Browsing by Author "Muhammad Mushab Zakaria"

This thesis presents the study of the post-cure behaviour of thick carbon-epoxy composite laminates manufactured to Airbus specification AIMS05-27-002 cured using autoclave. The post-cure behaviour being studied is the warpage of flat laminates and the springback of curved or angled laminates. These laminates resemble the geometrical features found on the actual structure of the rib of the flap of Airbus A350. The gap left from previous works using similar material and process is the uncertainty that exists within the predictive finite element model for the thicker range of the laminates. This study thus attempted to close this gap by first extending the warpage and springback database from the thin laminates (16 layers and below) to thick laminates of up to 28 layers. Then the predictive model is iterated several times by changing the initial stress within the first layer of the elements until the warpage in the simulation matches that of the experiment respectively. The outcomes are a warpage and springback database covering both thin (less than 20 layers) and thick (more than 20 layers) range of thickness as well as a predictive finite element model that is able to predict the behaviour of laminates of the actual aircraft structure. Three categories of specimen were defined; flat unidirectional, curved unidirectional, and flat symmetrical. The effect of four parameters were studied; specimen size, corner angle, laminate thickness, and ply configuration. The results of the analysis indicate that the effect of specimen size for flat unidirectional and aspect ratio for curved unidirectional specimens is negligible. Quadratic response can be found for the effect of specimen size for curved unidirectional and aspect ratio for flat unidirectional. The effect of specimen thickness is linear for all specimen types. The accuracy of the predictive finite element model is mostly above 70% and peaking above 90% around the middle range of each parameter. Overall, the results of the experiment are satisfactory and the practicality of the predictive model is acceptable but more works are required in order to further improve the accuracy of the model.