A Case Study in Design of Expermints: Improving the Manufacture of Viscose Fiber

This case study illustrates how statistical techniques help improve the quality of a manufacturing process. The study shows how well designed experiments advance one's understanding of the underlying process, and it emphasizes the importance of experimentation for attaining knowledge. The investigation involves the production of a certain viscose fiber. The results of a simple fractional factorial experiment with ten factors in 32 runs point to useful strategies for improving fiber strength and fiber elongation. The study also investigates the relationship between the strength and the thickness of the fiber.     

Reduction of Yarn Packing Defects Using Six Sigma Methods: A Case Study

This article originated when an Indian textile company identified packing rejection of yarn cones as its major quality problem and decided to use Six Sigma methods to correct the problem. At the end of its manufacturing process, yarn is wound into conical-shaped packages called cones, and it is shipped to customers in this format. Customers were rejecting cones due to unacceptable weight variation. Pareto charts revealed the major “counts” (a measure of yarn fineness) that were experiencing this problem. Technological deliberations led to identifying variation in yarn length, yarn count, empty yarn container weight, and moisture content of yarn as the critical parameters for this rejection. Statistical hypothesis testing established that the observed weight was significantly more than the set weight of yarn at the assembly winding stage. In addition, a significant difference in gross yarn weight between left and right sides of a machine was found at this stage. This occured despite the attachment of electronic length measuring devices (LMDs) on all assembly winding machines. The gage capability analysis of LMDs, performed on the yarn length at two assembly winding machines, revealed inadequate capability. In addition, for the polyester yarn of count 4/12^'s, a relation was found between gross yarn weight and length of yarn through regression analysis. This relationship was used to arrive at the optimum parameter level.