For an analysis focus rather than design/manufacture/chemistry/NDT, etc., "Mechanics of Fibrous Composites" by M.H. Datoo (Elsevier) is good if a bit dry. He builds carefully up from simple isotropic cases to fully anisotropic ones in simple steps, and then addresses slightly more abstruse matters like edge stresses.
"Mechanics of Composite Materials" by R.M. Jones (Taylor & Francis) has always (well, it has since it was published in 1975, anyway) been something of a standard analysis reference for classical laminate theory (CLT). Many earlier microcomputer-based analysis codes were based on it. This has recently been updated ('98), and bulked up by a third compared with the original. I haven't seen the new one. The old one may be becoming hard to find.
The treatment of FE software by composites books tends to be fairly general, and the treatment of layered composites by general FE books equally so. However, there are now a bunch of books aimed specifically at FE modeling of composites. Alas, I've never seen a one of them. However, I'm familiar with three of the authors of "Finite Element Modelling of Composite Materials and Structures" by F.L. Matthews, G.A.O. Davies, D. Hitchings and C. Soutis (Woodhead/CRC), and I'd be very surprised if it was a dud.
However, modeling layered materials tends to be an additional (often tacked-on) pre-processing facility in FE codes, and many quirks abound, individual to FE packages. For instance, if you model a sandwich plate with layered faceplates using one major package, then you can't trust the ordinary "plate" stress output; you have to use the "layered" stress output. However, the package will still happily output the plate stresses for you, wrong though they are (they're generally still ok-ish for monolithic areas). Also, for instance, it's common for FE packages to recover average layer stresses at ply centerlines. Most specialised CLT-based plate analysis packages output the layer stresses at the ply extreme fibers (for a four-ply laminate this can make a big difference). In addition, when building a laminate, an offset from the local nodal plane can usually be specified. What happens to this during large displacement non-linear analysis can be rather package-dependent.
If you need a slightly broader work, Michael Chun-Yung Niu's book "Composite Airframe Structures" is worth a look, though it has so much advice (sometimes somewhat contradictory) that I often find myself no wiser for having consulted it. However, it has a lot of background info relevant to design and analysis. As usual with Niu, there are a large number of photos and drawings.
I recently checked my copy of the "Primer on Composite Materials Analysis," as recommended by JimMetalsCeramics above, and found it useful-looking, particularly in how they help you visualise plate deformations with coupling between extension and shear or twist, etc.
I didn't mention sandwiches: for sandwich panels it's hard to beat H.G. Allen's "Analysis and Design of Structural Sandwich Panels," Pergamon, 1969. It's out of print, but it's worth looking out for a second hand copy. This also goes for MIL-HDBK-23, which is supposed to have been subsumed into MIL-HDBK-17, but somehow doesn't seem to have quite made it there yet, while still being cancelled itself. The more modern "An Introduction to Sandwich Construction" by Zenkert is good, but owes much to its predecessors and perhaps lacks their approachability.
While composites usually make design life a bit more complicated and interesting, there are some things the books tend not to tell you. For instance, a composite design can affect team balance and the like, with (typically) the ratio of stress engineers to designers going up. (Traditionally 1 stress:4 design in UK aerospace (probably nearer 1:3 in more recent years), on one all-composite project we were nearly at parity for a large part of the detail design.) Tooling designers also become even more important.
Does anyone have any other favorite composites references they use??
