section properties
| Section | I | S = I/c | Notes |
|---|---|---|---|
| Rectangle b × h | b h³ / 12 | b h² / 6 | depth cubed, width linear |
| Solid round | π d⁴ / 64 | π d³ / 32 | J = 2I |
| Hollow round | π (do⁴ - di⁴) / 64 | divide by do/2 | best stiffness per mass |
| Hollow rectangle | (B H³ - b h³) / 12 | divide by H/2 | box sections |
| Triangle, base b | b h³ / 36 | about the centroid | centroid at h/3 |
Transverse shear
Bending stress peaks at the outer fibre where shear is zero, and shear peaks at the neutral axis where bending stress is zero. They rarely combine. The exception is the web to flange junction of an I-beam, where both are significant and a von Mises check is warranted.
Load, shear and moment
Standard deflection cases
| Case | δ max | M max | Slope |
|---|---|---|---|
| Cantilever, point load at tip | P L³ / 3EI | P L, at wall | P L² / 2EI |
| Cantilever, UDL | w L⁴ / 8EI | w L² / 2, at wall | w L³ / 6EI |
| Cantilever, moment at tip | M L² / 2EI | M, constant | M L / EI |
| Simply supported, central load | P L³ / 48EI | P L / 4, centre | P L² / 16EI |
| Simply supported, UDL | 5 w L⁴ / 384EI | w L² / 8, centre | w L³ / 24EI |
| Fixed both ends, central load | P L³ / 192EI | P L / 8, at ends | 0 at ends |
| Fixed both ends, UDL | w L⁴ / 384EI | w L² / 12, at ends | 0 at ends |
| Propped cantilever, UDL | w L⁴ / 185EI | w L² / 8, at fixed end | 0 at fixed end |
Fixing both ends of a uniformly loaded beam cuts maximum deflection to one fifth and maximum moment to two thirds, and moves the peak moment from midspan to the supports.
Deflection methods
Scaling and ratios
Plastic bending
The shape factor is the reserve between first yield and a fully plastic hinge. An I-beam has almost none, because its material already sits at the extreme fibre, which is exactly why it is efficient elastically.
Worked chain: sizing a beam
Numbers worth memorizing
| Quantity | Steel | Aluminium | Ratio |
|---|---|---|---|
| E (GPa) | 200 | 69 | 2.9 |
| Density (kg/m³) | 7850 | 2700 | 2.9 |
| Specific stiffness E/ρ | 25.5 | 25.6 | 1.0 |
| Deflection, same geometry | 1× | 2.9× | - |
| Bending stress, same geometry | 1× | 1× | - |
Specific stiffness is identical, so equal-mass equal-stiffness swaps are impossible in pure tension. In bending aluminium wins, because the thicker section it allows raises I faster than the lower E costs you.
Concepts that decide the interview
- Depth is worth cubed, width only linear. Doubling depth gives 8 times the stiffness and 4 times the strength; doubling width gives 2 times both. Quote the exponents, not just the direction.
- Length beats everything. Deflection scales with L³ or L⁴ while stress scales with only L or L². When something is too floppy, look at span before section.
- Material at the neutral axis does nothing for bending but is exactly where shear peaks. That designs the I-beam: flanges carry moment through the Ad² term, the web carries shear.
- You can put a hole through a beam near the neutral axis at midspan, but not near the supports, where shear is maximum.
- Bending stress in a determinate beam is independent of E, because M comes from statics alone. Deflection is not. In indeterminate beams load redistributes toward the stiffer members in proportion to EI.
- Deflection, not stress, usually sizes the part, and far more so in robotics where positional accuracy rather than failure is the constraint.
- End fixity is a large, free win, exactly like bracing a column. Change the boundary conditions before you change the part.
- Superposition requires linear elastic material and small deflections. Once a fibre yields or geometry changes materially, adding standard cases is no longer legitimate.