Models for predicting ring or rotor yarn hairiness are built using a back-propagation neural network algorithm. These models are based on fiber property input measured by three different systems, hvi, afis, and fmt. We compare the prediction results from the different models, which reveal that yarn hairiness measurements from hvi data are superior to other models. The optimum model is based on the availability of all three measurement systems. We also study the impact of each fiber property on yarn hairiness. The dominant effect is fiber length. Each of the remaining properties has a different degree of impact on ring or rotor yarn hairiness.
Get full access to this article
View all access options for this article.
References
1.
AndersonJ. A., “An Introduction to Neural Networks,” The MIT Press, Massachusetts Institute of Technology, Cambridge, Massachusetts, 1994.
BarellaA.ManichA. M., Cotton Spun Yarns of Equal Twist Multiplier: Influence of Spinning Process, Yarn Linear Density and Fiber Properties on Yarn Hairiness, J. Textile Inst.79, 2 (1988).
PillayK. P. R., A Study of the Hairiness of Cotton Yarns—Effect of Fiber and Yarn Factors, Textile Res. J.34, 663 (1964).
7.
ViswanathanG.MunshiV. G.UkidveA. V.ChandranK., A Critical Evaluation of the Relation Between Fiber Quality Parameters and Hairiness of Cotton Yarns, J. Textile Inst.80, 707 (1989).
8.
WaltonW., Use of a Yarn Hairiness Meter and Results Showing the Effect of Some Spinning Conditions on Yarn Hairiness, J. Textile Inst.59, 365 (1968).
9.
Using Neuralworks—A tutorial for Neuralworks Professional II/Plus and Neuralworks Explore, NeuralWare, Inc., Technical Publications Group, Pittsburgh, PA, 1995.
10.
Texas Cotton Quality Evaluation of Crop of 1993, 1994, 1995, International Textile Center, Lubbock, Texas.
