A numerical procedure enabling accurate descriptions of strain rate-sensitive flow of polycrystals within crystal visco-plasticity theory

The plastic deformation of polycrystalline metals is carried by the motion of dislocations on specific crystallographic glide planes. According to the thermodynamics theory of slip, in the regime of strain rates, roughly from 10(-5)/s to 10(5)/s, dislocation motion is thermally activated. Dislocations must overcome barriers in order to move, and this concept defines critical activation stresses tau(s)(c) on a slip system s that evolve as a function of strain rate and temperature. The fundamental flow rule in crystal visco-plasticity theory that involves tau(s)(c) in order to activate slip has a power-law form: (gamma) over dot(s) = (gamma) over dot(0) (vertical bar tau(s)vertical bar/tau(s)(c))(n) sign (tau(s)). This form is desirable because it provides uniqueness of solution for the active slip systems that accommodate an imposed strain rate; however, it also introduces a strain rate dependence, which in order to represent the actual behavior of polycrystalline materials deforming in relevant conditions of temperature and strain-rate usually needs to be described by a high value of the exponent n. However, since until now the highest value of n was limited by numerical tractability, the use of the power-law flow rule frequently introduced an artificially high rate-sensitivity. All prior efforts to correct this extraneous rate sensitivity have only lessened its effect and unfortunately also at the expense of substantial increases in computation time. To this day, a solution for the power-law exponent reflecting true material behavior is still sought. This article provides a novel method enabling the use of realistic material strain rate-sensitivity exponents to be used within the crystal visco-plasticity theory without increasing computation time involved in polycrystal simulations. Calculations are performed for polycrystalline pure Cu and excellent agreement with experimental measurement is demonstrated. (C) 2016 Elsevier B.V. All rights reserved.