Question 5 of 8: Plastic Instability and Stretch-Forming a Magnesium Sheet

Question 5: Plastic Instability and Stretch-Forming a Magnesium Sheet (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

5.1 — (a) Deriving the Considère necking condition ε = n

Given. A material work-hardens according to $\sigma = K\epsilon^n$ in true stress and true strain, deformed at constant volume in uniaxial tension.

Find. Show that plastic instability (necking) begins at the true strain $\epsilon = n$.

Approach. Necking begins where the load-carrying capacity of the bar reaches a maximum, because beyond that point any further stretch reduces the cross-section faster than the material can strain-harden to compensate, so deformation localises. That maximum-load condition is Considère's criterion, and combining it with the given hardening law fixes the strain at which it occurs.

1. Write the load and find its stationary point. The axial load is $P = \sigma A$, where $A$ is the current cross-sectional area. Instability (maximum load, $dP = 0$) occurs where $$dP = A\,d\sigma + \sigma\,dA = 0 \quad\Longrightarrow\quad \frac{d\sigma}{\sigma} = -\frac{dA}{A}.$$
2. Bring in constant-volume flow. Plastic deformation conserves volume, $A L = A_0 L_0 = \text{const}$, so $dA/A = -dL/L = -d\epsilon$ (true strain $\epsilon = \ln(L/L_0)$, $d\epsilon = dL/L$). Substituting, $$\frac{d\sigma}{\sigma} = -(-d\epsilon) = d\epsilon \quad\Longrightarrow\quad \boxed{\dfrac{d\sigma}{d\epsilon} = \sigma} \quad \text{(Consid\`ere's criterion).}$$ Necking begins exactly where the slope of the true stress–true strain curve equals the current true stress — geometrically, where a line from the origin at strain $\epsilon=-1$ is tangent to the curve.
3. Apply it to the given power law. With $\sigma = K\epsilon^n$, the hardening rate is $d\sigma/d\epsilon = Kn\epsilon^{n-1}$. Setting this equal to $\sigma$ itself per Considère's criterion, $$Kn\epsilon^{n-1} = K\epsilon^{n} \quad\Longrightarrow\quad n = \epsilon.$$

Physically, $n$ is therefore not just a curve-fitting exponent: it is the uniform (pre-necking) true strain the material can sustain in a tension test, and a metal with a higher $n$ can be drawn or stretch-formed further before it localises — the property exploited deliberately in part (b).

5.2 — (b) Over-stretch required to reach 6.2 m after spring-back

Given. Magnesium sheet, 1.5 mm thick $\times$ 80 mm wide, original length $L_0 = 5$ m, target length after the stress is released $L_f = 6.2$ m, $E = 65$ GPa, $\sigma_y = 200$ MPa. No hardening data is supplied, so the sheet is modelled as elastic–perfectly-plastic: once yielding begins the stress plateaus at $\sigma_y$.

Find. The length $L_{\text{before}}$ the sheet must be pulled to, under load, so that after the load is released it settles at the target 6.2 m.

1. Separate the total strain into elastic and plastic parts. While the sheet is under load at the plateau stress $\sigma_y$, its total engineering strain is $\epsilon_{\text{total}} = \epsilon_{\text{elastic}} + \epsilon_{\text{plastic}}$. On unloading, only the elastic part recovers (spring-back); the plastic part is permanent and is exactly what sets the final, released length. So $$L_{\text{before}} = L_0(1+\epsilon_{\text{total}}), \qquad L_f = L_0(1+\epsilon_{\text{plastic}}) = L_0(1+\epsilon_{\text{total}}-\epsilon_{\text{elastic}}).$$
2. Quantify the recoverable (elastic) strain. At the moment of release the stress in the sheet is $\sigma_y$ (perfectly plastic plateau, no further hardening to release from), so Hooke's law gives the strain that snaps back: $$\epsilon_{\text{elastic}} = \frac{\sigma_y}{E} = \frac{200\ \text{MPa}}{65{,}000\ \text{MPa}} = 3.077\times10^{-3}.$$ Note the sheet's 1.5 mm $\times$ 80 mm cross-section plays no role here — it fixes the force needed ($\sigma_y \times$ area) but not the strain recovered, which depends only on stress and modulus.
3. Combine to get the pre-release length. Subtracting the two length expressions above, $L_{\text{before}} - L_f = L_0\,\epsilon_{\text{elastic}}$, so $$L_{\text{before}} = L_f + \frac{\sigma_y}{E}\,L_0 = 6.2\ \text{m} + (3.077\times10^{-3})(5\ \text{m}) = 6.2\ \text{m} + 0.01538\ \text{m}.$$ $$\boxed{L_{\text{before}} \approx 6.215\ \text{m} = 6215\ \text{mm}}$$

The sheet must therefore be over-pulled by about 15 mm — roughly a quarter of one percent of its length — beyond the 6.2 m target before the load is taken off, purely to pay back the elastic strain that will spring back on release. Because $\epsilon_{\text{elastic}}$ is small compared with the plastic stretch of $(6.2-5)/5 = 24\%$, the correction is a fine-tuning step rather than a dominant one, but it is exactly the calculation a stretch-forming operator performs to hit a net dimensional target.