CN110823966A - Grape wine SO based on electronic nose2Method for rapidly measuring concentration - Google Patents
Grape wine SO based on electronic nose2Method for rapidly measuring concentration Download PDFInfo
- Publication number
- CN110823966A CN110823966A CN201911259322.7A CN201911259322A CN110823966A CN 110823966 A CN110823966 A CN 110823966A CN 201911259322 A CN201911259322 A CN 201911259322A CN 110823966 A CN110823966 A CN 110823966A
- Authority
- CN
- China
- Prior art keywords
- wine
- concentration
- sample
- electronic nose
- rapidly measuring
- 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
- 235000009754 Vitis X bourquina Nutrition 0.000 title claims description 3
- 235000012333 Vitis X labruscana Nutrition 0.000 title claims description 3
- 235000014787 Vitis vinifera Nutrition 0.000 title claims description 3
- 240000006365 Vitis vinifera Species 0.000 title claims description 3
- 238000000034 method Methods 0.000 claims abstract description 34
- 238000013528 artificial neural network Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000000513 principal component analysis Methods 0.000 claims abstract description 4
- 238000012360 testing method Methods 0.000 claims description 19
- 238000012545 processing Methods 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 7
- 230000004044 response Effects 0.000 claims description 7
- 238000004140 cleaning Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000012847 principal component analysis method Methods 0.000 claims description 4
- 230000014509 gene expression Effects 0.000 claims description 3
- 238000005086 pumping Methods 0.000 claims description 3
- 238000011084 recovery Methods 0.000 claims description 2
- RAHZWNYVWXNFOC-UHFFFAOYSA-N sulfur dioxide Inorganic materials O=S=O RAHZWNYVWXNFOC-UHFFFAOYSA-N 0.000 description 17
- 239000000306 component Substances 0.000 description 14
- 241000005398 Figaro Species 0.000 description 8
- 235000013305 food Nutrition 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 5
- 238000001514 detection method Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 239000002778 food additive Substances 0.000 description 2
- 235000013373 food additive Nutrition 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000000844 anti-bacterial effect Effects 0.000 description 1
- 230000002421 anti-septic effect Effects 0.000 description 1
- 230000003078 antioxidant effect Effects 0.000 description 1
- 238000013124 brewing process Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000009614 chemical analysis method Methods 0.000 description 1
- 230000002301 combined effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000004821 distillation Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000796 flavoring agent Substances 0.000 description 1
- 235000019634 flavors Nutrition 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000004128 high performance liquid chromatography Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000004255 ion exchange chromatography Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 229910044991 metal oxide Inorganic materials 0.000 description 1
- 150000004706 metal oxides Chemical class 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003647 oxidation Effects 0.000 description 1
- 238000007254 oxidation reaction Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000002759 z-score normalization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
- G01N27/02—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance
- G01N27/04—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance by investigating resistance
- G01N27/12—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance by investigating resistance of a solid body in dependence upon absorption of a fluid; of a solid body in dependence upon reaction with a fluid, for detecting components in the fluid
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
- G01N27/26—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating electrochemical variables; by using electrolysis or electrophoresis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Electrochemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Analytical Chemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Mathematical Optimization (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Mathematical Physics (AREA)
- Computational Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Algebra (AREA)
- Bioinformatics & Computational Biology (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Investigating Or Analyzing Materials By The Use Of Fluid Adsorption Or Reactions (AREA)
Abstract
The invention discloses a wine SO based on an electronic nose2The method for rapidly measuring the concentration comprises S1, and taking SO with different concentrations2The wine sample of (a); s2, respectively collecting information of each wine sample through an electronic nose; s3, preprocessing all sample information, and extracting a plurality of characteristic parameters of each gas sensor; s4, constructing a regression model of the training data through principal component analysis and a neural network; and S5, obtaining the concentration of the wine sample to be tested through the regression model obtained in the step S4. Therefore, the wine SO based on the electronic nose of the invention2The method for rapidly measuring the concentration provides a thought for treating the problem of low-concentration key components in complex components by using an electronic nose, and can simply and rapidly realize the SO of the wine2And (4) measuring the concentration.
Description
Technical Field
The invention relates to the field of trace gas detection, in particular to an electronic nose-based wine SO2A method for rapidly measuring the concentration.
Background
Sulfur dioxide (SO)2) As a food additive, the product has antiseptic, antibacterial, and antioxidant effects, and especially in wine industry, the wine quality, flavor and maturity of wine and SO2The concentrations of (A) and (B) are closely related. During the brewing process of wine, if SO2If the amount of (3) is too small, the desired effect cannot be obtained. At the same time, excess SO2But also causes non-negligible damage to human health, and GB 2760-2014 established in China clearly stipulates SO in wine in the national food safety standard food additive use Standard2Should not exceed 0.25 g/L.
Thus realizing the SO of the wine2Accurate, rapid, on-line monitoring of the content is an important issue of concern. At present, the SO in wine is aimed at2The content determination methods are mainly divided into two categories. One is an instrumental analysis method, which is typically ion chromatography and high performance liquid chromatography, and these methods have high measurement accuracy and simple operation, but the high price becomes a key factor preventing the popularization of the methods. The other is a chemical analysis method, and the method comprises a distillation method of national food safety standard GB 5009.34-2016 (determination of sulfur dioxide in food safety national standard food), an oxidation method, a direct iodometry method and the like. The cost of the method is low, but the preparation and calibration of the medicine need high time cost, and the direct iodometry has the problem of small application range. Despite the numerous methods of determination, SO in wine production is realized and promoted2Real-time monitoring still presents major difficulties.
In recent years, the electronic nose has become a research hotspot in the field of odor/gas detection by virtue of the advantages of low cost, quick response, simple operation and the like, and has attracted extensive attention of researchers in other fields, including food safety, quality inspection, disease detection and the likeMeasurement, environmental monitoring, and the like. However, due to cross-sensitivity and repeatability problems, gas sensors of electronic nose core components have been used for qualitative analyses such as traceability, anti-counterfeiting, quality grading and odor detection in wine-related applications, and have performed few quantitative analysis tasks, such as the above-mentioned tasks relating to SO in wine2And (4) measuring the content.
Furthermore, the gas composition of wine is very complex, and SO2The content in the wine may be tens to hundreds of ppm (1ppm is one million), and the SO in the wine cannot be realized by only relying on a single gas sensor2And (5) quantification of concentration. Therefore, the SO in the wine is quantitatively analyzed by the electronic nose at the present stage2Concentration is an important consideration in the art.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: provides a kind of wine SO based on electronic nose2A method for rapidly measuring the concentration.
In order to solve the technical problem, firstly, a testing device is set up, and the testing device comprises a clean air source, a first mass flow controller, a second mass flow controller, a first one-way valve, a second one-way valve, a headspace bottle, a testing cavity, a gas sensor array and a control analysis module; the headspace bottle is stored with a wine sample to be tested and is provided with an air inlet pipe and an air outlet pipe, one end of the air inlet pipe in the headspace bottle is arranged below the liquid level of the wine sample, one end of the air outlet pipe in the headspace bottle is arranged above the liquid level of the wine to be tested, wherein,
the clean air source is respectively connected with the air inlets of a first mass flow controller and a second mass flow controller through air pipes, the air outlet of the first mass flow controller is connected with the air inlet of a first one-way valve through an air pipe, the air outlet of the second mass flow controller is connected with the air inlet pipe of a headspace bottle, the air outlet pipe of the headspace bottle is connected with the air inlet of a second one-way valve, the air outlet of the second one-way valve is connected with the air inlet of the first one-way valve, the air outlet of the first one-way valve is connected with the air inlet of the test chamber through an air pipe, and the air outlet of the test chamber discharges tail gas through an air pipe;
a gas sensor array is arranged in the test cavity and connected with the control analysis module;
and the control analysis module is used for processing and analyzing the data acquired by the gas sensor array.
The invention is based on the testing device and the wine SO based on the electronic nose2The method for rapidly determining the concentration comprises the following steps:
s1, taking SO with different concentrations2The wine sample of (a);
s2, respectively collecting information of each wine sample through an electronic nose;
s3, preprocessing all sample information, and extracting a plurality of characteristic parameters of each gas sensor;
s4, constructing a regression model of the training data through principal component analysis and a neural network;
and S5, obtaining the concentration of the wine sample to be tested through the regression model obtained in the step S4.
Compared with the prior art, the invention has the following technical effects that the wine SO can be simply and quickly realized2The concentration measurement provides a concept for treating the problem of low-concentration key components in complex components by using an electronic nose.
On the basis of the technical scheme, the invention can be further improved as follows.
Preferably, the volume of the wine sample is 10% to 40% of the headspace bottle volume and all samples remain the same amount.
Preferably, the specific steps of step S2 are as follows:
s21, cleaning the test cavity by clean air until the response of the sensor is stable or returns to a baseline;
s22, pumping clean air into the headspace bottle according to a set flow, and collecting information of a wine sample by using an electronic nose;
s23, cleaning the test cavity by using clean air with set flow;
s24, replacing the samples, and repeating the steps S21-S23 until all the samples are tested.
Preferably, the preprocessing of all the sample information in step S3 is to perform 5-point average filtering processing on all the sample information to reduce noise interference, and the specific method is to perform the processing by using the following formula:
where s is the gas sensor signal before filtering and s is the gas sensor information after filtering.
Preferably, in step S3, a plurality of characteristic parameters are extracted, including a steady-state maximum value, a steady-state differential value, a first-order differential maximum value, a first-order integral maximum value, a steady-state response time, and a recovery time, so as to select an optimal or approximately optimal combination.
Preferably, the specific steps of step S4 are as follows:
s41, carrying out standardization processing on all the characteristic parameters acquired in the step S3;
s42, compressing the sample space by using a principal component analysis method, and extracting principal components;
s43, using the neural network to make the first n principal component pairs SO with the sum of the contribution rates larger than the threshold value gamma2Regression modeling of concentration;
s44, verifying the effect of the model until the accuracy meets the index requirement;
and S45, outputting model parameters.
Preferably, in step S41, if the unit and magnitude of the characteristic parameter are different, the characteristic parameter is normalized by using the z-score method, and the calculation formula is as follows:
x′＝(x-μ)/σ (2)
in the above formula, x and x are the raw data and normalized data of the characteristic parameter, respectively, and μ and σ are the mean and standard deviation of the raw data of the characteristic parameter, respectively.
All the normalized characteristic parameters form a new sample set X
Preferably, in step S42, the principal component analysis method includes the following specific steps:
s42-1, calculating the covariance matrix C of the normalized sample set X in the following way:
in the formula, N is the sample size,
s42-2, solving the eigenvalue and the corresponding eigenvector of the covariance matrix C, extracting the eigenvectors corresponding to the first k eigenvalues according to the sequence from big to small and arranging the eigenvectors into a matrix P, and calculating to obtain a new data set Z, wherein the calculation mode is as follows:
Z＝PX (4)
preferably, the neural network mentioned in step S43 is an operation model, and generally comprises an input layer, an implicit layer and an output layer, wherein the layers are connected with each other through neurons, each neuron comprises a linear combination and a nonlinear activation function, and their expressions can be expressed as:
c＝ωTa+b (5)
d＝g(c) (6)
specifically, in the above equation, a is the input, d is the output, and is also the input into the next layer of neurons, ω is the weight, and b is the intercept. For example, the principal component extracted in step S42 is input to the input layer, and the final output layer outputs SO by calculation of the hidden layer2The concentration of (c). Different activation functions have different operation modes, and the ReLU function is used in the patent and is calculated as follows:
f(c)＝max(0，c) (7)
preferably, in step S44, the determination coefficient R is used2Mean square error MSE and mean absolute error MAE are used as the judgment standard of the regression effect, and the expressions are respectively as follows:
in the above formula yiAnd
respectively the true value and the predicted value of the ith sample,
is the arithmetic mean of all samples.
Compared with the prior art, the invention has the following technical effects:
wine SO based on electronic nose adopted by the invention2The method for quickly measuring the concentration can simply and quickly realize the SO of the wine2The concentration measurement provides a thought for treating the problem of low-concentration key components in complex components by using an electronic nose, and can be expanded to the application of other food safety detection aspects by combining the specific conditions of other foods.
Drawings
FIG. 1 is a flow chart of the method of the present invention
FIG. 2 is a typical response curve of a gas sensor in an example
FIG. 3 is a comparison of the average signal (polar coordinates) of the steady state maximum of the 16 sensors in the example
FIG. 4 is an average signal comparison (polar coordinates) of steady state differential values for 16 sensors in an example
FIG. 5 is the average signal comparison (polar coordinates) of the first order differential maxima of the 16 sensors in the example
FIG. 6 is a comparison of the total contribution of the first 20 principal components of four different combinations of feature parameters in an example
FIG. 7 is a schematic diagram of a feedforward neural network (using a single hidden layer as an example)
FIG. 8 is a prediction result of an optimal feature parameter combination in an example
FIG. 9 is a schematic structural diagram of a testing apparatus according to an embodiment of the present invention
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
Please refer to fig. 1, which shows an electronic nose-based wine SO of the present invention2Schematic structure of the method for rapid determination of concentration. The wine SO based on the electronic nose2The method for rapidly determining the concentration comprises the following steps:
step one, adding SO with different concentrations2Sampling the to-be-detected wine and injecting the to-be-detected wine into a headspace bottle;
in the example of pink wine, SO is added2The concentrations were 0, 40, 80, 120, 160, 200ppm, the sample size was 20ml, and the headspace volume was 50 ml.
Collecting the information of the wine sample through a gas sensor array of the electronic nose;
the electronic nose contains 16 gas sensors, including 15 metal oxide semiconductor (MOX) gas sensors and 1 electrochemical gas sensor, and the detailed information of these sensors is shown in table 1.
TABLE 1 gas sensor details
| Numbering | Model number | Number of | Manufacturer(s) | |
| S1、 | TGS2602 | 2 | Figaro | |
| S3、 | TGS2600 | 2 | Figaro | |
| S5 | 4SO2-20 | 1 | Honeywell | |
| S6 | TGS2603 | 1 | Figaro | |
| S7、 | TGS2610 | 2 | Figaro | |
| S9、 | TGS2611 | 2 | Figaro | |
| S11 | TGS2612 | 1 | Figaro | |
| S12 | TGS2620 | 1 | Figaro | |
| S13、 | TGS2630 | 2 | Figaro | |
| S15、 | WSP7110 | 2 | Wei hold |
In addition, all the MOX gas sensors are heated for 1 week through 5V voltage before testing, and the testing is carried out in a room-temperature and dry laboratory environment, and an upper computer program written by LabVIEW is used for controlling a gas circuit and a circuit. Before testing a wine sample, air is used for cleaning an air cavity where a gas sensor is located to enable the sensor to be recovered to a standard state, the whole cleaning process is divided into two stages, and the time and the flow are respectively 20min, 300sccm, 5min and 100 sccm; when the sample is tested, the sample is collected by pumping air into the headspace bottle, and the time and the flow are 3min and 100 sccm. The above procedure was repeated until all samples were tested.
Description of the data: there were 6 different SOs in the examples2Wine of different concentrations, 8 samples were prepared for each wine of different concentrations, and each sample was tested 3 times, and a total of 6 × 8 × 3 to 144 pieces of data were obtained. A typical response curve for a gas sensor is given in fig. 2.
Preprocessing all sample information, and extracting a plurality of characteristic parameters of each gas sensor;
all data are subjected to 5-point average filtering processing according to formula (1), and then the steady-state maximum value V of each sensor is extractedmaxSteady state differential value VdifAnd first order differential maximum DmaxThe average signals of the three characteristic parameters extracted by all the gas sensors after normalization are respectively shown in fig. 3, 4 and 5, specifically, in fig. 3, S10 denotes the 10 th sensor, whose measured SO2The corresponding value of the wine with the content of 0ppm is about 0.4, and SO2The corresponding value of the wine with the content of 40ppm is about 1.5, and SO2The corresponding value of the wine with the content of 80ppm is about-0.7, and SO2The corresponding value of the wine with the content of 120ppm is about-0.8, and SO2The corresponding value of the wine with the content of 160ppm is about-1.2, and SO2The corresponding value of the wine with the content of 200ppm is about-1.4. From the value of the characteristic, SO in wine2There is no clear relationship between the concentration of (A) and the magnitude of the characteristic, since SO is present in wine2The content of (a) is low and the gas sensor is cross-sensitive, and the difficulty of solving the problem is also achieved.
And step four, constructing a regression model of the training data through principal component analysis and a neural network, and evaluating the regression model by using the test data.
Of the three characteristic parameters extracted in the example, the first order differential maximum value DmaxSince the units and magnitudes of the two other features are different, the mean value of all the features in the new data after normalization is 0 and the variance is 1 by using the z-score normalization process.
Next, the influence of different feature parameter combinations on the regression result is tested, and the three feature parameters can be divided into four cases: full characteristic combination (V)max+Vdif+Dmax) And pairwise combinations of characteristic parameters (V)max+Vdif、Vmax+DmaxAnd Vdif+Dmax). The principal components of the four feature combinations are extracted separately by means of step 42 in the summary of the invention, and the total contribution rates of the first 20 principal components of the four feature combinations are shown in fig. 6.
The setting of the threshold γ suggests that it is greater than 85%, with a larger γ indicating that more information will be contained in the principal component. In the present example, the threshold γ was set to 97.5%, and therefore, according to the result in fig. 6, Vmax+Vdif+Dmax、Vmax+Vdif、Vmax+DmaxAnd Vdif+DmaxThe first 14, 12, 13 and 10 principal components are selected in turn for four cases and usedThe neural network respectively establishes the main components and the SO2The mapping relationship of the concentration.
In this example, a feedforward neural network, also called a multi-layer perceptron, is used, and a schematic diagram of the network structure is shown in fig. 7. The parameters of the neural network in this example are set as follows: the number of hidden layers is 2, the number of neurons of the hidden layers is 2 times of the input dimension, the maximum iteration number is 500, the weight updating mode is an L-BFGS algorithm, the parameter of a regularization term is 0.00001, the selected principal component is used as input data of a neural network, and SO in a wine sample2As a label (output layer output).
In the regression that follows in this example, because 48 sets of samples were used each time with 47 of the samples as the training set and the remaining 1 as the test set, 48 total training times ensured that measurements from the same sample did not occur in both the training and test sets, or that more severe overfitting could occur.
Table 2 summarizes the predicted results of four different characteristic parameter combinations, R in three indexes2The larger the MSE and MAE, the better the regression model. In Table 2, the combined effect of the steady state maximum and the steady state differential values is at MSE, MAE and R2The aspects are all optimal, so the example finally selects two characteristic parameters of the steady-state maximum value and the steady-state differential value.
TABLE 2 comparison of results for different combinations of characteristic parameters
| Characteristic parameter combination | MSE | MAE | R2 |
| Vmax+Vdif+Dmax | 191.17 | 8.61 | 0.959 |
| Vmax+Vdif | 69.33 | 5.60 | 0.985 |
| Vmax+Dmax | 304.35 | 11.94 | 0.934 |
| Vdif+Dmax | 369.55 | 12.53 | 0.920 |
The detailed test results are shown in FIG. 8, which more accurately measures the SO in the wine2The concentration of (c).
The above examples are only preferred embodiments of the present invention, and one of ordinary skill in the art would understand that: many changes, modifications, substitutions and alterations to these embodiments are within the scope of the present invention without departing from the principles and spirit of the invention.
Claims (9)
1. Grape wine SO based on electronic nose2A method for rapidly measuring a concentration, comprising:
s1, taking SO with different concentrations2The wine sample of (a);
s2, respectively collecting information of each wine sample through an electronic nose;
s3, preprocessing all sample information, and extracting a plurality of characteristic parameters of each gas sensor;
s4, constructing a regression model of the training data through principal component analysis and a neural network;
and S5, obtaining the concentration of the wine sample to be tested through the regression model obtained in the step S4.
2. Electronic nose based wine SO according to claim 12Method for the rapid determination of the concentration, characterized in that the volume of the wine sample is between 10% and 40% of the volume of the headspace bottle, and all samples remain in the same amount.
3. Electronic nose based wine SO according to claim 1 or 22The method for rapidly measuring the concentration is characterized in that the specific steps of the step S2 are as follows:
s21, cleaning the test cavity by clean air until the response of the sensor is stable or returns to a baseline;
s22, pumping clean air into the headspace bottle according to a set flow, and collecting information of a wine sample by using an electronic nose;
s23, cleaning the test cavity by using clean air with set flow;
s24, replacing the samples, and repeating the steps S21-S23 until all the samples are tested.
4. Electronic nose based wine SO according to claim 1 or 22The method for rapidly determining the concentration is characterized in that the preprocessing of all sample information in step S3 means that 5-point average filtering processing is performed on all sample information to reduce noise interference, and the specific method is to use the following formula for processing:
where s is the gas sensor signal before filtering and s' is the gas sensor signal after filtering.
5. Electronic nose based wine SO according to claim 1 or 22The method for quickly measuring the concentration is characterized in that a plurality of characteristic parameters including a steady state maximum value, a steady state differential value, a first order differential maximum value, a first order integral maximum value, a steady state response time, and a recovery time are extracted in step S3.
6. Electronic nose based wine SO according to claim 1 or 22The method for rapidly measuring the concentration is characterized in that the specific steps of the step S4 are as follows:
s41, carrying out standardization processing on all the characteristic parameters acquired in the step S3;
s42, compressing the sample space by using a principal component analysis method, and extracting principal components;
s43, using the neural network to make the first n principal component pairs SO with the sum of the contribution rates larger than the threshold value gamma2Regression modeling of concentration;
s44, verifying the effect of the model until the accuracy meets the index requirement;
and S45, outputting model parameters.
7. Electronic nose based wine SO according to claim 1 or 22The method for rapidly measuring concentration is characterized in that in step S41, if the unit and magnitude of the characteristic parameter are different, the characteristic parameter is normalized by using z-score method, and the calculation formula is as follows:
x′＝(x-μ)/σ (2)
in the above formula, X and X are the original data of the feature parameters and the normalized data, μ and σ are the mean and standard deviation of the original data of the feature parameters, respectively, and all the normalized feature parameters form a new sample set X.
8. Electronic nose based wine SO according to claim 1 or 22A method for rapidly measuring concentration, characterized in that, in step S42,the principal component analysis method comprises the following specific steps:
s42-1, calculating the covariance matrix C of the normalized sample set X in the following way:
in the formula, N is the sample size;
s42-2, solving the eigenvalue and the corresponding eigenvector of the covariance matrix C, extracting the eigenvectors corresponding to the first k eigenvalues according to the sequence from big to small and arranging the eigenvectors into a matrix P, and calculating to obtain a new data set Z, wherein the calculation mode is as follows:
Z＝PX (4)。
9. electronic nose based wine SO according to claim 1 or 22The method for rapidly measuring the concentration is characterized in that in step S44, a determination coefficient R is used2Mean square error MSE and mean absolute error MAE are used as the judgment standard of the regression effect, and the expressions are respectively as follows:
in the above formula yiAnd
respectively the true value and the predicted value of the ith sample,
is the arithmetic mean of all samples.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911259322.7A CN110823966A (en) | 2019-12-10 | 2019-12-10 | Grape wine SO based on electronic nose2Method for rapidly measuring concentration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911259322.7A CN110823966A (en) | 2019-12-10 | 2019-12-10 | Grape wine SO based on electronic nose2Method for rapidly measuring concentration |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110823966A true CN110823966A (en) | 2020-02-21 |
Family
ID=69544414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911259322.7A Pending CN110823966A (en) | 2019-12-10 | 2019-12-10 | Grape wine SO based on electronic nose2Method for rapidly measuring concentration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110823966A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113238004A (en) * | 2021-05-10 | 2021-08-10 | 云南中烟工业有限责任公司 | Method for predicting sour taste and sweet taste by using MLP neural network model |
| CN113552289A (en) * | 2021-07-14 | 2021-10-26 | 清华苏州环境创新研究院 | Atmospheric pollution tracing method based on Gaussian model |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101135689A (en) * | 2007-09-21 | 2008-03-05 | 华中科技大学 | Electric nose development platform |
| CN102866179A (en) * | 2012-09-13 | 2013-01-09 | 重庆大学 | Online recognition and inhibition method based on non-target interference smell in electronic nose of artificial intelligent learning machine |
| US20130236976A1 (en) * | 2010-11-30 | 2013-09-12 | Foss Analytical A/S | Determination of sulphur dioxide in a liquid |
| US20150000371A1 (en) * | 2013-05-07 | 2015-01-01 | Russell W. Greene | Beverage diagnostic and preservation devices and methods |
-
2019
- 2019-12-10 CN CN201911259322.7A patent/CN110823966A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101135689A (en) * | 2007-09-21 | 2008-03-05 | 华中科技大学 | Electric nose development platform |
| US20130236976A1 (en) * | 2010-11-30 | 2013-09-12 | Foss Analytical A/S | Determination of sulphur dioxide in a liquid |
| CN102866179A (en) * | 2012-09-13 | 2013-01-09 | 重庆大学 | Online recognition and inhibition method based on non-target interference smell in electronic nose of artificial intelligent learning machine |
| US20150000371A1 (en) * | 2013-05-07 | 2015-01-01 | Russell W. Greene | Beverage diagnostic and preservation devices and methods |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113238004A (en) * | 2021-05-10 | 2021-08-10 | 云南中烟工业有限责任公司 | Method for predicting sour taste and sweet taste by using MLP neural network model |
| CN113552289A (en) * | 2021-07-14 | 2021-10-26 | 清华苏州环境创新研究院 | Atmospheric pollution tracing method based on Gaussian model |
| CN113552289B (en) * | 2021-07-14 | 2024-01-23 | 清华苏州环境创新研究院 | Atmospheric pollution tracing method based on Gaussian mode |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Jia et al. | Electronic noses as a powerful tool for assessing meat quality: A mini review | |
| Hui et al. | Winter jujube (Zizyphus jujuba Mill.) quality forecasting method based on electronic nose | |
| Li et al. | Non-destructive evaluation of pork freshness using a portable electronic nose (E-nose) based on a colorimetric sensor array | |
| Huang et al. | Improved generalization of spectral models associated with Vis-NIR spectroscopy for determining the moisture content of different tea leaves | |
| CN101825567A (en) | Screening method for near infrared spectrum wavelength and Raman spectrum wavelength | |
| CN110823966A (en) | Grape wine SO based on electronic nose2Method for rapidly measuring concentration | |
| CN103712939A (en) | Pollutant concentration fitting method based on ultraviolet-visible spectrum | |
| CN111537845A (en) | Method for identifying aging state of oil paper insulation equipment based on Raman spectrum cluster analysis | |
| CN109060771B (en) | Consensus model construction method based on different characteristic sets of spectrum | |
| WO2020248961A1 (en) | Method for selecting spectral wavenumber without reference value | |
| CN109374530A (en) | The optoacoustic spectroscopy on-line monitoring method of sulfur hexafluoride gas decomposition product | |
| Zhang et al. | Standardization of metal oxide sensor array using artificial neural networks through experimental design | |
| Li et al. | Local dynamic neural network for quantitative analysis of mixed gases | |
| Li et al. | Quantitative analysis of aflatoxin B1 of peanut by optimized support vector machine models based on near-infrared spectral features | |
| Li et al. | Qualitative and quantitative analysis of volatile metabolites of foodborne pathogens using colorimetric-bionic sensor coupled robust models | |
| Wang et al. | SVM classification method of waxy corn seeds with different vitality levels based on hyperspectral imaging | |
| CN113295673B (en) | Laser-induced breakdown spectroscopy weak supervision feature extraction method | |
| CN113804657B (en) | Sparse self-encoder spectral feature dimension reduction method based on multiple regression combination | |
| Ye et al. | Water chemical oxygen demand prediction model based on the CNN and ultraviolet-visible spectroscopy | |
| CN211292687U (en) | Wine SO2Electronic nose device for concentration determination | |
| CN116399836A (en) | Cross-talk fluorescence spectrum decomposition method based on alternating gradient descent algorithm | |
| Zhu et al. | A UV–Vis spectroscopic detection method for cobalt ions in zinc sulfate solution based on discrete wavelet transform and extreme gradient boosting | |
| Zhu et al. | Rapid freshness prediction of crab based on a portable electronic nose system | |
| CN113836813A (en) | Blast furnace tuyere water leakage detection method based on data analysis | |
| CN114088874A (en) | Edible vinegar total acid content detection method based on liquid phase visualization array |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |