Application of the Fuzzy Inference System to Evaluate the Quality of Air Textured Warp Yarn

Abstract

It has become one of the indispensable conditions to continuously improve the quality and achieve the quality standards in order to adapt to the increasingly competitive environment in the textile industry. However, the textile production process like many other industrial processes involves the interaction of a large number of variables. For a standard quality production, the relation between raw material properties, process parameters, and environmental factors must be established conclusively. The physical properties of air textured warp yarn that affect the quality of the yarn, construct the strength of the yarn. After the production process, different values of each yarn sample are revealed from the strength tests performed during the quality control process. Six criteria that affect the quality of the yarn and identify the strength of the yarn are defined as a result of strength tests. Those criteria are count, tenacity, elongation shrinkage, resistance per kilometer (RKM) and breaking force. The differences between the values of these criteria and linguistic variables cause uncertainty when defining the quality of the yarn. To take into consideration this uncertainty a fuzzy inference system (FIS) is developed using six criteria as inputs, 144 rules created, and the linguistic variables of air textured yarn (ATY) samples of a textile manufacturer. The quality level of the products according to the different membership functions are identified with the proposed FIS generated by MATLAB version 2015a and recommendations are made to the manufacturer.