A model based on cigarette density segmentation method for predicting burning cone drop rate of slim cigarettes
Burning cone drop is a potential safety problem and quality defect for slim cigarettes. In order to conveniently monitor the burning cone drop rate, a model based on cigarette density segmentation method was established and used for predicting the burning cone drop rate of cigarettes. This method is nondestructive and can predict the burning cone drop rate of cigarettes according to tobacco rod density and its variation rate.
The established model is based on a straight line ρ=A × ρ’+B in the Cartesian coordinate system with abscissa, the variation rate of tobacco rod density ρ' and ordinate, tobacco rod density ρ. The straight line splits the coordinate system into two regions. By measuring the density of a slim cigarette, the position of point (ρ', ρ) is determined, and the burning cone drop tendency can be predicted accordingly.
The accuracy of the model was validated. The results suggested that the predicted value and the measured value were basically near the 1:1 line. The coefficient of determination (R2) and the index of agreement (D index) between the predicted and measured values were 0.98 and 0.99 respectively, and the normalized root mean square error (nRMSE) was 11.8 %. The coefficient of variation of the predicted burning cone drop rate was 6.5 % for 10 consecutive replicates.The cigarette density segmentation model is suitable for predicting burning cone drop rate of slim cigarette samples with varied cut tobacco density or cut tobacco distribution and provides a valuable reference for the production and quality inspection of slim cigarettes.