Study On The Rotor Spinning Technology Of The Wool/Acrylic Blended Yarns

Rotor spinning is the most mature and widely used new spinning technology now, especially in the cotton spinning. However, few researches have been carried out on the rotor spinning technology of wool yarns, in particular wool/acrylic blended yarns. In this thesis, the wool/acrylic blended yarn of 37.48 tex is spun on a FA601A rotor spinning machine, and the rotor spinning technology of the wool/acrylic blended yarn is explored.In order to obtain the optimum combination of the spinning elements, the rotor spinning elements such as the opening roll and navel are optimized using fuzzy decision-making and fuzzy clustering methods. According to the relevant qualitative index of the spun yarns, the appropriate spinning elements are choosen, that is, the saw-toothed opening roller and navel of hard chromium plated steel (45#).After choosing the spinning elements, the processing parameters such as the rotor speed, opening roller speed and twist factor are optimized using the second order general rotatable and combinational design and optimization method. Experiments are arranged with the help of the second order general rotatable and combinational design. The regression equations between the relevant processing parameters and quality indices of yarns are built using the least square method. The effective regression equations are got by verifying the significance of the regression equations and regression coefficients. The optimum processing parameters are then obtained by applying the random direction searching method. The optimum processing parameters are the rotor speed of 36741 rpm, opening roller speed of 8000 rpm, and twist factor of 387. Yarns are spun according to the optimum processing parameters. The quality of yarns is generally better than previous experiments, which proves the effectiveness of the optimum processing parameters.Neural network, which is widely used in pattern recognition, information processing and fault diagnosis, is a branch of intelligence science. In this thesis, a neural network prediction model of the relevant quality indices of the wool/acrylic blended yarn is established. A back propagation (BP) neural network with one hidden layer is built. The number of hidden neurons is 3. The transfer function of hidden and output layer are tansig and purelin, respectively. The preferred train function is traingdx. The prediction results of neural networks are compared with those of regression equations. It is found that the prediction errors of neural networks are smaller than those of regression equations, which indicates the feasibility and effectiveness of predicting the quality of rotor spun yarn with neural networks.It is well known that a certain number of experimental data is required to establish a neural network. If experimental data of the yarn quality are unavailable, it is possible to establish a theoretical method for the yarn quality? An effort is made in this thesis. A theoretical method is established for the breaking strength of wool/acrylic blended yarn when the tensile properties of fibers are only known. The model considers thetensile properties of fibers—the average tensile curve, thedistribution of the breaking strength and breaking elongation of fibers and the basic specifications of the blended yarn and fibers. The model is utilized to predict the breaking strength of wool/acrylic blended yarns (30/70 and 55/45). Because some simplifications and assumptions are made during the establishment of the model, the prediction errors are about 20%. However, a theoretical model is after all established for the breaking strength of blended yarn only according to the tensile properties of fibers and without experiment data of the yarn quality. This is a beneficial exploration. And the theoretical model will be improved continuously with unremitting efforts.