Research Article, J Fashion Technol Textile Eng Vol: 2 Issue: 2
Predicting the Hairiness and Coefficient of Variation of Elastic Core-spun Yarns Produced on Sirofil-Spinning System using Artificial Neural Network
Hossein Hasani*, Mohsen Shanbeh and Fateme Reisi | |
1Department of Textile Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran | |
Corresponding author : Dr. Hossein Hasani Department of Textile Engineering, Isfahan University of Technology, Isfahan 8415683111, Iran Tel: +98-311-391-5040; Fax: +98-311-391-2444 E-mail: h_hasani@cc.iut.ac.ir |
|
Received: February 08, 2014 Accepted: April 16, 2014 Published: April 21, 2014 | |
Citation: Hasani H, Shanbeh M, Reisi F (2014) Predicting the Hairiness and Coefficient of Variation of Elastic Core-spun Yarns Produced on Sirofil-Spinning System using Artificial Neural Network. J Fashion Technol Textile Eng 2:2. doi:10.4172/2329-9568.1000102 |
Abstract
Predicting the Hairiness and Coefficient of Variation of Elastic Core-spun Yarns Produced on Sirofil-Spinning System using Artificial Neural Network
This study aims to predict the hairiness (Number of hairs ≥ 3mm) and coefficient of variation (%CVm) of elastic core-spun yarns produced on Sirofil spinning system using artificial neural network method. Different controllable factors in Sirofil spinning system such as distance between two strands, twist level of produced yarns, draw ratio and feeding angle of the elastane and feeding position of elastane part between two strands were considered as input data. The effectiveness of each controllable factor on these two quality responses was also determined. The results showed that an artificial neural network model with two hidden layers with seven neurons and output layer with two neurons gives the best predictive power of the hairiness and %CVm of Sirofil spun yarns. The findings revealed that the feeding position of elastane part was the most dominant parameter on both yarn hairiness and %CVm. Also, the feeding angle of elastane part and yarn twist level showed the least impact on the mentioned quality responses, respectively.