| The "Made in China 2025”strategy bring an unprecedented development opportunities and challenges for the textile industry.Especially,the main task of the textile industry was determined to be "improve the quality and control total amount".So,how to solve the problem of low quality and low added value of China’s textiles effectively is a great challenge for the development of the Chinese textile industry.The key reasons behind this problem is that the raw materials frequently undergo physical and chemical modification process in the spinning process,these reason make the spinning quality control haven’t achieve precise control,and the root of the problem is:(1)The knowledge association degree of spinning data is relatively low;(2)The influence factors of the spinning quality fluctuations are difficult to identify effectively;(3)The output value of theyarn quality is difficult to control accurately.The existed research on the spinning quality control was limited in spinning quality control model based on a single process,which realized the monitor of the abnormal quality data and event in the spinning process.However,the other problems,such as a low degree of the knowledge association,the mechanism of the yarn quality fluctuation is difficult to identify effectively and spinning quality is difficult to control accurately,has not been solved thoroughly.In order to improve the spinning quality in cotton spinning digital workshop,the thesis focus on the key problems of the knowledge association among multi-process in spinning,the identification of key factors effected quality fluctuation,and the intelligent control method for spinning quality.Hence,the specific innovation work was described as follows.(1)Association method for spinning data in cotton spinning process based on mass loss function.Aiming at the problem of low knowledge assocaiton among the multi-process in cotton spinning digital workshop,an integration model for cotton textile workshop was constructed based on multi-agent theory to achieve the function collaboration among each subsystem.Then,the conflicts among the heterogeneous data are analyzed and studied,and the integrated analysis model for cotton spinning digital workshop is established.Furthermore,a knowledge association method based on mass loss function is proposed,and the results show that the knowledge association method is helpful to realize knowledge association for spinning data.(2)Identification method for the factors affecting spinning quality abomomal flucation.The main quality indexs affecting yarn quality fluctuation are selected from the national quality standard of the yarn.Furthermore,the yarn breaking strength index was selected as the key quality index,and the identification method for abnormal fluctuation factors of spinning quality is proposed based on softmax regression mode.So,the results show that the identification method is conducive to identify the factors affecting quality fluctuation effectively.(3)Control model for quality spinning based on knowledge association among multi-process.A spinning quality prediction model based on fireworks algorithm improved BP neural network is proposed after identifying the key quality indexs and influence factors affecting the abornomal flucation of the spinning quality,and the results show that the prediction accuracy is up to 97.88%.Then,the quality control model for spinning quality based on multi-process knowledge association is established,and the results were shown that the nonconforming rate of the yarn production was decreased by 23.48%by comparing the results with the control model ignoring multi-process knowledge association.So the results also show the quality control model is conducive to realize the the spinning quality control effectively.The research result of this thesis is not only conducive to solve the problem of "the data is rich but knowledge is poor" in cotton digital workshop,and provide the theoretical support to realize the spinning quality control based on data driven,but also is conducive to solve the problem of spinning quality hard to control effectively,and provide technical support for improving spinning quality management level in cotton digital workshop. |
PDF Full Text Request