Electronic nose system for identifying mixed gas based on sensing array and application

Abstract

The invention relates to an electronic nose system for identifying mixed gas based on a sensing array and application. The electronic nose system is composed of a gas distribution system, a sensing array and a signal module, wherein the gas distribution system is composed of dynamic gas distribution equipment and a gas source; the sensing array is composed of a plurality of sensors; and the signal module carries out processing and mode identification on the signal through the upper computer development system. And the electronic nose system obtains response-recovery curve data and performs further processing and pattern recognition on the response-recovery curve data, so that accurate recognition of the low-concentration ethanol, ammonia gas and toluene mixed gas is realized. According to the invention, the mixed gas of low-concentration ethanol, ammonia gas and toluene can be accurately identified, the gas concentration can be accurately judged, and the method has a potential application prospect in monitoring ethanol, ammonia gas and toluene gas in environments such as atmosphere, chemical plants and homes.

Description

Technical Field

[0001] The present invention belongs to the technical field of gas identification, and in particular relates to an intelligent electronic nose system and application thereof, which is based on a combination of an eight-channel sensor array of Co3O4 and SnO2 oxides and is used for identifying low-concentration ethanol, ammonia and toluene gases and mixed gases. Background Art

[0002] There are many polluting gases in the atmosphere, such as ethanol, ammonia, formaldehyde, toluene, acetone, etc., which seriously threaten human health and environmental quality. However, traditional detection methods, such as calorimetry, gas chromatography, gas chromatography-mass spectrometry, selective ion flow tube mass spectrometry, etc., are costly and inconvenient to detect gaseous pollutants. In the development of today's technology, electronic nose systems have been widely used in air quality monitoring, food testing, breath detection and other fields. They can monitor gaseous pollutants in real time and have the advantages of low cost, fast response speed and simple structure. The electronic nose system consists of three parts: gas distribution system, sensor array and signal module, which can accurately identify the target gas. Among them, the sensor array is the core part of the electronic nose system. A single sensor has selectivity problems for similar gases in chemical or physical properties, and the combination of multiple sensors is easier to obtain the response characteristics of the target gas. Among many sensors, sensors made of metal oxide semiconductor materials have the advantages of portability, easy production and miniaturization. Therefore, researchers optimize the sensor array by exploring metal oxide sensors to improve the recognition accuracy of the electronic nose system for the target gas. Therefore, developing and designing a high-performance sensor array is of great significance for the intelligent electronic nose system.

[0003] The electronic nose device achieves accurate identification of the target gas by coupling the sensor array with the algorithm model, which improves the electronic nose system's ability to extract features and recognize patterns, reduces the impact of cross-sensitivity on gas recognition, and thus improves the ability to recognize mixed gases of ethanol, ammonia, and toluene. In the process of identifying target gases by the electronic nose system, the influence of the feature value extraction method and the machine learning method needs to be further considered. Therefore, the selection of feature extraction algorithms and machine learning algorithms in the signal module is crucial to ensure the model's effective generalization ability for new data and the effective and accurate identification of the feature information of mixed gases.

[0004] Through the above analysis, the problems and defects of the prior art are as follows: (1) Traditional gas detection methods are costly and inconvenient. Existing electronic nose technology can usually only accurately identify a single gas. However, there are many gases in the atmosphere, which may make it difficult to identify characteristic gases in specific application scenarios, delaying the development and application of electronic nose systems.

[0005] (2) The chemical characteristics and spatial structures of VOC gases are similar, especially in the case of low-concentration mixed gas pollutants. At the same time, it is more difficult to identify the mixed components of inorganic gases and VOC gases, and current research still faces challenges in accurately identifying mixed pollutants. Summary of the invention

[0006] In view of the problems existing in the existing electronic nose system, the present invention provides an intelligent electronic nose system and application for mixed gas recognition based on a sensor array of Co3O4 and SnO2 oxides. The system has good recognition ability for low-concentration ethanol, ammonia and toluene mixed gases.

[0007] In order to achieve the above object, the technical solution of the present invention is as follows: (1) Build a gas distribution system, use intelligent dynamic gas distribution equipment, provide target gas through gas cylinders, and simulate mixed gas.

[0008] (2) Design and develop a high-performance sensor array, prepare Co3O4 and Mn-Co3O4 sensors, and combine them with TGS2620 (SnO2), TGS2600 (SnO2), TGS2609 (SnO2) and TGS2602 (SnO2) sensors to construct an eight-channel sensor array.

[0009] (3) The signal acquisition module collects the original response-recovery curve through the host computer development system, selects characteristic values ​​from the response-recovery curve, and further standardizes the data.

[0010] (4) The data set is randomly divided into a training set and a test set. The training set is trained using KNN, SVM, LR and RF algorithms to obtain a gas type and concentration recognition model. The gas type and concentration recognition model is then tested using the test set.

[0011] Furthermore, the preparation method of the Co3O4 sensing material in step (2) is: dissolving Co(NO3)2·6H2O in deionized water, then slowly dripping N2H4·H2O into the Co(NO3)2 solution, washing with deionized water and ethanol and drying to obtain a green product, and then calcining in a muffle furnace to obtain the Co3O4 material.

[0012] Furthermore, the preparation method of the Mn-Co3O4 sensing material in step (2) is: dissolving Co(NO3)2·6H2O, Mn(NO3)2·4H2O and HMT in deionized water to obtain a uniform solution, transferring the uniform solution to a polytetrafluoroethylene reactor for hydrothermal reaction, and then washing and drying the obtained green product, and calcining it in a muffle furnace to obtain the Mn-Co3O4 material.

[0013] Furthermore, the preparation method of the Co3O4 and Mn-Co3O4 sensors in step (2) is as follows: the synthesized Co3O4 and Mn-Co3O4 sensing materials are dispersed in deionized water, and ultrasonically formed into a uniform solution. Then, the electrode substrate is welded to a metal bracket (9.4×0.6×4 mm). Finally, the dispersed solution is dripped onto the surface of the 1×1.5 mm electrode substrate to obtain the Co3O4 and Mn-Co3O4 sensors.

[0014] Furthermore, the characteristic values ​​include maximum resistance, response value, curve integral area, response time, response recovery time, rise time, stabilization time, baseline resistance, response resistance, average resistance ratio and slope.

[0015] Furthermore, the maximum resistance value is the maximum resistance value of the response curve.

[0016] Further, the response value is obtained by the following formula: Response = (n-type) / (p-type).

[0017] Furthermore, the integral area of ​​the curve is calculated using the Newton-Cotes quadrature formula: .

[0018] in, Indicates that the integral area is divided equally into equal parts; is the integration step length, which is the distance between the horizontal coordinates of two adjacent points on the curve; and The first and Function value.

[0019] Furthermore, the response time is the time required for the resistance to decrease or increase to 90% of a stable value when the gas sensor detects gas.

[0020] Furthermore, the response recovery time is the time required for the ventilation balance state of the gas sensor to recover to a 10% signal.

[0021] Furthermore, the rise time is the time it takes for the resistance of the gas sensor to rise.

[0022] Furthermore, the stabilization time is the time during which the response curve of the gas sensor maintains equilibrium.

[0023] Furthermore, the baseline resistance is the average resistance of the gas sensor in a balanced state in the air.

[0024] Furthermore, the response resistance is an average resistance of the gas sensor in a state of equilibrium in the target gas.

[0025] Furthermore, the average resistance ratio is a ratio of a response resistance to a baseline resistance.

[0026] Further, the slope is a slope of 10-100% in the response rising stage and a slope of 10-100% in the response recovery falling stage.

[0027] Furthermore, the data standardization process in step (3) adopts the following formula: .

[0028] in, X ij For sample j Standardized features of i , X ij For sample j Features i , μ i Features i The mean of σ i Features i The standard deviation of .

[0029] Compared with the prior art, the advantages and positive effects of the present invention are not only profound but also multi-dimensional, which are specifically reflected in the following aspects: 1. Innovative design of intelligent electronic nose system.

[0030] The core of the present invention is its innovative design of a highly intelligent electronic nose system. This system integrates advanced sensor technology, signal processing algorithms and artificial intelligence learning mechanisms to achieve accurate identification of harmful or critical gases such as low-concentration ethanol, ammonia and toluene and their mixed gases. In low-concentration environments, traditional methods are often difficult to effectively distinguish due to insufficient sensitivity, while the intelligent electronic nose of the present invention can keenly capture and analyze the subtle features of these gases, providing good accuracy and reliability for environmental monitoring, industrial safety, medical health and other fields.

[0031] 2.Excellent selectivity of Co3O4 and SnO2 oxide sensor arrays.

[0032] The sensor array based on Co3O4 and SnO2 oxides developed by the present invention has demonstrated extraordinary selective recognition capabilities for ethanol, ammonia and toluene gases by virtue of its unique material properties and structural optimization. These two oxide materials not only have excellent chemical stability and gas sensitivity on their own, but their combined use greatly enhances the system's recognition capability for target gases through synergistic effects, effectively reduces cross-interference, and ensures the accuracy of the detection results.

[0033] 3. Significant improvement in feature value selection and recognition accuracy.

[0034] In the data processing stage, the present invention screened and extracted a series of highly representative characteristic values ​​by deeply studying the interaction mechanism between gas molecules and the sensor response characteristics. These characteristic values ​​not only fully reflect the subtle differences in gas types and concentrations, but also significantly improve the recognition accuracy of mixed gas characteristic information through complex mathematical models and algorithm optimization. This breakthrough enables the system to maintain stable detection performance even under complex and changing environmental conditions.

[0035] 4. Powerful empowerment of machine learning algorithms.

[0036] To ensure the model's ability to effectively generalize to new data, the present invention introduces advanced machine learning algorithms. These algorithms can not only effectively learn and memorize the characteristic patterns of known gases, but also continuously adapt to new gas samples and detection environments through continuous learning and self-optimization. This powerful adaptive ability enables the gas detection system of the present invention to provide strong technical support for future gas detection tasks.

[0037] 5. Broad application prospects and significant economic benefits.

[0038] The present invention shows great application potential in many fields such as air quality monitoring, food safety testing, and respiratory health analysis. In terms of air quality monitoring, it can monitor and warn of harmful gas exceeding the standard in real time to protect public health; in the field of food testing, it can accurately identify volatile organic compounds in food to ensure food safety; in medical respiratory testing, it can provide an important basis for the early diagnosis and treatment of respiratory diseases. These applications not only improve the quality and safety of social life, but also bring good economic benefits to related industries by improving detection efficiency and accuracy, and promote the technological progress of gas detection systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Schematic diagram of the intelligent electronic nose system of the present invention.

[0040] Figure 2The scanning electron microscope pictures and transmission electron microscope pictures of Co3O4 and Mn-Co3O4 materials of the present invention are shown in FIG.

[0041] Figure 3 The X-ray diffraction diagram and Raman spectrum diagram of Co3O4 and Mn-Co3O4 materials of the present invention are shown.

[0042] Figure 4 It is the XPS spectrum diagram of Co3O4 and Mn-Co3O4 materials of the present invention.

[0043] Figure 5 The figure is a response curve diagram of the sensor array of the present invention to ethanol, ammonia and toluene at different concentrations.

[0044] Figure 6 This is the original response-recovery curve diagram of the present invention.

[0045] Figure 7 This is a diagram showing the recognition accuracy of the intelligent electronic nose system for low-concentration ethanol, ammonia and toluene gases and their mixed gases under different algorithms of the present invention. DETAILED DESCRIPTION

[0046] In order to better illustrate the present invention and facilitate understanding of the technical solution of the present invention, the specific embodiments described in the present invention are only used to explain the present invention and are not used to limit the present invention.

[0047] Example 1.

[0048] Build an intelligent electronic nose system.

[0049] Schematic diagram of the electronic nose system and experimental equipment Figure 1 As shown in the figure. The system mainly consists of a gas distribution system, a sensor array and a signal module. The gas distribution system adopts an intelligent dynamic gas distribution device (Beijing BiDa Electronic Technology Co., Ltd.), which provides the target gas through a gas cylinder to simulate the mixed gas. The gas is diluted with zero-order air at a flow rate of 300 mL / min. The gas sensor array is interconnected with the gas chamber, which has a size of 105×40×36 mm, a volume of 30 mL, and a gas inlet and outlet. In addition, a total of 8 gas sensors are assembled and connected to the signal conditioning board. The heating output voltage range of the conditioning board is 0-5V. During the experiment, the working voltage is 5 V and the heating voltage is 3 V. The signal module collects the original response-recovery curve through the host computer development system, and further processes and performs pattern recognition on it.

[0050] Example 2.

[0051] Preparation of Co3O4 and Mn-Co3O4 materials.

[0052] 587.9 mg Co(NO3)2·6H2O was weighed with an analytical balance and dissolved in 30 mL deionized water. The mixture was stirred for 5 min under magnetic stirring. Then 365.0 μL N2H4·H2O was slowly dripped into the Co(NO3)2 solution and stirred at room temperature for 6 h. The green product was washed with deionized water and ethanol for 6 times. The product was then dried in a vacuum drying chamber at 60°C for 12 h and finally heated in a muffle furnace at 500°C at 5°C·min -1 The heating rate was calcined for 2 h to obtain Co3O4 material. 1425.8 mg Co(NO3)2·6H2O, 38.5 mg Mn(NO3)2·4H2O and 2124.1 mg HMT were weighed and dissolved in 50 mL deionized water using an analytical balance. The reaction solution was magnetically stirred for 10 min. The reaction solution was transferred to a 100 mL stainless steel autoclave and heated at 90 °C for 6 h. The green product was centrifuged and washed with deionized water and ethanol 6 times. The dried product was dried at 60 °C for 12 h and then heated in a muffle furnace at 3 °C·min -1 The heating rate was increased and annealed at 450℃ for 2 h to obtain Mn-Co3O4 material.

[0053] Figure 2 It can be seen that the prepared Co3O4 and Mn-Co3O4 materials exhibit nano-granular and nano-sheet structural characteristics, respectively; Figure 3 and Figure 4 It can be seen that X-ray diffraction patterns, Raman and XPS characterizations indicate that Co3O4 and Mn-Co3O4 materials were successfully prepared.

[0054] Example 3.

[0055] Design and construct metal oxide sensor arrays.

[0056] 2 mg of synthesized Co3O4 and Mn-Co3O4 sensing materials were dispersed in 200 µL of deionized water and ultrasonicated for 10 min to form a uniform solution. Then, the electrode substrate was welded to a metal bracket (9.4×0.6×4 mm). Finally, 0.5 µL of the dispersed solution was dropped onto the surface of the 1×1.5 mm electrode substrate and naturally air-dried for 30 min to obtain Co3O4 and Mn-Co3O4 sensors. Commercial sensors TGS2620 (SnO2), TGS2600 (SnO2), TGS2609 (SnO2) and TGS2602 (SnO2) were selected to combine with Co3O4 and Mn-Co3O4 sensors to construct a sensing array.

[0057] Example 4.

[0058] The response of the sensor array to ethanol, ammonia, and toluene at different concentrations.

[0059] At the optimal heating voltage of 5 V, different concentrations of ethanol, ammonia and toluene gases were introduced into the intelligent electronic nose system. The test results are as follows: Figure 5 As shown, the sensor array has excellent selectivity and sensitivity to ethanol, ammonia and toluene gases.

[0060] Example 5.

[0061] Extraction of characteristic values ​​from response-recovery curves.

[0062] The characteristic values ​​of the present invention include maximum resistance, response value, curve integral area, response time, response recovery time, rise time, stabilization time, baseline resistance, response resistance, average resistance ratio and slope. The characteristic values ​​are obtained from the original response-recovery curve, such as Figure 6 shown.

[0063] Example 6.

[0064] Recognition of low-concentration ethanol, ammonia, toluene and their mixed gases by the intelligent electronic nose system.

[0065] Figure 7 It can be seen that under the conditions of mono- and binary mixed gases of different concentrations of ethanol, ammonia and toluene, the average accuracy of the KNN, SVM, LR and RF model algorithms after five cross-validations were 97.2%, 91.6%, 86.1% and 88.8% respectively. Under mono-gas, the accuracy of the KNN model was the highest, reaching 100%, the accuracy of the SVM and LR models both reached 95.8%, and the accuracy of the RF model was 87.5%. The intelligent electronic nose system has high accuracy in realizing low-concentration gas recognition, and therefore has broad application prospects in advanced gas detection technology.

[0066] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. An electronic nose system for mixed gas recognition based on a sensor array, characterized in that: The intelligent electronic nose system consists of three parts: gas distribution system, sensor array and signal module.

2. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 1, characterized in that: The mixed gas includes ethanol, ammonia and toluene gases, and the concentrations of the ethanol, ammonia and toluene gases are 2 ppm, 5 ppm, 7 ppm and 10 ppm respectively.

3. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 1, characterized in that: The mixed gas combination is ethanol and ammonia gas, ethanol and toluene gas, ammonia gas and toluene gas.

4. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 1, characterized in that: The sensor array described above is an eight-channel array consisting of TGS2620 (SnO2), TGS2600 (SnO2), TGS2609 (SnO2), TGS2602 (SnO2), Co3O4 and Mn-Co3O4 sensors.

5. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 1, characterized in that: The signal module comprises the following steps: (1) Collect the original response-recovery curve through the host computer development system; (2) Selecting characteristic values ​​from the response-recovery curve and further standardizing the data; (3) Divide the data set into training set and test set; (4) KNN, SVM, LR and RF algorithms are selected to train the training set to obtain a gas type and concentration recognition model, and the gas type and concentration recognition model is tested using a test set.

6. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 4, characterized in that: The preparation method of Co3O4 sensing material comprises the following steps: (1) Dissolve 2 mmol Co(NO3)2·6H2O in 30 mL deionized water and stir magnetically for 5 min. Then, slowly drop 6 mmol N2H4·H2O (80%) into the Co(NO3)2 solution and stir at room temperature for 6 h. (2) washing with deionized water and ethanol to obtain a green product, which was then dried in an oven at 60°C; (3) The green product was placed in a muffle furnace at 5°C min -1 The material was heated to 500 °C at a heating rate of 100 °C and calcined for 2 h to obtain Co3O4 material.

7. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 4, characterized in that: The preparation method of Mn-Co3O4 sensing material is characterized by comprising the following steps: (1) Dissolve 4.85 mmol Co(NO3)2·6H2O, 0.15 mmol Mn(NO3)2·4H2O, and 15 mmol hexamethylenetetramine (HMT) in 50 mL deionized water to form a uniform mixed solution; (2) transferring the uniformly mixed solution to a polytetrafluoroethylene reactor for hydrothermal reaction, washing with deionized water and ethanol to obtain a green product, and then drying it in an oven at 60° C.; (3) The green product was placed in a muffle furnace at 3°C ​​min -1 The material was heated to 450 °C at a heating rate of 100 °C and calcined for 2 h to obtain Mn-Co3O4 material.

8. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 5, characterized in that: The characteristic values ​​include maximum resistance, response value, curve integral area, response time, response recovery time, rise time, stabilization time, baseline resistance, response resistance, average resistance ratio and slope.

9. The electronic nose system for mixed gas identification based on a sensor array as claimed in claim 5, characterized in that: Standardization The following formula is used to standardize the eigenvalues; in, X ij For sample j Standardized features of i , X ij For sample j Features i , μ i Features i The mean of σ i Features i The standard deviation of .

10. Application of the electronic nose system for mixed gas recognition based on a sensor array according to any one of claims 1 to 9 in air quality monitoring, food safety testing, and respiratory health analysis.

Images