Threshold and flow loci were constructed experimentally at 650°C in the axial-shear stress plane for the cobalt-based alloy Haynes 188. These loci suggest that the stress-dependence of the dissipation potential is on the second deviatoric stress invariant, J2, alone. The functional form of the dissipation is critical to associative potential-based viscoplasticity models because it dictates the direction of the inelastic strain rate vector through normality. The directions of the inelastic strain rate vectors for proportional loading were in generally good agreement with the normality rule. Data suitable for characterizing the material parameters of a viscoplasticity model are presented. Serrated yielding was often present for axial, torsional, and combined axial-torsional loading. By applying proportional and nonproportional load histories at identical equivalent strain rates, J2 was found to generally reduce the data to a single curve.

The effect that various loads have on a 6092/SiC/17.5p-T6 particulate reinforced aluminum composite is determined. In addition to the mechanical response from tensile, compressive, and shear loading, yield loci in the axial-shear stress plane are constructed using axial-torsional loading of a thin-walled tube. Yield loci are determined by multiple yield probes of a single specimen using a 40 x 10-6 equivalent offset strain definition of yielding. Cyclic tensile straining to increasingly higher amplitudes indicated a modulus reduction of 16% prior to fracture, strongly suggesting accumulation of internal damage, but no change in the elastic Poisson's ratio was observed. Cyclic compressive loading resulted in no observable change in modulus. Cyclic shear loading led to a minimal shear modulus reduction of approximately 6%. The initial yield locus in the axial-shear stress plane had an eccentricity in the compressive stress direction that is known as a strength differential. The strength-differential was measured to be 55% and is believed to be associated with thermal residual stresses from heat treatment. After shear prestraining subsequent yield loci were constructed. Hardening was observed to be primarily kinematic.

A theoretical model capable of predicting the thermomechanical response of continuously reinforced, titanium matrix composite (TMC) laminates, subjected to multiaxial loading has been developed. The model is based on micromechanics and employs classical lamination theory to determine inelastic response. The constitutive relationships for each lamina are determined from a micromechanics analysis that is performed numerically using the finite element method. Matrix viscoplasticity, thermal stresses, and damage to the fiber/ matrix interfacial zone are explicitly included in the model.

A detailed set of experiments has been conducted in an attempt to characterize the inelastic flow behavior of materials used in aeronautics applications. These experiments have shown that aged (precipitate-hardened) Inconel 718 exhibits an initial nonlinear elastic behavior as well as a strength differential in tension and compression at room temperature. This nonlinear elastic behavior correlates reasonably well with a second order stress-strain law that was developed to account for interactions between dislocations and interstitial solute atoms. Flow loci in the axial-shear stress plane at 25 and 650°C have been determined from yield locus data using both inelastic power and equivalent inelastic strain definitions of flow. Flow loci are more theoretically meaningful than yield loci for describing the time-dependent material response. These flow loci are especially important if they are proportional to the dissipation potential since the normality rule is associated with dissipation and not necessarily with a particular yield or flow definition employed in an experiment. A threshold function that depends only on the second deviatoric stress invariant, J2, is inadequate for predicting the threshold surface (initial yield locus) for Inconel 718 due to the difference in flow behavior in tension and compression. Threshold functions including all three stress invariants (I1, J2, and J3) with each term having units of stress raised to the first or third power were equally successful in fitting the initial flow locus. Additionally, both F=aI1+bJ21/2-1 and F=b3J23/2+c3J3-1 are shown to represent the data very well. The outward normals for these two representations of the threshold surface are consistent with experimentally determined directions of the inelastic strain rate vectors.

An experimental program has been implemented to determine small offset yield loci for the nickel-base superalloy In conel 718 (IN718) under axial-torsional loading at elevated temperatures. Initial and subsequent yield loci were determined for solutioned IN718 at 23 and 454°C and for aged (precipitation-hardened) IN718 at 23 and 649°C. The shape of the initial yield loci for solutioned and aged IN718 agrees well with the von Mises prediction. However, in general, the centers of initial yield loci are eccentric to the origin due to a strength-differential (S-D) effect that increases with temperature. Subsequent yield loci exhibit kinematic and distortional hardening, and solutioned IN718 shows a weak cross effect. This work demonstrates that it is possible to determine yield surfaces for metallic materials at temperatures up to at least 649°C using multiple probes of a single specimen. The experimental data are first of their kind for a superalloy at these temperatures and will enable a better understanding of multiaxial material response.

The concept of a yield surface is central to the mathematical formulation of classical plasticity theories. However, at elevated temperatures, material response can be highly rate-dependent, which is beyond the realm of classical plasticity. Viscoplastic theories have been developed for just these conditions. In viscoplastic theories, the flow law is given in terms of inelastic strain rate rather than the inelastic strain increment used in rate-independent plasticity and in unified viscoplastic theories no yield criterion is used. Thus, surfaces of constant inelastic strain rate or flow surfaces, which describe hardening behavior, are considered to be more useful than yield surfaces.