Mahesh Sivasambu

This study involved laboratory unidirectional (UDR) and reverse (RR) cold rolling of steel, and corresponding direct (and indirect) observations of surface (and sub-surface) microstructures. Though both processes had identical strain mode of plane strain compression (PSC), the as-rolled surface grains showed clear differences in imposed mesoscopic shear strains. Further, the surface microstructure and its orientation sensitivity differed remarkably between the two processes. RR had more dislocation density, grain misorientations and non-crystallographic microbands, but exhibited insignificant differences between different crystallographic orientations. These effects appeared significant in low carbon interstitial free (IF) steel, but noticeably less so for high strength low alloy (HSLA) grade. The crystallographic textures of both the processes were identical in the mid-thickness section. However, the surface textures differed noticeably irrespective of the steel grade. These were captured quantitatively with a crystal plasticity model, and by introducing parametrically positive (UDR) and negative (RR) local shear strains for the respective surfaces. In summary, this study established, quantitatively, the defining role of local shear strain on the developments of as-rolled microstructure and crystallographic texture of steel.

The microstructure of an Al-Cu-Li alloy sheet is characterised in the solution treated, underaged, and peak aged tempers. Its mechanical response under uniaxial tension is measured at (Formula presented.), (Formula presented.), and (Formula presented.) along (Formula presented.), (Formula presented.), and (Formula presented.) to the rolling direction. The anisotropic stress-strain curves are interpreted using a polycrystal plasticity model, which implements three hardening modes suggested by the microstructure, viz., matrix hardening, hardening due to isotropic precipitates, and hardening due to anisotropic precipitates. Phenomenological activation theory based analysis of the work hardening suggests that the physical mechanism underlying work hardening due to anisotropic precipitates remains constant over the temperature range studied, while the mechanism underlying matrix hardening varies strongly with temperature.

Macroscopic shear bands (MSB) may develop during hot working of metallic materials. They are well-understood as a physical manifestation of flow instability, and the processing regimes wherein they form are well-charted. However, the microstructural transitions that occur between the onset of flow instability and MSB inception are not fully understood. In order to elucidate them, several aluminum alloy specimens were subjected to strip testing in a thermomechanical simulator (Gleeble™) at 298 K and 573 K. Prominent MSB were observed along the diagonals of the strip volume of only Aluminum-6 wt% Magnesium alloy specimen deformed at 573 K. Comparing the experimental grain morphology and crystallographic textures with those from plastic flow models revealed that MSB inception occurred only after ∼0.20 homogeneous plane strain deformation. However, classical flow instability was predicted at much smaller strain. This ‘delay’ was explained experimentally by showing that clusters of neighbouring severely deforming, fragmenting, mostly soft-oriented grains gradually developed due to lattice rotations along the specimen diagonals, and that MSB inception corresponded to the formation of a percolating network of such grains spanning the specimen. Further, clear experimental evidence revealed that differential dynamic recovery between hard- and soft-oriented grains was essential for MSB formation.

The two-dimensional stress fields induced by a set of infinitely many parallel edge dislocations are difficult to estimate as those of individual dislocations decay slowly. A simple numerical method to compute them is proposed. The method is based on series summation using a convergence factor, (Formula presented.) that decays rapidly with radial distance r from the field point, and letting the positive parameter (Formula presented.) numerically through Richardson extrapolation. The present method is more general than a lattice summation method with explicit spurious stress cancellation that is widely used in the literature. Furthermore, the spurious long-range stresses are cancelled in the present method without explicit evaluation.

A linear, periodic, three-dimensional shear-lag model of unidirectionally-reinforced composites that allows for fibre breakage, and matrix failure is proposed. Matrix failure can take the form of matrix splitting or interfacial debonding. A computationally efficient scheme for its solution is developed. This scheme exploits the translation invariance of the elastostatic fields due to failed elements in the periodic cell, and is asymptotically faster than the classical eigensolution-based approach. The new computational scheme is used to illustrate the influence of matrix failure on the elastostatic fields induced by small clusters of fibre breaks in several test problems. Monte Carlo simulations of fracture in model three-dimensional composite specimen with Weibull-distributed fibre segment strengths are also performed. Matrix failure is found to considerably alter fracture development, to weaken the median specimen, and to reduce the variability in composite strength.

Mechanical properties at ambient and cryogenic temperatures of Al-Cu-Li alloy are required for design and fabrication of liquid hydrogen and liquid oxygen tanks of satellite launch vehicles. In the present work, bead-on-sheet, friction stir welding was carried out with three different rotation speeds. The yield and strain hardening behaviors of the welds were evaluated in temperature range of 20 K to 298 K. Both yield stress and strain hardening ability in the specimen increased with decrease in testing temperature. The dependence of yield stress on temperature was modeled on the basis of thermally activated dislocation mobility, while that of strain hardening was modeled on the temperature dependence of dynamic recovery rate parameter. The recovery parameter followed an Arrhenius-type relationship with temperature. The model parameters determined from the experimental data were further used to simulate the stress–strain curves at different sub-zero temperatures for the friction stir welds.

The transverse fracture of model unidirectional composite specimen, comprising up to 220 fibers with random strengths, is studied using Monte Carlo simulations. The load sharing from broken to intact fibers is assumed to obey power-law scaling ∼r-γ with distance r from the fiber break. Fiber breaks are assumed to interact in order to remain traction free. The pattern of fiber breaks that propagate catastrophically is interpreted through cluster analysis. The empirical strength distributions obtained from the simulations are interpreted using two probabilistic models of brittle fracture available in the literature. These point to a transition from the brittle to the tough fracture mode as γ↓2. The transitional γ is approximately equal to that reported in the literature for noninteracting fiber breaks.

Friction stir welds (FSWs) can be considered as an ensemble of elements of material with composite microstructures connected in series. In the present study, bead-on-plate FSW runs were made on an Al–Cu–Li alloy with varying rotation speeds ranging from 400 to 800 rpm. Microstructure of the FSW region was investigated by using optical microscope, electron backscattered diffraction (EBSD) and transmission electron microscope (TEM). Thermal stability of various precipitates was evaluated with differential scanning calorimetry (DSC) measurements. Strength variation across FSW cross sections was mapped by microhardness measurements. Average as well as local mechanical properties were evaluated using a digital image correlation (DIC) technique. Irrespective of the process parameters, FSW samples showed similar tensile and strain hardening behaviour along with serrations in stress-strain curves while local strength values showed increasing trend with rotation speed. The FSW alloy produced at intermediate rotation speed exhibited different mechanical behavior and is correlated with the resultant substantial changes in the microstructure. Strain localization occurred at the boundary of nugget zone and thermo-mechanically affected zone which led to failure of the FSW tensile specimens within weld regions. Fractography investigation revealed that the failure is initiation controlled, that is void nucleation at coarse precipitate-matrix interfaces.

The evolution of the yield surface, as predicted by a polycrystal plasticity model of a face-centered cubic material, is studied. Grains in the model polycrystal are endowed with a classical hardening law, which accounts for interaction among the slip system through a hardening matrix. In the literature, the elements of the hardening matrix are assumed non-negative. In the present work, the effect of negative elements in the hardening matrix on the evolution of the yield function, particularly, during monotonic tensile and shear deformation, is systematically studied. In particular, it is shown that certain parametric values simulate a substantial kinematic hardening, similar to experimental observations. The greatest kinematic hardening is obtained when the latent hardening ratio of the reverse slip systems is taken to be −1.2.

Fracture of three-dimensional unidirectional composites is studied through Monte Carlo fracture simulations on model composites. Fracture develops in the model composites by the failure of fibre segments wherein the tensile stress exceeds a Weibull-distributed random strength, and by the failure of the fibre–matrix interfaces wherein the shear stress exceeds a deterministic interfacial strength, T. The size of the weakest-link failure event is inferred from empirical strength distributions obtained from the simulations. It is found to diverge or converge with composite size for T below or above a threshold value, respectively. The threshold is identified as the tough–brittle fracture mode transition. The mechanistic cause underlying the transition is also identified.