CN104076010A - Method for detecting honey quality in refining process - Google Patents
Method for detecting honey quality in refining process Download PDFInfo
- Publication number
- CN104076010A CN104076010A CN201310101453.9A CN201310101453A CN104076010A CN 104076010 A CN104076010 A CN 104076010A CN 201310101453 A CN201310101453 A CN 201310101453A CN 104076010 A CN104076010 A CN 104076010A
- Authority
- CN
- China
- Prior art keywords
- honey
- near infrared
- data
- sample
- refining process
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a method for detecting the sample quality in the honey refining process. The method comprises the following steps that: 1, samples in the honey refining process are collected, and content determination data of index ingredients is obtained; 2, near infrared spectrogram data of honey samples is collected; 3, the near infrared spectrogram data collected in the second step and the data obtained through measurement in the first step are used for building an associated quantitative model by adopting a partial least square method; and 4, the near infrared spectrogram data of samples to be detected is obtained, and the content values of the index contents are obtained by the model in the third step. The method has the advantages that the fast nondestructive quality control on the integral quality in the honey refining process can be realized; the prediction precision meets basic requirements of the industrial process analysis; the method is fast, reliable, simple and convenient; obvious advantages which cannot be realized by a traditional chemical analysis method are achieved; and high practical application values are realized.
Description
Technical field
The present invention relates to a kind of quality determining method of honey, be specifically related to a kind of in refining process the quick quality determining method of honey.
Background technology
Honey in China as medicinal, with a long history, be recorded in the earliest Shennong's Herbal, Ming Dynasty Compendium of Material Medica that Li Shizhen (1518-1593 A.D.) is shown also has the record of pair honey.Its taste is sweet, and property is flat, has bowl spares emergency, moistens the lung and relieve the cough, ease constipation is dry, effect of removing toxic substances.Honey is extremely extensive in clinical middle application, except separately as health food and medicinal, honey is also used as the important auxiliary material of Chinese medicine preparation (honey is processed) and traditional Chinese medicine pill after refining.
The impact that in refined honey process, the quality of refined honey is subject to several factors is as heat time, temperature, evaporation capacity, gas flow rate etc., and the quality of the different refined honeys of technique of refining honey is also different.At present, Chinese Pharmacopoeia and each province and city concocted specification also clearly do not specify concrete refining process, refining process disunity, the standard that refined honey is ununified, uncontrollable refined honey is refined any degree the best, could play the satisfied process of preparing Chinese medicine of synergy and sweet needs for honeyed bolus with medicine.
The quality of honey refining is the key link of preparing honeyed bolus.If refine inadequately, or use refining degree too high honey, honeyed bolus is prone to elephant skin, stiff, the quality problems such as break at lay up period.When honey is processed, the quality of refined honey directly affects the quality of concocting medicine materical crude slice, and then affects the curative effect of medicine.Therefore, the refining process of honey and quality control are of crucial importance.
Currently available technology is not for the detection method in honey refining process.
Summary of the invention
The object of the present invention is to provide the quality determining method in a kind of honey refining process, the method is fast and convenient, detects index few.
Honey is nectar, secretion or the honeydew of honeybee herborization, after self secretion is combined, through fully brewageing the crude sweet material forming, principal ingredient is glucose and fructose, containing in addition maltose, sucrose, moisture and other micro constitutent (as amino acid, protein, organic acid, vitamin, Flavonoid substances and acetylcholine), is a kind of natural nutrition health food.Honey is used as the important auxiliary material of Chinese medicine preparation (honey is processed) and traditional Chinese medicine pill after refining, itself and the physicochemical property generation great changes of initial honey.The quality of the refined honey in refined honey process is subject to the impact of several factors as heat time, temperature, evaporation capacity, gas flow rate etc., if refine inadequate, or using refining degree too high honey, honeyed bolus is prone to elephant skin, stiff, the quality problems such as break at lay up period.When honey is processed, the quality of refined honey directly has influence on the quality of concocting medicine materical crude slice, and then affects the curative effect of medicine.At present those skilled in the art are according to the existing detection method method convenient, fast, that effectively detect refined honey process product that is in no position to take possession of.Those of ordinary skill in the art will be often that glucose and fructose are that starting point is determined detection method from honey principal ingredient.Be different from the general understanding of prior art and those of ordinary skill, the invention provides a kind of very unique scheme.
A kind of detection method technical scheme convenient, fast, that effectively detect each refined honey process product provided by the invention is:
To a quality determining method for sample in honey refining process, the method comprises the steps:
Step 1: collect sample in honey refining process, obtain the assay data of index components;
Step 2: gather honey sample near infrared light spectral data;
Step 3: the data acquisition that the near infrared spectrum data that step 2 is gathered records with step 1 is set up associated quantitative model by partial least square method; And
Step 4: obtain testing sample near infrared light spectral data, utilize the model of step 3 to obtain index components content value.
In honey refining process of the present invention, sample can be the honey after intermediate material or the refining in honey, honey refining process.
Further, described in step 1, index components is moisture and 5 hydroxymethyl furfural;
Further, the honey sample in step 1 be honey respectively at refining under 100,110,120 DEG C of constant temperatures, the refining time is that 0-10h((comprises 0min) every sampling in 15 minutes from 0min) honey sample that obtains;
Further, the assay of 5 hydroxymethyl furfural adopts Syrups by HPLC, and concrete grammar comprises: chromatographic condition: taking (4-10): methanol-water (70-120) is as mobile phase isocratic elution; Column temperature: 20-40 DEG C; Flow velocity: 0.5-1.2ml/min.Chromatographic condition is more preferably: taking the methanol-water of 6:94 as mobile phase isocratic elution; Column temperature: 30 DEG C; Flow velocity: 0.8ml/min.
Further, in step 2 near infrared ray method: adopt the saturating reflection accessory that light path is 2.0mm, sweep limit is 4000-10000cm
-1, resolution is 8cm
-1.
Further, in step 2, near infrared ray, adopt commercially available near infrared spectrometer, as Antaris II ft-nir spectrometer (Thermo company of the U.S.).
Further, before in step 3, near infrared spectrum data is associated with content data, spectrum is carried out to pre-service, and described preprocess method is selected from any one or a few in following method: first order derivative (1D), second derivative (2D), Savitzky Golay level and smooth (SG), polynary scatter correction (MSC).
Further, for different index components, best preprocess method: moisture data adopt the method for first order derivative, level and smooth, polynary scatter correction coupling to process; The method of 5 hydroxymethyl furfural data preprocessing method first order derivative, level and smooth, polynary scatter correction coupling is processed.
Described in step 1, index components is for comprising moisture and 5 hydroxymethyl furfural, glucose, fructose.
Glucose, fructose content detection method adopt high efficiency liquid phase method, and wherein mobile phase is: acetonitrile-water.
Partial least square method (PLS) is a kind ofly can process the regression modeling method of multiple dependent variables to multiple independents variable.The basic thought of partial least squares regression: be provided with p independent variable x1, x2 ... xp} and q dependent variable y1, y2 ... yq}.In order to study the statistical relationship of dependent variable and independent variable, observe n sample point, form thus the tables of data of independent variable and dependent variable: X={x1, x2 ..., xp}n × p; Y={y1, y2 ..., yq}n × q.Partial least squares regression extracts respectively composition t1 and u1 in X and Y, wherein t1 be x1, x2 ... the linear combination of xp}, u1 be y1, y2 ... the linear combination of yq}.Extracting when this two compositions, for the needs of regretional analysis, there are following two requirements: (1) t1 and u1 should carry their variation information in tables of data separately as wide as possible; (2) degree of correlation of t1 and u1 can reach maximum.These two requirements show: t1 and u1 be representative data Table X and Y as well as possible, and meanwhile, the composition t1 of independent variable has again very strong interpretability to the ingredient u 1 of dependent variable.After first composition t1 and u1 are extracted, partial least squares regression is implemented respectively recurrence and the Y recurrence to t1 of X to t1.If regression equation has reached satisfied precision, algorithm stops; Otherwise, extract utilizing residual, information after residual, information after X is explained by t1 and Y are explained by t1 to carry out the second composition of taking turns.And so forth, until can reach a more satisfied precision.The number of optimum composition (being latent variable) generally adopts the mode of cross validation, can be divided into: 1. stay a cross-validation method; 2. cross-validation method in batches; 3. divide sample cross-validation method; 4. random sample cross-validation method etc.If finally to X extracted altogether m composition t1, t2 ..., tm, partial least squares regression will by implement yk (k=1,2 ..., q) to t1, t2 ..., tm recurrence, be then expressed as yk about former variable x1, x2 ..., xp regression equation.For example set up equation for dependent variable y1: y1=β 1t1+ β 2t2+ ... β rtr+E.Wherein E is regression error.After regression model is determined, need to be to model evaluation.Model performance often adopts RMSEC(Root mean square error of calibration, proofread and correct root-mean-square error), RMSECV(Root mean square error of cross validation, cross validation root-mean-square error), RMSEP(Root mean square error ofprediction, predicted root mean square error), coefficient R and RPD(Relative prediction deviation, relatively prediction deviation) etc. index evaluate.
Have not yet to see the research report about the quick quality control of honey refining process, the quick method of quality control of near infrared spectrum of the honey refining process that this patent is set up can be realized the quick nondestructive quality control to refined honey process total quality, precision of prediction meets the basic demand that industrial process is analyzed, method is quick, reliable, easy, there is the remarkable advantage that traditional chemical analytical approach does not have, be of very high actual application value.
Brief description of the drawings
Fig. 1 a5-HMF reference substance chromatogram
Fig. 1 b honey sample chromatogram
The correlativity of Fig. 2 moisture measured value and predicted value
The correlativity of Fig. 3 5-HMF measured value and predicted value
Fig. 4 moisture is with refining time, temperature variation
Fig. 5 5-HMF content is with refining time, temperature variation
Fig. 6 fructose content is with refining time, temperature variation
Fig. 7 glucose content is with refining time, temperature variation
Embodiment
Embodiment 15-HMF, moisture are set up respectively content assaying method
1 content assaying method
In refined honey, sample 5-HMF, moisture are set up respectively content assaying method.
1.1 instruments and reagent
5 hydroxymethyl furfural reference substance, Chinese pharmaceutical biological product is examined and determine research institute (lot number 111626-200906) is provided
Methyl alcohol, analyze pure, Beijing Chemical Plant
Methyl alcohol, chromatographically pure, fisher scientific
Acetonitrile, chromatographically pure, fisher scientific
Electronic balance BS-124S, Beijing Sai Duolisi instrument system company limited
LC-20AT Shimadzu high performance liquid chromatograph, Japanese Shimadzu company
DF-101S oil bath pan, Zhengzhou Greatwall Scientific Industrial & Trading Co., Ltd.
WAY-2S type Abbe refractometer, Shanghai Precision Scientific Apparatus Co., Ltd
MP501A ultra thermostat, YI HENG TECHNICAL.CO., LTD
1.25-HMF content assaying method
The preparation of reference substance solution: get 5 hydroxymethyl furfural reference substance appropriate, accurately weighed, add methyl alcohol and make the solution of every 1ml containing 0.0188mg, to obtain final product.
The preparation of need testing solution: get honey 2.5g, accurately weighed, put in conical flask, add 10% methanol solution, be settled in 25ml volumetric flask, shake up, (0.45 μ is m), for subsequent use to cross miillpore filter.
Chromatographic condition: instrument: LC-20AT Shimadzu high performance liquid chromatograph, Japanese Shimadzu company; Chromatographic column: Eclipse XDB-C18 post (4.6mm × 150mm, 5 μ are m); Mobile phase: methanol-water (6:94); Column temperature: 30 DEG C; Flow velocity: 0.8ml/min.
Methodological study result: equation of linear regression is Y=9837069.20x-6843.09(r=0.9999), be good linear relation at 188～3.76 μ g; Minimum detects and is limited to 0.376 μ g; Negative sample is noiseless; Precision RSD is 1.42%; Stability RSD is 0.60%; Average recovery empirical average value is that 97.28%, RSD is 1.86%; Replica test RSD is 0.83%.
5-HMF reference substance chromatogram and honey sample chromatogram are shown in Fig. 1.
1.3 moisture determination
Instrument: Abbe refractometer, ultra thermostat.
The preparation of sample: uncrystallized sample firmly stirs it.Have the sample of crystallization, can sample bottle capping plug is tight after, be placed in that to be no more than the water-bath of 60 DEG C warm, after sample all melts, stir evenly, be cooled to rapidly room temperature and use in order to inspection.In the time melting, must be noted that to prevent moisture intrusion.
Running program:
(1) Abbe refractometer and ultra thermostat are connected, and the temperature of ultra thermostat is adjusted to required temperature.
(2) correction of refractometer: before working sample refraction index, first proofread and correct the refraction index of refractometer by table 1 with fresh distilled water.
Table 1 distilled water shading index
| Temperature, DEG C | Refraction index | Temperature, DEG C | Refraction index |
| 14 | 1.3335 | 25 | 1.3325 |
| 16 | 1.3333 | 26 | 1.3324 |
| 18 | 1.3332 | 28 | 1.3322 |
| 20 | 1.3330 | 30 | 1.3319 |
| 22 | 1.3328 | 38 | 1.3308 |
| 24 | 1.3326 | 40 | 1.3305 |
Regulate and be just 40 DEG C by the water flow temperature of refractometer.Separately refractometer two sides prism, dips in distilled water with absorbent cotton and wipes only (the necessary moment dips in dimethylbenzene or ether is wiped only).Then use clean absorbent cotton (or lens wiping paper) to wipe dry, after prism bone dry, dip 1～2 of distilled water with glass bar, drip on prism below rapid closing prism, alignment light source, observe with eyepiece, rotation hand wheel, the refraction index of water while making refraction index on scale just be 40 DEG C, in observation telescope, whether bright-dark cut is in the middle of objective lens cross curve.If there is deviation, regulate spanner to rotate indicating value adjusting screw with annex square hole, make bright-dark cut be transferred to central authorities.After adjustment, in the time of working sample, do not allow to rotate again the screw regulating.
Sample determination: first by prism scrub, measure precision in order to avoid leave other material impacts before mensuration.Dip 1～2, the sample that mixes with glass bar, drip on prism below, rapid closing prism, leaves standstill the several seconds, to treat that sample reaches 40 DEG C.Alignment light source, is observed by eyepiece, and rotation compensation device spiral, makes bright-dark cut clear; Rotary scale pointer spiral, makes its bright-dark cut just by the intersection point of cross curve on objective lens, reads the refraction index on scale, checks temperature simultaneously, should be just 40 DEG C.
Result is calculated: moisture is pressed formula (1) and calculated: X=100-[78+390.7 (n-1.4768)] (1)
In formula: moisture in X-sample, %;
The refraction index of n--sample in the time of 40 DEG C.
The permissible error of parallel experiment is 0.2%.
As surveyed and read refraction index 20 DEG C time, can be scaled by table 2 percent of moisture.To survey while reading refraction index in room temperature, the refraction index can be converted to 20 DEG C by formula (2) time.
Refraction index (20 DEG C)=n+0.00023(t-20) (2)
In formula: the refraction index of n--in the time of room temperature t DEG C;
Temperature when t-read refraction index.
Note: as disputable, be used in 40 DEG C of method inspections of measuring refraction index.
Table 2 honey moisture converts
| Refraction index, 20 DEG C | Moisture, % | Refraction index, 20 DEG C | Moisture, % | Refraction index, 20 DEG C | Moisture, % |
| 1.5044 | 13.0 | 1.4935 | 17.2 | 1.4830 | 21.4 |
| 1.5038 | 13.2 | 1.4930 | 17.4 | 1.4825 | 21.6 |
| 1.5033 | 13.4 | 1.4925 | 17.6 | 1.4820 | 21.8 |
| 1.5028 | 13.6 | 1.4920 | 17.8 | 1.4815 | 22.0 |
| 1.5023 | 13.8 | 1.4915 | 18.0 | 1.4810 | 22.2 |
| 1.5018 | 14.0 | 1.4910 | 18.2 | 1.4805 | 22.4 |
| 1.5012 | 14.2 | 1.4905 | 18.4 | 1.4800 | 22.6 |
| 1.5007 | 14.4 | 1.4900 | 18.6 | 1.4795 | 22.8 |
| 1.5002 | 14.6 | 1.4895 | 18.8 | 1.4790 | 23.0 |
| 1.4997 | 14.8 | 1.4890 | 19.0 | 1.4785 | 23.2 |
| 1.4992 | 15.0 | 1.4885 | 19.2 | 1.4780 | 23.4 |
| 1.4987 | 15.2 | 1.4880 | 19.4 | 1.4775 | 23.6 |
| 1.4982 | 15.4 | 1.4875 | 19.6 | 1.4770 | 23.8 |
| 1.4976 | 15.6 | 1.4870 | 19.8 | 1.4765 | 24.0 |
| 1.4971 | 15.8 | 1.4865 | 20.0 | 1.4760 | 24.2 |
| 1.4966 | 16.0 | 1.4860 | 20.2 | 1.4755 | 24.4 |
| 1.4961 | 16.2 | 1.4855 | 20.4 | 1.4750 | 24.6 |
| 1.4956 | 16.4 | 1.4850 | 20.6 | 1.4745 | 24.8 |
| 1.4951 | 16.6 | 1.4845 | 20.8 | 1.4740 | 25.0 |
| 1.4946 | 16.8 | 1.4840 | 21.0 | ? | ? |
| 1.4940 | 17.0 | 1.4835 | 21.2 | ? | ? |
The foundation of the honey refining process quick Environmental Evaluation Model of embodiment 2 based on moisture and 5-HMF
1 near infrared spectra collection
Get the about 5g of honey sample, be placed in the sealed bag without near infrared absorption, drain air, sealing.Sealed bag is placed in to integrating sphere window, taking the built-in background of instrument as reference.
Spectra collection condition is as follows:
Instrument: Antaris II ft-nir spectrometer (Thermo company of the U.S.)
Annex: adopt the saturating reflection accessory that light path is 2.0mm
Spectral scan scope: 4000-10000cm-1
Resolution: 8cm-1
Scanning times: 32 times, each sample gathers respectively 3 times, and result is with log(1/R) value represents.
The foundation of 2 models
Adopt the TQ analyst V8 software matching with instrument, moisture and 5-HMF content data that the near infrared spectrum of collection is recorded with embodiment 1 are associated, and adopt PLS method to set up quantitative model.
The calibration set and checking collection of sample divided in layering, and wherein calibration set comprises 84 samples, and checking collection comprises 39 samples.
The preprocessing procedures of attempting comprises first order derivative (1D), second derivative (2D), SavitzkyGolay level and smooth (SG), polynary scatter correction (MSC).In experiment, adopt 1D, 1D+SG, 1D+SG+MSC, 2D, 2D+SG, 2D+SG+MSC to compare, found optimum preprocess method.
Adopt different latent variable because of subnumber, the predictive ability of model has larger difference.Selecting applicable latent variable is also one of effective ways that make full use of spectral information and filter out noise because of subnumber.Evaluation index is PRESS, and its value is less, illustrates that the estimated performance of model is better.
With PRESS evaluation index, determine that latent variable is because of subnumber.Investigate the performance of model with forecast set root-mean-square error RMSEP, cross validation root-mean-square error RMSECV, prediction residual quadratic sum (PRESS) and coefficient of determination R.
The comparison of table 3 different pretreatments method calibration model result
RMSEP:Root mean square errorofprediction predicted root mean square error
Rp2: the forecast set coefficient of determination
RMSECV:Root mean square error of cross validation cross validation root-mean-square error
RMSEP/RMSECV: predicted root mean square error/cross validation root-mean-square error
RPD: prediction deviation relatively
Latent Factor: the latent variable factor
PRESS:predicted residual error sum square Prediction sum squares
Best quantitative model the results are shown in Table 4, and between model predication value and measured value, correlativity is shown in Fig. 2-3.
Index component quantifying model in table 4 honey
Moisture through 1D+SG+MSC method process after, select latent variable because of subnumber be 7, set up quantitative model.The related coefficient of calibration model is 0.9995, and checking collection mean square deviation (RMSEP) and cross validation mean square deviation (RMSECV) are 0.155 and 0.174, and both ratios are that 0.8908, RPD value is 22.04.Fig. 2 shows that the correlativity between model predication value and measured value is good.
5-HMF is after 1D+SG+MSC method is processed, adopt PLS method to set up optimum calibration model, latent variable is because of subnumber 14, the related coefficient of calibration model is 0.9981, checking collection mean square deviation (RMSEP) and cross validation mean square deviation (RMSECV) are 132 and 168, both ratios are that 0.7857, RPD value is 11.39.Fig. 3 shows that the correlativity between model predication value and measured value is good.
3 conclusions
Because the scope of quantitative model calibration set has contained different temperatures, sample under the different refining time, the model scope of application is wider.Experimental result shows, the quick nondestructive that near-infrared spectrum technique can be applied to honey refining process detects, and application prospect is good.
The method for measuring research simultaneously of embodiment 3 fructose, glucose content
1.1 instruments and reagent
Fructose reference substance, Chinese pharmaceutical biological product is examined and determine research institute (lot number 100231-200904) is provided
Glucose reference substance, Chinese pharmaceutical biological product is examined and determine research institute (lot number is provided
110833-200904) methyl alcohol, analyze pure, Beijing Chemical Plant
Methyl alcohol, chromatographically pure, fisher scientific
Acetonitrile, chromatographically pure, fisher scientific
Electronic balance BS-124S, Beijing Sai Duolisi instrument system company limited
Agilent1100 high performance liquid chromatograph, Alltech3300ELSD detecting device
DF-101S oil bath pan, Zhengzhou Greatwall Scientific Industrial & Trading Co., Ltd.
The preparation of 1.2 reference substances and need testing solution
The preparation precision of reference substance solution takes fructose reference substance 31.25mg, and glucose reference substance 30.56mg, puts in 50ml volumetric flask, adds methanol constant volume to scale, shakes up, and obtains the hybrid standard product solution of fructose and glucose.
The about 0.2g of honey sample is got in the preparation of need testing solution, accurately weighed, put in small beaker, add 10% methanol solution appropriate, stir sample is dissolved completely with glass bar, be transferred in 100ml measuring bottle, and then wash beaker and glass bar three times cleansing solution is transferred in volumetric flask with appropriate 10% methanol solution, be settled to scale, shake up.(0.45 μ is m), for subsequent use to cross miillpore filter.
1.3 chromatographic condition
The present invention respectively flow phase ratio (ratio of acetonitrile and water is respectively 80:20,78:22,75:25), drift tube temperature (120,90,85,80 DEG C), gas flow rate (2.8mL/min, 3mL/min) investigates, according to chromatographic peak degree of separation size, determine following chromatographic condition.
Instrument: Agilent1100 high performance liquid chromatograph, Alltech3300ELSD detecting device; Mobile phase: acetonitrile-water (75:25); Chromatographic column: APS-2Hypersil nh 2 column (4.6mm × 250mm, 5 μ are m); Flow rate of mobile phase: 1ml/min; Drift tube temperature: 80 DEG C; Gas flow rate: 3ml/min; Gain: 1.Sample size 5 μ l.
1.4 methodological study
The methodological study of fructose: equation of linear regression is Y=0.8964x-2.2585(r=0.9993), 1.372～13.72 μ g are good linear relation; Negative sample is noiseless; Precision RSD is 0.59%; Stability RSD is 0.39%; Average recovery empirical average value is that 97.83%, RSD is 3.28%; Repeated experiment RSD is 1.5%.
Glucose methodological study: equation of linear regression is Y=0.9699x-2.1600(r=0.9996), be good linear relation at 1.2048～12.048 μ g; Negative sample is noiseless; Precision RSD is 0.91%; Stability RSD is 0.66%; Average recovery empirical average value is that 108.89%, RSD is 3.02%; Repeated experiment RSD is 1.3%.
2 Quality Control Model based on fructose, glucose detection Index Establishment
2.1 spectra collection conditions
Get the rear about 5g of honey sample of refining, be placed in the sealed bag without near infrared absorption, drain air, sealing.Sealed bag is placed in to integrating sphere window, taking the built-in background of instrument as reference.Near infrared light spectrogram when this problem has compared saturating reflection accessory light path and is 1.0mm and 2.0mm, in the time that light path is 2.0mm, effect is better.Other spectra collection conditions are as follows:
Instrument: Antaris II ft-nir spectrometer (Thermo company of the U.S.)
Annex: adopt the saturating reflection accessory that light path is 2.0mm
Spectral scan scope: 4000～10000cm
-1
Resolution: 8cm
-1
Scanning times: 32 times, each sample gathers respectively 3 times, and result is with log(1/R) value represents.
2.2 modeling algorithm
Adopt the TQ analyst V8 software matching with instrument, near infrared spectrum and the chemical score that adopts the present embodiment fructose, glucose detection method to record are set up to quantitative model.PLS method is the perfect adaptation of multiple linear regression, canonical correlation analysis and principal component analysis (PCA), is also PLS method most widely used algorithm in near-infrared spectrum analysis, therefore adopts PLS method to set up quantitative model.The calibration set and checking collection of sample divided in layering, and wherein calibration set comprises 82 samples, and checking collection comprises 41 samples.Investigate the performance of model with forecast set root-mean-square error RMSEP, cross validation root-mean-square error RMSECV, prediction residual quadratic sum (PRESS) and coefficient of determination R.
The selection of 2.3 preprocessing procedures
Conventional preprocessing procedures comprises the methods such as first order derivative (1D), second derivative (2D), Savitzky Golay level and smooth (SG), polynary scatter correction (MSC).1D, 2D method can farthest reduce spectrum peak skew and drift, conventionally adopt SG smoothly derivative spectrum to be carried out to smoothing processing; Polynary scatter correction is for eliminating multiple spectrum deviation.Preprocess method can be used in conjunction with, and adopts (1D, 1D+SG, 1D+SG+MSC, 2D, 2D+SG, 2D+SG+MSC) to compare herein, finds optimum preprocess method.
The comparison of table 5 different pretreatments method calibration model result
RMSEP:Root mean square errorofprediction predicted root mean square error
Rp2: the forecast set coefficient of determination
RMSECV:Root mean square error of cross validation cross validation root-mean-square error
RMSEP/RMSECV: predicted root mean square error/cross validation root-mean-square error
RPD: prediction deviation relatively
Latent Factor: the latent variable factor
PRESS:predicted residual error sum square Prediction sum squares
Determining of the 2.4 latent variable factors
While adopting PLS method to set up quantitative correction model, adopt different latent variable because of subnumber, the predictive ability of model has larger difference.Selecting applicable latent variable is also one of effective ways that make full use of spectral information and filter out noise because of subnumber.Evaluation index is PRESS, and its value is less, illustrates that the estimated performance of model is better.As calculated, the latent variable of fructose, glucose is respectively 10,8 because of subnumber.
The foundation of 2.5 each component quantifying models
Select best preprocess method and latent variable because of subnumber, set up PLS model for fructose, glucose, result is as shown in table 6.Good relationship between predicted value and the actual value of each index model.
Each component modeling and predicting the outcome in table 6 honey
The comparison of embodiment 4 embodiment 2 and embodiment 3 detection methods
When PLS method is set up model, conventional evaluation index comprises forecast set root-mean-square error RMSEP, cross validation root-mean-square error RMSECV, coefficient of determination R and RPD etc.
Wherein n is checking collection number of samples, and yi is the chemical score of the sample i that records,
for the estimated value of sample i of the model prediction of setting up.
Wherein n is calibration set number of samples, and yi is the chemical score of the sample i that records,
for the estimated value of the sample i with the model prediction of setting up after Rejection of samples i.
Predicted value and measured value correlation calculations formula as above.
by the mean value of survey chemical score.
The value of RMSEP and RMSECV is less and more approaching, illustrates that the model of setting up is better.Coefficient of determination R is larger, illustrates that correlativity is stronger, and model is better.
RPD=SD/SEP, wherein SD is the standard deviation of checking collection chemical score, SEP is the standard deviation of checking collection.When the value of RPD is larger, the estimated performance of model is good.
In refined honey process, moisture, 5-HMF, fructose and glucose are set up quantitative model, its correction coefficient (R) is respectively 0.9995,0.9981,0.9369,0.9049, forecast set root-mean-square error (RMSEP) is respectively 0.155,132,18.4,20.1, and the model scope of application is respectively 8.99%～21.80%g/g, 0.0044～5.5721mg/g, 264.1140～447.8220mg/g, 305.4088～474.071mg/g.From above data, the quantitative model prediction accuracy of moisture, 5-HMF is higher than fructose and glucose, and the quantitative scope of moisture, 5-HMF is than glucose, the quantitative wide ranges of fructose.
Embodiment 5 refining temperatures, the impact of time on honey index components
1 sampling method
Get honey in right amount in flask, be placed in thermostatical oil bath, respectively at refining under 100,110,120 DEG C of constant temperatures, the refining time is 10h.From 0min, (comprise 0min) every sampling in 15 minutes, altogether 123, sample.
The collection of 2 refined honey sample chemical values
Moisture determination: utilize Abbe refractometer, detecting the moisture in above-mentioned sample by importing and exporting 3.4 lower methods of the honey method of inspection in the inspection and quarantining for import/export industry standard SN/T0852-2000 of the People's Republic of China (PRC), is 40 DEG C by the water flow temperature of refractometer.
The assay of 5-HMF, fructose and glucose: the content of 5-HMF, fructose and glucose in the content testing sample of recording according to " embodiment 15-HMF content assaying method " and " embodiment 3 fructose and glucose content assay method " respectively.
3 refining temperatures, the impact of time on each component content
3.1 impacts on water cut
Honey is in refining process, and the quantitative change that contains of moisture reduces gradually with the increase of refining time, and the less loss speed of moisture increases (see figure 4) with the increase of refined honey temperature.
3.2 impacts on 5-HMF content
Honey is in refining process, and the content of 5-HMF significantly increases.The content of 5-HMF increases with the increase of refining time, and refining temperature has appreciable impact to the rate of rise of 5-HMF content.Honey is refined respectively after 120min, 75min, 60min at 100,110,120 DEG C, and 5-HMF content exceedes the international standard 40 μ g/g(that limit the quantity and sees Fig. 5).
3.3 impacts on fructose, glucose content
Glucose and fructose content in refined honey process changes, and there is no obvious rule (seeing Fig. 6-7) but change.
3 conclusions
In honey refining process, fructose, glucose content change, and there is no a rule but change, obvious and regular the following of the content of moisture and 5-HMF.
Claims (10)
1. the quality determining method to sample in honey refining process, is characterized in that, the method comprises the steps:
Step 1: collect sample in honey refining process, obtain the assay data of index components;
Step 2: the near infrared light spectral data that gathers honey sample;
Step 3: the data acquisition that the near infrared spectrum data that step 2 is gathered records with step 1 is set up associated quantitative model by partial least square method; And
Step 4: obtain testing sample near infrared light spectral data, utilize the model of step 3 to obtain index components content value.
2. the method for claim 1, is characterized in that, index components is moisture and 5 hydroxymethyl furfural described in step 1.
3. method as claimed in claim 2, is characterized in that, the assay of 5 hydroxymethyl furfural adopts Syrups by HPLC, chromatographic condition: taking (4-10): methanol-water (70-120) carries out isocratic elution as mobile phase; Column temperature: 20-40 DEG C; Flow velocity: 0.5-1.2ml/min.
4. method as claimed in claim 2, is characterized in that, in step 2 near infrared ray method: adopt the saturating reflection accessory that light path is 2.0mm, sweep limit is 4000-10000cm
-1, resolution is 8cm
-1.
5. method as claimed in claim 2, it is characterized in that, before near infrared spectrum data is associated with content data in step 3, spectrum is carried out to pre-service, described preprocess method is selected from one or more in following method: first order derivative, second derivative, level and smooth, polynary scatter correction.
6. method as claimed in claim 5, is characterized in that, moisture data adopt first order derivative, level and smooth and polynary scatter correction method for combined use to enter to process.
7. method as claimed in claim 5, is characterized in that, 5 hydroxymethyl furfural for data acquisition first order derivative, level and smooth and polynary scatter correction method for combined use process.
8. method as claimed in claim 2, is characterized in that, index components is for also comprising glucose and/or fructose described in step 1.
9. method as claimed in claim 8, is characterized in that, glucose, fructose content detection method adopt high efficiency liquid phase method, and wherein mobile phase is: acetonitrile-water.
10. method as claimed in claim 8, is characterized in that, in step 2 near infrared ray method: adopt the saturating reflection accessory that light path is 2.0mm, sweep limit is 4000-10000cm
-1, resolution is 8cm
-1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310101453.9A CN104076010A (en) | 2013-03-27 | 2013-03-27 | Method for detecting honey quality in refining process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310101453.9A CN104076010A (en) | 2013-03-27 | 2013-03-27 | Method for detecting honey quality in refining process |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104076010A true CN104076010A (en) | 2014-10-01 |
Family
ID=51597424
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310101453.9A Pending CN104076010A (en) | 2013-03-27 | 2013-03-27 | Method for detecting honey quality in refining process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104076010A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104614323A (en) * | 2015-02-04 | 2015-05-13 | 南昌航空大学 | Scattering spectrum detection method of true and false honey |
| CN105115905A (en) * | 2015-07-23 | 2015-12-02 | 广东省中医药工程技术研究院 | Method used for standardizing processing technology of radix aconiti |
| CN106198446A (en) * | 2016-06-22 | 2016-12-07 | 中国热带农业科学院热带作物品种资源研究所 | The method of L-Borneol content near infrared spectrum quick test Herba Blumeae Balsamiferae leaf powder |
| CN107688008A (en) * | 2017-07-19 | 2018-02-13 | 北京化工大学 | A kind of quick, method of infrared spectrophotometry of accurate measurement animal glue drying time |
| CN109030410A (en) * | 2018-10-24 | 2018-12-18 | 吉林省现代中药工程研究中心有限公司 | The construction method of royal jelly near-infrared quantitative calibration models and the detection method of royal jelly |
| CN110292163A (en) * | 2019-07-29 | 2019-10-01 | 浙江中医药大学 | A kind of constant temperature refined honey method |
| CN112630180A (en) * | 2020-12-14 | 2021-04-09 | 华南农业大学 | Ultraviolet/visible light absorption spectrum model for detecting concentration of organophosphorus pesticide in water body |
| CN113390821A (en) * | 2021-05-21 | 2021-09-14 | 北京中医药大学 | Real-time monitoring equipment and method for quality of intelligent honey refining process |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0605662B1 (en) * | 1991-09-27 | 1998-06-10 | Foss Electric A/S | An infrared attenuation measuring system |
| JP2003503667A (en) * | 1999-07-06 | 2003-01-28 | エウセビ インピアンティ−エッセ．エッレ．エッレ． | Valve for high pressure gas cylinder |
| CN101363828A (en) * | 2008-09-26 | 2009-02-11 | 中国农业科学院蜜蜂研究所 | Method for appraising royal jelly quality |
| CN101413885A (en) * | 2008-11-28 | 2009-04-22 | 中国农业科学院蜜蜂研究所 | Near-infrared spectrum method for rapidly quantifying honey quality |
-
2013
- 2013-03-27 CN CN201310101453.9A patent/CN104076010A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0605662B1 (en) * | 1991-09-27 | 1998-06-10 | Foss Electric A/S | An infrared attenuation measuring system |
| JP2003503667A (en) * | 1999-07-06 | 2003-01-28 | エウセビ インピアンティ−エッセ．エッレ．エッレ． | Valve for high pressure gas cylinder |
| CN101363828A (en) * | 2008-09-26 | 2009-02-11 | 中国农业科学院蜜蜂研究所 | Method for appraising royal jelly quality |
| CN101413885A (en) * | 2008-11-28 | 2009-04-22 | 中国农业科学院蜜蜂研究所 | Near-infrared spectrum method for rapidly quantifying honey quality |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104614323A (en) * | 2015-02-04 | 2015-05-13 | 南昌航空大学 | Scattering spectrum detection method of true and false honey |
| CN105115905A (en) * | 2015-07-23 | 2015-12-02 | 广东省中医药工程技术研究院 | Method used for standardizing processing technology of radix aconiti |
| CN106198446A (en) * | 2016-06-22 | 2016-12-07 | 中国热带农业科学院热带作物品种资源研究所 | The method of L-Borneol content near infrared spectrum quick test Herba Blumeae Balsamiferae leaf powder |
| CN107688008A (en) * | 2017-07-19 | 2018-02-13 | 北京化工大学 | A kind of quick, method of infrared spectrophotometry of accurate measurement animal glue drying time |
| CN107688008B (en) * | 2017-07-19 | 2020-12-11 | 北京化工大学 | Infrared spectrum method for quickly and accurately measuring drying time of animal glue |
| CN109030410A (en) * | 2018-10-24 | 2018-12-18 | 吉林省现代中药工程研究中心有限公司 | The construction method of royal jelly near-infrared quantitative calibration models and the detection method of royal jelly |
| CN109030410B (en) * | 2018-10-24 | 2021-09-10 | 吉林省现代中药工程研究中心有限公司 | Construction method of royal jelly near-infrared quantitative correction model and royal jelly detection method |
| CN110292163A (en) * | 2019-07-29 | 2019-10-01 | 浙江中医药大学 | A kind of constant temperature refined honey method |
| CN112630180A (en) * | 2020-12-14 | 2021-04-09 | 华南农业大学 | Ultraviolet/visible light absorption spectrum model for detecting concentration of organophosphorus pesticide in water body |
| CN113390821A (en) * | 2021-05-21 | 2021-09-14 | 北京中医药大学 | Real-time monitoring equipment and method for quality of intelligent honey refining process |
| CN113390821B (en) * | 2021-05-21 | 2024-04-12 | 北京中医药大学 | Real-time monitoring equipment and method for quality of intelligent honey refining process |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104076010A (en) | Method for detecting honey quality in refining process | |
| CN104062263B (en) | The near-infrared universal model detection method of light physical property close fruit quality index | |
| Li et al. | Application of near infrared spectroscopy for rapid analysis of intermediates of Tanreqing injection | |
| CN102539566B (en) | Method for fast detecting content of dioscin in dioscorea zingiberensis by utilizing near infrared spectrum technology | |
| CN102106939B (en) | Quality control method for extract concentrated liquor of condensed pills of six ingredients with rehmannia root | |
| CN104237060B (en) | Multi-index quick detection method of honeysuckle | |
| CN102621092B (en) | Method for detecting Danhong injection ethanol precipitation process on line | |
| CN103776777B (en) | Method for identifying ginsengs with different growth patterns by using near infrared spectrum technology and determining content of components in ginsengs | |
| CN104568822A (en) | Method for rapidly and simultaneously detecting multiple indexes of fructus forsythia medicinal material | |
| CN104062260A (en) | Near-infrared on-line detection method used in production process of traditional Chinese medicine containing naringin | |
| CN104792652A (en) | Multi-index rapid detection method for radix astragali | |
| CN103808665A (en) | Method for determining content of multi-index components in purification process of Chinese buckeye seed extractive | |
| CN106198446A (en) | The method of L-Borneol content near infrared spectrum quick test Herba Blumeae Balsamiferae leaf powder | |
| CN108241033A (en) | The method of 6 quality index content of material and application in a kind of quick detection Radix Ophiopogonis alcohol extract | |
| CN104568813A (en) | Multi-index fast detection method of dogwood medicinal materials | |
| CN101661023A (en) | High-efficiency liquid-phase fingerprint analytical method for judging alcoholizing quality of flue-cured tobacco | |
| CN102028710B (en) | Method for measuring contents of indole alkaloids in cinobufagin alcohol precipitation liquid | |
| CN104865322A (en) | Rapid detection method for concentration process of Fructus Gardeniae extract liquor | |
| CN105372204B (en) | A kind of method for online detecting near infrared spectrum of Etimicin Sulfate column separation process | |
| Ye et al. | Rapid determination of chemical components and antioxidant activity of the fruit of Crataegus pinnatifida Bunge by NIRS and chemometrics | |
| CN102147360A (en) | Detection method for drug combination | |
| CN103335960A (en) | Rapid detection method of key indicators in cinobufagin extraction and concentration processes | |
| CN102106919B (en) | Method for controlling quality in extraction process of traditional Chinese medicine chuanhong blood circulation-activating capsules | |
| CN105784951A (en) | Multiple indicator rapid detection method for raw medicinal powder of condensed pill of six drugs with rehmannia | |
| CN103175804A (en) | Method for determining contents of flavonoid constituents in microcos paniculata based on near infrared spectrum technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20141001 |
|
| RJ01 | Rejection of invention patent application after publication |