Certification of Discontinuous Composite Material Forms for Aircraft Structures: Stiffness and Strength Predictions

The overall goal of this study is to help establish methods to certify prepreg-based discontinuous fiber composite (DFC) aircraft parts based on numerical analysis supported by a modest experimental database. Towards that end a stochastic (Monte Carlo-type) finite-element modeling approach has been developed. During a typical analysis the DFC structure of interest is divided into regions called Random Laminate Volume Elements (RLVEs). Nominal in-plane RLVE dimensions are 19 mm x 19 mm (0.75 x 0.75 in). A unique randomly generated nonsymmetric stacking sequence is assigned to each RLVE. All finite elements present within a given RLVE are assigned the same non-symmetric stacking sequence. Experimentally-observed variations in the stiffness of a DFC part are then simulated by performing many FE analyses, where a new random stacking sequence is generated for each RLVE during each analysis. Fracture predictions are obtained by increasing the load until a ply (i.e., a "chip") is predicted to fail in an element somewhere in the structure. Stiffness's of the failed ply are reduced, and loading is increased until the next ply failure is predicted to occur, etc. Final structural failure of the part is declared when all plies within a single element have failed. A typical analysis predicts that ply failures (i.e., "damage") will accumulate at many regions within a DFC part. That is, damage is not predicted to occur just at stress risers (e.g., at/near holes), but instead damage is predicted to evolve in a distributed manner throughout a typical DFC structure, in qualitative agreement with experimental measurements. Analyses of un-notched tension and compression coupon specimens are discussed in this paper, and predicted B-basis and B-Max measures of modulus and B-basis strengths are presented.