An empirical and analytical investigation of the cure related distortion of flat composite laminates

Manufacturing distortion of autoclave-processed symmetric composites has historically been difficult to predict, resulting in multiple tooling design iterations, increased cost, and delayed product development time. Previous research has identified two potential sources of distortion of autoclave cured, symmetric laminates laid up on flat tooling: part-tool interaction and a through-the-thickness resin volume fraction gradient. Part-tool interaction theory suggests that tooling strains caused by thermal expansion of the tool during autoclave heatup are transferred into the laminate fibers in layer(s) adjacent to the tooling. These tensile strains become "locked in" to the tool side layer(s) of the laminate after resin cure and subsequently result in laminate curvature once the laminate is released from the tooling after processing. A through-the-thickness resin volume fraction gradient in a laminate, characterized by higher resin volume fractions in tool side layer(s) than in bag side layer(s), can also result in laminate curvature. Such an asymmetric laminate will exhibit post-cure curvature because resin rich regions will shrink more than resin poor layers. The first phase of the present investigation determined the relative contributions of these two potential sources of distortion by (1) performing extensive acid digestion and optical photomicroscopy analyses on laminates to accurately characterize the through-the-thickness constituent content, and (2) manufacturing laminates under different experimental conditions proposed to affect the magnitude of part-tool interaction. Curvature results indicated strong effects of tooling material, part thickness, ply orientation, and presence of intermediate layer of Teflon film on measured curvature of cured laminates. Conclusions of the first phase of this investigation are that (1) the absence of any through-the-thickness resin volume fraction gradient indicates that laminate constituent content does not contribute to laminate distortion, and (2) part-tool interaction can result in substantial laminate curvature. The second phase of the present investigation involved empirical and analytical efforts to further understand part-tool interaction. Processing factors empirically investigated included magnitude of autoclave pressure and when in the processing cycle autoclave pressure was applied. Analytical analyses were aided by surface grinding studies that have allowed for the determination of a nonlinear function capable of describing the actual part-tool interaction related strain through the laminate thickness.