Research On Rapid Characterization Methods Of Fiber Materials

Natural fibers are important materials in textile industry and people's lives.The quality of natural fibers has been received much attention by manufactures and quality supervision departments.Currently,most of the national standard and industry standard used in the detection of fiber samples is the traditional manual testing method,which is time consuming and affected by the subjective consciousness of the testing personnel.The change of environmental humidity also has a great influence on the detection results because natural fibers are very prone to be affected with damp.In addition,bench top instruments in laboratory can not be suitable to detect fibers in site.So far,fibers testing is short of rapid and accurate technique for the process quality control and quality supervision in markets.This thesis aims at studying fast identification method of fur and quantitative analytical methods of duck down and scoured cashmere using near infrared spectroscopy combined with multivariate methods.(1)In this paper,the near-infrared spectrum was carried out to explore the feasibility of classification for animal fur.The principal component analysis was conducted on 90 samples of 5 kinds of furs,and the clustering tendency could be viewed from the scoring diagram of the first three principal components.A SIMCA-based model were constructed to discriminate the types of fur,which show the good classification performance of 100%.The results indicated that the method could realize the rapid and nondestructive classification for animal furs.(2)The qualitative analysis was conducted for five kinds of leather by using NIR combined with PCA-based model and a SIMCA-based model.The study found that the principle component analysis could accurately distinguish the artificial and animal leather,but not identify the specific types of animal leather.After applying wavelet denoising to the spectrum of leather,the recognition rate of the established SIMCA model was 100%,which was superior to the original SIMCA model.(3)A rapid quantitative analysis method for the down content of duck down was established by combining NIR with multivariate method.Based on the collection of 60 different down content of duck down,the quantitative model for down content was built using partial-least-square(PLS)regression.The SECV of partial least square model was 3.15,and the SEP 4.91.(4)The near-infrared spectrums of scoured cashmere were obtained by a portable NIR instrument.And the quantitative calibration model of pure cashmere content was established by partial-least-square(PLS)regression.However,pure cashmere and guard hair are very similar in their chemical compositions,and the spectra of the cashmere were heavily influenced by moisture.This paper adopted a new method to successfully eliminate the effect of H20 by subtracting spectra of water from the spectra of scoured cashmere.Then,we obtained the characteristic absorption wavelength which reflected the difference information between pure cashmere and guard hair by two-dimensional spectroscopy.Finally,the PLS model of pure cashmere content was established by adopting the selection of NIR characteristic wavelength bands.The SECV of partial least square model was 7.43,and the SEP was 7.76.The results indicated that a portable NIR instrument could achieve the rapid and on-site detection of scoured cashmere.