CN208171949U - Electric nasus system - Google Patents
Electric nasus system Download PDFInfo
- Publication number
- CN208171949U CN208171949U CN201820650509.4U CN201820650509U CN208171949U CN 208171949 U CN208171949 U CN 208171949U CN 201820650509 U CN201820650509 U CN 201820650509U CN 208171949 U CN208171949 U CN 208171949U
- Authority
- CN
- China
- Prior art keywords
- gas
- filter
- sensor array
- faims
- ion
- 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.)
- Active
Links
Abstract
The utility model discloses a kind of electric nasus system, electric nasus system, including:Air inlet and filter assemblies;FAIMS pneumatic filter;Gas sensor array;Data processor, data processor is connected with gas sensor array, to extract sensor array column information principal component according to sensor array signal, according to the sensor array signal principal component of extraction, establish neural network, the mass data in maintenance data library is trained it, and carries out pattern discrimination to sample to be tested gas in storage environment after determining neural network model and parameter.Electric nasus system according to the present utility model increases gas screening and filtering device, and reduction, which mixes gas interference component in environment, makes system filter out interference, and the precision of gas detection in storage environment can be improved.
Description
Technical field
The utility model relates to electronic analytical instrument fields, specifically, in particular to a kind of electric nasus system.
Background technique
Electric nasus system is the electronic system that smell is identified using the response of gas sensor array.It is by selectivity
Electrochemical sensor array and recognition methods appropriate composition instrument, can identify smell, can be obtained and the sensory evaluation of people
Consistent result.
With logistics fast development, storage is the storage center for undertaking multiple function by the development of single storage function, such as
Order processing, sorting, packaging, secondary operation, manufacturing etc..The quantity of goods and type stored in storage center is all increasingly
Staff that is more, participating in operation is also more and more.There are a variety of potential pollution sources in storage, directly affect cargo quality and
Service life influences the health of warehousing and storage activities personnel.
When carrying out storage environment monitoring using electric nasus system, there are " broad spectrum activity " of sensor measurement and " intersect
Gas in sensibility " problem and storage environment mixes the problems such as measuring accuracy is not high.
Utility model content
The utility model is intended to solve one of above-mentioned technical problem in the prior art at least to a certain extent.In view of
This, the utility model needs to provide a kind of electric nasus system, increases gas screening and filtering device, reduces and mixes in environment
Gas interference component makes system filter out interference.
The one side of the utility model provides a kind of electric nasus system, including:Air inlet and filter assemblies, the air inlet and mistake
Filter component is sampled warehouse gas to be measured, stores the gas sampled after to gas sampled pre-filtering;FAIMS gas
Filter, the FAIMS pneumatic filter are connected with the air inlet and filter assemblies, for receiving the air inlet and filtering group
The gas sampled of part storage, the FAIMS pneumatic filter carry out ion to the gas sampled by FAIMS sensor
Filtering;Gas sensor array, the gas sensor array are connected with the FAIMS pneumatic filter, the gas sensing
The resistance variations that device array response is generated by the gas sampled of ion filter, are converted by signal acquisition circuit and A/D
Generate corresponding sensor array signal;Data processor, the data processor are connected with the gas sensor array, with
According to the sensor array signal extract sensor array column information principal component, according to the sensor array signal of extraction it is main at
Point, neural network is established, the mass data in maintenance data library is trained it, and right after determining neural network model and parameter
Sample to be tested gas carries out pattern discrimination in storage environment.
The electric nasus system of embodiment according to the present utility model increases gas screening and filtering device, reduces and comes from ring
Mixing gas interference component in border makes system filter out interference, and the precision of gas detection in storage environment can be improved, meanwhile, increase
Detection accuracy, and gas source can be positioned, eliminate pollution and leakage source.
In addition, can also have following additional technology special according to the electric nasus system of the utility model above-described embodiment
Sign:
One embodiment according to the present utility model, the air inlet and filter assemblies include:Aspiration pump, the aspiration pump are used
In the extraction gas sampled;Filter device, the entrance of the filter device are connected with the outlet of the aspiration pump;Detection dress
Set, the entrance of the detection device is connected with the outlet of the filter device, the detection device include pressure gauge, flowmeter and
Thermometer is measured with pressure, flow velocity and the temperature of the gas sampled flowed out to the filter device;Gas storage dress
It sets, the gas storage device is connect with the outlet of the detection device;Control module, the control module respectively with the pumping
Air pump is connected with the filter device.
One embodiment according to the present utility model is equipped with the first valve between the aspiration pump and the filter device
Door is equipped with the second valve, in the outlet of the gas storage device between the detection device and the gas storage device
Place is equipped with third valve, and first valve, second valve and the third valve are connected with the control module.
One embodiment according to the present utility model, the filter device include:Active carbon layer, the silicon rubber successively arranged
Stratum granulosum and molecular sieve layer.
One embodiment according to the present utility model, FAIMS pneumatic filter include:FAIMS drift tube;Ion source, institute
State the ionization area that ion source is located at the FAIMS drift tube, the reactive ion phase that the gas sampled is generated with the ion source
Interaction forms product ion;
Peripheral circuit module, the peripheral circuit module act on the product ion, and the peripheral circuit module is used
In generating compensating electric field and asymmetric electric field, to realize the variety classes ion isolation to the product ion.
One embodiment according to the present utility model, the peripheral circuit module include matched non-with FAIMS drift tube
Circuit occurs for balancing waveform, circuit, ionic current amplifying circuit and auxiliary electrode circuit occur for offset voltage.
One embodiment according to the present utility model, the peripheral circuit module further comprise banishing with the ion-conductance
The data processing display unit of big circuit connection.
One embodiment according to the present utility model, the gas sensor array include:Mass flow controller, it is described
Mass flow controller is for controlling the flow of the gas sampled.
Another aspect according to the present utility model provides a kind of gas source of electric nasus system in storage and identifies and positioning
Method, including:Following steps:Gas sampling is sampled warehouse gas to be measured, pre-filtering and stores;Gas filtration,
Gas after sampling carries out ion filter by FAIMS pneumatic filter;Filtered gas is passed through electronics by gas detection
Nose extracts useful sensor information;Feature extraction, normalization carry out dimensionality reduction to the feature of extraction, extract sensor array letter
Cease principal component;Pattern-recognition establishes neural network according to the sensor array signal principal component of extraction, maintenance data library it is big
Amount data are trained it, determine neural network model, parameter;Mode is carried out to sample to be tested gas in storage environment to sentence
Not.
Gas source of the electric nasus system of embodiment according to the present utility model in storage identifies and localization method, reduces and
Mixing gas interference component from environment makes system filter out interference, and the precision of gas detection in storage environment can be improved, meanwhile,
Detection accuracy is increased, and gas source can be positioned, pollution and leakage source are eliminated.
One embodiment according to the present utility model is followed the steps below when carrying out pattern-recognition:By electronic nose
Collected data are stored in sensor array data library, and the sensing data of several samples is extracted from the database,
As data to be analyzed;A behavior sample is converted thereof into, the M row N column sample matrix of sensor reading is classified as;It calculates
The covariance matrix and mean value of sample matrix out;Find out the characteristic value and its corresponding feature vector of covariance matrix;By feature
Vector is sorted from large to small as corresponding eigenvalue, the corresponding feature vector of K characteristic value before taking;As sensor array
Signal principal component extracts sensor array signal data from database, establishes BP- neural network, imports sample data to nerve
Network model and parameter optimize, and finally test.
The additional aspect and advantage of the utility model will be set forth in part in the description, partially will be from following description
In become obvious, or recognized by the practice of the utility model.
Detailed description of the invention
Fig. 1 is the structural schematic diagram of the electric nasus system of embodiment according to the present utility model.
Fig. 2 is the electric nasus system workflow schematic diagram of embodiment according to the present utility model.
Fig. 3 is the air inlet of the electric nasus system of embodiment according to the present utility model and the structural schematic diagram of filter assemblies.
Fig. 4 is the structural block diagram of the FAIMS pneumatic filter of the electric nasus system of embodiment according to the present utility model.
Fig. 5 is the structural block diagram of the gas sensor array of the electric nasus system of embodiment according to the present utility model.
Fig. 6 is gas source identification and localization method stream of the electric nasus system of embodiment according to the present utility model in storage
Cheng Tu.
Fig. 7 is the electric nasus system of embodiment according to the present utility model in gas source identification and localization method in storage
Pattern-recognition constructs flow chart.
Fig. 8 is the electric nasus system of embodiment according to the present utility model in gas source identification and localization method in storage
Principal component analysis flow chart.
Fig. 9 is gas source identification and localization method of the electric nasus system of embodiment according to the present utility model in storage
BP neural network constructs flow chart.
Figure 10 is the position view that the electric nasus system of embodiment according to the present utility model is arranged in storage.
Specific embodiment
The embodiments of the present invention are described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning
Same or similar element or element with the same or similar functions are indicated to same or similar label eventually.Below by ginseng
The embodiment for examining attached drawing description is exemplary, it is intended to for explaining the utility model, and should not be understood as to the utility model
Limitation.
In the description of the present invention, it should be understood that term " center ", " longitudinal direction ", " transverse direction ", " length ", " width
Degree ", " thickness ", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom" "inner", "outside", " suitable
The orientation or positional relationship of the instructions such as hour hands ", " counterclockwise " is to be based on the orientation or positional relationship shown in the drawings, merely to just
In description the utility model and simplify description, rather than the device or element of indication or suggestion meaning there must be specific side
Position is constructed and operated in a specific orientation, therefore should not be understood as limiting the present invention.
In addition, term " first ", " second " are used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance
Or implicitly indicate the quantity of indicated technical characteristic.Define " first " as a result, the feature of " second " can be expressed or
Implicitly include one or more of the features.The meaning of " plurality " is two or two in the description of the present invention,
More than, unless otherwise specifically defined.
In the present invention unless specifically defined or limited otherwise, term " installation ", " connected ", " connection ", " Gu
It is fixed " etc. terms shall be understood in a broad sense, for example, it may be being fixedly connected, may be a detachable connection, or integral;It can be
Mechanical connection, is also possible to be electrically connected;It can be directly connected, two can also be can be indirectly connected through an intermediary
The interaction relationship of connection or two elements inside element.It for the ordinary skill in the art, can basis
Concrete condition understands the concrete meaning of above-mentioned term in the present invention.
In the present invention unless specifically defined or limited otherwise, fisrt feature the "upper" of second feature or it
"lower" may include that the first and second features directly contact, and also may include that the first and second features are not direct contacts but lead to
Cross the other characterisation contact between them.Moreover, fisrt feature includes above the second feature " above ", " above " and " above "
One feature is right above second feature and oblique upper, or is merely representative of first feature horizontal height higher than second feature.First is special
Sign is directly below and diagonally below the second feature including fisrt feature under the second feature " below ", " below " and " below ", or only
Indicate that first feature horizontal height is less than second feature.
As shown in Figure 1, the electric nasus system 100 of embodiment according to the present utility model, including:Air inlet and filter assemblies
10, FAIMS pneumatic filter 20, gas sensor array 30 and data processor 40.
Specifically, air inlet and filter assemblies 10 can be sampled warehouse gas to be measured, to the pre- mistake of gas sampled
Gas sampled is stored after filter, steel cylinder storage gas sampled can be used for example.
FAIMS pneumatic filter 20 is connected with air inlet and filter assemblies 10, for receiving air inlet and filter assemblies 10 are deposited
The gas sampled of storage, FAIMS pneumatic filter 20 carry out ion filter to gas sampled by FAIMS sensor.
Gas sensor array 30 is connected with FAIMS pneumatic filter 20, and ion is passed through in the response of gas sensor array 30
The resistance variations that the gas sampled of filtering generates generate corresponding sensor array by signal acquisition circuit and A/D conversion and believe
Number.
Data processor 40 is connected with gas sensor array 30, to extract sensor array according to sensor array signal
Information principal component establishes neural network, the mass data pair in maintenance data library according to the sensor array signal principal component of extraction
It is trained, and carries out pattern discrimination to sample to be tested gas in storage environment after determining neural network model and parameter.
The electric nasus system 100 of embodiment according to the present utility model, increases gas screening and filtering device, and reduction comes from
Mixing gas interference component in environment makes system filter out interference, and the precision of gas detection in storage environment can be improved, meanwhile, increase
Add detection accuracy, and gas source can have been positioned, eliminates pollution and leakage source.
Referring to fig. 2, it is to be understood that embodiment according to the present utility model, 100 course of work of electric nasus system can be with
It is substantially as follows:
1. gas sampling, warehouse gas to be measured is sampled, pre-filtering and is stored.
2. gas filtration, the gas after sampling carries out ion filter by FAIMS sensor.
3. filtered gas is passed through electronic nose, extracts useful sensor information by gas detection.
4. feature extraction, normalization.Dimensionality reduction is carried out to the feature of extraction, extracts sensor array column information principal component.
5. pattern-recognition, according to the sensor array signal principal component of extraction, establish neural network, maintenance data library it is big
Amount data are trained it, determine neural network model, parameter.
6. sample to be tested gas carries out pattern discrimination in pair storage environment, nutrient quality is judged.
As shown in figure 3, the electric nasus system 100 of embodiment according to the present utility model, air inlet and filter assemblies 10 include:
Aspiration pump 11, filter device 12, detection device 13, gas storage device 14 and control module 15.
Specifically, aspiration pump 11 is for extracting gas sampled.The outlet phase of the entrance of filter device 12 and aspiration pump 11
Even.The entrance of detection device 13 is connected with the outlet of filter device 12, and detection device 13 may include pressure gauge, flowmeter and temperature
Degree meter, is measured with pressure, flow velocity and the temperature of the gas sampled flowed out to filter device 12.Further, gas stores
Device 14 can be connect with the outlet of detection device 13.Control module 15 is connect with aspiration pump 11 and filter device 12 respectively, with
Realize the control to aspiration pump 11 and filter device 12.
It is understood that the first valve 16 is equipped between aspiration pump 11 and filter device 12 for the ease of control,
It is equipped with the second valve 17 between detection device 13 and gas storage device 14, is equipped with third in the exit of gas storage device 14
Valve 18, the first valve 16, the second valve 17 and third valve 18 are connected with control module 15.
It should be noted that air inlet and filter assemblies 10 mainly include pumping for the gas collection in storage environment
Pump 11, filter device 12, detection device 13, gas storage device 14 and control module 15, wherein aspiration pump 11 is for ring of storing in a warehouse
Gas extraction in border.Filter device 12 is filter, including the active carbon layer, silicone rubber particles layer and molecular sieve successively arranged
Layer, for the impurity in pre-filtering gas.Detection device 13 includes pressure gauge, flowmeter and thermometer, for measuring aspiration pump
Pressure, flow velocity and the temperature of gas is discharged.Gas storage device 14 can be steel cylinder, the sample gas storage for being finally collected into
It deposits.
Wherein, the first valve 16 is equipped between aspiration pump 11 and filter device 12, between detection device 13 and storage device
Equipped with the second valve 17, gas storage device 14 is equipped with air inlet and air outlet, is equipped with third valve 18 at gas outlet, controls mould
Block 15 can be used for controlling the switch of aspiration pump, and power and the first valve 16, the second valve 17 and third valve 18 are opened
It closes.
Air inlet and 10 sampling step of filter assemblies are simply described below below:
Before sampling, the pipe inner wall of sampling will remove lubricant, grease, solid slag particle and other pollutants.It also wants simultaneously
The residual gas and trace amounts of moisture in pipeline are removed, prevents surface from chemically reacting, generate absorption and is polluted by residual gas.
(1) processing on surface can chemically (pickling, alkali cleaning, passivation or with other similar chemical surfactant
Processing) or mechanical means (such as ultrasonic method) purified.Metal catheter is processed by shot blasting, to grass tube silane
Its adsorptivity can be reduced by changing reagent processing etc..
(2) depending on the cleaning method of cleaning pipe is with amount, pressure existing for gas.General 10 times of gas with upper volume
Cleaning.
Pressure reducer, valve and conduit have certain dead volume, the use of simple cleaning operation are not very effectively, because of residual
Gas and trace amounts of moisture are stopped and are slowly diffused into transported gas in dead volume, using the cleaning for increasing and decreasing pressure repeatedly
Operating effect can be more preferable.
(3) leak test should be carried out to system after the cleaning of leak test conduit.Being segmented leak test is effective good method.Leak testing process
It is general as follows.
1. system pressure or decompression are then shut off outlet, the variation of pressure gauge (or flowmeter) is observed.When pressure gauge exists
When decline is no more than 0.1MPa in 0.5h or flowmeter float on it is as air tight when falling to zero.
2. system is pressurizeed, all tie points are smeared with surfactant (such as dodecane imitates aqueous sodium persulfate solution), are generated
Bubble person is air leakage point.
Samples selecting, it is the gas in fruit warehouse that gas is collected in this experiment, and bleeding point is selected in doubt region.
3. sampling, opens the first valve 16, the second valve 17 and third valve 18, starting aspiration pump 11 just starts gas
Sample acquisition stops acquisition gas when pressure gauge show value reaches 6MPa, closes aspiration pump 11, and successively closes third valve
18, the second valve 17, the first valve valve.Air inlet/outlet valve is fastened in the connection for disconnecting steel cylinder and filter device 12, completes gas
Body is collected.
As shown in figure 4, one embodiment according to the present utility model, the FAIMS pneumatic filter 20 of electric nasus system 100
Including:
FAIMS drift tube 21, ion source 22 and peripheral circuit plate module 23.
Specifically, ion source 22 can be located at the ionization area of FAIMS drift tube 21, gas sampled and ion source 22 are generated
Reactive ion interact to form product ion.Peripheral circuit module 23 acts on product ion, peripheral circuit module 23
It can be used for generating compensating electric field and asymmetric electric field, to realize the variety classes ion isolation to the product ion.
It should be noted that High-Field asymmetric waveform ion mobility spectrometry (High Field Asymmetric Ion
Mobility Spectrometry, FAIMS) be it is a kind of work under atmospheric pressure environment, using gaseous ion in high electric field
Nonlinear motion come carry out trace materials detection and isolated technology.
FAIMS gas filtration modular system includes ion source, drift tube and peripheral circuit three parts, wherein:
Drift tube is the core component of FAIMS, and ion is separated and detected in drift tube.
Ion source is FAIMS pith, and under test gas is ionized into ion and injects drift tube by ion source.Ion source
Performance all has a great impact to response sensitivity and resolution ratio, and ionization source has certain selectivity.
Peripheral circuit includes high-frequency and high-voltage asymmetric electric field and compensating electric field generation circuit needed for FAIMS work.
The FAIMS course of work is as follows:
Enter FAIMS drift tube by the filtered gas of first order filtration system.
When by ionization area, the reactive ion generated with ion source interacts to form product ion sample gas, then
Into the filtering area for having asymmetric electric field and compensating electric field, realized under collective effect of the compensating electric field from asymmetric electric field different
Type ion isolation.
Sample ions after separation reach detection zone, and ion is detected in electrode and becomes with after under the action of deflection voltage
Neutral gas then passes to sensor array.
By adjusting the parameter of asymmetric electric field and compensating electric field, bandwidth selection molecular weight can be set in 20-200 model
It encloses and is selectively passed through by gas, so that reducing into sensor array gaseous species, reduce network analysis difficulty.
Further, one embodiment according to the present utility model, peripheral circuit module include matching with FAIMS drift tube
Asymmetric waveform circuit occurs, circuit, ionic current amplifying circuit and auxiliary electrode circuit occur for offset voltage.It is preferred that
Ground, peripheral circuit module 23 may further include the data processing display unit connecting with the ionic current amplifying circuit.
Preferably, one embodiment according to the present utility model, gas sensor array 30 may include:Mass flow control
Device processed, mass flow controller is for controlling the flow of the gas sampled.
As shown in figure 5, one embodiment according to the present utility model, sample gas is after the filtering of FAIMS filter
Gas, after FAIMS is filtered, sample gas constituents complexity reduce, be conducive to carry out qualitative and quantitative analysis to it.
It is imported in gas sensor array by the filtered sample gas of FAIMS, sensor responds gas and generates resistance
The response signal (sensor array signal) in generation time domain is converted in variation by signal acquisition circuit and A/D.The signal carries out
After pretreatment, before carrying out pattern-recognition, pretreatment appropriate is carried out to sensor signal in electric nasus system.
Flow-control module is mass flow controller (MFC), controls gas flow by MFC, to control into survey
The tested gas concentration of chamber is tried, tested gas is customized.
Sensor array, using discrete gas sensor array, for being detected to sample gas constituents.
By taking fruit stores as an example, due in fruit storing process major gaseous component have esters, alcohols, aldehydes, ethylene, with
And because of gases such as hydrogen sulfide, alkane, the Ammonias of corruption generation.Sensor selected by the sensor array see the table below:
The response signal of sensor is tentatively improved, is then fed by signal conditioning circuit module, signal conditioning circuit
Data collecting card carries out analog-to-digital conversion, realizes display and storage in a computer.
Data collecting card selects the NI USB-6366 data collecting card of NI company, has the synchronous mould in 8 tunnels under the channel 2MS/s
Quasi- input, 16 bit resolutions, effect are analog quantity to be converted to digital quantity, and data feeding computer is stored.
Sensor array signal pretreatment, can generate one and time phase when each sensor reacts with tested gas
The response curve of pass, due to the relationship of data volume, in order to simplify follow-up mode identification process, common way takes sensor
Steady-state response is analyzed and is handled.The purpose of Signal Pretreatment has filtering, Baseline Survey, drift compensation, Information Compression and returns
One change etc..Common processing method has a score ratio method, wavelet transformation technique baseline drift inhibit field have good effect, point
Formula differential method, relative mistake point-score can compensate the temperature effect of sensor.
Course of work description, gas sampling:Firstly, injecting pure air 2min to test chamber, intracavitary survival gas is excluded
Clean experimental situation is created in interference to experiment.Then it is passed through into test chamber by the filtered sample gas of FAIMS,
The concentration for being passed through gas is controlled by MFC.Signal acquisition:When system brings into operation, NI USB-6366 data is called to adopt
The dynamic link of truck simultaneously starts data acquisition by data collecting card subprogram.Temperature modulation and signal condition, pass through temperature
Modulating system controls suitable temperature so that sensor distinguishes obviously the response of gas with various.Pass through signal conditioning circuit pair
Signal amplifies, and removes dryness, and measured signal is converted to voltage value by A/D.Acquisition terminates and data storage, when program setting
Acquisition time terminate or external user single machine panel on " stopping " button when exit the program, acquisition terminates.Acquisition terminates, data
It is stored in specified file.
Referring to Fig. 6, gas source of the electric nasus system of embodiment according to the present utility model in storage identifies and positioning side
Method, including:Following steps:
Gas sampling is sampled warehouse gas to be measured, pre-filtering and stores.
Gas filtration, the gas after sampling carry out ion filter by FAIMS pneumatic filter.
Filtered gas is passed through electronic nose, extracts useful sensor information by gas detection.
Feature extraction, normalization carry out dimensionality reduction to the feature of extraction, extract sensor array column information principal component.
Pattern-recognition establishes neural network according to the sensor array signal principal component of extraction, maintenance data library it is a large amount of
Data are trained it, determine neural network model, parameter.
Pattern discrimination is carried out to sample to be tested gas in storage environment.
Gas source of the electric nasus system of embodiment according to the present utility model in storage identifies and localization method, reduces and
Mixing gas interference component from environment makes system filter out interference, and the precision of gas detection in storage environment can be improved, meanwhile,
Detection accuracy is increased, and gas source can be positioned, pollution and leakage source are eliminated.
Further, an embodiment according to the present utility model, gas source of the electric nasus system in storage identifies and positioning
Method is followed the steps below when carrying out pattern-recognition:
It is stored in sensor array data library by the collected data of electronic nose, if extracting dry sample from the database
This sensing data, as data to be analyzed.
A behavior sample is converted thereof into, the M row N column sample matrix of sensor reading is classified as.
Calculate the covariance matrix and mean value of sample matrix.
Find out the characteristic value and its corresponding feature vector of covariance matrix.
By feature vector sorting from large to small by corresponding eigenvalue, the corresponding feature vector of K characteristic value before taking.
Referring to Fig. 7-10, gas source of the electric nasus system of embodiment according to the present utility model in storage identifies and positioning
Method is followed the steps below when carrying out algorithm for pattern recognition:PCA extracts sensor array signal principal component, usually exists
Before pattern-recognition, the more characteristic parameters of crossing of extraction are subjected to dimensionality reduction, reduce the complexity of analysis, principal component analysis (PCA) is one
The multivariate statistical analysis technology of kind common data compression and feature extraction, can effectively remove the linear correlation between data
Property, but the nonlinear correlation feature between data is had ignored, therefore use NLPCA, nonlinear transformation first is made to observation data, is introduced
High-order statistic, then principal component is analyzed.
Algorithm flow is as follows:
1. being stored in sensor array data library by the collected data of electronic nose, extracted from the database several
The sensing data of sample, as data to be analyzed
2. converting thereof into a behavior sample, it is classified as the M row N column sample matrix of sensor reading
3. calculating the covariance matrix and mean value of sample matrix
4. finding out the characteristic value and its corresponding feature vector of covariance matrix
5. by feature vector sorting from large to small by corresponding eigenvalue, the corresponding feature vector of K characteristic value before taking
The feature vector that constituent analysis part obtains is the input of pattern recognition module, these main components are as nerve
The input at network, is trained neural network by great amount of samples data, finally establishes neural network and carries out to sample gas
Pattern-recognition
When carrying out BP neural network pattern-recognition, used neural network algorithm is BP (Back
Propagation) neural network, i.e. error-duration model error back propagation.The learning process of algorithm by information forward-propagating and
Two process compositions of backpropagation of error.Each neuron of input layer is responsible for receiving from the defeated of previous stage fuzzy membership function
Information out, and pass to each neuron of middle layer;Middle layer is internal information process layer, is responsible for information transformation, is become according to information
The demand of change ability, middle layer can be designed as single hidden layer or more hidden layer configurations;It is each that the last one hidden layer is transmitted to output layer
The information of neuron, after further treatment after, complete the forward-propagating treatment process that once learns, outwardly exported by output layer
Information processing result Ym.Optimization information is exported according to processing result Ym, takes different intervening measure and means.
The construction method of BP neural network is as follows:Using sensor array by PCA processed feature vector, X n as
The input vector of artificial neural network;Input data of the Xn as each neuron of the input layer of the BP neural network,
And each neuron of middle layer is single hidden layer or more hidden layer configurations, it is each that the last one hidden layer of the middle layer is transmitted to output layer
Forward-propagating treatment process of the information of neuron once to be learnt, and the output layer output information processing result is
Ym.Optimize intermediate layer parameter with output information processing result.
Specifically, sample gas constituents are complicated, thus detect many of gas componant and concentration to storage environment into
Row differentiates.It chooses gas sample sensor array data and pattern-recognition verification is carried out to electric nasus system, be inferred in unknown
Storage environment gas componant and concentration under the conditions of mixed gas.
Gas source location algorithm is briefly described below, the embodiments of the present invention are described below in detail, by the utility model
Device is arranged in several positions in warehouse, is monitored in real time to the gas componant in warehouse.This detection device firstly the need of
One auxiliary gas source determines a reference position, establishes three-dimensional coordinate system, to position to target gas source, simultaneously
Correction can also be synchronized to each device in warehouse.The initialization that monitoring system is completed by auxiliary gas source, assists gas
Source is known gas ingredient, and the size judgment means distance of the gas concentration detected according to each device assists the position of gas source
It sets, then to assist gas source to establish a three-dimensional space coordinates as origin.Further to need to detect object gas carry out
Monitoring, determines the position of target gas source.
Monitoring device work specific steps include:Parameter initialization is carried out to device power, then starts to carry out gas
Sampling.Known auxiliary gas source is detected firstly, adjusting FAIMS ionization area's voltage, all devices are carried out by communication clock
Synchronized sampling, the data of acquisition are uploaded to host computer by data collecting card and carry out data processing, first at progress PCA dimensionality reduction
Reason is handled the data input BP neural network after dimensionality reduction after allowing, by the output result of Processing with Neural Network and auxiliary gas
The physical location in source is compared, until exporting result in error range.Then three-dimensional system of coordinate is established in space, adjusted
FAIMS ionization area's voltage is measured in real time storage gas, if having detected that object gas constituent concentration changes, according to
Abovementioned steps are handled, and finally determine gas source position.
As sensor array signal principal component, sensor array signal data are extracted from database, establish BP- mind
Through network, imports sample data and neural network model and parameter are optimized, finally test.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is contained at least one embodiment or example of the utility model.In the present specification, to the schematic table of above-mentioned term
It states and is necessarily directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be with
It can be combined in any suitable manner in any one or more of the embodiments or examples.In addition, those skilled in the art can incite somebody to action
Different embodiments or examples described in this specification are engaged and are combined.
Although the embodiments of the present invention have been shown and described above, it is to be understood that above-described embodiment is
Illustratively, it should not be understood as limiting the present invention, those skilled in the art are in the scope of the utility model
Inside it can make changes, modifications, alterations, and variations to the above described embodiments.
Claims (8)
1. a kind of electric nasus system, which is characterized in that including:
Air inlet and filter assemblies, the air inlet and filter assemblies are sampled warehouse gas to be measured, to the pre- mistake of gas sampled
The gas sampled is stored after filter;
FAIMS pneumatic filter, the FAIMS pneumatic filter are connected with the air inlet and filter assemblies, for receiving institute
The gas sampled of air inlet and filter assemblies storage is stated, the FAIMS pneumatic filter is taken by FAIMS sensor to described
Sample gas carries out ion filter;
Gas sensor array, the gas sensor array are connected with the FAIMS pneumatic filter, the gas sensor
The resistance variations that array response is generated by the gas sampled of ion filter are produced by signal acquisition circuit and A/D conversion
Raw corresponding sensor array signal;
Data processor, the data processor are connected with the gas sensor array, to be believed according to the sensor array
It number extracts sensor array column information principal component and neural network is established, with number according to the sensor array signal principal component of extraction
It is trained according to the mass data in library, and to sample to be tested gas in storage environment after determining neural network model and parameter
Carry out pattern discrimination.
2. electric nasus system according to claim 1, which is characterized in that the air inlet and filter assemblies include:
Aspiration pump, the aspiration pump is for extracting the gas sampled;
Filter device, the entrance of the filter device are connected with the outlet of the aspiration pump;
Detection device, the entrance of the detection device are connected with the outlet of the filter device, and the detection device includes pressure
Meter, flowmeter and thermometer, in terms of pressure, flow velocity and the temperature of the gas sampled to flow out to the filter device carry out
Amount;
Gas storage device, the gas storage device are connect with the outlet of the detection device;
Control module, the control module are connect with the aspiration pump and the filter device respectively.
3. electric nasus system according to claim 2, which is characterized in that between the aspiration pump and the filter device
Equipped with the first valve, it is equipped with the second valve between the detection device and the gas storage device, is stored in the gas
The exit of device be equipped with third valve, first valve, second valve and the third valve with the control
Module is connected.
4. electric nasus system according to claim 2, which is characterized in that the filter device includes:The work successively arranged
Property layer of charcoal, silicone rubber particles layer and molecular sieve layer.
5. electric nasus system according to claim 1, which is characterized in that FAIMS pneumatic filter includes:
FAIMS drift tube, drift tube are the core components of FAIMS, are separated in drift tube for realizing ion;
Ion source, the ion source are located at the ionization area of the FAIMS drift tube, and the gas sampled and the ion source generate
Reactive ion interact to form product ion;
Peripheral circuit module, the peripheral circuit module act on the product ion, and the peripheral circuit module is for producing
Raw compensating electric field and asymmetric electric field, to realize the variety classes ion isolation to the product ion.
6. electric nasus system according to claim 5, which is characterized in that the peripheral circuit module includes floating with FAIMS
It moves and manages matched asymmetric waveform generation circuit, circuit, ionic current amplifying circuit and auxiliary electrode electricity occur for offset voltage
Road.
7. electric nasus system according to claim 6, which is characterized in that the peripheral circuit module further comprises and institute
State the data processing display unit of ionic current amplifying circuit connection.
8. electric nasus system according to claim 6, which is characterized in that the gas sensor array includes:Quality stream
Amount controller, the mass flow controller is for controlling the flow of the gas sampled.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201820650509.4U CN208171949U (en) | 2018-05-03 | 2018-05-03 | Electric nasus system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201820650509.4U CN208171949U (en) | 2018-05-03 | 2018-05-03 | Electric nasus system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN208171949U true CN208171949U (en) | 2018-11-30 |
Family
ID=64369939
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201820650509.4U Active CN208171949U (en) | 2018-05-03 | 2018-05-03 | Electric nasus system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN208171949U (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108398533A (en) * | 2018-05-03 | 2018-08-14 | 北京物资学院 | Electric nasus system and its air source discriminating in storage and localization method |
CN109738584A (en) * | 2018-12-30 | 2019-05-10 | 盐城工学院 | A kind of electric nasus system |
CN111044683A (en) * | 2019-12-25 | 2020-04-21 | 华中科技大学 | Electronic nose technology capable of realizing congenital recognition and acquired training and application thereof |
-
2018
- 2018-05-03 CN CN201820650509.4U patent/CN208171949U/en active Active
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108398533A (en) * | 2018-05-03 | 2018-08-14 | 北京物资学院 | Electric nasus system and its air source discriminating in storage and localization method |
CN109738584A (en) * | 2018-12-30 | 2019-05-10 | 盐城工学院 | A kind of electric nasus system |
CN111044683A (en) * | 2019-12-25 | 2020-04-21 | 华中科技大学 | Electronic nose technology capable of realizing congenital recognition and acquired training and application thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhang et al. | A miniaturized electronic nose with artificial neural network for anti-interference detection of mixed indoor hazardous gases | |
CN208171949U (en) | Electric nasus system | |
CN108398533A (en) | Electric nasus system and its air source discriminating in storage and localization method | |
CN101226166B (en) | Low-power consumption hand-hold electric nasus system for testing on-site | |
CN102128877B (en) | Device for qualitatively, quantitatively and quickly detecting pesticide residue on site | |
CN1194227C (en) | Fast non-destructive detection method and device of food smell based on gas sensor array technology | |
CN106841308B (en) | Portable intelligent electronic nose system and smell identification method | |
CN101788517A (en) | Pungent traditional Chinese medicine odor fingerprint map construction system and method based on bionic olfaction | |
CN106053551A (en) | Multi-channel multi-type sensor capability test system | |
CN111596002A (en) | LEL and V0C intelligent online analysis and treatment system for detecting waste gas | |
CN201955325U (en) | Hand-held device for quickly detecting pesticide residues | |
Qian et al. | A novel multisensor detection system design for low concentrations of volatile organic compounds | |
CN103389323B (en) | Method for evaluating ages of precious medicinal materials quickly and losslessly | |
CN109406719A (en) | A kind of odor detection system of carrier gas sample introduction | |
CN105866190A (en) | Electronic nose device for food detection and detection method | |
CN109030611A (en) | Method for distinguishing is sentenced in a kind of quick, the lossless fingerprint analysis of green tea fragrance and odor type based on ion mobility spectrometry | |
CN106093134A (en) | The compensation method of metal oxide sensor array response drift | |
CN201867391U (en) | Device for qualitatively, quantitatively and fast detecting pesticide residue in site | |
CN100412543C (en) | Gas detecting apparatus and method based on field ionization effect | |
CN110470798B (en) | Temperature compensation method for portable electronic nose enrichment device | |
AU2020102518A4 (en) | A method of gas sensor calibration based on linear optimization | |
CN105929016A (en) | Value determination method for selenium content of quality control sample of tobacco leaf | |
CN206020330U (en) | A kind of intelligent water quality trace heavy metal on-line monitoring of internet of things oriented and early warning system | |
Mandenius et al. | Sensor fusion with on‐line gas emission multisensor arrays and standard process measuring devices in baker's yeast manufacturing process | |
CN102282932A (en) | Method for rapidly detecting germination rate of vegetable seeds |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GR01 | Patent grant | ||
GR01 | Patent grant |