Strain to the onset of serrated flow (epsilon(c)) remains an important meas
ure of formability for aluminum alloys, especially in the 5XXX-series (Al-M
g) alloys, since serrated flow can signal undesirable surface appearance as
well as premature fracture. For this reason, extensive experimental and th
eoretical work has been performed in order to identify the key parameters t
hat affect the onset of serrated flow. In particular, it has been determine
d that thermal treatments applied after cold rolling can favorably impact t
he epsilon(c) of an Al-Mg alloy (5182 can end stock).
Recent developments of dislocation density-related constitutive modeling pr
ovide a unique opportunity to theoretically address this important problem.
Using the Ananthakrishna model described and extensively tested previously
, results are presented here to show that recovery processes (as simulated
by a reduction in the initial density of dislocations) can cause significan
t increases in epsilon(c), in agreement with the aforementioned experimenta
l results. Smaller initial dislocation densities can result in a considerab
le delay of serrated behavior. Using standard computational tools used in n
on-linear dynamics (phase portraits and bifurcation diagrams), the influenc
e of recovery (as mimicked by initial dislocation density changes) has also
been examined on Luders behavior, and results which confirm previous semiq
uantitative considerations in the literature have been obtained. The model
in its present form is limited to homogeneous straining and, thus, cannot t
reat the propagative instabilities often observed in experiments. Therefore
, the next step in the theoretical development will be to include effects o
f nonhomogeneous strain.