3D textile composite reinforcements can generally be modelled as continuum media. It is shown that the classical continuum mechanics of Cauchy is insufficient to depict the mechanical behavior of textile materials. A Cauchy macroscopic model is not capable of exhibiting very low transverse shear stiffness, given the possibility of sliding between the fibers and simultaneously taking into account the individual stiffness of each fibre. A first solution is presented which consists in adding a bending stiffness to the tridimensional finite elements. Another solution is to supplement the potential of the hyperelastic model by second gradient terms. Another approach consists in implementing a shell approach specific to the fibrous medium. The developed Ahmad elements are based on the quasi-inextensibility of the fibers and the bending stiffness of each fiber.

In this study, an epoxy resin, YZ-05, designed specifically for carbon fiber reinforced polymer prepregs, was characterized and modeled. A degassing method for the highly viscous YZ-05 was established for specimen preparation. To numerically model the behavior of YZ-05 in curing, several components, including the heat transfer, curing kinetics, and viscoelastic constitutive law ones, were developed, and the corresponding material properties to be input were tested. As YZ-05 shows severe creep at high temperature under mechanical loading, the combination of digital image correlation (DIC) and thermography technique was utilized to obtain its thermal expansion and chemical shrinkage. In the aspects of viscoelastic behavior, stress relaxation tests were performed based on time–temperature superposition principle with a numerical method to calculate shift factors. After development and input identification, these modeling components for YZ-05 resin were integrated and the modeling results were validated using experiments, where bending of resin beams was induced during curing.

Manufacturing process would modify the yarn angle of woven composites. To have a high-fidelity mechanical analysis, it is necessary to study its potential impact. In this article, experimental tests were specifically designed and conducted to investigate the influence of yarn angle on the mechanical behaviors of cured woven composites. The results show that yarn angle variation has a significant impact on the elastic modulus, Poisson’s ratio, and yield stress. The tensile deformation along yarn directions and shear deformation between two yarns were also investigated and their coupling with yarn angle variation was analyzed. It shows that when the yarn angle variation (from 90°) is less than 10°, the influence of the tension–shear coupling on the tensile behavior along yarn directions and shear behavior between two yarn directions is small and can be neglected. When the yarn angle variation is more than 20°, the influence of tension–shear coupling is significant and must be accounted for in constructing the constitutive law for the cured non-orthogonal woven composites.

To enhance the prediction accuracy of forming simulations of carbon fiber unidirectional (UD) thermoset prepregs, particularly the wrinkle predictions, bending behavior of UD prepregs in the forming process needs to be accurately characterized and modelled. In this paper, an improved bending test system was proposed to characterize the bending behavior of UD prepregs at different forming temperatures and loading rates. It is found that both the temperature and the loading rate have significant impact on the bending stiffness of UD prepregs. An obvious nonlinear bending behavior can be noted in the fiber direction. The loading rate sensitivity of bending stiffness in the fiber direction decreases with temperature increasing. To simulate the bending deformation, a superimposed membrane-shell element model was employed to achieve the decoupling of tensile-bending deformation. Membrane element simulated the in-plane tensile and in-plane shear deformations of UD prepregs, while shell element primarily accounted for the bending deformation. Fiber orientation dependent nonlinear bending stiffness was accounted in the model. Deformation simulations were conducted for vertical cantilever bending, horizontal cantilever bending, and compression wrinkling of UD prepregs. Good agreement is noted between simulation and experiment, demonstrating efficiency of the model and validity of measured bending stiffness.

This paper focuses on the ordering and transportation of raw materials for companies. This paper establishes a logistic regression equation to fit the relationship between supply quantity and order quantity based on the supply characteristic index system. In this paper, a multi-objective BP neural network model is established to solve the supply solution with the lowest purchase and storage cost, the highest supplier reliability score, as the objective function. A gray prediction model is built based on the past loss rate data of the forwarder to obtain a prediction formula to predict the loss rate. Finally, a linear programming model is built to solve for the optimal operation solution to meet the demand based on the predicted loss rate.