Analytical research on intellectual control of yarning characteristics for cotton collocation and rotor spinning

This research deals with the analysis of the characteristics of cotton collocation and the quality characteristics of rotor spinning. To achieve appropriate cotton collocation, the Taguchi method was employed in the experimental design in which the cotton characteristics which were measured with a high volume instrument were used to design the experiment plan according to the relevant orthogonal array. To satisfy the requirements of the modern textile industry, a Rieter open-end (OE) machine was used in the factory to spin the rotor yarn; the yarns thus produced were then measured using an Uster Tester-3, to obtain relevant yarn characteristics. For the characteristic analysis, intellectual control theory was employed to assess the factorial analysis, and the results indicated that, using neural network training and test, the mean square error obtained from the network was below 0.1. Genetic algorithms were also applied to seek one set of optimization characteristics from among cotton quality characteristics. In this way, the cotton properties and the OE rotor yarn characteristics prediction model was structured according to the result obtained by the intellectual control system. The final yarn characteristics were determined by the known cotton collocation conditions. The cotton properties and the OE rotor yarn characteristic prediction model combine the characteristics of neural networks and genetic algorithms by using the evolve perceptron neural network as the basic framework. Compared with the application of a back propagation neural network, it possesses better prediction accuracy and faster convergence. The result indicated that this research is applicable configuring the yarn characteristic prediction model system.