Investigating the impact of fiber and yarn structure on yarn tensile properties: A computational approach with artificial neural networks

The properties of yarns strongly depend on the fibers and spinning techniques used to manufacture them. It is imperative to investigate the influence of diverse fibers and yarn structures on the tensile properties of the yarns. Simultaneously, it is equally important to forecast these properties to mitigate costs, reduce waste, prevent faults, and achieve the desired performance in the final products. This study used three different fibers, namely conventional polyester, coolmax, and thermolite, and four different spinning techniques, namely conventional ring, compact, siro, and siro-compact spinning, to develop ring-spun yarns. The effect of fiber and spinning technique was investigated on lea strength (usually taken as count of yarn x lea strength (CLSP)), single yarn strength, and breaking elongation of the developed yarns. An artificial neural network model (ANN) was designed to predict the properties of the yarns using varying fiber types and spinning techniques. The yarns made of conventional polyester fibers resulted in the highest strength of the yarns, whereas coolmax fibers imparted maximum breaking elongation to the yarn compared with other fibers. Conventional ring spun yarns showed minimum strength, whereas siro yarns showed the highest strength among all the yarn samples. The ANN model predicted the CLSP, single yarn strength as well, and breaking elongation of the yarn with the highest accuracy as indicated by the R2-value, which suggests the model can be used reliably and accurately for the prediction of tensile properties of the yarns for different fibers and yarn structures.