PREDICTION OF THE RESPONSE OF METAL MATRIX COMPOSITE LAMINATES UNDER MULTIAXIAL LOADING

S. Subramanian, M. A. Foringer and S. R. Soni

ABSTRACT

In this paper, a simple micromechanics model is proposed to predict the response of metal matrix composites under multiaxial loading. The model includes the effects of residual thermal stresses, interphasial yielding and matrix plasticity. In this work, the concentric cylinders model (CCM) developed by Pagano and Tandon [1] has been modified to include effects that are commonly observed in metal matrix composites (MMC). The matrix region is divided into five layers, and the stresses are determined in each of these layers and the fiber and interphase regions using the CCM. Interfacial debonding is modeled using a cylindrical interphase region and evaluating the yielding behavior of this region under thermo-mechanical loading. The nonlinear response of the MMC is predicted by considering progressive yielding of the various matrix layers. An iterative scheme is used to predict the onset and progression of plasticity in each matrix region. At any applied external load (strain), the volume averaged stresses are estimated in each of the constituent region. The properties of any region that undergoes yielding are altered using the nonlinear stress-strain response of the matrix material. This procedure is repeated under the same applied load, until the solution converges. The model predicts the onset of interfacial debonding and onset and progression of matrix plasticity. The response of multi-directional laminates is predicted using the micromechanics model described above with the classical lamination theory (CLT).

Results indicate that the predicted response of unidirectional and multidirectional laminates under thermo-mechanical loading agree well with experimental data. The onset of interfacial debonding and plasticity is predicted well by the model for SCS6/Ti 15-3 composites. In addition, the predicted response of SCS6/Ti 15-3 composites at room and elevated temperatures agree well with the experimental data.