Residual stress analysis of graphite/polyimide composites using the concept of metallic inclusions

The purpose of this thesis is to investigate the use of metal particles (Al, Ag, Nb) embedded between the first and second plies of 6-ply unidirectional and 4-ply 8-harness satin weave cloth carbon/polyimide laminates, as strain sensors for the determination of residual and applied stresses by x-ray diffraction. XRD measurements were made using a Siemens D500 diffractometer with parallel-beam optics a solid state detector and Cu K a radiation. Specimens were subjected to bending loads while irradiated, using a 4-point bending device mounted on the D500 goniometer.; Finite Element calculations were performed on a specimen with an isolated spherical particle located at half the distance between neutral axis and the surface of the specimen for the 4-ply laminate and two thirds the distance for the 6-ply laminate. ANSYS v.5.2 was used with tetrahedral Solid92 elements.; Eshelby calculations were done using the Eshelby tensor for a spherical inclusion embedded in an infinite homogeneous anisotropic matrix, the known strain matrix for bending and the matrices for thermal expansion of the composite and the metal inclusion.; FEM and Eshelby method results were found to be equivalent for an isolated particle in a large volume of matrix, i.e. a volume fraction of filler approaching zero. For XRD measurements, a certain minimum concentration of filler was required in order to have enough diffracted x-ray intensity to obtain measurable peak positions within acceptable limits of errors. For multiple inclusions, the slopes of strains and stresses versus outer pin displacement inside the inclusions do not differ significantly from those in single inclusions, however a remarkable change is in the intercept. This is due to a constant stress-strain field that is added to each particle single solution, because of the multiple inclusion interaction.; Strains and stresses obtained by XRD in the embedded particles were sensitive to the residual stresses in the as-cured laminates and responded linearly to stresses applied by 4-point bending, qualitatively as expected. Particle strains were smaller than applied strains measured by strain gage or calculated from bending displacement, due to the larger stiffness of the particles compared with the matrix. Particle strains in the fiber direction in the unidirectional laminates, predicted by FEM, show the same trends with bending displacement as those measured by XRD. Quantitative agreement is acceptable even if the distribution of diffracting particles in the laminates is less than optimal for modeling.; Stress transfer factors were introduced and measured experimentally to relate the stresses in the matrix with those in the filler and the fibers.