Abstract:

Objective Fiber-reinforced polymer composites comprise aligned, random, or woven fibers in polymeric matrix. The commonly used woven fabric-reinforced composites made of high-performance fibers are frequently used in applications that require durability in aggressive environments. To achieve deep-level exploitation for engineering applications, modelling of the fabric's microgeometry and the mechanical response is crucial. However, research to characterize these microstructures highly depends on its microscope images and assumes constant yarn cross-sectional shape, which is not the case for most woven fabric types. Since the fiber architecture of woven fabrics has a profound effect on their mechanical properties, a dynamic simulator capable of modelling fabric weaving process considering textile mechanics is necessary.
Method A dynamic textile weaving simulator was established to link weaving actions to fabric patterns and microstructures. Digital element approach (DEA) was implemented under the framework of the software package digital fabric and mechanics analyzer (DFMA). This method established the geometrical model of the key components of a loom. Yarn interlacing motion was guided by weaving matrix specified by steps. Shedding, weft insertion, beat-up, let-off motion, and take-up actions are modelled. The inter-fiber contact force, fiber forces (tensile, shear, and bending), and boundary conditions in weft direction are considered utilizing the central difference algorithm. The weaving process of four unit-cells in the warp direction of a 10-layer three-dimensional orthogonal woven fabric was explicitly modelled at filament-level to derive for its microgeometry.
Results It took 160 steps in total to complete the process, each cell takes 40 steps to weave. During each step, heddles lift or lower the connected warp or binder yarn to form a space between the fell and heddle. Then, the shuttle moves across the encircled space and layers on weft yarn followed by the beat-up motion. The results show that the micro-geometry of the unit-cell is affected by neighboring cells and subjected to change during the weaving process. Take cell two as an example. In step 80, the right edge of the weft yarns was lined up with the reed at the fell, leaving an empty triangular region encircled by binder yarns. After weaving cell three (step 120), the empty space had disappeared and was filled with filaments, causing a reduction in cell thickness and width. The microstructure of cell two ceases to change at approximately step 160 or beyond. The thickness and length of cell two decrease and its microstructures converge with further weaving steps. During yarn-interlacing and shedding motion, the microstructure of the weft yarns next to the fell changes drastically. When some of the warp yarns are raised and the rest are lowered, the fabric next to the fell rips open between the warp yarns in the up and down positions. A V-shaped passage much larger than the yarn cross-sectional area was formed, causing filaments inside the passage to scatter in random directions in an extremely loose state. These filaments were disrupted again by the beating motion before deforming back into a racetrack cross-sectional shape. When the weft yarns on top were pushed to the fell, the filaments circled by warp yarns and the fell bundled up tightly together to form a triangular cross-sectional shape. It changes into a semi-lenticular shape as the weaving process continues. The microstructure, thickness, crimp angle, and cell width of cell two are measured and compared to the actual specimen. The discrepancies are 1.37%, 0.75% and 0.39%, respectively.
Conclusion A dynamic textile weaving simulator which explicitly models shedding, weft insertion, beat-up, let-off motion, and take-up actions was established. This method fully models loom kinetics and kinematics and is capable of generating multiple cells in consecutive. The weaving process of a 10-layer three-dimensional orthogonal woven fabric was successfully modelled step-by-step at the filament level. Four cells in warp direction were produced. The simulation process reveals the revolution of weft yarn microstructure during shedding and beating motion and therefore concluded that the fabric microgeometry changes during weaving. It takes approximately 80 steps (the number of steps to produce two cells in the lengthwise direction in this case) for the microstructure of a newly woven cell to converge. The stress, as well as the cross-sectional shape of weft yarns are approximately symmetrical from top to bottom. The cross sections of the second cell in the warp direction closely match the microscopy images and accurately capture the main characteristics of the fabric, as regard to the fabric thickness, yarn crimp, yarn path, and cross-sectional shapes. This work provides a reliable method for weaving process study and the findings give valuable insights into fabric design and manufacture instruction.

Key words: weaving process, digital element approach, three-dimensional orthogonal woven fabric, micro-geometry, stress distribution