科研之友
微信
新浪微博
Facebook
分享链接
摘要
Based on the characteristic of histograms for fiber length, a finite mixture model has been applied in this paper to characterize cotton fiber length probability distribution, which made it possible to describe the entire fiber length distribution with a parametric continuous function. With the analysis of previous researches, a mixture of Weibull distributions was chosen to generate the parametric probability density function. In this paper, five cotton fiber samples were collected from different spinning processes: bale, card mat, carded sliver, combed sliver, and finished sliver. The fiber length frequency histograms by weight, the histograms by number and some length measures were tested by USTER AFIS Pro. Two-component and Three-component mixture distributions were adopted to fit these histograms, and some length measures were calculated. Then, comparisons have been made between the calculated values of fiber length measures and the tested ones. The results showed that both the finite mixture models based on different component Weibull distributions could fit the cotton fiber length frequency histograms very well, which would be beneficial to pursue the process-related factors in cotton mills for end products.
