Date of Award
Open Access Dissertation
Because of their superior mechanical and environmental properties compared to traditional metals, fiber-reinforced composite materials have earned a widespread acceptance for different structural applications. The tailoring potential of composites to achieve high specific stiffness and strength has promoted them as promising candidates for constructing lightweight structures. From that aspect, designers have tackled the problem of designing composite laminates, which is inherently challenging due to the presence of non-linear, non-convex, and multi-dimensional optimization problems with discrete and continuous design variables. However, despite their increased usage, the possible improvements that can be achieved by composite laminates have not been fully exploited. With the introduction of new manufacturing technologies such as advanced fiber placement, engineers now have the capability to harness the full potential of nonconventional variable stiffness composite laminates using in-plane fiber steering. This can be a blessing as well as a curse for the designer, where the additional improvements can be attained at the expense of an increased complexity of the design problem. To circumvent this difficulty, this research aims to develop appropriate design tools to help unlock the advancements achieved by nonconventional variable stiffness laminates. The purpose is to adopt an efficient design optimization methodology to abandon the traditional usage of straight fiber composite laminates in the favor of exploring the structural improvements that can be achieved by steered laminated composite structures, subject to manufacturing constraints and industry design guidelines. This represents a remarkable step in the development of energy-efficient light-weight structures and in their certification.
The complexity of the optimization problem imposes the need for an efficient multi-level optimization approach to achieve a global optimum design. In this work, the importance of including a design-manufacturing mesh is demonstrated in each optimization step of the multi-level optimization framework. In the first step (Stiffness Optimization), a theoretical optimum stiffness distribution parameterized in terms of lamination parameters is achieved that accounts for optimum structural performance while maintaining smoothness and robustness. The design-manufacturing mesh allows the spatial stiffness distribution to be expressed as a B-spline or NURBS surface defined by the control points of the design-manufacturing mesh. The fiber angle distribution is then obtained in the second optimization step (Stacking Sequence Retrieval) to match the optimum stiffness properties from the first optimization step while accounting for the maximum steering constraint and laminate design guidelines to attain manufacturability and feasibility. A bilinear sine angle variation is presented to obtain smooth fiber angle distributions, and the maximum steering constraint is derived to guarantee a certain degree of manufacturability at the second optimization step. Using the design-manufacturing mesh, a constant curvature arc solution is developed in the third optimization step (Fiber Path Construction) to generate manufacturable fiber paths with piecewise constant curvature arcs that match the optimal fiber orientation angles from the second optimization step while locally satisfying the maximum curvature constraint. To minimize gaps and overlaps obtained due to fiber steering, a design-for-manufacturing tool is developed to generate tow-by-tow descriptions of the steered plies in the form of manufacturing boundaries for the AFP machine with optimized cut and restart positions.
The design of cylindrical shells under bending with a specified cutout is chosen as an aerospace application to demonstrate the effectiveness of using nonconventional variable stiffness laminates compared to traditional conventional laminates. The presence of the cutout in the cylindrical shell imposes severe stress concentrations yielding a need to use variable stiffness laminates that have continuously varying fiber orientation angles to redistribute the stresses and obtain a structurally optimal design. A design-manufacturing mesh was introduced to perform the buckling load optimization, where both circumferential and longitudinal stiffness variations were considered to physically understand the importance of the stiffness tailoring mechanism in efficient load redistribution and local reinforcements around the regions of the cutouts. The multi-level optimization framework is utilized to obtain a manufacturable fiber-steered laminate that improves the buckling load significantly. The design-for-manufacturing tool developed then generates the tow-level information in the form of exported AFP boundaries. The designed cylindrical shell is imported into CATIA V5® for composite design programming to demonstrate the applicability of the design-for-manufacturing tool developed.
Albazzan, M.(2020). Efficient Design Optimization Methodology for Manufacturable Variable Stiffness Laminated Composite Structures. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/5731