Interpretational analysis for yarn quality control

One of the important requirements of textile production is the interpretation of the properties of the final product by the properties of the raw material. By this way, it will be possible to reach required quality level, work efficiently and save sources. The same of Kawabata?s method is adapted for predicting yarn properties using fiber properties and searched by different perspectives. In this study, in the direction of indicated main theme, transformation equations were obtained to reach properties of final product from properties of raw material by using statistical and mathematical methods and reliability of these equations were determined. Feasible regions are determined by using response surfaces and yarns in the form of cone and bobbin are discriminated succesfully by using discriminant analysis.Regular measurements of a cotton yarn producing factory are taken into account as data of properties of raw material and final product. Studies by using data chart method, regression analysis, artificial neural network, principal component analysis, central composite design and discriminant analysis are performed. Because of the properties of raw material are very many, component groups that are composed of effective properties are obtained by using Varimax method and reliability of results of these studies are compared with reliability of results of studies that are performed with all properties included. Response surfaces for one property of the final product by using properties of raw material are obtained by central composite design method and feasible regions are determined by overlapping contour plots of these response surfaces. For the yarn production, forms of bobbin and cone are discriminated seperately by using discriminant analysis.At the end of this study, transformation equations for predicting the properties of final product from the properties of raw material are obtained by using statistical and mathematical methods mentioned above and it is indicated that accuracy rate of interpretations made by these equations are high. By means of these equations, that would be used for the aim of providing quality control in factories, in the direction of indicated main theme, targets of reaching required quality level, working with efficiency and saving sources will be achieved.One of purposes of this study, adaption of the same of Kawabata?s method for predicting yarn properties using fiber properties, is reached. According to this purpose, data chart method, regression analysis, artificial neural network method and reduction of data by varimax are used. Another purpose of this study, determining of feasible regions by using response surfaces and discrimination of yarns in the form of bobbin and cone by using discriminant analysis is achieved.