Dezvoltǎri recente în ştiinţa şi tehnologia tricoturilor din bǎtǎturǎ: direcţii pentru noul mileniu (partea a II-a)

The paper presents recent developments of techniques for the control of yarn input tension and fault detection using advanced signal processing. A novel approach is proposed to detect and identify faults during the production of weft knitted fabric, based on real time measurement of yarn input tension. It is suggested that different physical situations (ex. various types of defective knitting elements) would result in different representations of the yarn input tension waveform, and thus leading to the possibility of correctly identifying and distinguishing knitting faults. The findings strongly support the development of an automatic system for the classification of knitting faults. Such a solution might well be a pattern recognition system (PRS). A tool of particular importance used in this study is cluster analysis.