Void nucleation is studied both experimentally and computationally with the aim of identifying a macroscopic criterion for nucleation by particle cracking. Three types of circumferentially notched cylindrical specimens made of a low-alloy steel were used, in order to vary the stress triaxiality in the notch region. The tensile tests were interrupted at various loads below the fracture load. The specimens were sectioned parallel to the loading axis, and the locations of cracked and uncracked titanium-nitride inclusions were identified. No evidence was found of void nucleation by inclusion debonding. Finite-element calculations were carried out for each specimen geometry using conventional isotropic-hardening plasticity theory. The ability of various potential void-nucleation criteria to predict the onset of void nucleation by inclusion cracking is explored.References 8, 21 through 23, 32 through 34, 37, and 38. In steels, most studies found that microvoids first form around ellipsoidal MnS inclusions in the steels through debonding of the steel-sulfide interface. Two reasons have been suggested to explain why these interfaces are so weak. One is that the difference in the coefficient of thermal expansion between these inclusions and the matrix is such that this interface is in tension. [35] The other is that sulfur segregated to this interface weakens it. [36] A significant amount of research has also addressed the issue of how these large voids link up to cause fracture. Most results indicate that small voids form around the carbides very late in the fracture process, and these small voids help link the large voids formed around the sulfides. [3,28,30,36] In most of these steels, the sulfides are arrayed in the material so that the fracture plane contains the long axis of the ellipsoidal inclusions. Relatively little work has been done for materials in which the long axis is perpendicular to the fracture plane, which would presumably make these inclusions less effective as void nuclei. In addition, there has been relatively less work investigating the role that inclusions other than sulfides play in nucleating the initial microvoids (e.g., References 8 and 23). Cox and Low [30] reported results on an 18-Ni maraging steel in which the nucleating inclusions were titanium carbo-nitrides. They found that these inclusions cracked when a critical stress was reached and that these cracked inclusions nucleated microvoid formation.Various void-nucleation criteria have been proposed based on experimental observations or micromechanical analyses. [34,[39][40][41][42][43] In order to formulate a physically based nucleation criterion, careful attention must be paid to the mechanism by which this process occurs. As suggested earlier, the two most common mechanisms would be either debonding of the interface between the inclusion and the matrix, as has been reported for MnS, or cracking of the second-phase inclusions, which has been reported for the titanium carbo-nitrides in maraging steel or Fe-Si inclusions in aluminum alloys...

A microstructure-level simulation model was recently developed to characterize machining behavior of heterogeneous materials. During machining of heterogeneous materials such as cast iron, the material around the machining-affected zone undergoes reverse loading, which manifests itself in permanent material softening. In addition, cracks are formed below and ahead of the tool. To accurately simulate machining of heterogeneous materials the microstructure-level model has to reproduce the effect of material softening on reverse loading (MSRL effect) and material damage. This paper describes procedures used to calculate the material behavior parameters for the aforementioned phenomena. To calculate the parameters associated with the MSRL effect, uniaxial reverse loading experiments and simulations were conducted using individual constituents of ductile iron. The material model was validated with reverse loading experiments of ductile iron specimens. To determine the parameters associated with fracture of each constituent, experiments and simulation of notched specimens are performed. During the validation stage, response of simulated ductile iron was in good agreement with the experimental data.

A microstructure-level model for simulation of machining of cast irons using the finite element method is presented. The model explicitly combines ferritic and pearlitic grains with graphite nodules to produce the ductile iron structure. The behaviors of pearlite, ferrite, and graphite are captured individually using an internal state variable model for the material model. The behavior of each phase is dependent on strain, strain rate, temperature, and amount of damage. Extensive experimentation was conducted to characterize material strain rate and temperature dependency of both ferrite and pearlite. The model is applied to orthogonal machining of ductile iron. The simulation results demonstrate the feasibility of successfully capturing the influence of microstructure on machinability and part performance. The stress, strain, temperature, and damage results obtained from the model are found to correlate well with experimental results found in the literature. Furthermore, the model is capable of handling various microstructures in other heterogeneous materials such as steels.

To understand the influence of cast iron microstructure on its machinability, a numerical model that simulates machined material on a microstructure scale was developed. This microstructure-level model assembles individual constituents into a composite material based on microstructural composition, grain size, and distribution. Extensive experimentation was performed to determine strain, strain rate, temperature, and load history dependent material properties. The purpose of this work is to validate the microstructure-level model on machining of ductile iron and two of its constituents: pearlite and ferrite. Orthogonal cutting experiments were conducted of the three materials. The measured chip morphology and machining forces were compared with the model predictions, and a good correlation between them was found.

No abstract