CN110426421A - Multi-component harmful gas detection device and detection method in a kind of kitchen environment - Google Patents
Multi-component harmful gas detection device and detection method in a kind of kitchen environment Download PDFInfo
- Publication number
- CN110426421A CN110426421A CN201910848479.7A CN201910848479A CN110426421A CN 110426421 A CN110426421 A CN 110426421A CN 201910848479 A CN201910848479 A CN 201910848479A CN 110426421 A CN110426421 A CN 110426421A
- Authority
- CN
- China
- Prior art keywords
- sensor
- voltage
- gas
- value
- module
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- 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
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Health & Medical Sciences (AREA)
- Pure & Applied Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Software Systems (AREA)
- Computational Mathematics (AREA)
- Databases & Information Systems (AREA)
- Electrochemistry (AREA)
- Biochemistry (AREA)
- Probability & Statistics with Applications (AREA)
- Operations Research (AREA)
- Analytical Chemistry (AREA)
- Immunology (AREA)
- Algebra (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Pathology (AREA)
- Investigating Or Analyzing Materials By The Use Of Fluid Adsorption Or Reactions (AREA)
Abstract
The invention discloses multi-component harmful gas detection device and detection method in a kind of kitchen environment, which includes sample introduction module, power supply module, gas chamber and sensor module, signal processing module, AD conversion module, single-chip minimum system module, serial port module.In the external world at the uniform velocity in the case where gas chamber air inlet, sensor can fast and stable, response and restore voltage baseline, its voltage responsive exported can be received and processed by single-chip microcontroller, the Concentration Testing result of three kinds of components such as carbon monoxide, methane and formaldehyde into host computer output gaseous mixture.The apparatus function is perfect, small in size, easy to operate, and accuracy, stability, repeatability preferably, can be used as in core apparatus investment kitchen environment in the manufacturing of harmful gas detecting instrument.
Description
Technical field
The present invention relates to multi-component harmful gas detection dresses in pernicious gas detection field more particularly to a kind of kitchen environment
It sets and detection method.
Background technique
For a long time, due to various factors such as combustion gas, finishing, geographical location, ventilation habits, toxic in indoor environment has
Evil gas concentration is likely in the short time significantly larger than safety standard.And the toxic and harmful gas of high concentration, human body will cause
Extremely disadvantageous influence, gently then cause it is uncomfortable, go to bits, it is heavy then cause the major accidents such as fire, explosion.
The detection of current toxic and harmful gas is usually applied to industry and outdoor occasion, such as tunnel, mine, factory, and
The function of alarm is often only provided, and cannot detect the concentration of a variety of independent component gases in COMPLEX MIXED gas, it is difficult to be met
The growing demand of people.And commercial apparatus on the market it is big, expensive, inconvenient due to volume the disadvantages of, it is uncomfortable
Close the detection for carrying out long term air quality in daily life.It is, therefore, desirable to provide a kind of quick, cheap, convenient, miniaturization
Toxic and harmful gas detection device and detection technique further realize the concentration of each target gas component of intelligent distinguishing.
Summary of the invention
Present invention aims in view of the deficiencies of the prior art, provide multi-component harmful gas detection in a kind of kitchen environment
Device and detection method, solve in common detection means it is complicated for operation, expensive, instrument is huge, can not carry out multicomponent
The disadvantages of measurement.
The purpose of the present invention is achieved through the following technical solutions: multi-component harmful gas inspection in a kind of kitchen environment
Survey device, which is characterized in that the device include sample introduction module, power supply module, gas chamber and sensor module, signal processing module,
AD conversion module, single-chip minimum system module, serial port module；
The sample introduction module includes air pump, dried chimney filter and heating room, at the uniform velocity leading to toward gas chamber and sensor module
Enter the gas after drying, filtering and temperature control；The air pump is used to make gas flow rate when cleaning when sample introduction with sensor consistent
It is and constant, it is ensured that the voltage signal of sensor output is relatively stable before and after sensor cleaning；The dried chimney filter, it is internal
Using discoloration silicone filler, for filtering particulate contaminant and steam in sample introduction gas, and silica gel is indicated by color change
Performance condition；The air pump, dried chimney filter and heating room are sequentially connected；
The power supply module is used to provide required different voltages for sample introduction module and single-chip minimum system module, by+
The power supply of 12V Switching Power Supply；
The gas chamber and sensor module include gas chamber and sensor array, and the gas chamber main body is rectangular parallelepiped structure, gas
There is a gas detection cell above chamber interior, gas detection cell side is provided with air inlet, and the air inlet is connected with air inlet pipe, gas
Indoor bottom is provided with fixation and replacement of 6 mos sensor circular grooves for sensor, is also provided with 3mm high in the bottom of gas chamber
The bottom empty slot of degree, and the side opposite with air inlet in the bottom of gas chamber is provided with outgassing groove, the bottom empty slot and outlet
The bottom of slot connection, the outgassing groove is equipped with Temperature Humidity Sensor site, is provided with venthole, the outlet in the side of outgassing groove
Hole connects escape pipe, is discharged after converging for gas to be detected from the bottom empty slot of gas chamber bottom through escape pipe；The sensing
Device array includes 6 mos sensors and 1 Temperature Humidity Sensor；The mos sensor is arranged in mos sensor circular groove,
The head of mos sensor is located in gas detection cell, consistent with the height of air inlet；The Temperature Humidity Sensor is arranged in outlet
At the Temperature Humidity Sensor site of the bottom of slot, for detecting after heating the actual temperature of gas of flows through sensor array and wet
Degree.
The signal processing module is filtered using differential amplifier circuit and RC, for the output signal to sensor array into
Row differential amplification and filtering processing make its output voltage stabilization after being passed through 1000 ml/min pure air of constant flow rate
In 1.0V hereinafter, to allow sensor output voltage to have a greater change section；
The AD conversion module is used to be rapidly completed the AD of the voltage signal of the sensor after differential amplification and filtering processing
Conversion, is input to single-chip minimum system module；
The single-chip minimum system module is mounted with the algorithm model for differentiating under test gas concentration, is used for received letter
Analytic operation number directly is carried out in slave computer and obtains final result, and host computer is sent to by serial port module；
The serial port module turns usb signal line by serial ports using USART serial ports and sends the result detected to host computer
Further, the air pump is use for laboratory minipump, realizes that flow velocity is adjusted by PWM wave, entire to detect
Air pump is all run always in the process；Air pump can also be replaced by distributing instrument or other sample introduction instruments.
Further, the heating room is made of copper pipe, heating sheet, heat-conducting glue, fixed frame etc.；It heats room and uses high-performance
Heat-resisting nylon 3D printing forms, and inside can be packed into the copper pipe of 4 outer diameter 6mm internal diameter 4mm.
Further, the power supply module generates+5V voltage using voltage stabilizing chip UA7805 and REG1117-3.3 respectively
With+3.3V voltage, and join a resistance using two paster LED lamp strings and be followed by exporting between ground in voltage stabilizing chip, for referring to
Show whether voltage conversion is normal.
Further, the gas chamber material is resin, and plenum interior is filled with most of useless using following 8000 resins
Shorten the time of cleaning sensor so that plenum interior gas volume further decreases in space；The air inlet pipe and an air outlet pipe of gas chamber
Size is outer diameter 6mm, internal diameter 4mm, and the silicone tube as air circuit connection is good to guarantee having a size of outer diameter 8mm, internal diameter 4mm
Air-tightness.
Further, a kind of combination of the sensor array are as follows: MP-9, MP-4, MP503, TGS821, TGS816,
TGS2602 and SHT20 digital hygro sensor.
Further, the pernicious gas of the sensor array detection is carbon monoxide, methane and the first in kitchen environment
Aldehyde, the interference gas hydrogen and ethyl alcohol of detection.
Multi-component harmful gas detection method in a kind of kitchen environment using above-mentioned detection device, which is characterized in that should
Method the following steps are included:
(1) sensor preheats
The sensor needs of detection device are preheated to not a half hour in the case where no air inlet, keep sensor sufficiently pre-
Heat.
(2) sensor cleans
Single-chip minimum system module connects the air pump in sample introduction module by relay, makes it by normal air with 1000
The constant flow rate of ml/min is passed through gas chamber, cleaning sensor array, and removal impurity interference continues at least 30 minutes, Zhi Daochuan
After the output voltage of sensor array is steady, and each sensor voltage drop to preset voltage baseline value hereinafter, into
Enter the state for capableing of sample introduction and detection, executes step (3).
(3) sample introduction and detection, specifically includes the following steps:
(3.1) sample introduction and sensor characteristic values are acquired, specifically: will it is dry in sample introduction module, filter and temperature control after
Under test gas is passed through gas chamber and is detected, and continues 5 minutes, during which single-chip minimum system module samples sensor array output
Voltage signal, constant duration sampling, is converted to digital signal for voltage signal by signal processing module and AD conversion module,
And then characteristic value data and the preservation of sensor are obtained, the characteristic value data of sensor includes baseline value, voltage maximum before sample introduction
Positive slope, voltage peak area, restores voltage baseline time, temperature profile value and Humidity Features value, single-chip microcontroller at voltage responsive peak value
Minimum systematic module is mounted with BP artificial neural network；
Before baseline value is single detection process before the sample introduction, by the sensor output voltage of differential amplification processing
Value, is denoted as Bi, i=1,2, and 3,4,5,6；The voltage responsive peak value is to handle in single detection process by differential amplification
The maximum voltage value of sensor output, is denoted as Pi, i=1,2, and 3,4,5,6；The voltage peak area is that single detects sample introduction process
In, voltage responsive rises, the opposite voltage difference of baseline value before sample introduction of the output voltage of sensor it is cumulative with；The voltage is most
Big positive slope is the gradient maxima of response curve when voltage responsive rises during the sample introduction of single detection；The recovery
Voltage baseline time is that the sampled point since the cleaning process that single detects counts, the electricity of baseline before voltage is dropped to and detected
When pressure, then it is assumed that cleaning is complete, using the number of sampled point at this time as the time for restoring voltage baseline；The temperature profile value and
Humidity Features value are as follows: the Temperature Humidity Sensor numerical value extracted in sensor peak value, two sampled points adjacent with pre-and post-peaking
Temperature Humidity Sensor numerical value carry out 3 points average, obtain final temperature characteristic value and Humidity Features value.
Baseline value before the sample introduction of sampling and voltage responsive peak value are handled, i-th of peak value of response moment sensing is obtained
Before device resistance and sample introduction when baseline sensor resistance ratio, i.e. resistance ratio is denoted as Ri, formula is as follows:
(3.2) characteristic value that sensor is acquired and handled is as training set, BP ANN, specifically:
The input layer of BP artificial neural network 26, intermediate hidden neuron 9, output layer neuron 3,3 output layers
Neuron respectively represents the gas concentration value of carbon monoxide, methane and formaldehyde, each neuron of input layer and middle layer
Initial weight and threshold value generate at random, and loss function uses mean square error, are the concentration calculation value and reality of three kinds of detection gas
The mean square error of value and.Wherein, the determination method of intermediate hidden neuron number is as follows: rule of thumb formula determines that formula is such as
Under:
Wherein n be input neuron number, m be output neuron number, l be middle layer neuron number, a be [1,
10] constant in range.9 kinds of situations that intrerneuron number is 6~14 are chosen, several times three are calculated to each case respectively
The concentration calculation value of kind of detection gas and the mean square error of actual value and, and be averaged respectively, each case calculating mean square error
The number of difference is identical, obtains 9 average values, and wherein the corresponding intrerneuron number of minimum value is 9, as best intermediate hidden
Layer neuron number.
Input layer-middle layer activation primitive is tansig, therefore the output mid_output of each middle layer neuron can
It is calculated by following formula.Wherein, n indicates multiple inputs and its obtained knot that respectively adds up after weight, threshold calculations of single neuron
Fruit.
Mid_output=tansig (n)=2/ (1+exp (- 2*n)) -1
And middle layer-output layer activation primitive is purelin, therefore the output output of each output layer neuron can be by
Following formula calculates.
Output=purelin (n)=n
By the ratio of the sensor resistance before the sensor resistance and sample introduction of 6 voltage responsive peak value moments when baseline, 6
Voltage maximal positive slope, 6 voltage peak areas, 6 recovery voltage baseline times, 1 temperature profile value and 1 Humidity Features value
It is input to as training set in the BP artificial neural network of single-chip minimum system module loading, passes through Regularization algorithms
It is trained BP artificial neural network, loss function is less than 10-3When deconditioning, it is final to determine each of input layer and middle layer
The weight and threshold value of a neuron obtain trained BP artificial neural network.
(3.3) 26 characteristic values of the gas to be detected of sensor acquisition are input in BP artificial neural network, are passed through
BP artificial neuron neural computing exports 3 kinds of gas concentrations to be detected, and calculated result is sent to by serial port module
Position machine.
(4) state is cleaned
After the completion of detection, normal air is passed through detection gas chamber with the constant flow rate of 1000 ml/mins by sample introduction air pump, clearly
Wash sensor array 10 minutes or more, until each sensor restore voltage baseline value hereinafter, then repeat step (3.3) and
(4) sample introduction can be continued and detected, until artificially stopping detecting.
Further, the single-chip minimum system module can also be loaded into the BP people of classical Principal Component Analysis Algorithm optimization
Artificial neural networks, classical Principal Component Analysis Algorithm carry out pre-optimized place for 26 input feature vector values of BP artificial neural network
Reason calculates the covariance matrix of former input, and characteristic root is principal component variance, and characteristic root is arranged from big to small, takes preceding 4 spies
Coefficient of the feature vector as principal component corresponding to root is levied, cumulative proportion in ANOVA reaches 85% or more, can explain former defeated
Enter characteristic value most information.According to 26 characteristic values of coefficient and input, 4 principal component F can be obtainedi=ai0+ai1x1+ai2x2
+…+ai26x26, i=1,2,3,4 are used as new feature value, wherein aijFor the coefficient being calculated, i=1,2,3,4, j=1,2 ...
26, x1~x26For 26 characteristic values of input.BP artificial neural network is input layer 4 at this time, intermediate hidden layer nerve
9, member, it is output layer neuron 3, manually refreshing by Regularization algorithms training BP using 4 principal components as training set
Through network, the weight and threshold value of each neuron of determining input layer and middle layer obtain trained BP artificial neuron
Network, for sample introduction and detection.
Further, the single-chip minimum system module can also be loaded into the BP people of classical partial least squares algorithm optimization
Artificial neural networks regard 26 characteristic values input in training set as independent variable, and 3 kinds of gas actual concentrations outputs to be detected are made
For dependent variable, the extract component t1 and u1 in independent variable and dependent variable, t1 are independent variable x to classical partial least squares algorithm respectivelyi
Linear combination, i.e. t1=a0+a1x1+a2x2+…+a26x26, u1 is the linear combination of dependent variable, i.e. u1=β0+β1x1+β2x2+…
+β26x26, the covariance of both t1 and u1 is made to reach maximum.Note t1 and u1 is first composition.With the residual error for having extracted first composition
Continue to extract second composition, repeats the step, take first four ingredient, Cross gain modulationEach extract component
Coefficient a between available ti and each input independent variableij, i=1,2,3,4, j=1,2 ... 26.According to classical minimum two partially
Four groups of coefficients that multiplication obtains carry out pre-optimized processing for 26 input feature vector values of BP artificial neural network, calculate four
A ingredient Fi=ai0+ai1x1+ai2x2+…+ai26x26, i=1,2,3,4 are used as new feature value, wherein x1~x26It is 26 of input
Characteristic value.BP artificial neural network is input layer 4 at this time, intermediate hidden neuron 9, output layer neuron 3,
Using 4 ingredients as training set, by Regularization algorithms BP ANN, determining input layer and centre
The weight and threshold value of each neuron of layer, obtain trained BP artificial neural network, for sample introduction and detection.
The beneficial effects of the present invention are:
First, there is the synchronous detection function of multicomponent gas, especially under kitchen environment existing for certain interference gas
In the case of, the quick detection of multicomponent gas still can be accurately carried out to object gas, enhance the practicability of device.
Second, unique detection air chamber structure design.Plenum interior is filled with most of useless using following 8000 resins
Space, so that maximum gas volume further decreases, the time of cleaning sensor also can further shorten.
Third, the process design of dynamic sample detection.Static state diffusion tends not to meet the needs of quickly detecting, can not
The data for ensuring to measure are accurate enough, therefore the present invention is by the way of fixed flow rate dynamic detection, and to cleaning and detection
Time is optimized, and overcomes this problem well.
Fourth, device miniaturization and intelligence.Each modular structure of instrument is simple, connection is optimized with layout, small volume,
Complicated algorithm model can be embedded in slave computer and carry out data operation, realize the detection to kitchen pernicious gas, and will be in result
Reach host computer or other display equipment.Instrument integrated operation is easy, and function is more perfect, and intelligent level is high.
Fifth, proposing Principal Component Analysis Algorithm and the combination of BP artificial neural network and partial least squares algorithm and BP people
Two kinds of unified algorithms associated with artificial neural networks, Principal Component Analysis Algorithm and partial least squares algorithm are directed to BP artificial neural network
The input feature vector value part of network optimizes.
Detailed description of the invention
Fig. 1 is electrical block diagram；
Fig. 2 is gas circuit heating device structure chart；
Fig. 3 is the left and right visual angle chart at the bottom of of gas chamber；
Fig. 4 is that gas chamber temperature changes with time figure；
Fig. 5 is that gas chamber humidity changes with time figure.
Fig. 6 is gas sensor output signal primitive character value schematic diagram
In figure, 1. sample introduction modules；2. power supply module；3. gas chamber and sensor module；4. signal processing module；5.AD conversion
Module；6. single-chip minimum system module；7. serial port module；8. air inlet pipe；9. air inlet；10.MOS sensor circular groove；11. temperature
Humidity sensor site；12. bottom empty slot；13. outgassing groove；14. venthole；15. escape pipe.
Specific embodiment
Invention is further described in detail in the following with reference to the drawings and specific embodiments.
As shown in Figure 1, a kind of circuit structure of multi-component harmful gas detection device shows in a kind of kitchen environment of the present invention
It is intended to, including sample introduction module 1, power supply module 2, gas chamber and sensor module 3, signal processing module 4, AD conversion module 5, drying
Pipe 5, single-chip minimum system module 6, serial port module 7.The sample introduction module 1 includes air pump, filter screen, dried chimney filter, adds
Hot cell.Air pump is successively passed through dried chimney filter by relay driving, with gas of the constant flow rate extraction through filter screen filtration, adds
Hot cell, eventually arrives at gas chamber, and air pump therein can also be replaced by distributing instrument or other sample introduction instruments.It heats room and uses high-performance
Room is heated made of heat-resisting nylon 3D printing, inside can be packed into the copper pipe of 4 outer diameter 6mm internal diameter 4mm.The power supply module by
The power supply of+12V Switching Power Supply, then generates+5V voltage and+3.3V using voltage stabilizing chip UA7805 and REG1117-3.3 respectively
Voltage, required different voltages are provided for detection device, and two paster LED lamp strings join a resistance and are followed by exporting in voltage stabilizing chip
Between ground, whether instructed voltage conversion is normal.The gas chamber and sensor module 3 include gas chamber and sensor array, described
Gas chamber formed using the following 8000 resin 3D printings, gas chamber size has that smoothness is high, steady performance.Gas chamber into
Tracheae and outlet pipe size are outer diameter 6mm, internal diameter 4mm, and the silicone tube as air circuit connection is having a size of outer diameter 8mm, internal diameter
4mm, to guarantee good air-tightness.Multiple sensors, which can be placed, in gas chamber bottom constitutes sensor array, the sensor
Array uses 6 kinds of commercial metal oxide-semiconductor sensors through screening, and can make cross response to complicated gas componant,
Including pernicious gas and harmless interference gas.In addition it is additionally added Temperature Humidity Sensor SHT20, gas indoor temperature and humidity is timed
Measurement.The signal processing module 4 is filtered using differential amplifier circuit and RC, can be located in advance to sensor output voltage
Reason, its variation range of flexible modulation.The AD conversion module has low-power consumption, multi-pass using the outer ADC128S052 chip of piece
The AD conversion of 6 road sensor passage voltage signals can be rapidly completed in the advantages that road.The single-chip minimum system uses
MSP430F169 chip has many advantages, such as that low in energy consumption, arithmetic speed is fast, resourceful in piece, single-chip microcontroller is most as microcontroller
Mini system module 6 is mounted with algorithm model, measured signal directly can be carried out analytic operation and obtained most to terminate in slave computer
Fruit, and Concentration Testing data are sent to host computer by UART serial ports.Serial port module 7 turns USB by serial ports using USART serial ports
Signal wire connects host computer, can send detection progress situation and final detection data to host computer in detection overall process.
As shown in Fig. 2, in a kind of kitchen environment of the present invention multi-component harmful gas detection device a kind of heating chamber structure
Schematic diagram.The heating room is formed by high performance heat resistant nylon 3D printing, and intermediate groove can place outer diameter 6mm internal diameter 4mm's
Hollow copper tubing 4, for the heat of heating sheet rapidly and sufficiently to be passed to internal gas flow.It is square that heating sheet is fixed on top
In shape slot, bottom increases contact area and thermal coefficient, to realize good by heat conductive silica gel and copper pipe and heating means touch
Good is thermally conductive.
As shown in figure 3, in a kind of kitchen environment of the present invention multi-component harmful gas detection device a kind of air chamber structure figure.
Gas chamber bottom is connected by 705 transparent RTV silicon rubber with circuit board sensors region, guarantees its good adhesive and airtight
Property.
When detection device starting is for detecting, the sensor for needing will test device needs in the case where no air inlet in advance
Heat at least half an hour, it is fully warmed-up sensor.Then gas chamber, cleaning sensing are passed through with the constant flow rate of 1000 ml/mins
Device array, removal impurity interference, continues at least 30 minutes, after the output voltage of sensor array is steady, and each biography
Sensor voltage all drops to preset voltage baseline value hereinafter, into the state for capableing of sample introduction and detection, can be under test gas
Concentration Testing is carried out, does not need generally to stop after detection device starting, detection can be carried out continuously.When being carried out continuously detection, clearly
Washing the sensor array time can shorten, and ensure that within 10 minutes or so that each sensor voltage drops to preset voltage
Below baseline value.
In detection process, the mixed gas after dry and particle filtering enters gas chamber through air inlet pipe 8, with 1000 ml/mins
The flow velocity of clock continues air inlet 5 minutes.The too short then sensor response of sample injection time is too long still in acute variation, causes the energy
The excessive temperature and humidity measurement inaccuracy and sensor response of will lead to of consumption and sample introduction flow velocity with the time is wasted in switch mode
When be easy to appear and significantly cross response spike, too small, the gas concentration error that will lead to preparation increases and sensor response is inadequate
Obviously；Air inlet 9 is suitable with sensor head height, it is ensured that air-flow can flow fast through each sensor；Mos sensor circle
Slot 10 can steadily place sensor；Air-flow can be with flows through sensor surface, by bottom empty slot 12, and from outgassing groove 13
Temperature Humidity Sensor site 11 is passed through, and is then discharged from venthole 14 and escape pipe 15, and Temperature Humidity Sensor site 11, which is located at, to be passed
After sensor array, i.e., the end of gas circuit in gas chamber, therefore the data measured can accurately reflect sensor array air-flow nearby
Whole temperature and humidity situation of change, as shown in Figure 4 and Figure 5.6 timing sampling of single-chip minimum system module passes during detection
The voltage signal of sensor array output, is converted to digital letter for voltage signal by signal processing module 4 and AD conversion module 5
Number, and then the characteristic value data of sensor and the data of the Temperature and Humidity module of gas chamber and preservation are obtained, the original characteristic value number of sensor
According to including baseline value, voltage maximal positive slope, voltage responsive peak value, voltage peak area, recovery voltage baseline time, temperature before sample introduction
Humidity etc., before baseline value is single detection process before the sample introduction, by the sensor output voltage value that differential amplification is handled,
It is denoted as Bi, i=1,2,3,4,5,6；The voltage responsive peak value is in single detection process, by the sensing of differential amplification processing
The maximum voltage value of device output, is denoted as Pi, i=1,2, and 3,4,5,6；During the voltage peak area detects sample introduction for single,
Voltage responsive rises, the opposite voltage difference of baseline value before sample introduction of the output voltage of sensor it is cumulative with；The voltage is maximum
Positive slope is the gradient maxima of response curve when voltage responsive rises during the sample introduction of single detection；The recovery electricity
Pressing baseline time is that the sampled point since the cleaning process that single detects counts, the voltage of baseline before voltage is dropped to and detected
When, then it is assumed that cleaning is complete, using the number of sampled point at this time as the time for restoring voltage baseline；The temperature profile value and wet
Spend characteristic value are as follows: the Temperature Humidity Sensor numerical value extracted in sensor peak value, two sampled points adjacent with pre-and post-peaking
Temperature Humidity Sensor numerical value carries out at 3 points and averages, and obtains final temperature characteristic value and Humidity Features value.
Operation is combined to baseline value before the sample introduction of sampling and voltage responsive peak value, is obtained i-th of the peak value of response moment
Before sensor resistance and sample introduction when baseline sensor resistance ratio, i.e. resistance ratio is denoted as Ri, formula is as follows:
After characteristic value is handled, be left resistance ratio, voltage maximal positive slope, voltage peak area, restore voltage baseline time,
Temperature and humidity etc., single-chip minimum system module 6 are mounted with BP artificial neural network, and by treated, sensor characteristic values are used as instruction
White silk sample, BP ANN, specifically: the input layer of BP artificial neural network 26, intermediate hidden layer mind
Through 9, member, output layer neuron 3,3 output layer neurons respectively represent the gas concentration of carbon monoxide, methane and formaldehyde
The initial weight and threshold value of each neuron of value, input layer and middle layer generate at random, and loss function uses mean square error,
For the concentration calculation value of three kinds of detection gas and the mean square error of actual value and.Wherein, the determination of intermediate hidden neuron number
Method is as follows: rule of thumb formula determines, formula is as follows:
Wherein n be input neuron number, m be output neuron number, l be middle layer neuron number, a be [1,
10] constant in range.9 kinds of situations that intrerneuron number is 6~14 are chosen, several times three are calculated to each case respectively
The concentration calculation value of kind of detection gas and the mean square error of actual value and, and be averaged respectively, each case calculating mean square error
The number of difference is identical, obtains 9 average values, and wherein the corresponding intrerneuron number of minimum value is 9, as best intermediate hidden
Layer neuron number.
Input layer-middle layer activation primitive is tansig, therefore the output mid_output of each middle layer neuron can
It is calculated by following formula.Wherein, n indicates multiple inputs and its obtained knot that respectively adds up after weight, threshold calculations of single neuron
Fruit.
Mid_output=tansig (n)=2/ (1+exp (- 2*n)) -1
And middle layer-output layer activation primitive is purelin, therefore the output output of each output layer neuron can be by
Following formula calculates.
Output=purelin (n)=n
By the ratio of the sensor resistance before the sensor resistance and sample introduction of 6 voltage responsive peak value moments when baseline, 6
Voltage maximal positive slope, 6 voltage peak areas, 6 recovery voltage baseline times, 1 temperature profile value and 1 Humidity Features value
It is input to as training set in the BP artificial neural network of the loading of single-chip minimum system module 6, is calculated by Bayesian regularization
Method is trained BP artificial neural network, and loss function is less than 10-3When deconditioning, finally determining input layer and middle layer
The weight and threshold value of each neuron obtain trained BP artificial neural network.
26 characteristic values of sensor acquisition and treated gas to be detected are input in BP artificial neural network,
By BP artificial neuron neural computing, 3 kinds of gas concentrations to be detected are exported, and calculated result is sent out by serial port module 7
It send to host computer.
Single-chip minimum system module 6 can also load the BP artificial neural network of classical Principal Component Analysis Algorithm optimization,
Classical Principal Component Analysis Algorithm carries out pre-optimized processing for 26 input feature vector values of BP artificial neural network, calculates former defeated
The covariance matrix entered, characteristic root are principal component variance, and characteristic root is arranged from big to small, is taken corresponding to preceding 4 characteristic roots
Coefficient of the feature vector as principal component, cumulative proportion in ANOVA reaches 85% or more, can explain that former input feature vector value is big
Partial information.According to 26 characteristic values of coefficient and input, 4 principal component F can be obtainedi=ai0+ai1x1+ai2x2+…+ai26x26,
I=1,2,3,4 are used as new feature value, wherein aijFor the coefficient being calculated, i=1,2,3,4, j=1,2 ... 26, x1~x26For
26 characteristic values of input.BP artificial neural network is input layer 4 at this time, intermediate hidden neuron 9, output layer
It neuron 3, is determined using 4 principal components as training set by Regularization algorithms BP ANN
The weight and threshold value of each of input layer and middle layer neuron, obtain trained BP artificial neural network, are used for sample introduction
With detection.
Single-chip minimum system module 6 can also be loaded into the BP artificial neural network of classical partial least squares algorithm optimization,
It regard 26 characteristic values input in training set as independent variable, regard 3 kinds of gas actual concentrations outputs to be detected as dependent variable, warp
The extract component t1 and u1 in independent variable and dependent variable, t1 are independent variable x to allusion quotation partial least squares algorithm respectivelyiLinear combination,
That is t1=a0+a1x1+a2x2+…+a26x26, u1 is the linear combination of dependent variable, i.e. u1=β0+β1x1+β2x2+…+β26x26, make t1
Reach maximum with the covariance of both u1.Note t1 and u1 is first composition.Continue extraction with the residual error for having extracted first composition
Binary repeats the step, takes first four ingredient, Cross gain modulationThe available ti of each extract component
With the coefficient a between each input independent variableij, i=1,2,3,4, j=1,2 ... 26.It is obtained according to classical Partial Least Squares
Four groups of coefficients carry out pre-optimized processing for 26 input feature vector values of BP artificial neural network, calculate four ingredient Fi
=ai0+ai1x1+ai2x2+…+ai26x26, i=1,2,3,4 are used as new feature value, wherein x1~x26For 26 characteristic values of input.
At this time BP artificial neural network be input layer 4, intermediate hidden neuron 9, output layer neuron 3, by 4 at
Be allocated as training set, by Regularization algorithms BP ANN, determining input layer and middle layer it is every
The weight and threshold value of one neuron obtain trained BP artificial neural network, for sample introduction and detection.
For carbon monoxide 240ppm, a variety of slave computers of the mixed gas sample of methane 8400ppm and formaldehyde 0ppm are calculated
Method test case is as shown in the table, the threshold voltage of 6 in detection device sensor be respectively as follows: 0.6V, 1.0V, 0.7V,
0.3V、0.5V、0.8V。
When the algorithm being loaded into single-chip minimum system module 6 is Partial Least Squares Regression or principal component regression, detection
Carbon monoxide and methane error it is very big, only formaldehyde be 0, more accurately, when what is be loaded into single-chip minimum system module 6
Algorithm be BP artificial neural network when, the error of three kinds of gas values of detection is reduced, be closer to actual value, when single-chip microcontroller most
The algorithm being loaded into mini system module 6 is principal component analysis+BP artificial neural network or offset minimum binary+BP artificial neural network
When, the result of detection is more accurate, the error very little with actual value.
Cleaning is completely the same in gas chamber with the airflow channel of sample introduction, and scavenging period is longer than sample injection time, and 10 minutes
Above until making each sensor restore voltage baseline value or less.Too long scavenging period will cause the wave of time and resource
Take, too short still have residual, baseline to have little time to be restored to initial level to form apparent baseline drift it will cause part under test gas
It moves.It can be continued to test after the completion of cleaning, until artificially stopping detecting.
In kitchen environment of the present invention a kind of sensor combinations of multi-component harmful gas detection device be MP-9, MP-4,
MP503, TGS821, TGS816, TGS2602 and SHT20 digital hygro sensor, detectable each target gas levels model
It encloses are as follows:
Carbon Monoxide Detection range: 10~1000ppm；
CH_4 detection range: 500~10000ppm；
Formaldehyde examination range: 1~50ppm.
The concentration range that interference gas allows under normal detection is respectively as follows: hydrogen 10-100ppm, ethyl alcohol 10-50ppm.
The above examples only illustrate the technical idea of the present invention, and this does not limit the scope of protection of the present invention, all
According to the technical idea provided by the invention, any changes made on the basis of the technical scheme each falls within the scope of the present invention
Within.
Claims (10)
1. multi-component harmful gas detection device in a kind of kitchen environment, which is characterized in that the device include sample introduction module (1),
Power supply module (2), gas chamber and sensor module (3), signal processing module (4), AD conversion module (5), single-chip minimum system
Module (6), serial port module (7).
The sample introduction module (1) includes air pump, dried chimney filter and heating room, at the uniform velocity toward gas chamber and sensor module (3)
The gas being passed through after drying, filtering and temperature control；Gas flow rate one when the air pump is used to make when sample introduction and sensor cleans
It causes and constant, it is ensured that the voltage signal of sensor output is relatively stable before and after sensor cleaning；The dried chimney filter, it is interior
Portion for filtering particulate contaminant and steam in sample introduction gas, and indicates silicon by color change using discoloration silicone filler
Glue performance condition；The air pump, dried chimney filter and heating room are sequentially connected.
The power supply module (2) is used to provide required difference electricity for sample introduction module (1) and single-chip minimum system module (6)
Pressure is powered by+12V Switching Power Supply.
The gas chamber and sensor module (3) include gas chamber and sensor array, and the gas chamber main body is rectangular parallelepiped structure, gas chamber
Inner upper has a gas detection cell, and gas detection cell side is provided with air inlet (9), and the air inlet (9) is connected with air inlet
It manages (8), the indoor bottom of gas is provided with fixation and replacement of 6 mos sensor circular grooves (10) for sensor, in the bottom of gas chamber
It is also provided with the bottom empty slot (12) of 3mm height, and is provided with outgassing groove in the bottom of gas chamber and air inlet (9) opposite side
(13), the bottom empty slot (12) is connected to outgassing groove (13), and the bottom of the outgassing groove (13) is equipped with Temperature Humidity Sensor position
Point (11) is provided with venthole (14) in the side of outgassing groove (13), and the venthole (14) connects escape pipe (15), for be checked
The gas of survey from the bottom empty slot (12) of gas chamber bottom converge after through escape pipe (15) be discharged；The sensor array includes 6
Mos sensor and 1 Temperature Humidity Sensor；The mos sensor is arranged in mos sensor circular groove, the head of mos sensor
It is consistent with the height of air inlet (9) in gas detection cell；The temperature of the bottom of outgassing groove is arranged in the Temperature Humidity Sensor
At humidity sensor site (11), for detect heating after flows through sensor array gas actual temperature and humidity.
The signal processing module (4) is filtered using differential amplifier circuit and RC, for the output signal to sensor array into
Row differential amplification and filtering processing adjust output signal variation range.
The AD that the AD conversion module (5) is used to be rapidly completed the voltage signal of the sensor after differential amplification and filtering processing turns
It changes, is input to single-chip minimum system module (6).
The single-chip minimum system module (6) is mounted with the algorithm model of detection under test gas concentration, is used for received letter
Analytic operation number directly is carried out in slave computer and obtains final result, and host computer is sent to by serial port module (7).
The serial port module (7) uses USART serial ports, turns usb signal line by serial ports and sends the result detected to host computer.
2. multi-component harmful gas detection device in a kind of kitchen environment according to claim 1, which is characterized in that described
Air pump is use for laboratory minipump, realizes that flow velocity is adjusted by PWM wave, air pump is all run always in entire detection process；
Air pump can also be replaced by distributing instrument or other sample introduction instruments.
3. multi-component harmful gas detection device in a kind of kitchen environment according to claim 1, which is characterized in that described
Heating room is made of copper pipe, heating sheet, heat-conducting glue, fixed frame etc.；Heating room is formed using the 3D printing of high performance heat resistant nylon, interior
Portion can be packed into the copper pipe of 4 outer diameter 6mm internal diameter 4mm.
4. multi-component harmful gas detection device in a kind of kitchen environment according to claim 1, which is characterized in that described
Power supply module (2) generates+5V voltage and+3.3V voltage using voltage stabilizing chip UA7805 and REG1117-3.3 respectively, and uses
Two paster LED lamp strings join a resistance and are followed by exporting between ground in voltage stabilizing chip, whether normal are used to indicate voltage conversion.
5. multi-component harmful gas detection device in a kind of kitchen environment according to claim 1, which is characterized in that described
Gas chamber material is resin, and plenum interior is filled with most of wasted space using following 8000 resins, so that plenum interior gas
Volume further decreases, and shortens the time of cleaning sensor；The very little air inlet pipe (8) and escape pipe ruler (15) of gas chamber is outer diameter
6mm, internal diameter 4mm, the silicone tube as air circuit connection is having a size of outer diameter 8mm, internal diameter 4mm, to guarantee good air-tightness.
6. multi-component harmful gas detection device in a kind of kitchen environment according to claim 1, which is characterized in that described
A kind of combination of sensor array are as follows: gas sensor MP-9, MP-4, MP503, TGS821, TGS816, TGS2602 and number
Word Temperature Humidity Sensor SHT20.
7. multi-component harmful gas detection device in a kind of kitchen environment according to claim 1, which is characterized in that described
The pernicious gas of sensor array detection is carbon monoxide, methane and formaldehyde in kitchen environment, the interference gas hydrogen of detection
And ethyl alcohol.
8. multi-component harmful gas detection method in a kind of kitchen environment using above-mentioned detection device, which is characterized in that the party
Method the following steps are included:
(1) sensor preheats
The sensor needs of detection device are preheated to not a half hour in the case where no air inlet, are fully warmed-up sensor.
(2) sensor cleans
Single-chip minimum system module (6) by relay connect sample introduction module (1) in air pump, make its by normal air with
The constant flow rate of 1000 ml/mins is passed through gas chamber, cleaning sensor array, and removal impurity interference continues at least 30 minutes, directly
To sensor array output voltage it is steady after, and each sensor voltage drop to preset voltage baseline value with
Under, into the state for capableing of sample introduction and detection, execute step (3).
(3) sample introduction and detection, specifically includes the following steps:
(3.1) sample introduction and sensor characteristic values are acquired, specifically: will be dried in sample introduction module (1), filter and temperature control after to
Survey gas is passed through gas chamber and is detected, and continues 5 minutes, and during which single-chip minimum system module (6) sampling sensor array exports
Voltage signal, constant duration sampling, voltage signal is converted to by signal processing module (4) and AD conversion module (5)
Digital signal, and then obtain characteristic value data and the preservation of sensor, the characteristic value data of sensor include baseline value before sample introduction,
Voltage maximal positive slope, voltage peak area, restores voltage baseline time, temperature profile value and Humidity Features at voltage responsive peak value
Value, single-chip minimum system module (6) are mounted with BP artificial neural network；
Before baseline value is single detection process before the sample introduction, by the sensor output voltage value that differential amplification is handled, note
For Bi, i=1,2,3,4,5,6；The voltage responsive peak value is in single detection process, by the sensor of differential amplification processing
The maximum voltage value of output, is denoted as Pi, i=1,2, and 3,4,5,6；The voltage peak area is electricity during single detects sample introduction
Pressure response rises, the opposite voltage difference of baseline value before sample introduction of the output voltage of sensor it is cumulative with；The voltage maximum is just
Slope is the gradient maxima of response curve when voltage responsive rises during the sample introduction of single detection；The recovery voltage
Baseline time is that the sampled point since the cleaning process that single detects counts, the voltage of baseline before voltage is dropped to and detected
When, then it is assumed that cleaning is complete, using the number of sampled point at this time as the time for restoring voltage baseline；The temperature profile value and wet
Spend characteristic value are as follows: the Temperature Humidity Sensor numerical value extracted in sensor peak value, two sampled points adjacent with pre-and post-peaking
Temperature Humidity Sensor numerical value carries out at 3 points and averages, and obtains final temperature characteristic value and Humidity Features value.
Baseline value before the sample introduction of sampling and voltage responsive peak value are handled, i-th of sensor electricity of peak value of response moment is obtained
Resistance with sample introduction before baseline when sensor resistance ratio, i.e. resistance ratio is denoted as Ri, formula is as follows:
(3.2) characteristic value that sensor is acquired and handled is as training set, BP ANN, specifically: BP people
The input layer of artificial neural networks 26, intermediate hidden neuron 9, output layer neuron 3,3 output layer nerves
Member respectively represents the gas concentration value of carbon monoxide, methane and formaldehyde, each neuron of input layer and middle layer it is initial
Weight and threshold value generate at random, and loss function uses mean square error, are the concentration calculation value and actual value of three kinds of gas to be detected
Mean square error and.
Input layer-middle layer activation primitive is tansig, therefore the output mid_output of each middle layer neuron can be under
Formula calculates.Wherein, n indicates the multiple inputs and obtained result cumulative after its respectively weight, threshold calculations of single neuron.
Mid_output=tansig (n)=2/ (1+exp (- 2*n)) -1
And middle layer-output layer activation primitive is purelin, therefore the output output of each output layer neuron can be by following formula
It calculates.
Output=purelin (n)=n
By the ratio of the sensor resistance before the sensor resistance and sample introduction of 6 voltage responsive peak value moments when baseline, 6 voltages
Maximal positive slope, 6 voltage peak areas, 6 recovery voltage baseline times, 1 temperature profile value and 1 Humidity Features value conduct
Training set is input in the BP artificial neural network of single-chip minimum system module (6) loading, passes through Regularization algorithms
It is trained BP artificial neural network, finally the weight and threshold value of determining input layer and each neuron of middle layer, obtain
To trained BP artificial neural network.
(3.3) 26 characteristic values of the gas to be detected of sensor acquisition are input in BP artificial neural network, by BP people
Work nerve neural computing exports 3 kinds of gas concentrations to be detected, and calculated result is sent to by serial port module (7)
Position machine.
(4) state is cleaned
After the completion of detection, normal air is passed through detection gas chamber with the constant flow rate of 1000 ml/mins by sample introduction air pump, and cleaning passes
Sensor array 10 minutes or more, until each sensor restores voltage baseline value hereinafter, then repeating step (3.3) and (4)
It can continue sample introduction and detect, until artificially stopping detecting.
9. according to the method described in claim 8, it is characterized in that, the single-chip minimum system module (6) can also be loaded into
The BP artificial neural network of classical Principal Component Analysis Algorithm optimization, classical Principal Component Analysis Algorithm is for BP artificial neural network
26 input feature vector values carry out pre-optimized processing, calculate the covariance matrix of former input, and characteristic root is principal component variance, will
Characteristic root arranges from big to small, takes coefficient of the feature vector as principal component corresponding to preceding 4 characteristic roots, according to coefficient with it is defeated
4 principal component F can be obtained in 26 characteristic values enteredi=ai0+ai1x1+ai2x2+…+ai26x26, i=1,2,3,4 is as new feature
It is worth, wherein aijFor the coefficient being calculated, i=1,2,3,4, j=1,2 ... 26, x1~x26For 26 characteristic values of input.At this time
BP artificial neural network is input layer 4, intermediate hidden neuron 9, output layer neuron 3, by 4 principal components
As training set, by Regularization algorithms BP ANN, determining input layer and middle layer it is each
The weight and threshold value of a neuron obtain trained BP artificial neural network, for sample introduction and detection.
10. according to the method described in claim 8, it is characterized in that, the single-chip minimum system module (6) can also be loaded into
The BP artificial neural network of classical partial least squares algorithm optimization regard 26 characteristic values input in training set as independent variable,
It regard 3 kinds of gas actual concentrations outputs to be detected as dependent variable, classical partial least squares algorithm is respectively in independent variable and dependent variable
Middle extract component t1 and u1, t1 are the linear combination of independent variable, and u1 is the linear combination of dependent variable, make the association side of both t1 and u1
Difference reaches maximum.Note t1 and u1 is first composition.Continued to extract second composition with the residual error for having extracted first composition, repeats the step
Suddenly, first four ingredient is taken, while obtaining the t in four ingredientsiWith the coefficient a between each input independent variableij, i=1,2,3,
4, j=1,2 ... 26.The four groups of coefficients obtained according to classical Partial Least Squares are special for 26 inputs of BP artificial neural network
Value indicative carries out pre-optimized processing, calculates four ingredient Fi=ai0+ai1x1+ai2x2+…+ai26x26, i=1,2,3,4 is as new
Characteristic value, wherein x1~x26For 26 characteristic values of input.BP artificial neural network is input layer 4 at this time, intermediate
Hidden neuron 9, output layer neuron 3, using 4 ingredients as training set, pass through Regularization algorithms training BP
Artificial neural network determines the weight and threshold value of each neuron of input layer and middle layer, it is artificial to obtain trained BP
Neural network, for sample introduction and detection.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910848479.7A CN110426421B (en) | 2019-09-09 | 2019-09-09 | Device and method for detecting multi-component harmful gas in kitchen environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910848479.7A CN110426421B (en) | 2019-09-09 | 2019-09-09 | Device and method for detecting multi-component harmful gas in kitchen environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110426421A true CN110426421A (en) | 2019-11-08 |
| CN110426421B CN110426421B (en) | 2020-10-16 |
Family
ID=68418842
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910848479.7A Active CN110426421B (en) | 2019-09-09 | 2019-09-09 | Device and method for detecting multi-component harmful gas in kitchen environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110426421B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112819158A (en) * | 2021-02-05 | 2021-05-18 | 凌坤(南通)智能科技有限公司 | Gas identification method based on optimized BP neural network |
| CN113537334A (en) * | 2021-07-09 | 2021-10-22 | 长安大学 | Gas-liquid multi-component real-time intelligent detection method |
| CN114002303A (en) * | 2021-12-31 | 2022-02-01 | 中国农业科学院农业资源与农业区划研究所 | Calibration method for gas sensing in cold-chain logistics and multi-source sensing device |
| CN114965872A (en) * | 2022-04-27 | 2022-08-30 | 重庆科技学院 | Multi-sensor data fusion electronic nose and method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101000318A (en) * | 2006-12-28 | 2007-07-18 | 中国科学院合肥物质科学研究院 | Sensor array based on temperature control and gas detection method |
| CN202854118U (en) * | 2012-10-27 | 2013-04-03 | 西安科技大学 | Embedded networking intelligent multi-parameter gas detection system |
| CN205562572U (en) * | 2016-04-14 | 2016-09-07 | 怀化学院 | Portable sample quickly pretreatment device for monitoring environment and food |
| CN106289394A (en) * | 2016-08-04 | 2017-01-04 | 苏州云白环境设备股份有限公司 | A kind of Wearable real time environment gas controlling device and monitoring method thereof |
| CN106770738A (en) * | 2016-12-03 | 2017-05-31 | 浙江大学 | The expiratory air multi-analyte immunoassay instrument and detection method of a kind of gas concentration lwevel amendment |
-
2019
- 2019-09-09 CN CN201910848479.7A patent/CN110426421B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101000318A (en) * | 2006-12-28 | 2007-07-18 | 中国科学院合肥物质科学研究院 | Sensor array based on temperature control and gas detection method |
| CN202854118U (en) * | 2012-10-27 | 2013-04-03 | 西安科技大学 | Embedded networking intelligent multi-parameter gas detection system |
| CN205562572U (en) * | 2016-04-14 | 2016-09-07 | 怀化学院 | Portable sample quickly pretreatment device for monitoring environment and food |
| CN106289394A (en) * | 2016-08-04 | 2017-01-04 | 苏州云白环境设备股份有限公司 | A kind of Wearable real time environment gas controlling device and monitoring method thereof |
| CN106770738A (en) * | 2016-12-03 | 2017-05-31 | 浙江大学 | The expiratory air multi-analyte immunoassay instrument and detection method of a kind of gas concentration lwevel amendment |
Non-Patent Citations (4)
| Title |
|---|
| XIYANG SUN等: "An optimized multi-classifiers ensemble learning for identification ofginsengs based on electronic nose", 《SENSORS AND ACTUATORS A: PHYSICAL》 * |
| YOU WANG等: "Discrimination of Different Species of Dendrobium with an Electronic Nose Using Aggregated Conformal Predictor", 《SENSORS》 * |
| 王敏等: "基于智能电子鼻的冰箱冷藏食品新鲜度原位检测技术", 《传感技术学报》 * |
| 高凡等: "穿戴式实时环境气体监测系统", 《中国生物医学工程学报》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112819158A (en) * | 2021-02-05 | 2021-05-18 | 凌坤(南通)智能科技有限公司 | Gas identification method based on optimized BP neural network |
| CN112819158B (en) * | 2021-02-05 | 2024-02-27 | 凌坤(南通)智能科技有限公司 | Gas identification method based on optimized BP neural network |
| CN113537334A (en) * | 2021-07-09 | 2021-10-22 | 长安大学 | Gas-liquid multi-component real-time intelligent detection method |
| CN114002303A (en) * | 2021-12-31 | 2022-02-01 | 中国农业科学院农业资源与农业区划研究所 | Calibration method for gas sensing in cold-chain logistics and multi-source sensing device |
| CN114002303B (en) * | 2021-12-31 | 2022-04-05 | 中国农业科学院农业资源与农业区划研究所 | Calibration method for gas sensing in cold-chain logistics and multi-source sensing device |
| CN114965872A (en) * | 2022-04-27 | 2022-08-30 | 重庆科技学院 | Multi-sensor data fusion electronic nose and method |
| CN114965872B (en) * | 2022-04-27 | 2023-10-13 | 重庆科技学院 | Electronic nose and method for multi-sensor data fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110426421B (en) | 2020-10-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110426421A (en) | Multi-component harmful gas detection device and detection method in a kind of kitchen environment | |
| CN103630851B (en) | A kind of LED radiating module entire thermal resistance measuring method and measuring system | |
| CN105911228B (en) | A kind of portable electric nose system and its control method | |
| Walsby | A portable apparatus for measuring relative gas vacuolation, the strength of gas vacuoles, and turgor pressure in planktonic blue‐green algae and bacteria 1 | |
| CN207601552U (en) | air quality integrated detecting device | |
| CN108469455A (en) | A kind of design and method of the portable electric nose system differentiated for strawberry freshness | |
| CN104965009B (en) | A kind of Hydrogen Concentration Detector | |
| Qian et al. | Design of multisensor electronic nose based on conformal sensor chamber | |
| CN108918602A (en) | A kind of semiconductor gas sensor Performance Test System and test method | |
| Qian et al. | A novel multisensor detection system design for low concentrations of volatile organic compounds | |
| CN203758946U (en) | Device capable of simultaneously determining combustion heat and combustion rate of cigarette | |
| CN106770986A (en) | A kind of embedded electronic nose detecting system for detecting meat quality | |
| CN107607436A (en) | A kind of cigarette dynamic resistance to suction computational methods based on standard aspiration pattern | |
| Tzeng et al. | Design and implement a cost effective and ubiquitous air quality monitoring system based on ZigBee wireless sensor network | |
| CN202083655U (en) | Automatic calibration device for catalytic-combustion-type methane gas detector | |
| CN105891431A (en) | Linseed oil quality detection method | |
| CN210465319U (en) | Thermal conductivity gas analyzer | |
| CN106443031B (en) | Bionic olfactory detection and analysis device and its determination method based on dynamic air-distributing | |
| CN206074625U (en) | Bionic olfactory detection and analysis device based on dynamic air-distributing | |
| CN103558265A (en) | Automatic testing system with gas sensitive element | |
| CN101216480A (en) | Microenvironment humidity analog machine for textile anion test | |
| CN107976469A (en) | A kind of soil nutrient device for fast detecting based on Artificial Olfactory | |
| CN115112722A (en) | Gas sensor integrated system and method based on dynamic temperature modulation | |
| Rogers et al. | Feedback-enabled discrimination enhancement for temperature-programmed chemiresistive microsensors | |
| CN106353374A (en) | Equipment for environment humidity detection |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |