Optimization of the stacking sequence for maximization of the eigen frequency of composite structures.
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
Insinööritieteiden korkeakoulu |
Master's thesis
Authors
Date
2024-01-22
Department
Major/Subject
Mechanics of Materials
Mcode
Degree programme
Master's Programme in Mechanical Engineering (MEC)
Language
en
Pages
98
Series
Abstract
Carbon fiber composites, particularly unidirectional and woven forms, play a vital role in various structural applications. While extensive research exists on optimizing the stiffness and eigen frequencies of basic structures like plates, cylinders, and cones, there is a noticeable gap in literature concerning the optimization of complex composite structures, such as drone frames. The eigen frequency and dynamic response of a structure are influenced by factors like boundary conditions, geometric parameters (thickness, aspect ratio, shape), and the dynamic environment it operates in. In this study, optimization efforts have focused on thin composite plates and cylinders. Analytical and numerical methods based on classical lamination theory and first-order shear deformation theory were employed to derive the stiffness matrices, followed by optimization using genetic algorithms (GAs) with the objective of finding the stack sequence that maximizes the first mode frequency. GAs, inspired by natural selection, excel in navigating complex, non-convex design spaces and are well-suited for problems involving poorly understood or nonlinear objective functions. The choice of GA for optimizing drone frames is motivated by its ability to efficiently explore large design spaces and find optimal combinations that might be elusive through other means. The focus of the optimization is to ensure that none of the drone's eigen frequencies fall within critical operational ranges, particularly during hover, where blade passage frequency is a prominent forcing frequency. The research methodology involved identifying problematic frequencies through bench and flight tests, constructing a mathematical model in Ansys ACP, and validating it by comparing numerical and real drone frame modes. Subsequently, the validated model underwent optimization using GA. Key outcomes of the study include the observation that boundary conditions significantly affect both the eigen frequency and the convergence rate of optimization. Additionally, while higher-order deformation theories provide more accurate predictions, computational costs can be prohibitive. Therefore, choosing an appropriate theory should be based on the structure's geometry. For instance, classical lamination theory is effective for thin-walled structures, providing reasonably accurate results while potentially overestimating stiffness for moderately thick to thick-walled structures. Lastly, the study found that symmetric laminates generally exhibit higher stiffness compared to asymmetric laminates. This research contributes valuable insights into the optimization of complex composite structures, specifically drone frames, using a combination of analytical, numerical, and evolutionary methods.Description
Supervisor
Romanoff, JaniThesis advisor
Vuorio, JaakkoKeywords
optimization, genetic algorithm, carbon fiber composite, finite element method, drone frame, material modelling