A deep learning empowered smart representative volume element method for long fiber woven composites
نویسندگان
چکیده
In response to the global trend of carbon reduction over last few years, various industries, including aviation and automobile have gradually begun research, design, production fiber composite materials. These excellent mechanical properties, such as being lightweight, high strength, rigidity, which provide weight energy savings in applications across many fields. When used a load-beam structure, weave pattern determines primary properties material. Therefore, diverse products components can be carried out using different patterns weaving manufacturing according an application’s requirements. The woven composites obtained by simulation analysis software, reduce unnecessary waste during design manufacturing. However, difficulties arise due complexity method. With continuous improvement computer technology recent years enormous amount training data available, research teams implement artificial intelligence (AI) technology, has been widely overcome long-standing obstacles For example, problems involved prediction protein folding sequences physics structural materials all resolved AI. We convolutional neural network (CNN), deep learning method, establish model that utilizes representative volume element for predictive significantly streamlines computational analyzing materials, resulting substantial processing time compared conventional methods. Unlike traditional finite simulations, necessitate intricate boundary conditions interactions on case-by-case basis, our simplifies these complex procedures accommodates wide range scenarios. This offers advantages industrial manufacturing, particularly mass
منابع مشابه
3D Representative Volume Element Reconstruction of Fiber Composites via Orientation Tensor and Substructure Features
To provide a seamless integration of manufacturing processing simulation and fiber microstructure modeling, two new stochastic 3D microstructure reconstruction methods are proposed for two types of random fiber composites: random short fiber composites, and Sheet Molding Compounds (SMC) chopped fiber composites. A Random Sequential Adsorption (RSA) algorithm is first developed to embed statisti...
متن کاملMechanical properties of woven glass fiber-reinforced composites.
The aim of this investigation was to measure the flexural and compressive strengths and the corresponding moduli of cylindrical composite specimens reinforced with woven glass fiber. Test specimens were made by light-curing urethane dimethacrylate oligomer with woven glass fiber of 0.18-mm standard thickness. Tests were conducted using four reinforcement methods and two specimen diameters. Flex...
متن کاملMeso-scale Modeling of Tension Analysis of Pure and Intra-ply Hybrid Woven Composites Using Finite Element Method
One of the key issues associated with using of composites in various applications is their tensile behavior. The tensile behavior of a composite material is strongly influenced by the properties of its constituents and their distribution. This paper focuses on gaining some insights into the tensile process of pure and hybrid woven composite reinforced with brittle and ductile yarns. For this pu...
متن کاملA Level-Set Based Representative Volume Element Generator and XFEM Simulations for Textile and 3D-Reinforced Composites
This contribution presents a new framework for the computational homogenization of the mechanical properties of textile reinforced composites. A critical point in such computational procedures is the definition and discretization of realistic representative volume elements (RVEs). A geometrically-based weave generator has been developed to produce realistic geometrical configurations of the rei...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Frontiers in Materials
سال: 2023
ISSN: ['2296-8016']
DOI: https://doi.org/10.3389/fmats.2023.1179710