Definition Engineering stress and strain
$$\sigma_n = \frac{F}{A_0} \qquad \varepsilon_n = \frac{l - l_0}{l_0} = \frac{\Delta l}{l_0} \qquad A_0 = \frac{\pi d_0^2}{4}$$
- $F$ is the load, $A_0$ the original cross-section, $l_0$ the original gauge length [Session 11 Slide 9]. Stress in MPa ($\text{N/mm}^2$) and strain with no units (never in % inside a formula).
- Shear: $\tau = F/A_0$, $\gamma = a/h = \tan\theta$, $\tau = G\gamma$ [Session 11 Slide 7].
- Three-point bend (modulus of rupture): $\text{MOR} = \dfrac{3FL}{2bh^2}$ [Session 11 Slide 6].
Theorem Hooke's law and elastic deformation
$$\sigma = E\,\varepsilon \qquad \Delta l = \varepsilon\,l_0 = \frac{F\,l_0}{A_0\,E} \qquad l_0 = \frac{\Delta l\,A_0\,E}{F}$$
- $E$ is the slope of the straight part of the curve; it measures stiffness, depends on the bond strength, and $E_{\text{ceramics}} > E_{\text{metals}} > E_{\text{polymers}}$ [Session 11 Slides 12-13].
- Typical $E$: steel $207$ to $210$ GPa, aluminium $69$ to $70$ GPa, alumina $390$, silica $69$ [Session 11 Slides 13, 17]. Specific modulus: $E/\rho$.
- Valid only in the elastic zone: beyond the elastic limit read the table or the curve (Problems T5-01 item i, T5-07).
- Largest elastic length: from $\Delta l = F l_0/(A_0E)$ solved for $l_0$ with the largest elastic load (Problem T5-06).
Theorem Poisson's ratio and the shear modulus
$$\nu = -\frac{\varepsilon_{\text{transverse}}}{\varepsilon_{\text{longitudinal}}} \qquad \varepsilon_{\text{transverse}} = \frac{\Delta d}{d_0} \qquad E = 2G\,(1 + \nu)$$
- Uniaxial tension: the width shrinks, $\Delta d = -\nu\,\varepsilon_{ax}\,d_0 < 0$ [Session 11 Slide 19].
- Typical $\nu$: metals $0.25$ to $0.4$, polymers $0.3$ to $0.45$, rubber $0.49$; the ideal incompressible limit is $\nu = 0.5$.
- From a measured diameter change: $\varepsilon_{ax} = -\dfrac{\Delta d/d_0}{\nu}$, then $\Delta l = \varepsilon_{ax}\,l_0$ and $E = \sigma/\varepsilon_{ax}$ (Problem T5-07).
Algorithm Yield strength by the 0.2 percent offset
- Draw the elastic line $\sigma = E\varepsilon$ and its parallel through $\varepsilon = 0.002$: offset line $\sigma_{off}(\varepsilon) = E\,(\varepsilon - 0.002)$.
- The yield strength $\sigma_y$ is the stress where the curve crosses the offset line [Session 11 Slide 28].
- With a table: compute $g = \sigma_n - E(\varepsilon_n - 0.002)$ for each row; the crossing is on the segment where $g$ changes sign.
- On that segment (rows $a$ and $b$), the curve is the line $\sigma = \sigma_a + m\,(\varepsilon - \varepsilon_a)$ with $m = \dfrac{\sigma_b - \sigma_a}{\varepsilon_b - \varepsilon_a}$. Equate it to the offset line and solve:
$$\varepsilon^* = \frac{\sigma_a - m\,\varepsilon_a + 0.002\,E}{E - m} \qquad \sigma_y = E\,(\varepsilon^* - 0.002)$$
- Load and elongation at yield: $F_y = \sigma_y A_0$, $\Delta l_y = \varepsilon^* l_0$ (note: $\varepsilon^*$, the total strain, not $\varepsilon^* - 0.002$).
- Discontinuous yield (low-carbon steels): $\sigma_y = \dfrac{\sigma_{max} + \sigma_{min}}{2}$ of the upper and lower yield points [Session 11 Slide 28].
Example (Problem T5-01, rows 7 and 8, $E = 196\,160$ MPa): $m = 29\,749$ MPa, $\varepsilon^* = 0.005786$, $\sigma_y = 742.6$ MPa.
Derivation The offset intersection
In short: put the straight segment of the table and the offset line equal and solve for the strain at the crossing.
$\sigma_{off}(\varepsilon) = E\,(\varepsilon - 0.002)$
line parallel to the elastic one, shifted by the plastic strain $0.002$
$\sigma(\varepsilon) = \sigma_a + m\,(\varepsilon - \varepsilon_a)$
between table rows the curve is taken as a straight segment (linear interpolation)
$\sigma_a + m(\varepsilon - \varepsilon_a) = E\,(\varepsilon - 0.002)$
the yield point is on both lines
$\sigma_a - m\,\varepsilon_a + 0.002\,E = (E - m)\,\varepsilon$
collect the terms in $\varepsilon$
$\varepsilon^* = \dfrac{\sigma_a - m\,\varepsilon_a + 0.002\,E}{E - m}$
solve; $E - m > 0$ because the segment is flatter than the elastic line
Theorem Strength, ductility, toughness and resilience
$$\sigma_{UTS} = \frac{F_{max}}{A_0} \qquad EL\,(\%) = \frac{l_f - l_0}{l_0}\times100 \qquad RA\,(\%) = \frac{A_0 - A_f}{A_0}\times100$$
$$U_r = \int_0^{\varepsilon_y}\sigma\,d\varepsilon = \frac12\,\sigma_y\,\varepsilon_y = \frac{\sigma_y^2}{2E} \qquad \sigma_w = \frac{\sigma_y}{N} \qquad N_{eff} = \frac{\sigma_y}{\sigma_{service}}$$
- UTS: maximum nominal stress; with constant $A_0$ it is at the maximum load. Necking follows [Session 11 Slide 30].
- Elongation after fracture is measured on the broken pieces: the elastic part is recovered. If the last record before fracture is $(\varepsilon_f,\ \sigma_f)$, then $\varepsilon_p = \varepsilon_f - \sigma_f/E$ and $EL = \varepsilon_p \times 100$ [Session 11 Slide 32].
- Reduction in area needs the final area: $A_f = \pi d_f^2/4$; in general $EL\,(\%) \neq RA\,(\%)$ [Session 11 Slide 33].
- Toughness: area under the whole curve (energy per volume, MPa $=$ MJ/m$^3$) [Session 11 Slide 34]. By trapezoids over a table: $U_t \approx \sum \tfrac12(\sigma_i + \sigma_{i+1})(\varepsilon_{i+1} - \varepsilon_i)$.
- Resilience: area under the elastic line up to $\sigma_y$; high $\sigma_y$ and low $E$ means high resilience (springs) [Session 11 Slide 35].
- Working stress: $N$ is between $1.2$ and $4$, usually $2$ when the statement gives none [Session 11 Slide 29]. A cable is safe if $N_{eff} \ge N$ (Problem T5-07).
Theorem True stress and true strain
$$\sigma_r = \frac{F}{A_i} = \sigma_n\,\frac{A_0}{A_i} \qquad \varepsilon_r = \int_{l_0}^{l}\frac{dl}{l} = \ln\frac{l}{l_0}$$
- Before necking (constant volume, $l_0A_0 = lA_i$): $\sigma_r = \sigma_n\,(1 + \varepsilon_n)$ and $\varepsilon_r = \ln(1 + \varepsilon_n)$ [Session 11 Slide 37].
- After necking: use the area, $\sigma_r = F/A_i$ and $\varepsilon_r = \ln\dfrac{A_0}{A_i} = 2\ln\dfrac{d_0}{d_i}$ [Session 11 Slide 38].
- True stress is larger than engineering stress, true strain is smaller than engineering strain, $\sigma_r > \sigma_n$ and $\varepsilon_r < \varepsilon_n$ (for tension).
- From the reduction in area (Problem T5-12): $\dfrac{A_f}{A_0} = 1 - \dfrac{RA}{100}$, then $\sigma_{f,r} = \dfrac{\sigma_f}{A_f/A_0}$ and $\varepsilon_r = \ln\dfrac{A_0}{A_f}$.
- The corrected curve after necking accounts for the triaxial stress in the neck.
Algorithm Analysis of a tensile test from a load-elongation table
- Area: $A_0 = \pi d_0^2/4$ (mm$^2$). Rows: $\sigma_n = F/A_0$ (MPa) and $\varepsilon_n = \Delta L/l_0$ for every point; check the unit of the load column ("N/100" means hundreds of newtons).
- Young's modulus: points that lie on the straight start, $E = \sum\sigma_i\varepsilon_i\big/\sum\varepsilon_i^2$ (least squares through the origin), or the secant $\sigma/\varepsilon$ of one early point.
- Yield strength: offset construction above (algebraic on the sign-change segment).
- UTS: $F_{max}/A_0$ with the largest load of the table.
- Ductility: $\varepsilon_p = \varepsilon_f - \sigma_f/E$ (last row before fracture), $EL = \varepsilon_p\times100$.
- Resilience: $U_r = \sigma_y^2/2E$.
- Load or elongation at yield: $F_y = \sigma_y A_0$, $\Delta l_y = \varepsilon^* l_0$.
- Working stress: $\sigma_w = \sigma_y/N$, $N = 2$ if no value is given.
- Load at a given elongation: if $\Delta l$ is inside the elastic zone use Hooke; otherwise interpolate linearly between the two rows of the table (the interpolation card).
Pattern Strengthening mechanisms
All mechanisms work by blocking the motion of dislocations; the price is ductility and, in metals, conductivity [Session 12 Slides 4-14].
| Mechanism | What blocks the dislocations | Quantity | Effect on $\sigma_y$ and UTS | Effect on ductility and conductivity |
|---|
| Strain hardening (cold work) | tangles of new dislocations, $10^6$ to $10^{12}\ \text{m/m}^3$ | $\%\text{CW} = \dfrac{A_0 - A_d}{A_0}\times100$ | up | ductility down, conductivity down a little |
| Solid solution | strained lattice around solute atoms (smaller solute: tension; larger: compression) | solute content | up | ductility down, conductivity down a lot |
| Grain size reduction | grain boundaries (smaller $d$, more barriers) | $\sigma_y$ linear in $d^{-1/2}$ | up | ductility generally kept |
| Precipitation (age hardening) | dispersed particles of a second phase | smaller, denser, better dispersed | up (Al-Cu-Mg: UTS $469$ MPa against $45$ pure Al) | conductivity down |
- Extension: the Hall-Petch relation $\sigma_y = \sigma_0 + k_y\,d^{-1/2}$ is the straight line of the slide ($\sigma_0$ and $k_y$ are constants of the material, not given in the slides).
- Extension: for a wire drawn from diameter $d_0$ to $d_d$, $\%\text{CW} = \bigl[1 - (d_d/d_0)^2\bigr]\times100$.
- Cold work produces texture; a recrystallisation anneal restores ductility (Unit 4 diagrams give the temperatures).
Theorem Hardness numbers
Hardness is the resistance to a localised plastic indentation [Session 12 Slides 15-30].
| Test | Indenter | Formula | Units and notes |
|---|
| Brinell HB (HBW) | ball of diameter $D$ (WC or steel) | $HB = \dfrac{2P}{\pi D\left(D - \sqrt{D^2 - D_i^2}\right)}$ | $P$ in kgf, $D$ and $D_i$ in mm, result in kgf/mm$^2$ |
| Brinell load | | $P = Q\,D^2$ | $Q = 30$ steels, $10$ bronze and brass, $5$ Al alloys, $2.5$ anti-friction alloys |
| Vickers HV | diamond pyramid, $136^\circ$ | $HV = \dfrac{2F\sin(136^\circ/2)}{d^2} = 1.854\,\dfrac{F}{d^2}$ | $F$ in kgf, $d$ the mean of the two diagonals in mm |
| Knoop KHN | diamond pyramid, $l/b = 7.11$ | $KHN = 14.2\,\dfrac{P}{l^2}$ | microhardness, load up to 1 kgf |
| Rockwell HR | diamond cone $120^\circ$ or steel ball | $HR = N - \dfrac{h}{S}$ | $h$ residual depth in mm; $S = 0.002$ mm; $N = 100$ for A, C, D; $N = 130$ for B, E, F, G, K; superficial $N$ and $T$ scales $S = 0.001$, $N = 100$ |
- Brinell inverse (indentation diameter for a given $HB$): $k = D - \sqrt{D^2 - D_i^2} = \dfrac{2P}{\pi D\,HB}$, then $D_i^2 = k\,(2D - k)$ (Problem T5-15).
- Rockwell in depth units: $HRC = 100 - 500\,t$ and $HRB = 130 - 500\,t$ with $t$ in mm [Session 12 Slide 29].
- Precautions: distance to the edge above $4\varnothing$; thickness above $2\varnothing$ (Brinell, Rockwell) or $1.5$ times the diagonal (Vickers).
- Hardness and strength both measure plastic deformation, with a different line per material family (chart of Session 12 Slide 30). Extension, for steels only: $\sigma_{UTS}\,[\text{MPa}] \approx 3.45\,HB$.
Pattern Thermal stress in a constrained bar (link with Unit 7)
Held at constant length, a bar heated by $\Delta T = T_f - T_0$ develops $\sigma = -E\alpha\,\Delta T$ (compression when heated, tension when cooled). If a load $F$ already acts, the wire stays stretched when $E\alpha\Delta T = F/A$, so $A = \dfrac{F}{E\,\alpha\,\Delta T}$ and $d = \sqrt{4A/\pi}$ (Problem T5-05: $d = 5.87$ mm). The full derivation is in Unit 7.
Trap Mistakes in mechanical properties
- Taking the yield strength at the first row below the offset line instead of intersecting with it.
- Using $\varepsilon^* - 0.002$ as the strain at yield: the total strain is $\varepsilon^*$ (elastic plus the $0.002$ offset).
- Measuring ductility at the last point of the table without removing the elastic part $\sigma_f/E$.
- Using engineering values for fracture after necking, or $\ln(1+\varepsilon)$ after necking (it only holds before).
- Using the nominal diameter instead of the instantaneous one for true stress.
- Putting the load in kgf into formulas that need newtons (or the reverse): Brinell and Vickers use kgf.
- Dividing by $A_0$ in mm$^2$ and then treating the result as Pa: it is MPa.
- Confusing resilience (elastic area, $\sigma_y^2/2E$) with toughness (whole area).
- A statement that says "aluminium" with steel-like data: use the data of the statement and state the discrepancy.