Estimation of the effect of cotton blend selection on yarn quality in yarn production

Yarn production processes are shaped by the interaction of various factors such as raw materials, machine performance, operating conditions, and workforce factors. This study examines the effects of raw material management, one of the most critical factors affecting yarn quality, on yarn properties and develops prediction models based on these features. The study was conducted in a yarn production facility located in Denizli and consists of two main parts. In the first section, the effects of raw material properties on yarn quality were modeled using multiple linear regression analysis. Prior to the multiple linear regression analysis, a correlation analysis was conducted, and variables with a correlation of 0.90 or higher were excluded from the model. SCI, MST, MIC, UNI, STR, and ELG were determined as independent variables. The quality parameters (Fine50, Thick50, Neps200, H, B-force, Elg, and Rkm) of 14/1, 16/1, and 20/1 yarn counts produced using 8 different blends were considered as dependent variables. Particularly for the Neps200 and B-force parameters, R² and adjusted R² values exceeding 80% were achieved. These findings demonstrate that the models developed to predict yarn quality have high accuracy and are suitable for use in production processes. In the second section of the study, bale arrangement prediction is addressed. To prevent the quality differences caused by randomly arranged bales, quality scores were calculated for each bale, and grouping and clustering steps were applied based on these scores. The weights of each parameter in the quality scores, calculated based on the Mic, Str, and +b parameters, were determined using the Analytic Hierarchy Process (AHP) method, assigning weights of 65% to Mic, 29% to Str, and 6% to +b. As a result, a total of 54 bales were organized into groups of three, reducing the maximum quality difference between bales from 0.84 in random arrangements to 0.006 as a result of the grouping and clustering efforts. This arrangement demonstrated positive outcomes in reducing issues arising from quality differences by ensuring a homogeneous quality distribution.