the use of statistical techniques to study the relationship between cotton fibers and yarn properties - a case study of uganda rotor spun yarn

Abstract

Due to current stringent customer demand for consistent, fast and better quality textile materials, Uganda’s textile industry has lost ground to the competition coming from European textile produced products. Also, cotton being a natural fiber, its properties varies from fiber to fiber, season to season, and bale to bale due to changes in climatic conditions, soil type, growing regions, harvesting and ginning methods, which ultimately affects processing and consistency in yarn quality as per customer requirements. Selection of suitable cotton fibers and spinning parameters for a particular yarn quality requirement can thus reduce on this inconsistency. The objectives of this study were to test the mechanical and physical properties of cotton fibers using a High Volume instrument (HVI), spinning and testing of cotton rotor spun yarns for both mechanical and physical properties and to model the effect of cotton properties and spinning parameters on rotor yarn properties using statistical techniques. Yarns were spun using a rotor machine at Nytil factory in Jinja Uganda. Using Taguchi experimental design, different machine speeds were selected. For every yarn sample spun, the cotton used for the spinning process was characterized using High Volume Instrument (HVI). The cotton fiber properties measured included fiber length 28.54 mm, uniformity Index 83%, short fiber content 6.7%, fiber strength 28.6 mm, fiber elongation 7%, micronaire 4.3, trash content 26, trash area 0.46%, reflectance 74.6% and yellowness 10.91. The resulting yarns were tested for properties of strength, elongation, evenness, imperfections, count and twist. From the experimental data, multiple regression analysis employing Analysis of Variance was used to establish the relationship between cotton and yarn parameters. Regression models were developed and used to predict yarn properties. The model results showed that micronaire, maturity and count had the most significant influence on yarn strength at an adjusted R-Square (R 2 ) value of 0.8094. For yarn elongation, yarn count, fiber elongation, length uniformity and short fiber content were the most significant factors at an adjusted R 2 of 0.5720.Yarn evenness was mostly affected by count, reflectance, short fibre content and trash content at an adjusted R 2 value of 0.8955. Thin places were significantly affected by count, rotor speed, roller speed and trash content at an adjusted R 2 value of 0.9396. Thick places were mostly affected by count,maturity and rotor speed at an adjusted R 2 of 0.7656 while neps were mostly affected by yellowness, short fibre content, twist and rotor speed at an adjusted R 2 of 0.7616. This work has thus proposed models which can predict yarn strength, elongation, evenness, thin places, thick places and neps. Therefore, given fiber properties, the spinner can save time and material by carrying out fewer pre-spinning tests. As a recommendation, using the developed models could aid in the selection of suitable cotton fiber properties and spinning parameters for different yarn properties in order to attain desirable productivity and quality levels of the resulting rotor spun yarns.