Answers are grounded in this paper and cite their sources.
Investigates shear modulus predictions in woven fabrics, suggesting improved modeling techniques.
Estimating the shear modulus of woven fabrics is challenging due to the accuracy and complexity of measurement methods. Additionally, traditional testing methods face limitations such as non-uniform stress distribution and buckling phenomena, highlighting the necessity of predictive modeling as an efficient alternative. This study introduces a dual modeling approach using linear regression and deep neural networks (DNN) to estimate the shear modulus of woven fabric in the weft and warp directions. The models use intrinsic yarn and fabric parameters, including yarn linear density, fiber type, spinning system, and fabric density of the weft yarn. The dataset is part of a previous study on 18 plain woven fabric samples and the Kawabata Evaluation System for Fabric shear measurement. Models were trained and evaluated using k-fold cross-validation. Linear regression revealed density and yarn count as major influencing parameters, while the DNN model significantly outperformed linear regression across all statistical metrics (MSE, R2, adjusted R2, and relative error). The proposed DNN structure captures nonlinear interactions among the input parameters, showing a robust alternative to complex and costly experimental shear testing methods. This predictive framework supports textile engineering in achieving target properties of fabrics in a more efficient and quick approach.
No takes yet. Share an insight, caveat, or question.
Mojiri et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: