Composite materials are composed of a mixture of two or more constituents, giving them mechanical and thermal properties which can be significantly better than those of homogeneous metals, polymers and ceramics. An important class of composite materials are filamentary composites which consist of long fibres embedded in a tough matrix. Materials of this type include graphite fibre/epoxy resin composites widely used in the aerospace industry, and glass fibre/polyester mixtures which have wide applicability in the marine and automotive markets. Because of their predominance in high-quality structures which need to be analyzed before manufacture, the term composite material will refer to a filamentary composite having a resin matrix in this document. Furthermore, it will be assumed that the composite is manufactured in distinct layers, which is appropriate for almost all filamentary composite materials.

By decreasing the characteristic size of the microstructure and providing large interface areas, the toughness of the composite material is improved significantly compared with that of a homogeneous solid made of the same material as the fibres. In addition, the manufacturing processes of many components can be simplified by applying the fibres to the component in a manner which is compatible with its geometry. These and other considerations mean that composite materials are an effective engineering material for many types of structure.

However, filamentary composite materials are often characterized by strongly anisotropic behavior and wide variations in mechanical properties which are a direct result of the manufacturing route for a component. In addition, the cost of a composite component is highly dependent on the way the fibres are applied to a surface. This means that designers must be aware of the consequences of manufacturing considerations from the beginning of the development phase.

Filamentary composite materials are usually placed in components as tows (bundles of individual fibres) or as fabrics which have been processed in a separate operation.

A large proportion of commercially-produced components are built up from layers of fibre tows laid parallel to each other. Each tow consists of a large number of individual fibres as each fiber is usually too thin to process effectively. For example, graphite tows typically contain between 1000 to 10000 fibres. Tows containing many fibres result in cheaper components at some expense of mechanical properties.

Composite structures built up from tows have the greatest volume fraction of fibres which usually lead to the most favorable theoretical mechanical properties. They are also characterized by extreme anisotropy. For example, the strength and stiffness of a resulting layer may be ten times greater in the direction of the fibres compared with an orthogonal direction.

Individual tows may also be woven or stitched into fabrics which are used to form the component. This method effectively allows much of the fibre preparation to be completed under controlled conditions, while components can be rapidly built up from fabric during the final stage of manufacture.

Composite structures built up from fabrics are generally easier to manufacture and exhibit superior toughness compared with those built up from tows, with some loss in ultimate mechanical properties.

Some processing methods allow the user to mix tows and fabrics to achieve optimum performance. An example of this is a composite I-beam, where the shear‑loaded web consists of a fabric, while the axially-loaded flanges have a high proportion of fibres oriented along the beam.

Composite structures are manufactured using a wide variety of manufacturing routes. The ideal processing route for a particular structure will depend on the chosen fiber and matrix type, processing volume, quality required, and the form of the component. All these issues should be addressed right from the beginning of the development cycle for a component or structure.

A feature of almost all the manufacturing processes is that the fibres are formed into the final structure in layers. The thickness of each layer typically ranges between 0.125 mm (0.005”) for aerospace-grade pre-pregs up to several millimeters for woven rovings (say, 0.25”). This means that a component is usually built up of a large number of layers which may be oriented in different directions to achieve the desired structural response.

Another consequence of layer-based manufacturing is that a laminated area is usually thin compared with its area. This means that the dominant loads are in the plane of the fibres, and that classical lamination theory (which assumes that through-thickness stresses are negligible) and shell finite elements can be used to conduct representative analyses. In contrast, in particularly thick or curved skins, inter-laminar tensile and shear stresses can be significant. This can seriously compromise static and fatigue strength and may require the use of through‑thickness reinforcement. Another consequence of thick laminates is that the analyst must use special thick shell or solid elements to model the stress fields correctly.

In the wet layup process, fibres are placed on a mould surface in fabric form and manually wetted-out with resin. Wet layup is widely used to make large structures, like the hulls of small ships.

This process is amenable to high production rates but results in wide variations in quality. In particular, the inability to control the ratio of fibres to resin means that the mechanical properties of the laminate will vary from point-to-point and structure-to-structure.

In this process, tows or fabrics are impregnated with controlled quantities of resin before being placed on a mould. Pre-preg layup is typically used to make high‑quality components for the aerospace industry.

This process results in particularly consistent components and structures. Because of this, pre-preg techniques are often associated with sophisticated resin systems which require curing in autoclaves under conditions of high temperature and pressure. However, the application of pre-preg layers to a surface is highly labor‑intensive, and can only be automated for a small class of simple structures.

Compression moulding describes the process whereby a stack of pre-impregnated layers are compressed between a set matched dies using a powerful press, and then cured while under compression. This method is often used to manufacture small quantities of high-quality components such as crash helmets and bicycle frames.

Due to the use of matched dies, the dimensional tolerances and mechanical properties of the finished component are extremely consistent. However, the requirement to trim the component after curing and the need for a large press means that this method is extremely expensive. Also, it is very difficult to make components where the plies drop off consistently within the component.

Here, dry fibres are built up into intermediate preforms using tows and fabric held together by a thermoplastic binder. One or more preforms are then placed into a closed mould, after which resin is injected and cured to form a fully-shaped component of high quality and consistency. The in-mould cycle time for RTM is of the order of several minutes, while that for SRIM is measured in seconds.

As fibres are manipulated in a dry state, these processes provide unmatched design flexibility. RTM produces good-quality components efficiently but incurs high initial costs for tooling and development. As a result, there is often a cross‑over point between pre-preg layup and RTM for the manufacture of high‑quality components like spinners for aero engines. At a lower level, SRIM is used for the manufacture of automotive parts which have a lower volume fraction of fibres.

In this method, tows are wet-out with resin before being wound onto a mandrel which is rotated in space. This process is used for cylindrical and spherical components such as pipes and pressure vessels.

Winding is inherently automated, so it allows consistent components to be manufactured cheaply. However, the range of component geometries amenable to this method is somewhat limited.

This development of filament winding utilizes a computer-controlled 5-axis head to apply individual tows to a mandrel rotating in space. This allows the manufacture of complex surfaces, such as entire helicopter body shells with speed and precision.

Of course, the equipment required for manufacture is extremely expensive, being of the order of $1 million. In addition, the possibilities for fibre placement are so controllable that no component can possibly make use of the capabilities of the process at present. However, the development of CAE tools for optimized design of composite structures will increase its usefulness in the future.