Research On The Application Of Artificial Neural Network In Textile Area

The research of artificial neural network started from the idea of "mindlike machine" posed by McCulloch and Pitts in 1943. With the development of the research on brain and computer, the subject of the research has been tranfered from "mindlike machine" to ~'study machine". In recent years, the artificial neural network has been applied in a wide field from domestic electric appliances to industrial objects. In this thesis, it is discussed that the artificial neural network was applied in textile area, especially in the aspects of quality control of textile products, estimate of the hand characteristics of fabrics and the faults distinguishing of fabrics.1. Firstly, the developing history of artificial neural network was briefly introduced, then the general models and algorithm of artificial neural network, its suitable areas of application and its flaw were introduced. In addition, the situation of application of artificial neural network in textile area in China and abroad recently was summarized.2. The principle and the way of quality control of woolen top using artificial neural network were discussed. The possibility was discussed that the conventional mechanic carding machine was modified with a computer automatic controlling system based on the artificial neural network. Compared with the general automatic controlling system, computer automatic controlling system based on the artificial neural network was put forward to the experiment. The result of the practical measurement showed a precious quality control based on the artificial neural network.3.The objective hand estimation of fabrics has been the subject of research in textile area, and a great achievements in wool textiles, cotton textiles and other middle thick textiles have been made. Because of the complication of silk textiles and their fine texture, there have not been some effective methods of objective hand estimation for silk textiles. In this thesis, the objective hand estimation for silk textiles was studied with artificial neural network, the method of objective hand estimation for silk textiles with based on artificial neural network was put forward, and appeared superior to the other traditional methods. In addition, the hand of silk textiles were studied when different raw silk materials were used. It resulted in that it is valuable to estimate the hand of silk textile with artificial neural network. There was influence to the hand of silk textile when automatic reeled raw silk was used as warp, and there was almost no influence to the hand of silk textile when automatic reeled raw silk was used as weft.4. In the thesis, the new achievements and development of artificial neural network applied in textile field in the world were concluded, the application areas and choices of artificial neural network in textile field were also systematically introduced. There will be a great perspective of artificial neural network applied in textile field.