Useful life prediction of woolen hand-knotted carpets using multivariate multiple regression

In this study, the main purpose is to predict the useful life of woolen hand-knotted carpets using multivariate multiple regression. Thickness loss of surface pile yarns and compression toughness index were chosen as representative of the compression properties. Also, color difference index (Delta E) of pile yarns, tuft size index, and evenness of texture index were considered as representative of the appearance characteristics. Eighteen woolen hand-knotted carpet samples (symmetric knot) with different structural specifications were produced. The carpet samples were subjected to 4000, 8000, and 12000 drum revolutions (wear factor) using a Hexapod tumbler tester and functional properties of samples were investigated in original and worn out carpet samples. At first, the effective variables were selected using multivariate test, and then multivariate analysis of variance was used for evaluating the significance of obtained models. Optimal separate equations of the functional properties on hand-knotted carpets were determined through multivariate multiple regression method. Reverse model of wear factor can be considered as a proper equation to predict the useful life of carpets.