A statistical prediction model for pilling grades of blended worsted fabrics based on fabric bending rigidity

Abstract

A laboratory statistical tool based on fabric bending rigidity and weft yarn count to predict fabric pilling grades of blended worsted fabrics is proposed. Multiple regression analysis is performed to obtain three dependent models as evaluation tools for pilling tendency. Three different materials of weft yarns; 60/40% blend of wool/polyester fibers, 35/65% blend of viscose/polyester fibers and 20/40/40% blend of viscose/polyester/acrylic fibers were considered. Moreover, four different counts of weft yarns, different ranges of picks per inch (44–50) and three structures of fabric weave (Plain, twill and satin) were used. Results showed that the first model recorded the best performance to predict fabric pilling grades when considering coefficient of regression analysis was (R2 = 0.92) and Mean Square Error was (MSE = 0.074). Moreover, measuring of fabric bending rigidity and counts of weft yarns is sufficient to be an indicator for predicting the fabric pilling tendency.

Keywords

• Prediction
• Worsted fabrics
• Pilling
• Bending rigidity
• Regression Analysis