Font Size: a A A

Research And Application Of Quality Prediction Techology For Textiles

Posted on:2010-02-04Degree:MasterType:Thesis
Country:ChinaCandidate:J H WuFull Text:PDF
GTID:2121360275955072Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
Textile quality is not only the life of a textile enterprise,but also the key factor in core competitiveness.In combination of intelligent technologies,textile production in the way that eventually stabilizes textile quality is a trend in textile industry.Intelligent production is gradually becoming the characteristic of modern textile production.The way that adopts various intelligent algorithms of good performance to instruct production has gained certain success in the textile domains and many research institutions are also making good efforts in such aspect.Most of recent research,however,stays theoretical.To convert research findings to practical products is an urgent challenge to solve.On the basis of in-depths research and analysis of cotton textile production and system requirements of textile quality prediction software,the paper propose a solution to textile quality prediction system intended for medium/small-sized cotton textile enterprises.The paper investigates availability of typical algorithms in textile industry,proposes a prediction model of textile quality based on artificial neural network(ANN),supported vector machine(SVM) and decision tree(DCS),and presents problems/solutions which these models will involve in cotton industry.A quality prediction system for textile of certain self-adaptation has been successfully developed through research in the paper.The software is designed to be reusable,flexible and reconfigurable through adoption of domain-oriented.NET development technologies,mature design patterns and plug-in mechanism.Under such framework,the model could be effectively managed and it could easily add prediction algorithm library to the system.In conclusion,the paper introduces an industrial instance to verify the effectiveness of the model and the system.In comparison of various models,the performance of each model and the conditions where it is best used are defined.The system is currently on the trail run in some textile enterprise,and has acquired certain effect.
Keywords/Search Tags:Quality Prediction, Data Ming, Intelligence Algorithm, Quality Control
PDF Full Text Request
Related items