Wet filament winding technology has been extensively used for the manufacture of rotationally symmetric parts made of fiber‐reinforced plastics (FRP). As the design and modeling of FRP‐parts require numerous assumptions, deviations between calculated and achieved mechanical properties are expected. One aspect that contributes to this discrepancy is the assumption of a homogeneous rectangular cross‐section of the fiber‐band. In this work, the fiber‐band geometry in a wet‐winding process of carbon fiber cylinders is analyzed. An Infrared‐optical system for the detection of the fiber bandwidth and winding‐angle is implemented. The influence of the winding speed and the resin temperature is analyzed. An image processing algorithm for the automatic measurement of the fiber's bandwidth and winding‐angle is developed. Manual and adaptive gamma corrections are implemented to improve image quality. A parameter study for the suitable selection of image processing parameters is performed.

Wet filament winding has been established as the primary manufacturing process for overwrapped pressure vessels. However, the prediction of the mechanical performance of pressure vessels is still not sufficiently accurate, mostly due to process‐induced deviations within the laminate structure. In this work, we present an image processing algorithm (IPA‐Delfin) that enables the real‐time generation of a digital reconstruction based on the actual geometry and position of the placed roving. A Type‐IV pressure vessel is analyzed under real manufacturing conditions. Our approach allows the measurement of the overlapping degree of roving within the laminate and the prediction of gaps. The results show that the fiber‐band width and the overlapping area vary significantly when the roving is placed directly on the vessel's polymer liner. A more constant and uniform fiber‐band width and overlapping area were observed when the filament winding took place on the previously placed roving. Hence, the IPA‐Delfin can be potentially used for the generation of digital twins of composite overwrapped pressure vessels.

The high fatigue resistance and low weight features make the unidirectional‐carbon fiber reinforced plastics (UD‐CFRP) attractive for cyclic loading applications. However, the complex damage mechanisms within the composite hinder the understanding of its fatigue response. An energy‐based fatigue model has been proposed for the fatigue behavior of epoxy polymers. However, this kind of approaches dependent on the accurate description of the material's behavior. Our work aims to build a representative volume element model (RVE) for the simulation of the cyclic shear loading behavior of UD‐CFRP, which could support the later extension of an energy‐based fatigue model. To precisely reproduce the micromechanical stress‐strain distribution within the matrix, a nonlinear viscoelastic model has been implemented and validated against experimental data. As a result, the simulated UD‐CFRP hysteresis, as well as the stiffness and strain energy density were accurately reproduced for two load ratios (R = 0.1 and R =  − 1). Additionally, the cyclic strain accumulation observed in the UD‐CFRP can be accurately described by the proposed model.

For fiber‐reinforced plastics, the strain‐rate dependent response is governed by the matrix behavior. In this work, the Goldberg model is considered for the epoxy matrix constitutive material model. Moreover, the strain‐rate dependency is achieved by direct influence on the elastic modulus, the inelastic strain, and the material strain to failure. In addition, an anisotropic damage response is implemented and extended through a strain‐rate dependent definition. Since the constitutive model relies on nonphysical parameters, a parameter study is further performed. Additional numerical investigations using a micro‐mechanical model are performed. Tension and shear loading conditions are evaluated and the influence of different strain rates is explored. Furthermore, the implemented anisotropic damage model is compared and discussed against an isotropic damage model.