Gas monitoring device and method for multiple components of VOCs (Volatile Organic Compounds)
By designing a device that includes a semiconductor-type gas sensor array and a data sampling and processing motherboard, the problem of complexity and high cost of VOCs gas detection in the prior art is solved, real-time identification and quantitative detection of multi-component VOCs gas is realized, and the The accuracy and stability of the detection are suitable for a variety of application scenarios.
Patent Information
- Application Number
- CN202410877364.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-02
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-07-02
AI Technical Summary
The existing gas monitoring system is difficult to detect multi-component VOCs gas in real time and accurately, and the sensors have poor selectivity and poor stability, resulting in complex detection and high cost, making it difficult to be applicable to practical application scenarios.
A device including a gas sensor array, a data sampling processing motherboard and a gas inlet circuit was designed. The interconnection of the semiconductor gas sensor array and the data sampling processing motherboard is achieved through microcontrollers and neural networks to realize real-time gas type identification and Concentration calculation simplifies the detection process and reduces costs.
The multi-component qualitative identification and quantitative detection of VOCs gas is realized, which improves the accuracy and stability of detection, reduces maintenance costs, and is suitable for applications in different industries and scenarios.
Smart Images

Figure CN118897045A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of VOCs multi-component gas monitoring, and in particular, relates to a gas monitoring device and method for VOCs multi-components. Background Art
[0002] At present, the application fields of real-time gas monitoring systems are constantly expanding, covering industrial safety, environmental monitoring, indoor air quality monitoring and other fields. Traditional gas detection methods such as gas chromatography, mass spectrometry, infrared spectroscopy absorption, chemical absorption and other methods have high accuracy in qualitative and quantitative analysis, but the detection process is complicated and not real-time, which is not suitable for the current demand for gas monitoring devices;
[0003] With the development of detection technology, various gas sensors are applied to gas monitoring devices. Commonly used sensor-based detection methods include semiconductor sensors, electrochemical sensors and various optical sensors. Although these sensors have a fast response speed and meet the real-time requirements, they have problems such as poor selectivity, poor stability and poor anti-interference. In addition, there are few studies on high-precision, anti-interference miniaturized gas monitoring devices based on sensors. Limited by the defects of existing gas sensor technology, it is difficult to design sensors with high accuracy and wide applicability.
[0004] Existing gas monitoring systems mainly detect single gases or TVOC, but the detection effect is not ideal. Electrochemical or optical gas sensors are usually used, which easily lead to high complexity and high maintenance costs, which is not conducive to large-scale mass production.
[0005] Limited by the defects of semiconductor gas sensors such as poor selectivity, the current application research on semiconductor gas sensors is still mainly in the laboratory stage, or applied to relatively simple application scenarios such as combustible gas leak alarm monitoring. How to use the working characteristics of semiconductor gas sensors to achieve accurate monitoring of VOCs multi-component gas types and concentrations in actual application scenarios is a key technical problem that needs to be solved. Based on the above content, the present invention proposes a gas monitoring device and method for VOCs multi-components. Summary of the invention
[0006] The purpose of the present invention is to provide a gas monitoring device and method for VOCs multiple components to solve the technical problems that existing VOCs gas monitoring devices and methods are difficult to detect specific VOCs types, have complex detection processes and high detection costs. The device realizes qualitative type identification, quantitative concentration calculation and online monitoring of single and multiple mixed components of VOCs gas. The device is simple to operate and highly integrated, and can be mass-produced at a relatively low cost.
[0007] In order to achieve the above object, the technical solution adopted by the present invention is:
[0008] A gas monitoring device for VOCs multi-components, comprising a gas sensor array, a data sampling and processing mainboard and a gas sampling gas path, wherein the gas sensor array and the data sampling and processing mainboard are connected to each other, the gas sensor array comprises a plurality of detection units, each detection unit comprises a semiconductor gas sensor, each semiconductor gas sensor is connected to a matching resistor and a filter capacitor, and each detection unit leads to a voltage signal source output; the voltage output of each detection unit of the gas sensor array and two input voltages for a heating electrode and a sensitive body electrode of the semiconductor gas sensor are connected to the data sampling and processing mainboard through a flexible flat cable;
[0009] The gas sampling air circuit comprises a diaphragm air pump and a sensor cavity, and the diaphragm air pump is respectively connected to the sampling interface and the sensor cavity through a silicone hose.
[0010] Preferably, the data sampling and processing mainboard includes an integrated data acquisition module, a main control module, a multi-channel data transceiver module and a multi-channel power supply module; the gas sensor array is connected to the data acquisition module, the data acquisition module is connected to the main control module through a serial bus protocol, the main control module is connected to the multi-channel data transceiver module via a serial data transceiver, and is connected to a physical interface to communicate with an external device, and the multi-channel power supply module is respectively connected to the data acquisition module, the main control module, and the multi-channel data transceiver module.
[0011] Preferably, the data acquisition module includes an analog-to-digital conversion chip, a voltage follower and a reference voltage source chip; the voltage signal source connected to the gas sensor array is connected to a multi-channel voltage follower, the output of the voltage follower is connected to the analog-to-digital conversion chip, and the output voltage of the sensor array is measured through the analog-to-digital conversion chip; at the same time, the voltage output of the reference voltage source chip is connected to the analog-to-digital conversion chip via a voltage follower to provide a reference voltage.
[0012] Preferably, the voltage follower is generally formed by connecting multiple operational amplifiers in a specific manner.
[0013] Preferably, the main control module includes a microcontroller and an air pump control circuit, and the microcontroller is connected to the analog-to-digital conversion chip and the air pump control circuit respectively;
[0014] The microcontroller is the core of the gas monitoring device and has the following logical functions:
[0015] 1) Initialize the microcontroller itself and the serial interface connected to the outside;
[0016] 2) Initialize the analog-to-digital conversion chip, perform self-calibration, and select the sampling rate and voltage gain;
[0017] 3) Reading the register sampling value of the analog-to-digital conversion chip, and processing the original sampling value to obtain the collected voltage value data;
[0018] 4) Performing gas detection calculations during the detection of target gas by the device;
[0019] 5) Output the calculated results.
[0020] Preferably, the multi-channel data transceiver module includes a TTL-USB module, a TTL-RS485 module and a TTL-CAN module. The microcontroller is connected to the TTL-USB module, the TTL-RS485 module and the TTL-CAN module through a serial data interface, respectively. The TTL-RS485 module and the TTL-CAN module are connected to the serial port interface on the board via a double-throw switch, and are used to receive data sent by an external expansion device on the serial bus, or to send data to the outside as a slave on the serial bus. In addition, the microcontroller is connected to the wireless device interface on the board to serve as a slave of the wireless gas monitoring equipment network or to communicate directly with the data monitoring platform server.
[0021] Preferably, the multi-channel power supply module includes a power switch power supply chip and a low voltage difference linear regulator. The adjustable voltage output by the switch power supply chip is divided into two paths and connected to the gas sensor array separately through zero-ohm resistors. The voltage output by the linear regulator is connected to each chip on the data sampling and processing mainboard to provide working voltage.
[0022] A gas monitoring method for VOCs multi-components includes the following steps:
[0023] S1, the voltage data output by the sensor array is processed by the data acquisition module and the main control module, and a one-dimensional sequence of voltage values is output after calculation, and the one-dimensional sequence of voltage values is processed to identify and pre-process the baseline voltage and response voltage in real time;
[0024] S2, calculating the response value of each channel sensor according to the baseline voltage, the response voltage and the value, and outputting the calculated response value and the baseline voltage value at the corresponding moment;
[0025] S3. Identify the gas type and calculate the gas concentration based on the output response and baseline voltage data.
[0026] Preferably, the identification and preprocessing of the baseline voltage and the response voltage in S1 specifically include the following steps:
[0027] A1. After the device has been preheated for the sensor for the preset time in the built-in program, it starts to collect voltage. First, the baseline voltage is initialized and the weighted average of the initial voltage values is taken as the starting baseline voltage value (V base );
[0028] A2. After completing the initialization of the baseline voltage, the subsequent sampling voltage (V in ) and the baseline voltage value (V base ) calculates the sensor response value based on the relationship, and judges the calculated response value. If the response value is within the preset threshold range, the sampling voltage (V in ) is determined as the baseline voltage, and the existing baseline voltage value (V base ) is updated by dynamically adjusting the weighted moving average algorithm (EMA) of the weight (A), and the expression is:
[0029] A=L*(V in / V base )
[0030] If V in / V base >1, then:
[0031] A=L*(V base / V in );
[0032] V base =(1-A)*V base +A*V in
[0033] In the formula, L represents the weight adjustment coefficient. The smaller L is, the higher the baseline smoothness is. Its value is determined according to the response characteristics of the sensor itself and is usually between 0.1 and 0.5.
[0034] A3. If the response value exceeds the preset threshold range, the sampled voltage is determined as the response voltage (V res ), and outputs the baseline voltage, response voltage (V res ) to the cache for subsequent calculations.
[0035] Preferably, the calculation of the response value of each channel sensor according to the baseline voltage and the response voltage in S2 specifically includes the following steps:
[0036] B1. Calculate the response value of the sensor based on the change in the voltage value on the sensor before and after the response, reduce the circuit complexity and error caused by the resistance measurement process, and the calculation expression of the sensor response value is:
[0037] r=R air / R gas
[0038] Where r represents the defined sensor response value; R air Indicates the resistance value of the sensor in the air, R gas Indicates the resistance value of the sensor in the target test gas;
[0039] According to the circuit voltage division principle:
[0040] R air =(V s -V base ) / (V s / R l )
[0041] R gas =(V s -V res ) / (V s / R l )
[0042] r=((V s -V base ) / V base )*(V res / (V s -V res ))
[0043] In the formula, R l Indicates the resistance value of the matching resistor to which the sensor is connected; V s Indicates the voltage value output by the power module to the electrode end of the sensor array.
[0044] Preferably, S3 specifically includes the following contents:
[0045] C1. In a specific gas test system, a standard gas sample set is used to calibrate the response characteristics of the sensor array, wherein the standard gas sample set includes air with different humidity and single target gas and mixed target gas with different concentrations. During the calibration process, the sensor array response value output and baseline voltage value output corresponding to each standard gas sample are obtained;
[0046] C2. Take all the response values of the sensor array and the data of the corresponding gas sample types and concentrations as input and output values to build and train the neural network. After the training reaches the target effect, convert the format of the trained model and output the network parameters as a C code array;
[0047] C3. Build a lightweight multi-layer perceptron that runs directly on the main control module of the device, including the input layer, hidden layer, output layer, and network parameters corresponding to each sensor array, so that the main control module can directly perform network operations and obtain gas monitoring results in real time.
[0048] Compared with the prior art, the present invention provides a gas monitoring device and method for VOCs multi-components, which has the following beneficial effects:
[0049] (1) The present invention proposes a gas monitoring device for VOCs multi-components, which is actually a real-time gas online monitoring device, which realizes the qualitative identification and quantitative detection of multi-component VOCs gases, and relies on a sensor array composed of semiconductor gas sensors to identify single or mixed components, making full use of the characteristics of semiconductor gas sensors such as fast response speed, small size, low power consumption, and maintenance-free, and overcoming their defects such as poor selectivity and susceptibility to interference.
[0050] (2) The sensor array proposed in the present invention adopts a separate design, that is, the sensor array and the array sampling and processing mainboard are arranged on different PCBs, and a cavity matching the sensor array is designed based on processing methods such as 3D printing or CNC, and the data output and power supply wiring interface of the sensor array are retained. On the one hand, this design isolates the main circuit part inside the device from the external gas to be tested, thereby improving the stability of the device and making the detection sensitivity more sensitive than natural diffusion sampling; on the other hand, the separate sensor array is convenient for recalibration or replacement of the sensor array after long-term use.
[0051] (3) Based on the gas monitoring device, the present invention further designs a gas monitoring method for VOCs multi-components, realizes automatic and real-time response feature extraction, simplifies the use process of the monitoring device, and can achieve the required functions without complex data preprocessing and algorithm calculation processes on the host computer. The intake air sampling, data processing and calculation, and the output of the test results are realized in one process, which greatly reduces the difficulty of using the device and is conducive to its application in different industries and different scenarios.
[0052] (4) The present invention designs an integrated multi-channel data transceiver module including wired and wireless. Based on the data flow logic design within the program, the proposed device has high scalability and multi-purpose applicability. It can not only work in conjunction with external equipment, but also be used to form a monitoring network, making it suitable for various situations such as single-point monitoring of the park, multi-point distributed monitoring, and mobile monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 This is a schematic diagram of the overall structure of a gas quantitative monitoring device for VOCs multi-components proposed by the present invention;
[0054] Figure 2 This is a structural schematic diagram of a data sampling and processing mainboard of a gas quantitative monitoring device for VOCs multi-components proposed by the present invention;
[0055] Figure 3This is a schematic diagram of the operation flow of the gas monitoring device proposed in Example 1 of the present invention;
[0056] Figure 4 The external picture and internal physical structure diagram of the gas monitoring device proposed in Example 1 of the present invention;
[0057] Figure 5 A field multi-point application test diagram and test results for the monitoring device proposed in Example 1 of the present invention;
[0058] Figure 6 This is a schematic diagram of the original voltage value data results of the test measurement data proposed in Example 2 of the present invention;
[0059] Figure 7 This is a schematic diagram of the result of extracting the response characteristics of the single channel voltage data proposed in Example 2 of the present invention;
[0060] Figure 8 Schematic diagram of the training results of the neural network model under two input parameters proposed in Example 2 of the present invention;
[0061] Fig. 9 It is the accuracy of testing the neural network model trained with the data before drift using the data after the baseline drift proposed in Example 2 of the present invention.
[0062] Description of the numbers in the figure:
[0063] 1. Gas sensor array; 2. Data sampling and processing main board; 3. Gas sampling air path; 4. Diaphragm air pump; 5. Sensor cavity; 6. Semiconductor gas sensor; 7. Matching resistor; 8. Data acquisition module; 9. Main control module; 10. Multi-channel data transceiver module; 11. Multi-channel power supply module; 12. Analog-to-digital conversion chip; 13. Voltage follower; 14. Reference voltage source chip; 15. Microcontroller; 16. Air pump control circuit; 17. TTL-USB module; 18. TTL-RS485 module; 19. TTL-CAN module. DETAILED DESCRIPTION
[0064] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.
[0065] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.
[0066] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0067] In the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0068] Embodiment 1:
[0069] See also Figure 1 ,like Figure 1 As shown, the present invention proposes a gas monitoring device for VOCs multi-components, including a gas sensor array 1, a data sampling and processing mainboard 2 and a gas sampling air path 3. The gas sensor array 1 and the data sampling and processing mainboard 2 are interconnected to measure the electrical signal generated by the semiconductor gas sensor. The gas sampling air path 3 includes a diaphragm air pump 4 and a sensor cavity 5. The diaphragm air pump 4 is respectively connected to the injection interface and the sensor cavity 5 through a silicone hose. In this embodiment, the gas sensor array 1 is placed in the sensor cavity 5, but the gas sensor array 1 should not be regarded as a part of the gas sampling air path 3.
[0070] Specifically, the gas sensor array 1 includes multiple detection units, each detection unit includes a semiconductor gas sensor 6, each semiconductor gas sensor is connected to a matching resistor 7 and a filter capacitor, and each detection unit leads to a voltage signal source output. The voltage output of each detection unit of the gas sensor array 1, and two input voltages for the heating electrode and the sensitive body electrode of the semiconductor gas sensor 6 are connected to the data sampling and processing mainboard 2 through a flexible cable.
[0071] In this embodiment, each semiconductor gas sensor 6 used has four pins, and the detailed circuit connection method is as follows: two pins are connected to the heating resistor of the sensor to provide the required operating temperature for the device; the other two pins are connected to the gas-sensitive resistor of the sensor, and the gas-sensitive resistor pins are connected to the matching resistor. The voltage signal drawn by the detection unit is the voltage divided on the matching resistor. When the resistance value on the sensor decreases, the voltage value will increase. In this embodiment, the main target VOC gases are all reducing gases, and n-type semiconductor sensors are used. When the concentration of the target gas increases, the sensor resistance decreases and the input voltage increases. The change in gas concentration can be intuitively reflected by the voltage change, which improves the ease of use of the device and the efficiency of subsequent data processing.
[0072] Specifically, the data sampling and processing mainboard 2 includes an integrated data acquisition module 8, a main control module 9, a multi-channel data transceiver module 10, and a multi-channel power supply module 11. The gas sensor array 1 is connected to the data acquisition module 8, and the data acquisition module 8 is connected to the main control module 9 through a serial bus protocol. The main control module 9 is connected to the multi-channel data transceiver module 10 via a serial data transceiver, and is connected to a physical interface to communicate with an external device. The multi-channel power supply module 11 is respectively connected to the data acquisition module 8, the main control module 9, and the multi-channel data transceiver module 10.
[0073] like Figure 2 As shown, the detailed composition of the data sampling and processing mainboard 2 of the device is shown.
[0074] The data acquisition module 8 includes an analog-to-digital conversion chip 12, a voltage follower 13, and a reference voltage source chip 14. The voltage signal source connected to the gas sensor array 1 is connected to the multi-channel voltage follower 13, and the output of the voltage follower 13 is connected to the analog-to-digital conversion chip 12; at the same time, the voltage output of the reference voltage source chip 14 is connected to the analog-to-digital conversion chip 12 via the voltage follower 13 to provide a reference voltage.
[0075] In this embodiment, the analog-to-digital conversion chip selects an ADC chip with multiple channels, an accuracy of 16 bits or more, and includes a programmable gain amplifier (PGA), and selects single-ended input and differential input according to the usage scenario, but this embodiment does not limit the type of analog-to-digital conversion method of the ADC chip; the voltage follower used is a voltage follower composed of operational amplifiers connected in a specific way.
[0076] The main control module 9 includes a microcontroller 15 and its minimum peripheral system, and an air pump control circuit 16. The microcontroller 15 is connected to the analog-to-digital conversion chip 12 and the air pump control circuit 16 respectively.
[0077] In this embodiment, the microcontroller (MCU) is the core of the main control module 9, which realizes the control and data processing functions. It should be noted that the microcontroller can also be replaced by other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), and field programmable gate arrays (FPGA) to realize the required control and data processing functions.
[0078] The multi-channel data transceiver module 10 includes a TTL-USB module 17, a TTL-RS485 module 18, and a TTL-CAN module 19. The microcontroller 15 is connected to the TTL USB module 17, the TTL-RS485 module 18, and the TTL-CAN module 19 through a serial data interface respectively. The TTL-RS485 module 18 and the TTL-CAN module 19 are connected to the serial port interface on the board through a double-throw switch, and are used to receive data sent by an external expansion device on the serial bus, or to send data to the outside as a slave on the serial bus. In addition, the microcontroller 15 is connected to the wireless device interface on the board to serve as a slave of the wireless gas detection device network or to communicate directly with the data monitoring platform.
[0079] In some embodiments, a multi-channel data transceiver module can be used to expand external devices. For example, a meteorological data acquisition device can be connected to an external device interface reserved on a data sampling and processing mainboard. By programming the microcontroller to read data from an external serial device, collaborative work of multiple devices can be achieved.
[0080] like Figure 3 The figure shows the working process of the device when performing gas detection.
[0081] The microcontroller in the data sampling and processing mainboard performs the following main logical functions: initializing the microcontroller itself and the serial interface connected to the outside; initializing the analog-to-digital conversion chip; reading the sampling values of the analog-to-digital conversion chip register, and processing the original sampling values to obtain the collected voltage value data; performing gas detection calculations during the target gas detection process of the device; and outputting the calculated results.
[0082] In this embodiment, the initialization program includes the initialization of the microcontroller, the initialization of the data buffer area, the initialization of the microcontroller peripherals, and the initialization and sampling parameter configuration of the ADC, including the configuration of the sampling rate, PGA multiple, and data format; the serial interfaces connected to the microcontroller and the outside include SPI, IIC, CAN, and MODBUS. The serial interfaces used need to be initialized in the initialization program, including the configuration of the interface pin parameters, the interface working mode, the data format, etc.; in addition, the sampling value of the analog-to-digital conversion chip is read sequentially from the sampling value register of each channel, and the data of different address registers are read each time.
[0083] Specifically, the gas detection calculation includes the following steps: judging and calculating the baseline voltage and response voltage of the sampling voltage; calculating the sensor response value; identifying the gas type and calculating the gas concentration.
[0084] The method for judging and calculating the baseline voltage and the response voltage is as follows: after the device preheats the sensor for a pre-set period of time, the voltage collection begins. First, the baseline voltage is initialized, and the weighted average of the initial several voltage values is taken as the starting baseline voltage value (V base ); After completing the initialization of the baseline voltage, the subsequent sampling voltage (V in ) and the baseline voltage value (V base ) calculates the sensor response value based on the relationship, and makes a threshold judgment on the calculated response. If the response value is within the preset threshold range, the sampling voltage (V in ) is classified as the baseline voltage, and the existing baseline voltage value (V base ) is updated by dynamically adjusting the weighted moving average algorithm (EMA) of the weight (A), and the expression is: A = V in / V base , if A>1, then A=1 / A, V base =(1-A)*V base +A*V in If the response value exceeds the preset threshold range, the sampled voltage is classified as the response voltage (V res ), and output the obtained baseline voltage and response voltage values to the cache for subsequent calculations.
[0085] The sensor response value is calculated based on the change in the voltage value on the sensor before and after the response. The specific calculation expression is: r = Rair / R gas Where r is the defined sensor response value, R air is the resistance value of the sensor in the air, R gas is the resistance value of the sensor in the target test gas, obtained by the circuit voltage division principle:
[0086] R air =(V s -V base ) / (V s / R l ), R gas =(V s -V res ) / (V s / R l )
[0087] r=((V s -V base ) / V base )*(V res / (V s -V res ))
[0088] Among them, Rl is the resistance of the matching resistor to which the sensor is connected, VS is the voltage value output by the power module to the gas sensitive resistor end of the sensor array, and Rl is eliminated during the calculation of the sensor response value. The above response value calculation process will be performed on each channel sensor in turn, and the calculated response values of each channel sensor will be stored in the cache for gas type identification and gas concentration calculation.
[0089] In this embodiment, the voltage value VS input to the gas sensitive resistor of the sensor is 5V. For the selection of matching resistors in each detection unit, the output voltage should be between 0.25V and 2.5V under different concentrations of target gas. In this range, the detection error of the ADC chip is small, and there is no need to amplify the voltage signal. The formula for dynamically adjusting the weight A is derived from actual test experience, and its function is to adjust the weight of the new input voltage by the degree of voltage fluctuation. When not in contact with the target gas, the resistance of the gas sensitive resistor of the sensor theoretically remains unchanged, that is, the response value r is 1; but in actual use, the resistance value fluctuates to a certain extent due to changes in environmental conditions and the characteristics of the sensor itself. The normal fluctuation range of the resistance value and the corresponding r range are determined according to the results of actual tests. This range is the judgment threshold of the baseline voltage and the response voltage. The set r range will also affect the lower limit of the detection concentration of the device, which is usually at the ppb level.
[0090] The method for identifying the gas type and calculating the gas concentration is as follows: in a specific gas test system, a standard gas sample is used to calibrate the response characteristics of the sensor array. The standard gas sample set includes air with different humidity and single target gas and mixed target gas with different concentrations. During the calibration process, the sensor array response value output and baseline voltage value output corresponding to each standard gas sample are obtained;
[0091] Taking all the response values of the sensor array and the data of the corresponding gas sample types and concentrations as input and output values, the neural network is constructed and trained. After the training reaches the target effect, the trained model is converted into a format and the network parameters are output as a C code array.
[0092] A lightweight multilayer perceptron is built that can run directly on the main control module of the device, including an input layer, a hidden layer, an output layer, and network parameters corresponding to each sensor array, so that the main control module can directly perform network operations and obtain gas monitoring results in real time.
[0093] In this embodiment, the response value of each sensor is used as an input feature, and the identification of gas types is regarded as a multi-classification problem. If m sensors are used to classify n gases, the input layer should contain m nodes, and the output layer should contain n nodes, but the number of nodes in the hidden layer should be determined according to the actual training situation. In addition to the classification network, the hybrid neural network also includes a regression network for concentration calculation, and its structure is similar to that of the classification network. By inputting groups of data samples with actual gas concentration labels, and then using the loss function to measure the error between the predicted value and the true value, the network parameters are updated through the back propagation algorithm.
[0094] Embodiment 2:
[0095] Based on Example 1, but different in that, the present invention further designs a gas monitoring method for VOCs multi-components based on the proposed monitoring device, including the collection of sensor array outputs, extraction and preprocessing of response characteristics of the collected sequence, calculation of response values of each channel and output of calculation results, identification of gas types and calculation of gas concentrations.
[0096] In this embodiment, a set of data of 5 target VOC gases measured by an 8-channel MOS gas sensor array is subjected to a gas classification and identification process according to a common processing method and the processing method proposed in the present invention. The five VOC gases include styrene, toluene, methyl sulfide, dimethyl disulfide, and carbon disulfide. The sensor array and the monitoring device are placed in a static gas testing system, and a single target gas of 2, 5, 8, and 10 ppm is introduced respectively. The measured original voltage of each channel is connected to a PC host computer through a TTL-USB module to view and save the data.
[0097] like Figure 7 As shown, the raw voltage value data of this group of measurement data reflects the change in the voltage value output by the sensor array during the test of 5 target gases with different concentrations. The baseline voltage and response voltage identification and preprocessing methods are all processed for the raw voltage data shown in the figure.
[0098] like Figure 8 As shown, for Figure 7 The result of response feature extraction of the voltage data of channel 1 is shown to verify the performance of the baseline voltage, response voltage identification and preprocessing method in response feature extraction. In the figure, the original voltage data is represented by a solid line, and the processed voltage and the calculated response value are represented by a dotted line and scattered points respectively. Different judgment threshold ranges will be set for sensors with different response characteristics to achieve the best accuracy of response feature extraction. During the operation of the device, the response feature extraction process is performed on all channels.
[0099] like Fig. 9 As shown, after the obtained multi-channel response value data is output, the original voltage value and the extracted response value are used as the input value of the neural network, and the gas classification result is used as the output value. The feedforward neural network is trained based on MATALAB, where the network parameters are set as: 8 nodes in the input layer, 8 nodes in the hidden layer, and 6 nodes in the output layer (five target gases and air). The figure shows the training results of the model under two inputs.
[0100] In actual use scenarios, the MOS gas sensor used is susceptible to changes in the resistance of the sensor in pure air due to humidity, temperature, airflow, and sensor aging. This is reflected as a drift in the baseline voltage in the monitoring method of this device. To simulate this drift, the original voltage data of each channel is increased or decreased by different amounts, ranging from 0.1 to 0.2 V, and the corresponding features of the newly obtained voltage value data are re-extracted to output new response value data. Using the new voltage value and the new response value as input values and the gas type as output value, the two sets of models trained with the original data are input, and their accuracy is as follows: Fig. 9 shown.
[0101] Comparative verification shows that the model trained by the method proposed in the present invention has high robustness and still has high accuracy in gas classification tests on sensor data after drift.
[0102] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A gas monitoring device for VOCs multi-components, characterized in that: The invention comprises a gas sensor array (1), a data sampling and processing mainboard (2) and a gas sampling gas path (3), wherein the gas sensor array (1) and the data sampling and processing mainboard (2) are connected to each other, the gas sensor array (1) comprises a plurality of detection units, each detection unit comprises a semiconductor gas sensor (6), each semiconductor gas sensor (6) is connected to a matching resistor (7) and a filter capacitor, and each detection unit leads to a voltage signal source output; the voltage output of each detection unit of the gas sensor array (1) and two input voltages respectively used for the heating electrode and the sensitive body electrode of the semiconductor gas sensor (6) are connected to the data sampling and processing mainboard (2) through a flexible flat cable; The gas sampling circuit (3) comprises a diaphragm air pump (4) and a sensor cavity (5), and the diaphragm air pump (4) is connected to the sampling interface and the sensor cavity (5) respectively through a silicone hose.
2. A gas monitoring device for VOCs multi-components according to claim 1, characterized in that: The data sampling and processing mainboard (2) comprises an integrated data acquisition module (8), a main control module (9), a multi-channel data transceiver module (10) and a multi-channel power supply module (11); the gas sensor array (1) is connected to the data acquisition module (8), the data acquisition module (8) is connected to the main control module (9) via a serial bus protocol, the main control module (9) is connected to the multi-channel data transceiver module (10) via a serial data transceiver, and is connected to a physical interface to communicate with an external device, and the multi-channel power supply module (11) is respectively connected to the data acquisition module (8), the main control module (9) and the multi-channel data transceiver module (10).
3. A gas monitoring device for VOCs multi-components according to claim 2, characterized in that: The data acquisition module (8) comprises an analog-to-digital conversion chip (12), a voltage follower (13) and a reference voltage source chip (14); the voltage signal source connected to the gas sensor array (1) is connected to the multi-channel voltage follower (13), and the output of the voltage follower (13) is connected to the analog-to-digital conversion chip (12); the voltage output of the reference voltage source chip (14) is connected to the analog-to-digital conversion chip (12) via the voltage follower (13) to provide a reference voltage.
4. A gas monitoring device for VOCs multi-components according to claim 2, characterized in that: The main control module (9) comprises a microcontroller (15) and an air pump control circuit (16), wherein the microcontroller (15) is connected to the analog-to-digital conversion chip (12) and the air pump control circuit (16) respectively; The microcontroller (15) has the following logic functions: 1) Initializing the microcontroller (15) itself and the serial interface connected to the outside; 2) performing initialization, self-calibration, and selection of sampling rate and voltage value magnification of the analog-to-digital conversion chip (12); 3) reading the register sampling value of the analog-to-digital conversion chip (12), and processing the original sampling value to obtain the collected voltage value data; 4) Performing gas detection calculations during the detection of target gas by the device; 5) Output the calculated results.
5. A gas monitoring device for VOCs multi-components according to claim 2, characterized in that: The multi-channel data transceiver module (10) comprises a TTL-USB module (17), a TTL-RS485 module (18) and a TTL-CAN module (19); the microcontroller (15) is connected to the TTL-USB module (17), the TTL-RS485 module (18) and the TTL-CAN module (19) respectively through a serial data interface; the TTL-RS485 module (18) and the TTL-CAN module (19) are connected to a serial port interface on the board through a double-throw switch, and are used to receive data sent by an external expansion device on the serial bus, or to send data to the outside as a slave on the serial bus; in addition, the microcontroller (15) is connected to a wireless device interface on the board, so as to serve as a slave of a wireless gas monitoring device network or to directly communicate with a data monitoring platform server.
6. A gas monitoring device for VOCs multi-components according to claim 2, characterized in that: The multi-channel power supply module (11) comprises a power switch power supply chip and a low voltage difference linear regulator. The adjustable voltage output by the switch power supply chip is divided into two paths and connected to the gas sensor array (1) separately via a zero-ohm resistor. The voltage output by the linear regulator is connected to each chip on the data sampling and processing mainboard (2) to provide a working voltage.
7. A gas monitoring method for VOCs multi-components implemented by the device according to any one of claims 1 to 6, characterized in that: The following steps are involved: S1, the voltage data output by the sensor array is processed by the data acquisition module (8) and the main control module (9), and a one-dimensional sequence of voltage values is output after calculation, and the one-dimensional sequence of voltage values is processed to perform real-time identification and preprocessing of the baseline voltage and the response voltage; S2, calculating the response value of each channel sensor according to the baseline voltage, the response voltage and the value, and outputting the calculated response value and the baseline voltage value at the corresponding moment; S3. Identify the gas type and calculate the gas concentration based on the output response and baseline voltage data.
8. A gas monitoring method for VOCs multi-components according to claim 7, characterized in that: The identification and preprocessing of the baseline voltage and the response voltage in S1 specifically include the following steps: A1. After the device has been preheated for the sensor for the preset time in the built-in program, it starts to collect voltage. First, the baseline voltage is initialized, and the weighted average of several initial voltage values is taken as the starting baseline voltage value; A2. After completing the initialization of the baseline voltage, the sensor response value is calculated based on the relationship between the subsequent sampling voltage and the baseline voltage value, and the calculated response value is judged. If the response value is within the preset threshold range, the sampling voltage is judged as the baseline voltage, and the existing baseline voltage value in the program is updated through the weighted sliding average algorithm with dynamic adjustment of weights. The expression is: A=L*(V in / V base ) If V in / V base >1, then: A=L*(V base / V in ); In base =(1-A)*V base +A*V in Where A represents the dynamic adjustment weight; V in Indicates the sampling voltage; V base represents the baseline voltage value; L represents the weight adjustment coefficient, the smaller L is, the higher the baseline smoothness is; A3. If the response value exceeds the preset threshold range, the sampled voltage is determined as the response voltage, and the baseline voltage and the response voltage are output to the cache for subsequent calculations.
9. A gas monitoring method for VOCs multi-components according to claim 7, characterized in that: The calculation of the response value of each channel sensor according to the baseline voltage and the response voltage in S2 specifically includes the following steps: B1. Calculate the response value of the sensor based on the change in the voltage value on the sensor before and after the response, reduce the circuit complexity and error caused by the resistance measurement process, and the calculation expression of the sensor response value is: r=R air / R gas Where r represents the defined sensor response value; R air Indicates the resistance value of the sensor in the air, R gas Indicates the resistance value of the sensor in the target test gas; According to the circuit voltage division principle: R air =(V s -V base ) / (V s / R l ) R gas =(V s -V res ) / (V s / R l ) r=((V s -V base ) / V base )*(V res / (V s -V res )) In the formula, R l represents the resistance value of the matching resistor (7) to which the sensor is connected; V s Indicates the voltage value output by the power module to the electrode end of the sensor array.
10. A gas monitoring method for VOCs multi-components according to claim 7, characterized in that: The S3 specifically includes the following contents: C1. In a specific gas test system, a standard gas sample set is used to calibrate the response characteristics of the sensor array, wherein the standard gas sample set includes air with different humidity and single target gas and mixed target gas with different concentrations. During the calibration process, the sensor array response value output and baseline voltage value output corresponding to each standard gas sample are obtained; C2. Take all the response values of the sensor array and the data of the corresponding gas sample types and concentrations as input and output values to build and train the neural network. After the training reaches the target effect, convert the format of the trained model and output the network parameters as a C code array; C3. Build a lightweight multilayer perceptron that runs directly on the main control module (9) of the device, including an input layer, a hidden layer, an output layer, and network parameters corresponding to each sensor array, so that the main control module (9) can directly perform network operations and obtain gas monitoring results in real time.
Citation Information
Patent Citations
Baseline shift adaptive compensation detecting method used for combustible gas detector
CN101482531A
Sensing system and method for the estimation of analyte concentration
CN109781794A
Equipment for monitoring content and components of VOCs (volatile organic compounds) in ambient air
CN115267053A
Photoionization detector sensor performing separation and detection in same physical zone using time division mode
CN117665087A
Gaseous detecting system of VOCs
CN207067059U
Cited By
Integrated multi-sensor electronic nose system
CN119827716A
Electronic nose gas detection system and detection method
CN119861174A
Electronic nose gas analysis system based on Internet of Things technology
CN119936118A
Gas recognition device supporting on-chip training
CN121027421A
Gas sensor, gas monitoring method and robot
CN121068708A
