Abstract
Silk is a natural animal protein with special odour. Polyester yarns, similar in appearance to silk, are often used in silk production. Electronic nose (e-nose) technology can distinguish different gases by different pattern recognition methods. In this paper, e-nose technology was applied to obtain data sets of silk/polyester yarns with different mixing ratios, and cluster analysis was carried out for silk samples. The experimental results show that the results of e-nose response are related to the headspace, sample quality, and headspace generation time etc.; as the response time increases, the relative standard deviation of each sensor turns to be smaller, and the response value is more stable. There are significant differences in the fluctuations between the ten metal sensors, and different samples have significant effects on the response of sensors No. 2, 7, and. 9. The e-nose sensor has a higher prediction accuracy when identifying different ratios of silk/polyester yarns at the accuracy rate up to 90.286%.
