CN112098605A - High-robustness chemical sensor array soft measurement method - Google Patents
High-robustness chemical sensor array soft measurement method Download PDFInfo
- Publication number
- CN112098605A CN112098605A CN202010997648.6A CN202010997648A CN112098605A CN 112098605 A CN112098605 A CN 112098605A CN 202010997648 A CN202010997648 A CN 202010997648A CN 112098605 A CN112098605 A CN 112098605A
- Authority
- CN
- China
- Prior art keywords
- sensor array
- sub
- matrix
- chemical sensor
- chemical
- 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
- 239000000126 substance Substances 0.000 title claims abstract description 68
- 238000000691 measurement method Methods 0.000 title claims abstract description 8
- 238000005259 measurement Methods 0.000 claims abstract description 43
- 238000012549 training Methods 0.000 claims abstract description 19
- 238000003491 array Methods 0.000 claims abstract description 13
- 238000000605 extraction Methods 0.000 claims abstract description 12
- 238000012847 principal component analysis method Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 44
- 238000000034 method Methods 0.000 claims description 26
- 239000013598 vector Substances 0.000 claims description 16
- 230000004044 response Effects 0.000 claims description 11
- 238000000354 decomposition reaction Methods 0.000 claims description 8
- 230000001186 cumulative effect Effects 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 4
- 235000003181 Panax pseudoginseng Nutrition 0.000 claims description 2
- 244000131316 Panax pseudoginseng Species 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 5
- 239000007789 gas Substances 0.000 description 46
- 238000004458 analytical method Methods 0.000 description 11
- 238000003909 pattern recognition Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 5
- 238000007781 pre-processing Methods 0.000 description 4
- 238000000513 principal component analysis Methods 0.000 description 4
- 230000007774 longterm Effects 0.000 description 3
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 description 2
- 229910002091 carbon monoxide Inorganic materials 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 235000019645 odor Nutrition 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000005398 Figaro Species 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000000712 assembly Effects 0.000 description 1
- 238000000429 assembly Methods 0.000 description 1
- 238000007630 basic procedure Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 229920001940 conductive polymer Polymers 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000012843 least square support vector machine Methods 0.000 description 1
- 229910044991 metal oxide Inorganic materials 0.000 description 1
- 150000004706 metal oxides Chemical class 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000012567 pattern recognition method Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000003380 quartz crystal microbalance Methods 0.000 description 1
- 238000010897 surface acoustic wave method Methods 0.000 description 1
- 230000002277 temperature effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0062—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method or the display, e.g. intermittent measurement or digital display
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0062—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method or the display, e.g. intermittent measurement or digital display
- G01N33/0068—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method or the display, e.g. intermittent measurement or digital display using a computer specifically programmed
Landscapes
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medicinal Chemistry (AREA)
- Food Science & Technology (AREA)
- Combustion & Propulsion (AREA)
- Physics & Mathematics (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
Abstract
A high-robustness chemical sensor array soft measurement method relates to the technical field of gas detection equipment, aims at the problem that a traditional chemical sensor array soft measurement model in the prior art is poor in accuracy, and comprises the following steps: dividing a chemical sensor array into g sub-sensor arrays, wherein the number of various sensors in the chemical sensor array is consistent; step two: collecting output signals of each sub-sensor array to form a training sample set; step three: performing feature extraction on the training sample of each sub-sensor array by using a principal component analysis method to obtain a feature set of each sub-sensor array; step four: training a regression model of the sub-sensor array by using the feature set; step five: and measuring the measurement sample by using the trained regression model of each sub-sensor array, and selecting the median of the measurement result of the regression model of each sub-sensor array as the final measurement result.
Description
Technical Field
The invention relates to the technical field of gas detection equipment, in particular to a high-robustness chemical sensor array soft measurement method.
Background
The electronic nose system consists of a set of chemical sensors with partial specificity and an appropriate pattern recognition method, and can recognize and analyze simple or complex odors. The electronic nose system collects data of gas/smell through the chemical sensor array to form a training sample set and constructs a mode recognition model by utilizing the training sample set. The test gas/odor samples can be identified and analyzed according to the model. At present, the electronic nose system is widely applied to the fields of public safety, medical diagnosis, flammable and combustible detection, environmental pollution monitoring, food safety and the like.
Most of the research of the related scholars at present focuses on research on signal processing methods and pattern recognition techniques. Meanwhile, related scholars also apply different types of sensors to electronic nose systems, such as sensors of conductive polymers, metal oxides, surface acoustic waves, quartz crystal microbalances and the like. With the continuous expansion of the application field of the electronic nose system, the reliability of the system and the accuracy of an analysis result are of great importance under the long-term working condition. As the chemical sensor realizes the response to the target gas through complex physical and chemical reactions, the measurement quality of the chemical sensor array is reduced along with the aging of the gas sensitive element and the degradation of a circuit in the long-term use process. The fault of the chemical sensor can cause the response signal of the sensor array to generate mutation once, so that the accuracy of the gas identification and analysis result of the established pattern identification model is reduced.
The traditional soft measurement model utilizes the output signals of a group of chemical sensor arrays to perform baseline removal processing on the signals of the chemical sensor arrays through a signal preprocessing module, so that the influence of sensor drift is reduced. The pattern recognition model describes the concentration information of the target gas by using a feature extraction method, and then estimates the concentration of the mixed gas by using an appropriate regression model. If the chemical sensor array fails, the subsequent pattern recognition model can generate wrong recognition and analysis results on the array signals with failure information. Therefore, the reliability of the traditional chemical sensor array soft measurement model is poor.
Disclosure of Invention
The purpose of the invention is: aiming at the problem that the traditional chemical sensor array soft measurement model in the prior art is poor in accuracy, a high-robustness chemical sensor array soft measurement method is provided.
The technical scheme adopted by the invention to solve the technical problems is as follows:
a high robustness chemical sensor array soft measurement method comprises the following steps:
the method comprises the following steps: dividing a chemical sensor array into g sub-sensor arrays, wherein the number of various sensors in the chemical sensor array is consistent;
step two: collecting output signals of each sub-sensor array to form a training sample set;
step three: performing feature extraction on the training sample of each sub-sensor array by using a principal component analysis method to obtain a feature set of each sub-sensor array;
step four: training a regression model of the sub-sensor array by using the feature set;
step five: and measuring the measurement sample by using the trained regression model of each sub-sensor array, and selecting the median of the measurement result of the regression model of each sub-sensor array as the final measurement result.
Further, the specific steps of performing feature extraction by using a principal component analysis method in the third step are as follows:
suppose that
A matrix of measurements representing an array of sub-sensors, m representing the number of sensors and n representing the number of measurements, the steps of:
step three, firstly: performing drift compensation on an output signal of each sensor in the sensor array;
step three: standardizing the sub-sensor array after drift compensation to obtain a matrix X*；
Step three: calculating the matrix X*The covariance matrix Σ of;
step three and four: and performing singular value decomposition according to the obtained covariance matrix:
step three and five: obtaining the number k of the principal components by utilizing the cumulative variance percentage according to the covariance matrix sigma after singular value decomposition;
step three and six: determining a load matrix according to the determined number k of the principal components
The method comprises the steps that eigenvectors corresponding to the largest k eigenvalues in a covariance matrix sigma are formed;
step three, pseudo-ginseng: using load matrices
Obtaining a mapping matrix C:
step three eight: and (3) utilizing the mapping matrix C to extract features:
wherein the content of the first and second substances,
representing the results of the feature extraction on the chemical sensor array measurements.
Furthermore, the third step and the first step carry out drift compensation through a fractional difference method.
Further, the drift compensation performed by the fractional difference method is represented as:
xi＝(Vi max-Vi min)/Vi min
wherein, Vi maxMaximum output voltage, V, representing the sensor response signali minRepresenting the minimum output voltage of the sensor response signal.
Further, the step three-two matrixX*Expressed as:
wherein, XiRepresents Xn×pColumn vector of (E) (X)i) Represents the mean of the column vectors, D (X)i) Represents the variance of the column vector, i 1, 2.. p, p being the number of sensors in the sub-sensor array.
Further, in the third step, the covariance matrix Σ is represented as:
further, in the third and fourth steps, singular value decomposition is performed according to the obtained covariance matrix Σ and expressed as:
∑＝VΛVT
where Λ represents a diagonal matrix whose diagonal positions contain eigenvalues λ of the covariance matrix ΣiAnd arranging λ in descending order1≥λ2≥...≥λmAnd V is more than or equal to 0 and consists of a feature vector of sigma.
Further, the number k of the principal components obtained by using the cumulative variance percentage in the third step and the fifth step is expressed as:
when the CPV value is greater than 90, the value of k is the number of main components, where trace represents the number of traces of the matrix.
Further, when the CPV value is more than 90% and less than 95%, the value of k is the number of main components.
Further, the load matrix
Expressed as:
wherein u iskAnd expressing the selected feature vectors of sigma corresponding to the k principal components.
The invention has the beneficial effects that:
the invention provides a high-robustness chemical sensor array soft measurement model which has good fault tolerance of a chemical sensor and improves the accuracy and stability of an electronic nose system under a long-term working condition. The invention provides a high-robustness chemical sensor array soft measurement model which can keep the accuracy of an electronic nose system on gas identification and analysis results under the condition that a fault gas sensor exists in a chemical sensor array.
Drawings
FIG. 1 is a block diagram of a conventional chemical sensor array soft measurement model;
FIG. 2 is a block diagram of a soft measurement model of the high robustness chemical sensor array according to the present invention;
fig. 3 is a block diagram of an electronic nose system based on a MOS gas sensor array.
Detailed Description
The first embodiment is as follows: referring to the specific description of the present embodiment, the method for soft measurement of a highly robust chemical sensor array according to the present embodiment includes the following steps:
the method comprises the following steps: dividing a chemical sensor array into g sub-sensor arrays, wherein the number of various sensors in the chemical sensor array is consistent;
step two: collecting output signals of each sub-sensor array to form a training sample set;
step three: performing feature extraction on the training sample of each sub-sensor array by using a principal component analysis method to obtain a feature set of each sub-sensor array;
step four: training a regression model of the sub-sensor array by using the feature set;
step five: and measuring the measurement sample by using the trained regression model of each sub-sensor array, and selecting the median of the measurement result of the regression model of each sub-sensor array as the final measurement result.
In the application, the application is directed to a sensor array with the same number of types of sensors in the sensor array, and when the sensor array is divided into a plurality of sub-sensor arrays, for example, the sensor array has 12, 4 types, and 3 sensors in each type, the sensor array can be divided into 3 sub-sensor arrays, and each sub-sensor array includes one sensor in each of the 4 types.
Soft measurement model of traditional chemical sensor array
Fig. 1 shows a block diagram of a soft measurement model of a conventional chemical sensor array. Therefore, the traditional soft measurement model utilizes the output signals of a group of chemical sensor arrays to perform baseline removal processing on the signals of the chemical sensor arrays through a signal preprocessing module, so that the influence of sensor drift is reduced. The pattern recognition model describes the concentration information of the target gas by using a feature extraction method, and then estimates the concentration of the mixed gas by using an appropriate regression model. If the chemical sensor array fails, the subsequent pattern recognition model can generate wrong recognition and analysis results on the array signals with failure information. Therefore, the reliability of the traditional chemical sensor array soft measurement model is poor.
High-robustness chemical sensor array soft measurement model
The patent provides a high robustness chemical sensor array soft measurement model, and the model can keep the accuracy of an electronic nose system to gas identification and analysis results under the condition that a fault gas sensor exists in a chemical sensor array. Fig. 2 shows a block diagram of a proposed soft measurement model of a high-robustness chemical sensor array.
Chemical sensor arrays typically employ different types of chemical sensor compositions and each type of sensor has hardware redundancy. According to the above properties of the chemical sensor array, after the chemical sensor array signals are collected and signal pre-processed, a virtual chemical sensor array will be composed according to the kind of chemical sensors. The virtual chemical sensor array is divided into a plurality of sub-sensor arrays, and each sub-sensor array is composed of different chemical sensors in order to ensure that each sub-sensor array can have better broad-spectrum response characteristics. And acquiring enough training sample sets through each sub-sensor array, and constructing respective pattern recognition models. And (4) evaluating the estimated values of the different mode recognition models for the target gas, and selecting the median value as the result output of the soft measurement model. When a faulty sensor is present in the chemical sensor array, the analysis results of the sub-sensor array containing the faulty sensor may deviate from the analysis results of other pattern recognition models. In order to improve the robustness of the model, the median of each pattern recognition model is adopted as the analysis result of the chemical sensor array to be output.
Basic procedure
The basic steps of the high robustness chemical sensor array soft measurement model are as follows.
(1) Because the chemical sensor has drift caused by temperature effect, in order to reduce the influence of the drift, the sensor drift is compensated by adopting a split differential method. Signal preprocessing is performed on each chemical sensor by using a fractional difference method as shown in formula (1).
xi＝(Vi max-Vi min)/Vi min (1)
Wherein x isiRepresents the processed result, V, obtained after signal preprocessing of the chemical sensori maxMaximum output voltage, V, representing the sensor response signali minMinimum output voltage, V, representing sensor response signaliRepresenting a time-of-day value of the sensor response signal.
(2) Assuming that the chemical sensor array is composed of n chemical sensors and g different chemical sensors are used in total, the chemical sensor array may be divided into m sub-sensor arrays according to the proposed method to form a virtual chemical sensor array, where n is m · g.
(3) And carrying out sample collection experiments by using the measured gas with different concentrations according to an experiment plan, carrying out independent experiments aiming at different concentrations, and obtaining a training sample set corresponding to each sub-sensor array.
(4) Feature extraction is achieved using Principal Component Analysis (PCA). For the training samples of each sub-sensor array, assuming that the sub-sensor array is composed of p sensors, each training sample of the sub-sensor array is composed of n pre-processed measurement values to form Xn×p. To Xn×pCarrying out standardization processing to obtain a standardized matrix X*，
Wherein, XiRepresents Xn×pColumn vector of (E) (X)i) Represents the mean of the column vectors, D (X)i) Representing the variance of the column vector.
Decomposition of X into X by PCA
Wherein the content of the first and second substances,
referred to as the Principal Component Subspace (PCS),
referred to as Residual Subspace (RS);
in order to be a scoring matrix, the scoring matrix,
and k is the number of principal components of the PCA model.
Load matrix
Can be obtained by the following steps:
calculating X*Covariance ofThe matrix sigma is such that,
the covariance matrix sigma is subjected to a Singular Value Decomposition (SVD),
∑＝VΛVT (5)
wherein Λ is obtained by arranging the covariance matrix sigma eigenvalues in descending order (λ)1≥λ1≥λ2≥...≥λmNot less than 0), and V is composed of eigenvectors corresponding to all eigenvalues of the covariance matrix sigma.
Selecting the optimal number of the main components by adopting cumulative variance percentage (CPV),
forming a load matrix by using the first k column vectors of U
The loading matrix contains the gas information in the original sample and reduces the dimensionality of the sample.
And obtaining a feature set of each sub-sensor array through feature extraction, and training a regression model of the sub-sensor array by using the feature set. The invention adopts multivariable correlation vector machine (MVRVM), correlation vector machine (RVM) and least square support vector machine regression (LS-SVR) as a mixed gas concentration analysis model, and the basic principle of the MVRVM regression model can be referred to documents [1] Thayananthan A, Navararm R, Stenger B, et al.
And analyzing the measurement sample by using the trained regression model of each sub-sensor array, so as to obtain concentration values of different gases in the sample. In order to reduce the performance of the fault sensor on the electronic nose system, the median of all the sub-sensor array measurement values is selected as the output of the system.
Examples
Fig. 3 is a block diagram of an electronic nose system based on an MOS gas sensor array, which is mainly composed of an MOS gas sensor array, a data acquisition card, a power supply, a mass flow controller, and a PC. The MOS gas sensor array consisted of 16 commercially available MOS gas sensors of TGS series manufactured by FIGARO corporation of japan, and the specific model and the number of assemblies are shown in table 1.
The measured information can be converted into an analog signal by a corresponding measuring circuit. The data acquisition card is connected with a PC (personal computer) and adopts USB-6251 produced by NI company to acquire 16 paths of analog signals output by the MOS gas sensor array. The MOS gas sensor array is placed in a sealed organic glass gas chamber with the volume of 1L. The gas-sensitive material of the MOS gas sensing array is heated by adopting a +5V direct current power supply, so that the gas-sensitive sensor can fully react with the gas/smell to be detected. Methane (CH4) and carbon monoxide (CO) gases were selected as target gas samples, and the volume fractions of the two gases in the gas cell were controlled by two mass flow controllers. The MOS gas sensor array self-confirmation method program runs on a PC (personal computer) under an operating system with a main frequency of 2.27Hz, a memory of 4G and a 32-bit Window 7. Table 2 shows the mixed gas experimental samples collected by the experimental system. The MOS gas sensor response data samples for each concentration of the above experiment were obtained in five independent replicates.
TABLE 1 MOS GAS SENSOR ASSEMBLED TO A GAS SENSOR ARRAY
Table 2 sample of mixed gas experimental data
Taking the heating voltage open circuit fault of the No. 16 TGS2620 MOS gas sensor in the MOS gas sensor array as an example, at this time, the gas sensor lacks sufficient working temperature, so that the gas sensitive material can not perform chemical reaction with the gas to be detected, and the MOS gas sensor has constant output fault. Table 3 shows the proof relative error of concentration estimates for different regression models at constant output fault. Therefore, under the condition of constant fault, the identification and detection performance of the electronic nose experiment system is reduced, and particularly the gas concentration detection precision is obviously reduced. Table 4 is the average relative error of concentration estimates at constant output fault using the different regression models of the method of this patent. Therefore, the high-robustness chemical sensor array soft measurement model provided by the patent can obviously improve the reliability and accuracy of the analysis result of the electronic nose system.
TABLE 3 mean relative error of concentration estimation for different regression models at constant output failure
Table 4 mean relative error of concentration estimation under constant output fault using different regression models of the method of this patent
It should be noted that the detailed description is only for explaining and explaining the technical solution of the present invention, and the scope of protection of the claims is not limited thereby. It is intended that all such modifications and variations be included within the scope of the invention as defined in the following claims and the description.
Claims (10)
1. A high-robustness chemical sensor array soft measurement method is characterized by comprising the following steps:
the method comprises the following steps: dividing a chemical sensor array into g sub-sensor arrays, wherein the number of various sensors in the chemical sensor array is consistent;
step two: collecting output signals of each sub-sensor array to form a training sample set;
step three: performing feature extraction on the training sample of each sub-sensor array by using a principal component analysis method to obtain a feature set of each sub-sensor array;
step four: training a regression model of the sub-sensor array by using the feature set;
step five: and measuring the measurement sample by using the trained regression model of each sub-sensor array, and selecting the median of the measurement result of the regression model of each sub-sensor array as the final measurement result.
2. The method for soft measurement of the high robustness chemical sensor array according to claim 1, wherein the specific steps of feature extraction by using a principal component analysis method in the third step are as follows:
suppose that
A matrix of measurements representing an array of sub-sensors, m representing the number of sensors and n representing the number of measurements, the steps of:
step three, firstly: performing drift compensation on an output signal of each sensor in the sensor array;
step three: standardizing the sub-sensor array after drift compensation to obtain a matrix X*；
Step three: calculating the matrix X*The covariance matrix Σ of;
step three and four: and performing singular value decomposition according to the obtained covariance matrix:
step three and five: obtaining the number k of the principal components by utilizing the cumulative variance percentage according to the covariance matrix sigma after singular value decomposition;
step three and six: determining a load matrix according to the determined number k of the principal components
The method comprises the steps that eigenvectors corresponding to the largest k eigenvalues in a covariance matrix sigma are formed;
step three, pseudo-ginseng: using load matrices
Obtaining a mapping matrix C:
step three eight: and (3) utilizing the mapping matrix C to extract features:
wherein the content of the first and second substances,
representing the results of the feature extraction on the chemical sensor array measurements.
3. The method as claimed in claim 2, wherein the step three is a drift compensation method by a fractional difference method.
4. The method for soft measurement of the high robustness chemical sensor array according to claim 3, wherein the fractional difference method performs drift compensation as follows:
xi＝(Vi max-Vi min)/Vi min
wherein, Vi maxMaximum output voltage, V, representing the sensor response signali minRepresenting the minimum output voltage of the sensor response signal.
5. A highly robust chemical sensing according to claim 4The soft measurement method of the array is characterized in that the matrix X in the third step and the second step*Expressed as:
wherein, XiRepresents Xn×pColumn vector of (E) (X)i) Represents the mean of the column vectors, D (X)i) Represents the variance of the column vector, i 1, 2.. p, p being the number of sensors in the sub-sensor array.
6. The method as claimed in claim 5, wherein the covariance matrix Σ in step three is represented as:
7. the method as claimed in claim 6, wherein the step three or four is performed by performing singular value decomposition according to the covariance matrix Σ as:
∑＝VΛVT
where Λ represents a diagonal matrix whose diagonal positions contain eigenvalues λ of the covariance matrix ΣiAnd arranging λ in descending order1≥λ2≥...≥λmAnd V is more than or equal to 0 and consists of a feature vector of sigma.
8. The method as claimed in claim 7, wherein the number k of principal components obtained by using the cumulative variance percentage in the third step and the fifth step is expressed as:
when the CPV value is greater than 90, the value of k is the number of main components, where trace represents the number of traces of the matrix.
9. The method as claimed in claim 8, wherein the value of k is the number of main components when the CPV value is greater than 90% and less than 95%.
10. The method as claimed in claim 9, wherein the load matrix is a matrix of a plurality of load cells
Expressed as:
wherein u iskAnd expressing the selected feature vectors of sigma corresponding to the k principal components.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010997648.6A CN112098605A (en) | 2020-09-21 | 2020-09-21 | High-robustness chemical sensor array soft measurement method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010997648.6A CN112098605A (en) | 2020-09-21 | 2020-09-21 | High-robustness chemical sensor array soft measurement method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112098605A true CN112098605A (en) | 2020-12-18 |
Family
ID=73756413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010997648.6A Pending CN112098605A (en) | 2020-09-21 | 2020-09-21 | High-robustness chemical sensor array soft measurement method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112098605A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113030197A (en) * | 2021-03-26 | 2021-06-25 | 哈尔滨工业大学 | Gas sensor drift compensation method |
| CN115808504A (en) * | 2022-12-01 | 2023-03-17 | 哈尔滨工业大学 | Gas sensor online drift compensation method for concentration prediction |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1801136A (en) * | 2006-01-10 | 2006-07-12 | 华东理工大学 | Method for simultaneously determining smell kind and strength by machine olfaction |
| CN102944583A (en) * | 2012-11-30 | 2013-02-27 | 重庆大学 | Metal-oxide gas sensor array concentration detecting method based on drift compensation |
| CN105938116A (en) * | 2016-06-20 | 2016-09-14 | 吉林大学 | Gas sensor array concentration detection method based on fuzzy division and model integration |
| CN106569982A (en) * | 2016-07-11 | 2017-04-19 | 江南大学 | GPR online soft measurement method and system with singular point detection compensation function |
| CN108037183A (en) * | 2017-12-12 | 2018-05-15 | 杭州电子科技大学 | A kind of supersonic array gas sensor based on condenser type bridge structure |
| CN108287184A (en) * | 2017-12-29 | 2018-07-17 | 东北电力大学 | Paraffin odor Classified Protection based on electronic nose |
| CN109116042A (en) * | 2018-09-04 | 2019-01-01 | 中国农业大学 | A kind of network communication electronic nose detection system and detection method |
| CN109611077A (en) * | 2018-11-14 | 2019-04-12 | 渤海大学 | A kind of pumpingh well oil liquid gas-oil ratio online soft sensor method based on multi-model |
| CN105913079B (en) * | 2016-04-08 | 2019-04-23 | 重庆大学 | Electronic nose isomeric data recognition methods based on the study of the target domain migration limit |
| CN106841308B (en) * | 2016-12-26 | 2019-12-10 | 浙江工商大学 | Portable intelligent electronic nose system and smell identification method |
| CN111103325A (en) * | 2019-12-19 | 2020-05-05 | 南京益得冠电子科技有限公司 | Electronic nose signal drift compensation method based on integrated neural network learning |
| CN111126575A (en) * | 2020-01-09 | 2020-05-08 | 同济大学 | Gas sensor array mixed gas detection method and device based on machine learning |
-
2020
- 2020-09-21 CN CN202010997648.6A patent/CN112098605A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1801136A (en) * | 2006-01-10 | 2006-07-12 | 华东理工大学 | Method for simultaneously determining smell kind and strength by machine olfaction |
| CN102944583A (en) * | 2012-11-30 | 2013-02-27 | 重庆大学 | Metal-oxide gas sensor array concentration detecting method based on drift compensation |
| CN102944583B (en) * | 2012-11-30 | 2014-05-14 | 重庆大学 | Metal-oxide gas sensor array concentration detecting method based on drift compensation |
| CN105913079B (en) * | 2016-04-08 | 2019-04-23 | 重庆大学 | Electronic nose isomeric data recognition methods based on the study of the target domain migration limit |
| CN105938116A (en) * | 2016-06-20 | 2016-09-14 | 吉林大学 | Gas sensor array concentration detection method based on fuzzy division and model integration |
| CN106569982A (en) * | 2016-07-11 | 2017-04-19 | 江南大学 | GPR online soft measurement method and system with singular point detection compensation function |
| CN106841308B (en) * | 2016-12-26 | 2019-12-10 | 浙江工商大学 | Portable intelligent electronic nose system and smell identification method |
| CN108037183A (en) * | 2017-12-12 | 2018-05-15 | 杭州电子科技大学 | A kind of supersonic array gas sensor based on condenser type bridge structure |
| CN108287184A (en) * | 2017-12-29 | 2018-07-17 | 东北电力大学 | Paraffin odor Classified Protection based on electronic nose |
| CN109116042A (en) * | 2018-09-04 | 2019-01-01 | 中国农业大学 | A kind of network communication electronic nose detection system and detection method |
| CN109611077A (en) * | 2018-11-14 | 2019-04-12 | 渤海大学 | A kind of pumpingh well oil liquid gas-oil ratio online soft sensor method based on multi-model |
| CN111103325A (en) * | 2019-12-19 | 2020-05-05 | 南京益得冠电子科技有限公司 | Electronic nose signal drift compensation method based on integrated neural network learning |
| CN111126575A (en) * | 2020-01-09 | 2020-05-08 | 同济大学 | Gas sensor array mixed gas detection method and device based on machine learning |
Non-Patent Citations (1)
| Title |
|---|
| 陈寅生等: "自确认金属氧化物半导体气体传感器阵列及其应用研究", 《传感技术学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113030197A (en) * | 2021-03-26 | 2021-06-25 | 哈尔滨工业大学 | Gas sensor drift compensation method |
| CN115808504A (en) * | 2022-12-01 | 2023-03-17 | 哈尔滨工业大学 | Gas sensor online drift compensation method for concentration prediction |
| CN115808504B (en) * | 2022-12-01 | 2024-06-04 | 哈尔滨工业大学 | Online drift compensation method for gas sensor for concentration prediction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Richards et al. | Multivariate data analysis in electroanalytical chemistry | |
| Hou et al. | A portable embedded toxic gas detection device based on a cross-responsive sensor array | |
| Ma et al. | Statistical learning techniques and their applications for condition assessment of power transformer | |
| Wei et al. | Classification and prediction of rice wines with different marked ages by using a voltammetric electronic tongue | |
| Tiwari et al. | Identification of monofloral honey using voltammetric electronic tongue | |
| CN112098605A (en) | High-robustness chemical sensor array soft measurement method | |
| CN106841308B (en) | Portable intelligent electronic nose system and smell identification method | |
| Xu et al. | Comparison of machine learning algorithms for concentration detection and prediction of formaldehyde based on electronic nose | |
| Miszczyk et al. | Multispectral impedance quality testing of coil-coating system using principal component analysis | |
| Chen et al. | Fault detection, isolation, and diagnosis of status self-validating gas sensor arrays | |
| CN112462001A (en) | Gas sensor array model calibration method for data amplification based on condition generation countermeasure network | |
| Cai et al. | Determination of Chinese Angelica honey adulterated with rice syrup by an electrochemical sensor and chemometrics | |
| CN116985183A (en) | Quality monitoring and management method and system for near infrared spectrum analyzer | |
| Chambon et al. | A metallic oxide gas sensor array for a selective detection of the CO and NH3 gases | |
| Yu et al. | Pure milk brands classification by means of a voltammetric electronic tongue and multivariate analysis | |
| Yang et al. | A real-time fault detection and isolation strategy for gas sensor arrays | |
| CN208171949U (en) | Electric nasus system | |
| Li et al. | Voltammetric electronic tongue for the qualitative analysis of milk adulterated with urea combined with multi-way data analysis | |
| Sorrentino et al. | Determination of the distribution of relaxation times through Loewner framework: A direct and versatile approach | |
| Shaffer et al. | Multiway analysis of preconcentrator‐sampled surface acoustic wave chemical sensor array data | |
| Llobet et al. | Electronic nose simulation tool centred on PSpice | |
| Jin et al. | Chemometric analysis of gas chromatographic peaks measured with a microsensor array: methodology and performance assessment | |
| Song et al. | Design and implementation a real-time electronic nose system | |
| Ding et al. | Age identification of Chinese rice wine using electronic nose | |
| Men et al. | Data fusion of electronic nose and electronic tongue for discrimination of Chinese liquors |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201218 |