Evaluation of blending uniformity of redried tobacco strips based on online near-infrared spectroscopy and principal component analysis
-
Abstract
In order to promote the quality consistency of various blended tobacco modules and the blending uniformity of redried tobacco strips in threshing and redrying process, a method based on online near-infrared spectroscopy combined with multivariable statistical analysis technology for evaluating the blending uniformity of redried tobacco strips was developed by adopting the principal component analysis-Mahalanobis distance （PCA-MD） method of multivariable statistical analysis method. At the exit of redryer, an online near-infrared spectrometer is used to determine the spectra of redried tobacco strip samples. The original spectral information is preprocessed by Standard Normal Variate transformation （SNV） combined with Savitzky-Golay filt （SG） smoothing and second derivative （SD） preprocessing methods, and the blending uniformity of redried tobacco strips in blended tobacco modules of different batches was evaluated by the developed method. The test results showed that: 1） The blending uniformity coefficient of redried tobacco strips was 85.93%‒87.08%, which indicated that it was feasible to develop the evaluation method for the blending uniformity of redried tobacco strips with online near-infrared spectroscopy and PCA-MD of multivariable statistical analysis technology. 2） The blending uniformity coefficient of redried tobacco strips of each batch was higher than 85.00%, which suggested that the quality consistency of redried tobacco strips was better. Therefore, online near-infrared spectroscopy combined with multivariable statistical analysis technology can quickly and accurately evaluate the blending uniformity of redried tobacco strips, and provides a technical support for improving the process control level of intrinsic quality of redried tobacco strips during threshing and redrying.
-
-
DownLoad: