Textile composites can take many forms (eg. weaves, braids, knits, etc.), but in each case there is interlacing of the fiber bundles. This interlacing offers the potential for increased damage growth resistance. It also introduces waviness in the fibers and complicates prediction of composite properties. Figure 1 shows schematics which illustrate the complex microstructure for weaves. A variety of analyses for analyzing the textile composites have been proposed in recent years. These range from modified laminated plate theory (eg. Ishikawa and Chou (1982, 1983), and Naik (1992)) to fully three-dimensional models (eg. Whitcomb (1991), Dasgupta (1994)). In some cases there has been “verification” of an analysis by comparison with experiments. While such comparisons are essential, they do not provide a complete story. Because of uncertainties in the actual tow architecture, variation of fiber fraction, and in situ constituent properties, it is quite possible for analytical predictions and experimental observations to agree, but for the wrong reasons. A useful compliment to comparison of a single analysis to experimental results is a comparison of several analyses which are based on a wide range of simplifying assumptions. Since the “virtual specimen” is known with much greater accuracy than a real one, it is possible to identify with much greater certainty the validity (or error) of various approximations. Of course, this procedure has its own weaknesses. The main weakness is that the various approximations are being evaluated for an idealized geometric model of the composite (since the real geometry is too complex). Conclusions based on an idealized geometry do not necessarily hold for the “real thing”. In spite of this risk, it is proposed that significant insight can be obtained.

This content is only available via PDF.
You do not currently have access to this content.