Using Multivariate Statistics on Detection of Particular Signals during Production of Knitwear

This paper reports the recent developments in the pursuit to correctly locate, identify and distinguish faults during production of weft knitted fabrics. For this purpose a major textile parameter - yarn input tension (YIT) - is analyzed by means of signal processing techniques. An overview of the entire process of gathering the information and fault detection is presented. For the purpose of distinguishing faults, multivariate statistical methods, namely cluster and discriminant analysis are used, results presented and discussed. Finally, some conclusions are drawn from the obtained results and future developments are addressed