On-Line Control of Yarn Diameter and Evenness by Fuzzy Neural Networks in Melt Spinning
碩士 === 國立臺灣科技大學 === 材料科學與工程系 === 101 === In melt spinning, the yarn evenness plays a critical role in appearance, hairiness, strength, and productivity of yarn, and further affects its production, profit, disposal of the unqualified products. Larger variance of yarn also causes defects of yarn, whil...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2013
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/56819657453670952955 |
| Summary: | 碩士 === 國立臺灣科技大學 === 材料科學與工程系 === 101 === In melt spinning, the yarn evenness plays a critical role in appearance, hairiness, strength, and productivity of yarn, and further affects its production, profit, disposal of the unqualified products. Larger variance of yarn also causes defects of yarn, while yarn with smaller variance will yield more uniform quality. In this study, fuzzy neural network controllers maintain the mean value of yarn diameter at the desired value and reduce the yarn diameter variance. The control law is implemented in a laboratory scale of the melt spinning setup and the take-up roll speed is adjustable to regulate the yarn diameter. The diameter error and its variation are the inputs and the increased amount take-up roll speed is the output. The learning of fuzzy neural network adjusts the center and standard deviation of membership functions in the second layer and link weight values of the third to fourth layer to achieve the effect of convergence and learning. The fuzzy neural network controller can maintain the mean value of yarn diameter at the desired value and lower the variance of yarn diameter. The proposed approaches has been carried out successfully in on-line control of the yarn evenness.
|
|---|