Abstract
A Monte Carlo simulation technique, in which the microstructure is mapped onto a discReve, two-dimensional lattice with the use of the Q-state Potts model, is applied to the problem of dynamic recrystallization in a polycrystalline matrix. The response of the material to hot deformation is simulated by adding recrystallization nuclei and stored energy continuously with time. The simulations use an energy storage rate Schedule that models both work hardening and dynamic recovery of metals by way of a parabolic work hardening law. Recrystallization nuclei, having zero stored energy, were added at a rate which is related to the level of work hardening. Departing from a previously applied scheme, the nucleation rate is chosen to be proportional to the strain rate. Also considered is a more intuitive, local model in which the creation of a nucleus oceurs with a probability proportional to exp (−δE), where δE is the energy change associated with the introduction of an embryo into the matrix. The evolution of the stress and the mean grain size is similar to those obtained with use of the constant nucleation rate. It was found, however, that the relations between steady-state parameters are different from the results obtained earlier but are more consistent with the experimental literature. Careful numerical analysis of data obtained for the exponential nucleation model yield new power-law relations between steady-state parameters and nucleation attempt rate. The kinetics of dynamic recrystallization are analysed with use of the Johnson-Mehl-Avrami-Kolmogorov (JMAK) eauation which is shown to be sensitive to the rate of deformation. Higher than expected values of the JMAK kinetic exponent were obtained that are attributed to the use of the parabolic work hardening model in the simulations.