CN102628855A - Method for predicating smoke components in cured tobacco leaf - Google Patents

Abstract

The invention relates to a method for predicating smoke components in a cured tobacco leaf and belongs to the technical field of tobaccos. The method comprises the following steps: carrying out correlation analysis on each physical indexes of the cured tobacco leaf by utilizing SPSS (Statistical Product and Service Solutions), and carrying out stepwise regression analysis on the physical indexes and smoke index, thus obtaining regression equations of the physical indexes and smoke index of the cured tobacco leaf; and predicating the smoke index by utilizing each regression equation, wherein the physical index of the tobacco leaf is utilized to predicate total particle matters in the smoke, Y=10.568+0.098X, Y represents the total particle matters (mg/branch), and X represents the thickness of the tobacco leaf (mu m); the physical indexes of the tobacco leaf are utilized to predicate the tar yield of the smoke, Y=35.593-0.322X1-2.786X2, Y represents the tar yield (mg/branch), X1 represents the length-width ratio of the tobacco leaf (%), and X2 represents the tobacco shred filling value (cm<3>/g); and the physical indexes of the tobacco leaf are utilized to predicate the carbon monoxide amount, Y=15.08-0.163X, Y represents the carbon monoxide amount of the smoke (mg/branch), X represents the length-width ratio of the tobacco leaf (%), and the average predication relative error of three smoke indexes is less than 10%. The method has the following advantages: the method is simple and convenient, the average predication relative error is less than 10%, the method can be used for better guiding the cigarette formula work and reducing the workload of the cigarette formula.

Description

A kind of method of predicting the cured tobacco leaf smoke components

Technical field:

The present invention relates to a kind of method of predicting the cured tobacco leaf smoke components, belong to technical field of tobacco.

Background technology:

Cigarette composition is according to the product design requirement, with the tobacco compatibility of the different places of production, kind, grade together, and to obtain expecting the technology of target volume smoke product of exhaust gas volumn and style.Generally be rule of thumb to reach repeatedly experiment to allocate target product when doing compound design; In this process, generally only pay close attention to the place of production, kind, grade and the style of single-tobacco-typed cigarette; The exhaust gas volumn of final products needs to carry out the exhaust gas volumn detection according to GB again behind the rolling cigarette, during the consumption expenses of labour; If when compound design, can detect through simple tobacco leaf physical index; Dope the exhaust gas volumn of single-tobacco-typed cigarette according to physical index; Then can when compound design, accomplish to shoot the arrow at the target; According to the prediction exhaust gas volumn of single-tobacco-typed cigarette, mix the target product of exhaust gas volumn accord with expectation quickly, reduce workload and save time.The present invention provides a kind of detection through simple tobacco leaf physical index, and the method for prediction fume indication can play good directive function to the cigarette composition design.

List of references:

Ji Shuwen rises million ripples. tobacco physical detection [M]. and Henan Science Press, 1997

GB/T 19609-2004 cigarette is measured TPM and tar with conventional analysis with smoking machine

The non-scattering infrared method of mensuration of carbon monoxide in the GB/T 23356-2009 cigarette smoke gas phase

Summary of the invention:

The object of the present invention is to provide a kind of method that can simplify the prediction cured tobacco leaf smoke components of the miscellaneous work amount of filling a prescription.

The present invention carries out the mensuration of cured tobacco leaf physical index with reference to the method for Ji Shuwen, and the physical index of surveying comprises: Dan Yechong (g); Contain stalk rate (%); Length breadth ratio (%); Vane thickness (μ m); Heavy (the g/m of the food value of leaf ²); Filling value (cm ³/ g); Equilibrium moisture content (%); Pulling force (N).Each sample is rolled into 54mm length, and the cigarette of circumference 24.2mm detects its exhaust gas volumn according to GB.

The present invention carries out correlation analysis with SPSS to each cured tobacco leaf physical index, and physical index and fume indication are carried out stepwise regression analysis, and correlation analysis finds that TPM and cured tobacco leaf thickness are remarkable positive correlation; Tar content and cured tobacco leaf length breadth ratio are remarkable negative correlation, are remarkable negative correlation with tobacco fill value; Carbon monoxide in flue gas amount and cured tobacco leaf length breadth ratio are extremely significantly negative correlation; Through physical index and fume indication are carried out stepwise regression analysis, obtain the regression equation of cured tobacco leaf physical index and fume indication, with each regression equation fume indication is predicted; Wherein:

A. predict the flue gas TPM with the tobacco leaf physical index, Y=10.568+0.098X, Y props up for TPM mg/, and X is tobacco leaf thickness μ m, and the prediction average relative error is less than 10%;

B. with tobacco leaf physical index prediction tarry matters in smoke amount, Y=35.593-0.322X ₁-2.786X ₂, Y props up X for tar content mg/ ₁Be tobacco leaf length breadth ratio %, X ₂Be tobacco fill value cm ³/ g, the prediction average relative error is less than 10%;

C. with tobacco leaf physical index prediction carbon monoxide in flue gas amount, Y=15.08-0.163X, Y are that carbon monoxide in flue gas amount mg/ props up, and X is tobacco leaf length breadth ratio %, and the prediction average relative error is less than 10%.

The results of analysis of variance shows that each regression equation reaches the utmost point level of signifiance, and each regression coefficient reaches significantly or the utmost point level of signifiance.Carry out the fume indication prediction with each regression equation, average relative error is less than 10%.See for details like following table.

The variance analysis of table 1 physical index and TPM regression equation

Table 2 physical index and TPM regression coefficient and significance test

Table 3 physical index and TPM return the variable of getting rid of

The variance analysis of table 4 physical index and tar content regression equation

Table 5 physical index and tar content regression coefficient and significance test

Table 6 physical index and tar content return the variable of getting rid of

The variance analysis of table 7 physical index and carbon monoxide in flue gas amount regression equation

Table 8 physical index and carbon monoxide in flue gas amount regression coefficient and significance test

Table 9 physical index and carbon monoxide in flue gas amount return the variable of getting rid of

The invention has the advantages that: can predict the fume indication of cured tobacco leaf preferably with this method, and detect cured tobacco leaf thickness, when tobacco fill value is very easy for length and width, can play good directive function to compound design, simplifies the miscellaneous work amount of prescription.

Embodiment:

The method of prediction cured tobacco leaf smoke components of the present invention; With SPSS each cured tobacco leaf physical index is carried out correlation analysis; And physical index and fume indication carried out stepwise regression analysis; Obtain the regression equation of cured tobacco leaf physical index and fume indication, fume indication is predicted with each regression equation; Wherein:

A. predict the flue gas TPM with the tobacco leaf physical index, Y=10.568+0.098X, Y props up for TPM mg/, and X is tobacco leaf thickness μ m, and the prediction average relative error is less than 10%;

B. with tobacco leaf physical index prediction tarry matters in smoke amount, Y=35.593-0.322X ₁-2.786X ₂, Y props up X for tar content mg/ ₁Be tobacco leaf length breadth ratio %, X ₂Be tobacco fill value cm ³/ g, the prediction average relative error is less than 10%;

C. with tobacco leaf physical index prediction carbon monoxide in flue gas amount, Y=15.08-0.163X, Y are that carbon monoxide in flue gas amount mg/ props up, and X is tobacco leaf length breadth ratio %, and the prediction average relative error is less than 10%.

The inventor adopts said method to 20 in flue-cured tobacco sample; Detect each sample cured tobacco leaf physical index and predict its exhaust gas volumn with regression equation; Each sample is rolled into 54mm length, and the cigarette of circumference 24.2mm props up, and detects its exhaust gas volumn according to GB; And calculating the prediction relative error of each sample and the prediction average relative error of all samples, the result sees table 10 and table 11.

Table 10 cured tobacco leaf physical index detects data

Table 11 is with cured tobacco leaf physical index prediction exhaust gas volumn

Claims (1)

1. method of predicting the cured tobacco leaf smoke components; It is characterized in that this method carries out correlation analysis with SPSS to each cured tobacco leaf physical index; And physical index and fume indication carried out stepwise regression analysis; Obtain the regression equation of cured tobacco leaf physical index and fume indication, fume indication is predicted with each regression equation; Wherein:

A. predict the flue gas TPM with the tobacco leaf physical index, Y=10.568+0.098X, Y props up for TPM mg/, and X is tobacco leaf thickness μ m, and the prediction average relative error is less than 10%;

B. with tobacco leaf physical index prediction tarry matters in smoke amount, Y=35.593-0.322X ₁-2.786X ₂, Y props up X for tar content mg/ ₁Be tobacco leaf length breadth ratio %, X ₂Be tobacco fill value cm ³/ g, the prediction average relative error is less than 10%;

C. with tobacco leaf physical index prediction carbon monoxide in flue gas amount, Y=15.08-0.163X, Y are that carbon monoxide in flue gas amount mg/ props up, and X is tobacco leaf length breadth ratio %, and the prediction average relative error is less than 10%.