Variance Tolerancing and Decomposition in Short-Staple Spinning Processes. Part II: Simulations and Applications to Ring and OE Spun Yarns

Abstract

The variance tolerancing and decomposition method developed in Part I of this paper is applied to a large amount of fiber and yarn test data obtained from a spinning mill during a three-year period. The variabilities of single-end strengths found in six different ring spun and open-end yarns are successfully decomposed into random and processinduced variance components by tolerancing the variances of the matching fiber length and tensile properties. A total of 43,080 single-end yarn strength tests, 4,200,000 AFIS® fiber lengths, and over 70,000 Mantis® single-fiber tensile tests are performed for the required parameter estimation, simulation, and model validation. The results confirm that variance tolerancing through “effective gauge length” and other intrinsic components is effective for variance estimation and decomposition. Most significant is that the processinduced variances account for 69-82% of the total yarn strength variations, signifying the importance of and challenges in controlling and reducing process variances in spun yarn manufacturing.