CN106290639A - A kind of quantitative forecasting technique of Medicated cigarette milk flavoured type - Google Patents

Abstract

The invention discloses a kind of detection level based in Medicated cigarette 25 kinds of volatile ingredients and carry out the quantitative forecasting technique to Medicated cigarette milk flavoured type.The sample treatment first set up according to independent development and GC MS/MS analyze method and are measured the content of in tobacco shred 25 kinds of volatile ingredients, and then the content of mensuration be brought in regression equation calculating Y_MilkNumerical value, be the predictive value of this kind of Medicated cigarette milk flavoured type quantitative forecasting technique score, in the range from 0 10 points.Quantitative forecasting technique smoking result actual with the sense organ goodness of fit of the Medicated cigarette milk flavoured type of the present invention is preferable, it was predicted that mean absolute error be 0.1189 point.The technical support favourable for safeguarding the stable offer of Sensory Quality of Cigarette.

Description

A kind of quantitative forecasting technique of Medicated cigarette milk flavoured type

Technical field

The invention belongs to quantitative forecast and the control field of Medicated cigarette milk flavoured type, be specifically related to a kind of based in cigarette shreds 25 The detection level of kind of volatile ingredient, and combine regression coefficient and build the method for the regression model quantitative forecast to Medicated cigarette milk flavoured type Method.

Background technology

Stable and the lifting of cigarette product quality, it is critical only that the stable of organoleptic quality and promotes, how setting up Medicated cigarette sense The evaluation methodology of official's quality always is the study hotspot of industry.At present, expert is mainly passed through in the evaluation of cigarette sensory quality Sensory evaluating smoking carry out, and issued and implemented concerned countries and industry standard.Such as, GB 5606.4-2005 " Medicated cigarette the 4th Part: sense organ technology requires " and YC/T497-2014 " Medicated cigarette Chinese-style cigarette style sensory evaluation method ".But owing to the person of smokeing panel test exists There is subjectivity in smokeing panel test, make the repeatability of smoking result and stability there is certain deficiency, to such an extent as to evaluation result Reliability produces impact.

There is internal relation between the organoleptic feature of Medicated cigarette and its chemical composition, different chemical component can be special to Medicated cigarette sense organ Levy the impact produced in various degree.In recent years, Medicated cigarette sensory evaluation and the Study on Forecasting Method of chemically based composition there has also been relatively Many reports.Such as, the not magnificence etc. of Harbin Engineering University provides one and predicts volume by immune neural net algorithm The method (CN 101419209 A) of cigarette sensory evaluating smoking and fume indication, but its component substances including computation modeling in is only total Sugar, total nicotine, reducing sugar etc. more than 10 plant routine chemical components index, and it cannot represent the overall permanence of Medicated cigarette sense organ.Yizhong cigarette The Hu Shengguo etc. of grass (group) company limited has invented the BP neural network model of a kind of combination FUZZY SET APPROACH TO ENVIRONMENTAL, and uses this mould Medicated cigarette sensory evaluating smoking and fume indication are predicted (CN 1525394 A) by type, but it is for the chemical composition modeled and calculate Up to 110 remainders, Indexs measure is qualitative, the most loaded down with trivial details, and difficulty in computation is bigger.In turn result in it predict the outcome relatively low or not may be used The shortcoming such as lean on.

Summary of the invention

Present invention aims to and overcome the deficiencies in the prior art, it is provided that a kind of quantitatively pre-to Medicated cigarette milk flavoured type Survey method, to simplify the flow process of the quantitative forecast of Medicated cigarette milk flavoured type, thus complete Medicated cigarette milk flavoured type fast and accurately quantitative Prediction work.

The purpose of the present invention is achieved by the following technical programs.

The quantitative forecasting technique of a kind of Medicated cigarette milk flavoured type, comprises the following steps:

(1) β-caryophyllene in cigarette sample to be predicted, δ-decalactone, δ-dodecalactone, vanillin, 3-ethoxy are measured Base-4-hydroxy benzaldehyde, γ-amyl group butyrolactone, gamma-undecalactone, raspberry ketone, delta-undeca lactone, benzyl benzoate, gold Herba Albiziae benzylacetone, fenchene, 4-vinyl guaiacol, May bloom ketone, heliotropin, dihydro tonquin butter, eugenol, γ-pungent Lactone, cinnamyl cinnamate, 2-methyltetrahydrofuran-3-ketone, ethyl lactate, 2-methylpyrazine, furfuryl alcohol, alpha-angelica lactone, 2, The mass percentage content of these 25 kinds of materials of 6-lutidines, respectively with X₁To X₂₅Represent；

(2) by X₁To X₂₅Substitute in following regression equation respectively, calculate Y_MilkNumerical value:

Y_Milk=1.735X₁-1.353X₂–1.301X₃-1.094X₄+0.438X₅+0.018X₆+0.662X₇+0.301X₈+ 1.438X₉+1.018X₁₀+1.662X₁₁+1.301X₁₂+1.650X₁₃+1.672X₁₄+1.783X₁₅+1.457X₁₆-1.295X₁₇– 1.741X₁₈-1.613X₁₉–1.300X₂₀-1.363X₂₁–1.417X₂₂-1.412X₂₃-1.583X₂₄–1.561X₂₅-152.116；

(3) Y of gained is calculated_MilkNumerical value, be the score value of the quantitative forecast of this kind of Medicated cigarette milk flavoured type, in the range from 0-10 divides.

Table 1 below list for above-mentioned 25 kinds of substance to be measured, its quantitative forecast score to Medicated cigarette milk flavoured type Regression equation coefficient.

Table 1 regression equation coefficient

Relative to prior art, the invention have the advantages that

1, the present invention 25 kinds of volatile component contents from cigarette shreds, it is possible to be given more than sensory evaluating smoking method Objective, stable result of determination, eliminates the impact of artificial difference in sensory evaluating smoking.

2, use the material of the plurality of classes such as alkene, ester, aldehyde, ketone as depending on that the quantitative forecast score of Medicated cigarette milk flavoured type is predicted According to, more accurate than existing method.

3, use less material for modeling and Medicated cigarette milk flavoured type to be carried out quantitative forecast, have and detect speed faster Degree, the quantitative flow process relatively simplified, it is to avoid the problem of regression modeling overfitting, the result making quantitative forecast is accurate and reliable. The technical support favourable for safeguarding the stable offer of Sensory Quality of Cigarette.

Accompanying drawing explanation

Fig. 1 is the hyperbolic chart of 30 kinds of sample regression value of calculation and sensory evaluating smoking's result.

Detailed description of the invention

By the following examples the present invention is described in further detail, but specific embodiment is not to skill of the present invention The restriction of art scheme.

Embodiment 1

(box mark tar 8-12mg/ props up, and box mark nicotine 0.6-1.2mg/ props up, box mark to buy 30 kinds of domestic commercially available finished cigarettes CO 9-13mg/ props up, and 2015 respectively purchased from markets such as Kunming, Shanghai City, Changshas)；Collected sample is according to YC/T 497-2014 carries out sensory evaluation, determines that its milk style and features sensory evaluating smoking's score range is that 0.5-4.0 divides.

A, sample treatment and analysis

The cigarette of the same brand taking Xin Kaibao at random carries out sample and prepares, and sample preparation quick and precisely, and should guarantee sample Product are the most contaminated, and each laboratory sample prepares three parallel samples.Peel off cigarette paper and filter tip, take out tobacco shred, record weight, then It is respectively put in 3 100ml conical flasks, as sample.The internal standard of 1mL 60 μ g/mL it is separately added in ready sample Naphthalene, 9mL ether, close the lid rapidly and shake up, and is placed on concussion shaking table and shakes 2h, takes upper layer of extraction liquid through 0.22 μm with 10ml syringe Filtering with microporous membrane, analyzes for GC-MS/MS (gas chromatography-mass spectrum/mass spectrograph).

B, gas chromatography-mass spectrum/mass spectral analysis condition

Chromatographic column: DB-5MS (30m × 0.25mm i.d. × 0.25 μm d.f., Agilent company of the U.S.) elastic quartz Capillary chromatographic column；Injector temperature: 250 DEG C；Carrier gas: He, constant current mode, post flow 1mL/min, sample size: 1 μ L, shunting Ratio: 10:1；Heating schedule: 50 DEG C (2min), with the ramp of 5 DEG C/min to 250 DEG C (20min)；Transmission line temperature: 250 ℃；Ionization mode: EI；Ion source temperature: 170 DEG C；Ionization energy 70eV；Heater current: 80 μ A；Full scan monitoring Full scan Pattern, sweep limits: 10-500amu；Multiple-reaction monitoring MRM pattern, qualitative, quota ion is shown in Table 2 to Selection parameter.

25 kinds of volatile components and the qualitative ion pair of interior target in table 2 tobacco shred, quota ion to and collision energy

C, index assay

The retention time of test control, qualitative ion pair and quota ion pair, determine the target compound in sample.Work as examination Sample and standard specimen at identical retention time (± 0.2min) occur, and the standard that the relative abundance of each qualitative ion is suitable with concentration The ion relative abundance of solution is consistent, then can determine whether to exist in sample the measured object of correspondence.Each sample parallel measures 3 times.Root According to volatility and the content of semi-volatile organic compounds in the quota ion calculated by peak area sample of target compound in sample. In sample, the content of volatility and semi-volatile organic compounds is calculated by formula (1):

$C_s=\frac{(\frac{A_s}{A_i}+a)\×C_i\×V}{k\×m}\mathrm{...}\left(1\right)$

Wherein, Cs represents the content of a certain property material in sample, and unit is mg/kg；As be in sample volatility or The peak area of semi-volatile organic compounds, unit is U (integration unit)；Ai is the peak area of internal standard material, and unit is that U is (long-pending Divide unit)；Ci is the amount adding internal standard material, and unit is μ g/mL；M is for weighing tobacco quality, and unit is g；K is each volatility Or the standard working curve slope of semi-volatile organic compounds；A is the standard of each volatility or semi-volatile organic compounds Working curve intercept.

D, the β-caryophyllene measured according to method described in a-c in cigarette sample to be predicted, δ-decalactone, δ-dodecalactone, Vanillin, vanirom, γ-amyl group butyrolactone, gamma-undecalactone, raspberry ketone, delta-undeca lactone, benzene Formic acid benzylester, farnesyl acetone, fenchene, 4-vinyl guaiacol, May bloom ketone, heliotropin, dihydro tonquin butter, Eugenol, γ-octalactone, cinnamyl cinnamate, 2-methyltetrahydrofuran-3-ketone, ethyl lactate, 2-methylpyrazine, furfuryl alcohol, α- The content of 25 kinds of materials such as angelica lactone, 2,6-lutidines.The descriptive statistics of various content of material testing results is shown in Table 3.

The descriptive statistics of 25 kinds of volatile material testing results in table 3 tobacco shred

Note: statistical sample number is: 30

E, the content detecting obtained 25 kind material in d is brought in following regression equation respectively, calculates Y_MilkNumber Value:

Y_Milk=1.735X₁-1.353X₂–1.301X₃-1.094X₄+0.438X₅+0.018X₆+0.662X₇+0.301X₈+ 1.438X₉+1.018X₁₀+1.662X₁₁+1.301X₁₂+1.650X₁₃+1.672X₁₄+1.783X₁₅+1.457X₁₆-1.295X₁₇– 1.741X₁₈-1.613X₁₉–1.300X₂₀-1.363X₂₁–1.417X₂₂-1.412X₂₃-1.583X₂₄–1.561X₂₅-152.116；

F, the Y of calculating gained_MilkNumerical value, be the quantitative forecast score of this kind of Medicated cigarette milk flavoured type.Calculate 30 kinds of samples respectively The Y of product_MilkNumerical value, and with sensory evaluating smoking's result comparison, result is as shown in Figure 1.The curve of diamond indicia is that regression equation records in advance Point, and the square marks actual score of smokeing panel test that is sense organ.There it can be seen that use the Medicated cigarette milk style spy of forecast of regression model Levy sensory evaluating smoking's score preferable with the actual smoking result goodness of fit, it was predicted that average MAE (mean absolute error)=0.1189 point.

Claims (1)

1. a quantitative forecasting technique for Medicated cigarette milk flavoured type, comprises the following steps:

(1) β-caryophyllene in cigarette sample to be predicted, δ-decalactone, δ-dodecalactone, vanillin, 3-ethyoxyl-4-are measured Hydroxy benzaldehyde, γ-amyl group butyrolactone, gamma-undecalactone, raspberry ketone, delta-undeca lactone, benzyl benzoate, Acacia farnesiana Willd. Benzylacetone, fenchene, 4-vinyl guaiacol, May bloom ketone, heliotropin, dihydro tonquin butter, eugenol, γ-octalactone, Cinnamyl cinnamate, 2-methyltetrahydrofuran-3-ketone, ethyl lactate, 2-methylpyrazine, furfuryl alcohol, alpha-angelica lactone, 2,6-diformazan The mass percentage content of these 25 kinds of materials of yl pyridines, respectively with X₁To X₂₅Represent；

(2) by X₁To X₂₅Substitute in following regression equation respectively, calculate Y_MilkNumerical value:

Y_Milk=1.735X₁-1.353X₂–1.301X₃-1.094X₄+0.438X₅+0.018X₆+0.662X₇+0.301X₈+ 1.438X₉+1.018X₁₀+1.662X₁₁+1.301X₁₂+1.650X₁₃+1.672X₁₄+1.783X₁₅+1.457X₁₆-1.295X₁₇– 1.741X₁₈-1.613X₁₉–1.300X₂₀-1.363X₂₁–1.417X₂₂-1.412X₂₃-1.583X₂₄–1.561X₂₅-152.116；

(3) Y of gained is calculated_MilkNumerical value, be the score value of the quantitative forecast of this kind of Medicated cigarette milk flavoured type, in the range from 0-10 Point.