Experimental and numerical multi scale analysis of fiber reinforced composites

Increasing energy costs necessitate efficient lightweight constructions. Tailorable materials like fiber reinforced plastics (FRP) are gaining more and more importance. However, the characterization and experimental validation of advanced numerical composites analysis techniques is challenging. A significant outstanding quandary is in regards to the proper, or most beneficial, scale on which the constitutive damage model defining the mechanical response of the material should function. Treating composites as fully homogenized via the macro scale approach is computationally efficient, but it is extremely challenging to devise a predictive model on this scale. In contrast, the meso scale approach, wherein the composite tows are treated as effective anisotropic materials, are more computationally expensive. It further requires a complex damage model to handle the extreme anisotropy of the tow material. Advantageous is the better physical representation of the internal structure. The micro scale approach, involving modeling the composite to the scale of the constituents enables use of simpler models and avoids ad-hoc coupling rules for the various damage components. This approach can be very computationally demanding, and often, the constituent scale data best suited for characterizing the damage model are unavailable. In this dissertation, the hierarchical multi scale method is applied to predict the mechanical response of a carbon FRP plain weave composite. The validity of the method is verified by comparison of the numerically obtained material response with experimental data at all scales. The advantages and limitations of this technique are discussed. An anisotropic damage material model which is required for this approach, is presented.