Textile Surfaces Analysis and Modeling based on Statistical Methods: Variance Analysis and ...

摘要

Statistical analysis using between-variancenare usually used in various industries. In thisnpaper we propose to apply these statistical analysisntools to textile surfaces. Based on early theories [1–n9] statistical tools have been developed to take intonaccount periodic and random defects observed onnlinear textiles. Indeed, in the textile industry thenraw material presents a strong random unevennessnand moreover each processing step introduces itsnown periodic irregularity (e.g. faults due to an ellipticnroller or defective gear). A method for determiningnthese irregularities is developed whereby itnis possible to define the function of the variancemassnper unit area of the fibrous web if the overallnmass variance of the web produced during thenindustrial process is known. Indeed, with the helpnof the autocorrelation function, the between-areadensitynvariance function B(S) of the fibrous webncan be predicted and the shape of the B(S) curve isndetermined. The two types of irregularities havenalso been examined and analyzed. Random irregularitiesnwere developed based on the functions ofnthe most common distributions (isoprobable,nequiprobable, uniform distributions). Then, thenperiodic irregularities have been developed andngeneralized. Finally, we discussed the actual industrialncase known as a superposition of periodic andnrandom irregularities.