P. Rangaswamy

Comparison of residual strains measured by X-ray and neutron diffraction in a titanium (Ti–6Al–4V) matrix composite

Abstract This research compares matrix thermal residual strains measured in a continuous fiber reinforced SiC/Ti–6Al–4V titanium matrix composite (TMC) using X-ray and neutron diffraction with finite element predictions. The strain dependence on the strains for several reflections (105, 204, 300, 213 and 312) of the matrix were explored at the surface (X-ray) and in the bulk (neutron). To determine the longitudinal surface strains from the X-ray measurements for comparison with the neutron values, the e φψ versus sin 2 ψ plots were extrapolated to ψ =90°. Continuum micro-mechanics based multi-ply finite element models (FEM) simulating rectangular and hexagonal fiber distributions were explored for calculating average surface and bulk strains. For different reflections, the experimentally determined surface measured strains ranged from +1904±424 to +2974±321 μ e and the bulk measurements ranged from +2269±421 to +3022±1134 μ e . These values contrast with the single valued FEM prediction of+3200 μ e which was the same for both the surface and the bulk.

The influence of thermal-mechanical processing on residual stresses in titanium matrix composites

Abstract The effects of three distinct thermo-mechanical processes on the residual stress state in a uni-directionally reinforced SCS-6/Ti-6-2-4-2[0] 6 titanium-alloy matrix composite were predicted using a finite element model. For comparison the residual stresses were measured using X-ray and neutron diffraction. Reductions in stress were predicted by the models and both experimental techniques recorded a reduction compared to the as-fabricated material. While the numerically predicted trends qualitatively agreed with the neutron measurements quantitative agreement was not achieved. In the longitudinal direction the neutron results showed closer agreement to the calcualtion whereas in the transverse direction the X-ray results did. Nevertheless the changes did correlate with improvement in fatigue lifetimes.

Microstress evolution during in situ loading of a superalloy containing high volume fraction of γ′ phase

Pulsed neutron diffraction under in situ mechanical loading was used to monitor microstrain evolution in individual phases of a polycrystalline γ/γ′ superalloy, CM 247 LC. The load partitioning and yielding of differently oriented grains and phases were evaluated. The critical resolved shear stresses of individual phases were obtained and are compared with dislocation models.

Influence of residual stresses on the evolution of micro-structure during the partial reduction of NiAl2O4

Abstract Metal-ceramic microstructures were formed in situ by the partial reduction (i.e. the reduction of only one of the metallic elements) of the spinel compound NiAl 2 O 4 . Depending on reduction conditions, these microstructures consist of Ni particles embedded in an α-Al 2 O 3 or a multiphase matrix called ‘defect spinel’. The volume shrinkage that accompanies the reaction generates residual stresses which profoundly affect the microstructure evolution. Conversely, formation of metastable, intermediate phases, generation of porosity and cracking are all observed and may act to relax the residual stresses. Electron microscopy observations as well as both neutron and X-ray diffraction residual stress measurements are used to study the influence of residual stresses on the microstructure evolution during the reduction process.

Thermal expansion anisotropy in a Ti-6Al-4V/SiC composite

Abstract We studied the thermal expansion behavior of a Ti–6Al–4V/35 vol.% continuous SiC fiber composite using in situ high temperature neutron diffraction (ND). The lattice expansion of constituent phases within the composite was monitored from axial (parallel to the unidirectionally aligned fibers) and transverse (perpendicular to the fibers) directions during heating from room temperature (RT) to 1170 K. The phase-specific thermal expansion of the Ti–6Al–4V matrix and SiC fibers in the composite is discussed in the context of thermal load partitioning between the matrix and fibers. In the axial direction, the matrix and the fiber share the thermal load and co-expand up to about 800–900 K, above which the thermal load transfer becomes ineffective. In the transverse direction, the matrix and fibers expand independently over the whole temperature range. Using the Schapery model (J. Comp. Mater. 2 (1968) 380) and the rule-of-mixtures (ROM), the macroscopic thermal expansion of the composite is predicted and compared with the experimental results.

Complementary X-ray and neutron strain measurements of a carburized surface

Lattice spacing profiles through a carburized layer in a cylindrical 5120 steel specimen were recorded using X-ray and neutron diffraction. The neutron measurements were performed non destructively, whereas the X-ray measurements employed sequential layer removal by electropolishing. The X-ray measurements provided a base determination of residual stress and microstructure, against which the neutron results were critically compared. Several methods were considered for determining the unstrained lattice parameter variation, necessary for interpretation of the neutron results.

Experimental measurements and numerical simulation of stress and microstructure in carburized 5120 steel disks

Abstract A combined experimental and numerical study of residual stress and microstructure was performed on a carburized disk of 5120 steel. The disk-shaped specimen was carburized and quenched in agitated oil at 71°C. X-ray diffraction combined with electropolishing layer removal was used to determine stresses through the case of the disk, within ∼1 mm of the surface. Both martensite and retained austenite volume fractions were determined through one flat surface. Rietveld analysis was used to determine the lattice parameters of the constituents at sequential depths. Microstructure and residual stress profiles were compared to predictions. The numerical predictions were from ABAQUS; a finite element code coupled with a user-defined material subroutine (UMAT), that accounted for microstructure evolution. Measured retained austenite values varied from ∼25 vol.% at the surface to a maximum of ∼30% at 100 μm, then decreased to 4% at a depth of 600 μm. The numerical simulation predicted a maximum of 25 vol.% at the surface that monotonically decreased to 7% at a depth of 600 μm, and reached a minimum of 4% at 1.0 mm. The maximum measured compressive stress was 380 MPa at 550 μm, compared to the predicted value of 450 MPa at 330 μm. In addition, the carbon profile predicted from the numerical simulation was comparable to the profile obtained from the combustion burnout technique.

Residual stresses in continuous-tungsten-fibre-reinforced Kanthal-matrix composites

Residual stresses were measured in four Kanthal matrix-continuous-tungsten-fibre composites (with different tungsten fibre volume fractions V f = 10, 20, 30 and 70 vol.%) using neutron diffraction. Parallel to the fibres the stress in the Kanthal ranged from 40 MPa ( V f = 10 vol.%) to 1100 MPa ( V f = 70 vol.%) compared with m1877 MPa ( V f = 10 vol.%) to m400 MPa ( V f = 70 vol.%) for the tungsten. Perpendicular to the fibres the stress ranged from m52 MPa ( V f = 10 vol.%) to 620 MPa ( V f = 70 vol.%) in the Kanthal compared with m778 MPa ( V f = 10vol.%) to m195 MPa ( V f = 70 vol.%) in the tungsten. Assuming that the measured residual stresses were solely thermal in origin, predictions were made using concentric cylinder and finite-element models. In the absence of hardening data the assumed material behaviour was elastic-perfectly plastic and the predictions underestimated the measured stresses for all volume fractions. Nevertheless the model results were consistent with the experimental measurements. The transverse stress in the fibres is discussed in the context of the interface normal stress, which is significant to the global mechanical response.

Measurement and modeling of internal stresses at microscopic and mesoscopic levels using micro-Raman spectroscopy and X-ray diffraction

In this study, we use x-ray diffraction (XRD) and micro Raman spectroscopy (MRS) to measure internal strains in sensors embedded in polymer matrix composites. Two types of strain sensors embedded in either chopped graphite fiber/epoxy matrix composite (MRS) or unidirectional graphite fiber/polyimide matrix composite (XRD) were investigated. For XRD measurements, the sensors were in the form of spherical aluminum inclusions with diameters ranging from 1 to 20Pm. Due to large cross section area of an incident x-ray beam, only average stresses are reported using the XRD approach. Complementary to XRD experiments, MRS was pursued to measure internal strains in Kevlar-49 fibers embedded in chopped graphite fiber/epoxy matrix composite. In recent years, MRS as an experimental tool for micro-strain measurements has drawn considerable attention mostly due to its excellent spatial resolution. The resolution of MRS typically ranges between 1 and 10Pm, which means that strains can be measured in individual sensors. The principle of this method relies on a change of certain molecular vibration frequencies as a result of an applied stress. Several examples are presented and discussed to demonstrate the potential of combining micro and macro strain measurements and modeling to capture the stress distribution in heterogeneous materials.

Issues related to prediction of residual stresses in titanium alloy matrix composites

Recently, a detailed study of residual stresses on the as-processed SCS-6/Ti-24Al-11Nb [0] 8 composite, and SCS-6/Beta-21S composites in unidirectional [0] 4 , cross-ply [0/90] s , and quasi-isotropic [0/′ 45/90] s layups has been completed. In this study, residual stresses have been measured using X-ray diffraction (sin 2 ψ) technique. We have shown that the use of conventional unit cell models consisting of a quarter fiber surrounded by the matrix material to predict residual stresses for verification of experimental results is inadequate. Such models have predicted successfully the stresses at the fiber-matrix interface. However, experimental work to measure residual stresses have always been on surfaces far away from the interface region. In this paper, the approach taken in extending the conventional unit cell model to the concept of multifiber models to predict average stresses are presented. In this process, several modeling issues have been identified. These issues are ( I) use of conventional unit models for prediction of average surface residual stresses, (2) effect of orientation of the sub-surface plies on the residual stresses in the surface ply, (3) residual stresses in the interior plies, and (4) constituent material properties.