LEE- Savunma Teknolojileri-Yüksek Lisans
Bu koleksiyon için kalıcı URI
Gözat
Son Başvurular
1 - 5 / 9
-
ÖgeBir insansız hava aracına ait kompozit kanadın tasarımı ve yapısal optimizasyonu(Lisansüstü Eğitim Enstitüsü, 2024-05-21)Hava araçları, kısa mesafelerden uzun mesafelere kadar farklı uçuş gereksinimlerini karşılamak üzere kullanılmaktadır. Bu çeşitlilik, havacılık endüstrisinin sürekli olarak gelişmesine ve hava araçlarının çeşitli alanlarda etkili bir şekilde kullanılmasına olanak tanımaktadır. Teknolojik ilerlemeler, hava araçlarının yakıt verimliliği, güvenlik ve çevresel etkiler gibi alanlarda sürekli olarak geliştirilmesine olanak tanır. Elektronik sistemlerin ve uçuş kontrol teknolojilerinin ilerlemesi, hava araçlarının daha güvenli ve etkili bir şekilde uçmasına yardımcı olur. İnsansız Hava Araçları da (İHA'ları da) bu çeşitlilik içinde önemli bir yer tutmaktadır. İHA'lar, uzaktan kumanda veya önceden belirlenmiş programlar doğrultusunda otonom olarak uçabilen araçlardır. Bu araçlar, askeri operasyonlardan güvenlik gözetimine, tarımsal faaliyetlerden bilimsel araştırmalara kadar geniş bir yelpazede görevleri yerine getirmektedirler. İHA'ların tasarımında öne çıkan faktörlerden biri, hafif ve dayanıklı malzemelerin kullanımıdır. Bu bağlamda, genellikle karbon fiber, cam fiber ve epoksi reçine gibi bileşenlerden oluşan kompozit yapısal hafif malzemeler tercih edilmektedir. Hafif malzemelerin kullanımı, İHA'ların daha az enerji harcayarak uzun uçuş süreleri elde etmelerine yardımcı olur. Bu özellikler, İHA'ların uzun süreli ve maliyetinin düşük olması gereken görevlerde diğer hava araçlarına göre avantajlı olmalarını sağlar. Bu çalışma kapsamında bir insansız hava aracına ait kompozit kanadın uçuşa elverişli olacak şekilde tasarımı yapılmıştır ve bu tasarımın yapısal optimizasyonu gerçekleştirilmiştir. Uçuşa elverişli olabilmesi için öncelikle İHA görev profiline uygun olacak iki boyutlu bir kanat profili seçilmiştir. Bu kanat profilinin ilgili kaldırma kuvvetini karşılayabilecek şekilde 8 metre uzunluk, 1.2 metre kök genişliği olan üç boyutlu kanat tasarlanmıştır. Tasarlanan bu üç boyutlu kanadın Ansys Fluent yazılımı yardımı ile akışkanlar mekaniği analizi yapılmıştır. Bu nümerik çalışma sonucunda oluşan basınç dağılımı kanat kabuğu uzunluk yönünde 10 eşit parçaya bölünerek yüzeylerin belirlenen noktalarda oluşturduğu yük ve momentler elde edilmiştir. Bu yükler yapısal olarak iki kiriş ve iki kabuktan oluşan kanat geometrisine Abaqus yazılımında yayılı yük ve moment olarak tanımlanmıştır. Kanat yapısı kiriş uçlarından 6 serbestlik derecesinde tutularak sınır koşulu analiz modeline tanımlanmıştır. 40 bölgeye bölünmüş kanat geometrisine tek yönlü fiber takviyeli polimer matrisli kompozit malzeme ile serim yapılmış ve analiz Hashin hasar kriterine bağlı olarak çözdürülmüştür. Hasar kriterleri sınır değerinde altında kalmış ve dayanımı sağlamıştır. Kanat yapısı uçuşa el verişli ve dayanımlı olarak tasarlanmış ve yapısal optimizasyon çalışmasına geçilmiştir. Bu aşamada oluşturulan analiz modeli Python programlama dilinde hazırlanmış olan kod yardımı ile Abaqus yazılımı bir alt yazılım halinde çalıştırılarak optimize edilmiştir. Bu optimizasyon çalışmasında kanat yapısal ağırlığı, 45 analiz adımı sonunda %34.7 oranında hafiflemiştir. Yapılmış olan bu analiz çalışmalarında oluşan en kalın kompozit serimi, plaka halinde el ile serim metodu kullanılarak üretilmiştir. Üretilen bu plakadan çekme ve basma numuneleri kesilerek testlere tabii tutulmuştur. Çıkan sonuçlar doğrultusunda çalışmanın deneysel doğrulanması yapılmıştır.
-
ÖgeStudy of turbomachinery flows using open source analysis software(Graduate School, 2022-12-19)Construction of a proper mesh is a crucial stage of computational studies to obtain acceptable results as the quality of the mesh specifies accuracy, convergence and computational time. Solution is mainly restricted by imperfections on the geometry definition, low mesh quality and relatively coarse mesh structure. OpenFOAM as an open source software, which offers tools to generate grids for the solution domain in addition to execute CFD analysis. For complex geometries, mesh can be constructed by snappyHexMesh utility of OpenFOAM. snappyHexMesh creates an unstructured mesh over a structured state that is generated by blockMesh utility thus it presents an automatic mesh construction tool for engineering applications. In this study, NASA Rotor 37 that is an axial flow, transonic compressor delivering a pressure ratio of 2.1 at design conditions, is analyzed. OpenFOAM was used for both mesh construction and CFD analysis. An unstructured mesh was generated by snappyHexMesh. Since the solution domain contains periodic boundaries, cyclic boundary condition was used. Generation of cyclic BCs on an unstructured mesh are a difficult process by an open source tool as two periodic boundaries must be identical in order to overlap grids on these boundaries. The cyclic surfaces were produced as circular patterns of each other. Quality of the geometry file is also an important factor for mesh construction. Because of unstructured results of snappyHexMesh with non conformal grids, cyclic boundary condition option could not be used solitarily. Non conformal periodic boundaries are matched by cyclicAMI boundary condition options of OpenFOAM that states Arbitrary Mesh Interface and contructs the mesh matching for non overlapping grids. Integration of cyclicAMI was applied by an exclusive OpenFOAM dictionary, createPatchDict. Thus, for obtaining a mesh automatically, geometry files must have high quality, and periodic boundaries must be geometrically identical. Flow field of the rotor was evaluated in terms of total pressure and temperature ratios, and Mach number distributions. Spalart-Allmaras turbulence model was used. Computational results were compared with experimental measurements and other computational results. According to the results, method of preparation of the geometry and construction of the mesh affect accuracy considerably. Total temperature and pressure under predicted in comparison to experiment and other computational studies. It may be caused by rhoSimpleFoam that cannot capture shock, or by low quality geometry. Hence a proper solver and high quality geometry can give satisfying results.
-
ÖgeMiniature electrical propulsion system design for cube satellites(Graduate School, 2022-08-11)Cube satellites, also known as cubesats, are compact spacecraft that are made up from 10x10x10cm sized cubes. Each one of these cubes are named units or U for short. Based on mission requirements the size of the cubesat can range from 1U to 27Us. Ever since their establishment in 1999 they have been used for academic and educational purposes. Advancements in the miniature electronic now enables these cubesats to perform at a higher grade and be used for commercial and scientific missions. Their compact nature make them affordable and easy to access. This compactness also means that the power and mass budget is very limited compared to the bigger satellite classes. Thanks to these restraints very few cubesats with propulsion systems have been launched into space to date. A propulsion system has the potential to provide greater missions envelope, extended lifespan, precise control for close formation flying and space debris reduction. Propulsion systems are grouped under two main categories as chemical and electric propulsion systems. Compared to the electrical propulsion systems chemical systems provide greater thrust at the cost of reduced efficiency. Since greater efficiency is vital due to compact nature of the cubesat, electric propulsion systems constitute a tempting solution as a propulsion systems. Among them, RF ion thrusters are viable candidates due to their scalability and simple design. Ion thrusters provide greatest propellant consumption efficiency among electric propulsion systems which makes them very preferable. This study presents the design of an RF ion thruster fit to be used in a cubesat. Theoretical knowledge and calculations are presented and the system is calculated to provide 550 µN of maximum thrust and up to 3000 s of specific imoulse. Design and experimental details are provided and based on these designs the actual model of the thruster is manufactured. Manufactured model was then tested at the Space Technologies Laboratory of Bogazici University (BUSTLab). During the tests it was observed that the ions are successfully accelerated and thrust is generated. Measurements of actual thrust levels and ion beam characteristics are left as future work.
-
ÖgeAI-based visual odometry implementation on an embedded system(Graduate School, 2023-06-12)Navigation technology has always been a critical sub-field for defence technologies. The roots of navigation technology extend to early as 3000 BC year. The early examples of navigation technology were the observation of stars, bird following, etc. In modern days, Global Navigation Satellite Systems(GNSS) and Inertial Navigation System(INS) are widely used in defence technologies. The GNSS and INS systems are included nearly all the modern military platforms and smart munitions. However, they have also practical limitations. Therefore, a companion or alternative to these systems may provide flexibility in the navigational technology for military platforms. Visual odometry is an emerging technique for navigation technologies. The visual odometry technique is a dead reckoning technique, and it is the process of relative pose change estimation from camera images. Processing camera images for visual odometry is a computationally heavy operation. As the military platforms vary between a large scale of different sizes, their power and computational requirements also vary. Also, for expendable platforms like smart munitions, the cost is another important requirement. Embedded visual odometry(VO) implementation may provide a low-power, low cost and small-size alternative or companion positioning system to GNSS and INS. Hence the embedded systems are memory scarce, in this work, a new low-memory footprint neural network-based visual odometry method that is implementable on embedded systems is introduced and evaluated. In this work, firstly, the existing literature for geometry-based and deep learning-based methods was examined. Due to robustness and energy efficiency advantages, it was decided to realize a deep learning-based method, which can be further classified as supervised and unsupervised methods. The supervised learning methods generally require the involvement of other sensors than images and recurrent neural networks. These requirements come with additional computational and power consumption. As a very low-power and real-time system was targeted for this work, an unsupervised learning approach was selected as a training framework. With the ideas from the lightweight convolutional neural networks literature, a neural network namely TinyVO was designed. As monocular techniques provide more robustness and cost advantage than binocular ones, it was decided to use a monocular visual odometry technique in this work. The TinyVO network was trained with a well-known scale consistent unsupervised learning framework. After the training, TinyVO's performance on the KITTI dataset was evaluated. To deploy the neural network, MAX78002 artificial intelligence microcontroller has been chosen as the embedded platform. As MAX78002 supports 8 bits weights, the TinyVO was quantized to 8 bits using the provided MAX78002 software toolset. Also, the toolset supports the known answer test functionality. The known answer test was realized by using a sample from the KITTI dataset. Then, the energy and time consumption per inference values was measured. The resulting neural network, TinyVO enables a monocular visual odometry solution with a low-memory footprint, low power, and reasonable accuracy. To the best of found knowledge, this is the first study that provides a microcontroller-based visual odometry solution.
-
ÖgeAn artificial neural network approach to predict the results of strain gauge measurements in the tensile testing of unidirectional laminated composites(Graduate School, 2023)In the area of material science, obtaining material properties by using test methods is an exhausting and an expensive process. Especially in tensile testing, The specimen preparation is time-consuming, and buying specimens is also costly. In order to obtain reliable outcomes, a significant number of test samples incur damage. Therefore, the cost of testing systems is hardly affordable for many researchers, especially in the composite material development process for defense technologies area. New methods in the material testing area are necessary to avoid the economic burden of testing. The main idea of this thesis is to introduce an Artificial Neural Network (ANN) model for predicting a correlation between the device used for testing, and the strain gauges. Furthermore, the strain data obtained from the strain gauges is employed to derive the stress-strain curves essential for determining the material properties. In this thesis, the ANN model predicts the stress-strain curve. Thus, the different algorithms are modeled, and compared to select best algorithm for predicting a stress-strain curve obtained from different tests. Tensile testing is a crucial method for evaluating the mechanical properties of laminated composites. In this study, Artificial Neural Networks (ANN) were employed to analyze and summarize the tensile test results of laminated composites. The ANN models were trained using a dataset consisting of input variables such as displacement (mm), axial force (N), thickness (mm), length (mm), stress (MPa), and strain calculated from the displacement measured by the extensometer, when the output parameter is strain gauge readings. The objective was to develop a predictive model that could accurately estimate these mechanical properties based on the given input variables. Through an iterative training process, the ANN models were able to learn the complex relationships between the input variables and the tensile test results. Once trained, the models could make predictions for unseen laminated composite samples, providing valuable insights into their mechanical behavior without requiring for extensive physical testing. The accuracy and reliability of the ANN models were assessed through various statistical measures such as relative error, mean absolute error, root mean square error, and coefficient of determination and correlation. The results indicated that the developed ANN models were capable of accurately predicting the tensile properties of laminated composites based on the provided input variables. The use of ANN in this study offers several advantages. It provides a faster and more cost-effective alternative to traditional experimental testing, as the models can quickly analyze large amounts of data and provide predictions in real-time. Additionally, ANN models can capture complex nonlinear relationships between the input variables and the tensile properties, which may be challenging to identify using traditional analytical methods.