الفهرس | Only 14 pages are availabe for public view |
Abstract Fabric pilling propensity is one of fabric defects that affect the appearance of fabric and does not satisfy the consumer needs. This thesis presented a statistical model based on fabric construction to predict fabric pilling grades of blended of wool woven fabrics. Moreover, it is proposed that bending stiffness, weight and thickness of fabric are indicators for fabric pilling tendency. Three different materials of weft yarns of blended of wool woven fabrics were used; (60/40% blend of wool/polyester fibers, 35/65% blend of viscose/polyester fibers and 20/40/40% blend of viscose/polyester/acrylic) fibers. Warp yarns were made from 60/40% blend of wool/polyester fibers for a yarn count (Nwr 32/2). Different weft yarn counts, ranges of picks per inch and woven structures of fabric samples were used. Results showed that the fabric weave structure is the highest significant factor that affects fabric pilling grades. Bending stiffness of fabric is a good indicator to fabric pilling propensity. Four main models were derived to predict fabric pilling grades by using multiple regression analysis. Moreover, Artificial Neural Network (ANN) was applied in the last model to improve the performance of prediction. Results showed that, the model that used (ANN) to predict fabric pilling grades have higher efficiency and more reality compared to ones that used multiple regression analysis in prediction. |