A VOCs multi-component-oriented gas monitoring device and method

Abstract

The application discloses a kind of gas monitoring device and method for VOCs multi-component, belong to VOCs multi-component gas monitoring technical field.The device includes: gas sensor array, data sampling processing mainboard and gas sample gas path.Gas sensor array is connected with data sampling processing mainboard, and is placed in sensor cavity by gas sample gas path to deliver the gas to be measured;Each chip on mainboard is connected by PCB, carries out on-board communication based on various serial bus protocols, and is controlled and detected by host module to each functional module, calculation process;The method includes: pre-processing and response feature extraction to the sequence data obtained by sampling, sensor response value calculation, gas type and concentration identification, calculation.The application realizes real-time online monitoring for VOCs gas using multi-sensor array, including qualitative type identification and quantitative concentration calculation for single, mixed components of multiple VOC gases.

Description

Technical Field

[0001] This invention belongs to the field of VOCs multi-component gas monitoring technology. Specifically, this invention relates to a gas monitoring device and method for VOCs multi-components. Background Technology

[0002] At present, the application fields of real-time gas monitoring systems are constantly expanding, covering multiple fields such as industrial safety, environmental monitoring, and indoor air quality monitoring. Traditional gas detection methods such as gas chromatography, mass spectrometry, infrared spectroscopy absorption, and chemical absorption have high accuracy in qualitative and quantitative analysis, but the detection process is complex and not real-time, which is not suitable for the current needs of 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. Furthermore, there is limited research on high-precision, anti-interference miniaturized gas monitoring devices based on sensors. Due to the limitations of existing gas sensor technology, it is difficult to design sensors with high accuracy and wide applicability.

[0004] Existing gas monitoring systems mainly target single gases or detect TVOC, and the detection results are not ideal. They typically use electrochemical or optical gas sensors, which can lead to high complexity and maintenance costs, making them unsuitable for large-scale mass production.

[0005] Due to limitations such as poor selectivity in semiconductor gas sensors, current research on their applications remains largely confined to the laboratory stage or limited to relatively simple applications like combustible gas leak alarm monitoring. A key technical problem urgently needs to be solved: how to utilize the operating characteristics of semiconductor gas sensors to achieve accurate monitoring of the types and concentrations of multiple VOCs components in real-world applications. Based on the above, this invention proposes a gas monitoring device and method for multiple VOCs components. Summary of the Invention

[0006] The purpose of this invention is to provide a gas monitoring device and method for VOCs with multiple components. Existing VOCs gas monitoring devices and methods have difficulties in detecting specific VOCs types, have complex detection processes, and have high detection costs. This invention achieves qualitative identification, quantitative concentration calculation, and online monitoring of single and multiple mixed components of VOCs. The device is simple to operate, highly integrated, and can be mass-produced at a low cost.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A gas monitoring device for multi-component VOCs includes a gas sensor array, a data sampling and processing main board, and a gas injection path. The gas sensor array is interconnected with the data sampling and processing main board. The gas sensor array includes several detection units, each containing a semiconductor gas sensor. Each semiconductor gas sensor is connected to a matching resistor and a filter capacitor. Each detection unit outputs a voltage signal source. The voltage outputs of each detection unit in the gas sensor array, as well as two input voltages for the heating electrode and the sensing electrode of the semiconductor gas sensor, are connected to the data sampling and processing main board via flexible flat cables.

[0009] The gas injection path includes a diaphragm pump and a sensor cavity. The diaphragm pump is connected to the injection port and the sensor cavity respectively through a silicone tubing.

[0010] Preferably, the data sampling and processing motherboard 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 via a serial bus protocol, the main control module is connected to the multi-channel data transceiver module via a serial data transceiver, and connected to a physical interface for communication with external devices, and the multi-channel power supply module is connected to the data acquisition module, the main control module, and the multi-channel data transceiver module respectively.

[0011] Preferably, the data acquisition module includes an analog-to-digital converter 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, and the output of the voltage follower is connected to the analog-to-digital converter chip to measure the output voltage of the sensor array. At the same time, the voltage output of the reference voltage source chip is connected to the analog-to-digital converter chip via the voltage follower to provide a reference voltage.

[0012] Preferably, the voltage follower is typically constructed by connecting multiple operational amplifiers in a specific manner.

[0013] Preferably, the main control module includes a microcontroller and an air pump control circuit, wherein the microcontroller is connected to the analog-to-digital converter 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) Perform microcontroller self-initialization and serial interface initialization for external connections;

[0016] 2) Initialize the analog-to-digital converter chip, perform self-calibration, and select the sampling rate and voltage amplification factor;

[0017] 3) Read the sampled values ​​from the registers of the analog-to-digital converter chip and process the raw sampled values ​​to obtain the acquired voltage data;

[0018] 4) During the detection of the target gas by the device, gas detection calculations are performed;

[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, TTL-RS485 module, and TTL-CAN module respectively through a serial data interface. The TTL-RS485 module and TTL-CAN module are connected to the serial port interface on the board via a double-throw switch to receive data sent by external expansion devices 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 act as a slave in the wireless gas monitoring device network or to communicate directly with the data monitoring platform server.

[0021] Preferably, the multi-channel power supply module includes a power switching power supply chip, a low dropout linear regulator, and the adjustable voltage output by the power switching power supply chip is connected to the gas sensor array in two separate paths via zero-ohm resistors. The voltage output by the linear regulator is connected to each chip on the data sampling and processing motherboard to provide operating voltage.

[0022] A gas monitoring method for multi-component VOCs includes the following steps:

[0023] S1. The voltage data output by the sensor array is processed and calculated by the data acquisition module and the main control module to output a one-dimensional sequence of voltage values. The one-dimensional sequence of voltage values ​​is then processed to identify and preprocess the baseline voltage and response voltage in real time.

[0024] S2. Calculate the response values ​​of each channel sensor based on the baseline voltage and response voltage, and output the calculated response values ​​and the corresponding baseline voltage values ​​at the time.

[0025] S3. Based on the output response and baseline voltage data, identify the gas type and calculate the gas concentration.

[0026] Preferably, the identification and preprocessing of the baseline voltage and response voltage in S1 specifically includes the following steps:

[0027] A1. After the device has undergone the sensor warm-up time preset by the built-in program, it begins to acquire voltage. First, it initializes the baseline voltage by taking the weighted average of several initial voltage values, which is then used as the starting baseline voltage value (V). base );

[0028] A2. After completing the baseline voltage initialization, perform subsequent sampling on the voltage (V). in ) and baseline voltage value (V base The sensor response value is calculated based on a relational formula. The calculated response value is then evaluated. If the response value falls within a preset threshold range, the sampled voltage (V) is... in The value is determined to be the baseline voltage, and the existing baseline voltage value (V) in the program is compared. base The weighted moving average (EMA) algorithm is used to update the value by dynamically adjusting 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. Its value is determined based on the sensor's own response characteristics and is usually between 0.1 and 0.5.

[0034] A3. If the response value exceeds the preset threshold range, the sampled voltage will be determined as the response voltage (V). res ), and outputs baseline voltage and response voltage (V res It is stored in the cache for subsequent calculations.

[0035] Preferably, the calculation of the response values ​​of each channel sensor based on the baseline voltage, response voltage, and parameters in step S2 specifically includes the following steps:

[0036] B1. The sensor response value is calculated based on the change in voltage across the sensor before and after the response, reducing circuit complexity and errors caused by resistance measurement. The expression for calculating the sensor response value is:

[0037] r=R air / R gas

[0038] Where r represents the defined sensor response value; R air R represents the resistance value of the sensor in the air. gas This indicates the resistance value of the sensor in the target test gas;

[0039] Based on the voltage divider principle in circuits:

[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 This indicates the resistance value of the matching resistor to which the sensor is connected; V s This indicates the voltage value output by the power module to the electrode terminals of the sensor array's sensitive element.

[0044] Preferably, S3 specifically includes the following:

[0045] C1. In a specific gas testing system, the response characteristics of the sensor array are calibrated using a standard gas sample set. The standard gas sample set includes air with different humidity levels and single and mixed target gases 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. Using all the response values ​​of the sensor array and the corresponding gas sample types and concentrations as input and output values, construct and train a neural network. After the training achieves the target effect, convert the trained model into a format and output the network parameters as a C code array.

[0047] C3. Build a lightweight multilayer sensor that runs 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 calculations 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 multi-component VOCs, which has the following beneficial effects:

[0049] (1) This invention proposes a gas monitoring device for VOCs multi-components, which is actually a real-time online gas monitoring device. It realizes the qualitative identification and quantitative detection of multi-component VOCs gas. It relies on a sensor array composed of semiconductor gas sensors to identify single or mixed components. It makes full use of the characteristics of semiconductor gas sensors such as fast response speed, small size, low power consumption and maintenance-free operation, and overcomes their defects such as poor selectivity and susceptibility to interference.

[0050] (2) The sensor array proposed in this invention adopts a split design, that is, the sensor array and the array sampling and processing motherboard are arranged on different PCBs, and the cavity is designed to match the sensor array based on 3D printing or CNC processing methods, while retaining the data output and power supply cable interface of the sensor array. This design isolates the main circuit part inside the device from the external gas to be measured, which improves the stability of the device and the detection sensitivity is more sensitive than natural diffusion injection. On the other hand, the split sensor array facilitates the recalibration or replacement of the sensor array after long-term use.

[0051] (3) Based on the gas monitoring device, the present invention further designed a gas monitoring method for VOCs multi-components, which realizes automatic and real-time response feature extraction, simplifies the use process of the monitoring device, and can realize the required functions without complex data preprocessing and algorithm calculation on the host computer. It realizes the intake sampling, data processing and calculation, and output of detection results in one process, which greatly reduces the difficulty of using the device and is conducive to its application in different industries and 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 in the program, the proposed device has high scalability and multi-purpose applicability. It can work in conjunction with external devices or be used to form a monitoring network, making it suitable for various situations such as single-point monitoring, multi-point distributed monitoring, and mobile monitoring in the park. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the overall structure of a gas monitoring device for multiple components of VOCs proposed in this invention.

[0054] Figure 2 This is a schematic diagram of the data sampling and processing mainboard of a gas monitoring device for multiple components of VOCs proposed in this invention;

[0055] Figure 3This is a schematic diagram of the operation process of the gas monitoring device proposed in Embodiment 1 of the present invention;

[0056] Figure 4 The images show the external appearance and internal structure of the gas monitoring device proposed in Embodiment 1 of this invention.

[0057] Figure 5 The test diagram and test results of the monitoring device proposed in Embodiment 1 of the present invention are shown in the multi-point field application test diagram.

[0058] Figure 6 This is a schematic diagram of the original voltage value data results of the test measurement data proposed in Embodiment 2 of the present invention;

[0059] Figure 7 This is a schematic diagram of the result of response feature extraction of single-channel voltage data proposed in Embodiment 2 of the present invention;

[0060] Figure 8 This is a schematic diagram of the training results of the neural network model under two input parameters proposed in Embodiment 2 of the present invention;

[0061] Figure 9 The accuracy of testing the neural network model trained with data before the baseline drift, as proposed in Embodiment 2 of the present invention, is calculated based on the data input after the baseline drift occurs.

[0062] Explanation of the labels in the diagram:

[0063] 1. Gas sensor array; 2. Data sampling and processing motherboard; 3. Gas injection path; 4. Diaphragm 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 converter chip; 13. Voltage follower; 14. Reference voltage source chip; 15. Microcontroller; 16. Pump control circuit; 17. TTL-USB module; 18. TTL-RS485 module; 19. TTL-CAN module. Detailed Implementation

[0064] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0065] In the description of this invention, it should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0066] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0067] In the embodiments of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0068] Example 1:

[0069] Please see Figure 1 ,like Figure 1 As shown, this invention proposes a gas monitoring device for multi-component VOCs, including a gas sensor array 1, a data sampling and processing main board 2, and a gas injection path 3. The gas sensor array 1 and the data sampling and processing main board 2 are interconnected to measure the electrical signal generated by the semiconductor gas sensor. The gas injection path 3 includes a diaphragm pump 4 and a sensor cavity 5. The diaphragm pump 4 is connected to the injection port and the sensor cavity 5 respectively through a silicone hose. In this embodiment, the gas sensor array 1 is placed inside the sensor cavity 5, but the gas sensor array 1 should not be regarded as part of the gas injection path 3.

[0070] Specifically, the gas sensor array 1 includes multiple detection units, each containing a semiconductor gas sensor 6. Each semiconductor gas sensor is connected to a matching resistor 7 and a filter capacitor. Each detection unit outputs a voltage signal source. The voltage outputs of each detection unit in the gas sensor array 1, as well as the two input voltages used for the heating electrode and the sensing electrode of the semiconductor gas sensor 6, are connected to the data sampling and processing motherboard 2 via flexible flat cables.

[0071] In this embodiment, each semiconductor gas sensor 6 has four pins, and the detailed circuit connection is as follows: two pins are connected to the sensor's heating resistor to provide the required operating temperature; the other two pins are connected to the sensor's gas-sensitive resistor, and the gas-sensitive resistor pins are connected to a matching resistor. The voltage signal output by the detection unit is the voltage across the matching resistor. When the resistance value on the sensor decreases, this voltage value will increase. In this embodiment, the main target VOC gases are reducing gases, and an n-type semiconductor sensor is used. When the concentration of the target gas increases, the sensor resistance decreases, and the input voltage increases. This allows for a direct reflection of the gas concentration change through voltage variation, improving the ease of use of the device and the efficiency of subsequent data processing.

[0072] Specifically, the data sampling and processing motherboard 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. 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 external devices. The multi-channel power supply module 11 is connected to the data acquisition module 8, the main control module 9, and the multi-channel data transceiver module 10 respectively.

[0073] like Figure 2 The diagram shows the detailed composition of the data sampling and processing motherboard 2 of this device.

[0074] The data acquisition module 8 includes an analog-to-digital converter 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 converter chip 12. At the same time, the voltage output of the reference voltage source chip 14 is connected to the analog-to-digital converter chip 12 via the voltage follower 13 to provide a reference voltage.

[0075] In this embodiment, the analog-to-digital converter chip is a multi-channel ADC chip with a precision of 16 bits or more and including a programmable gain amplifier (PGA). Single-ended input and differential input are selected according to the application scenario. However, 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, an air pump control circuit 16, and the microcontroller 15 is connected to the analog-to-digital converter 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 control and data processing functions. It should be noted that the microcontroller can also be replaced by other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), or field-programmable gate arrays (FPGAs) 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, TTL-RS485 module 18, and TTL-CAN module 19 respectively through a serial data interface. The TTL-RS485 module 18 and TTL-CAN module 19 are connected to the serial port interface on the board through a double-throw switch to receive data sent by external expansion devices on the serial bus, or to send data to the outside as slave devices on the serial bus. In addition, the microcontroller 15 is connected to the wireless device interface on the board to act as a slave device in 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 a data sampling and processing motherboard with a reserved external device interface. Data can be read from external serial devices by programming the microcontroller, enabling collaborative work of multiple devices.

[0080] like Figure 3 The diagram shows the workflow of this device when performing gas detection.

[0081] The microcontroller in the data sampling and processing motherboard performs the following main logical functions: initializing the microcontroller itself and the serial interface connected to the external device; initializing the analog-to-digital converter chip; reading the sampled values ​​from the analog-to-digital converter chip registers and processing the raw sampled values ​​to obtain the acquired voltage data; performing gas detection calculations during the detection of the target gas; and outputting the calculated results.

[0082] In this embodiment, the initialization program includes the initialization of the microcontroller's internal components, the initialization of the data buffer, the initialization of the microcontroller's 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 connecting the microcontroller to the outside include SPI, IIC, CAN, and MODBUS. The initialization program needs to initialize the serial interfaces used, including the configuration of interface pin parameters, interface operating mode, and data format. In addition, the method for reading the sampled values ​​of the analog-to-digital converter chip is to sequentially read the sampled value register of each channel, reading data from different address registers each time.

[0083] Specifically, gas detection calculation includes the following steps: determining 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 determining and calculating the baseline voltage and response voltage is as follows: After the device preheats the sensor for a pre-set time, voltage acquisition begins. First, the baseline voltage is initialized by taking the weighted average of several initial voltage values ​​as the initial baseline voltage value (V). base After initializing the baseline voltage, the subsequent sampled voltages (V) are then processed. in ) and baseline voltage value (V base The sensor response value is calculated based on a relational formula. A threshold judgment is then applied to the calculated response. If the response value falls within a preset threshold range, the sampled voltage (V) is... in The baseline voltage is assigned to the existing baseline voltage value (V) in the program. base The weighted moving average (EMA) algorithm is used to update the value by dynamically adjusting 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 will be classified as the response voltage (V). res The obtained baseline voltage and response voltage values ​​are then output to the buffer for subsequent calculations.

[0085] The sensor response value is calculated based on the change in voltage across the sensor before and after the response. The specific calculation expression is: r = R air / R gas Where r is the defined sensor response value, R air R is the resistance value of the sensor in the air. gas The resistance value of the sensor in the target test gas is obtained from the voltage divider principle of the circuit:

[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, R l V is the resistance value of the matching resistor to which the sensor is connected. S R is the voltage value output from the power module to the gas resistor terminal of the sensor array. During the calculation of the sensor response value, R... l Once eliminated, the above response value calculation process will be performed sequentially on each channel sensor, and the calculated response values ​​of each channel sensor will be stored in a cache for gas type identification and gas concentration calculation.

[0089] In this embodiment, the voltage value V input to the gas-sensitive resistor of the sensor is... SFor the 5V setting, the selection of matching resistors in each detection unit should ensure that the output voltage is between 0.25V and 2.5V under different concentrations of target gas. Within this range, the detection error of the ADC chip is small, and gain amplification of the voltage signal is not required. The formula for the dynamically adjusted weight A is derived from actual testing experience, and its function is to adjust the weight of the new input voltage based on the degree of voltage fluctuation. Theoretically, when not in contact with the target gas, the resistance of the gas-sensitive resistor of the sensor remains unchanged, i.e., the response value r is 1. However, in actual use, this resistance value fluctuates due to changes in environmental conditions and the characteristics of the sensor itself. Based on the results of actual testing, the normal fluctuation range of its resistance value and the corresponding range of r are determined. This range is the judgment threshold for the baseline voltage and response voltage. The set range of r 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 gas types and calculating gas concentration is as follows: In a specific gas testing system, standard gas samples are 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] Using all the response values ​​of the sensor array and the corresponding gas sample types and concentrations as input and output values, a neural network is constructed and trained. After the training achieves 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 sensor that can run directly on the device's main control module is constructed, including an input layer, a hidden layer, an output layer, and network parameters corresponding to each sensor array, thereby enabling the main control module to directly perform network calculations and obtain gas monitoring results in real time.

[0093] In this embodiment, the response value of each sensor is used as the input feature, and the identification of gas types is treated 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. However, 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. Its structure is similar to that of the classification network. By inputting data samples with actual gas concentration labels, a loss function is used to measure the error between the predicted value and the true value, and the network parameters are updated through the backpropagation algorithm.

[0094] Example 2:

[0095] Based on Example 1 but with a difference, this invention further designs a gas monitoring method for VOCs multi-components based on the proposed monitoring device, including the acquisition of sensor array output, extraction and preprocessing of response features of the acquired sequence, calculation of response values ​​of each channel and output of calculation results, identification of gas types and calculation of gas concentration.

[0096] In this embodiment, a set of data from an 8-channel MOS gas sensor array measuring five target VOC gases is classified and identified using both common processing methods and the processing method proposed in this invention. The five VOC gases include styrene, toluene, dimethyl sulfide, dimethyl disulfide, and carbon disulfide. The sensor array and monitoring device are placed in a static gas testing system, and single target gases of 2, 5, 8, and 10 ppm are introduced respectively. The measured raw voltages of each channel are connected to a PC via a TTL-USB module to view and save the data.

[0097] like Figure 6 As shown, the raw voltage values ​​of this set of measurement data reflect the changes in voltage values ​​output by the sensor array during the testing of five target gases with different concentrations. The identification and preprocessing methods for baseline voltage and response voltage are all applied to the raw voltage data shown in the figure.

[0098] like Figure 7 As shown, this is for Figure 6 The results of response feature extraction from the voltage data of channel 1 are used 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, while the processed voltage and calculated response value are represented by dashed lines and scatter plots, respectively. Different judgment threshold ranges are set for sensors with different response characteristics to optimize the accuracy of response feature extraction. During device operation, this response feature extraction process is performed on all channels.

[0099] like Figure 8 As shown, after outputting the obtained multi-channel response value data, the original voltage value and the extracted response value were used as input values ​​for the neural network, and the gas classification result was used as the output value. The feedforward neural network was trained based on MATLAB, with the network parameters set as follows: 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 input conditions.

[0100] In practical applications, the MOS gas sensors used are susceptible to changes in resistance even in clean air due to factors such as humidity, temperature, airflow, and sensor aging. For the monitoring method in this device, this manifests as a drift in the baseline voltage. To simulate this drift, the raw voltage data of each channel were increased or decreased by varying amounts, ranging from 0.1 to 0.2 V. The newly obtained voltage values ​​were then re-extracted to extract relevant features and output as new response values. Using the new voltage and response values ​​as inputs and the gas type as the output, two models trained with the original data were input. The accuracy was as follows: Figure 9 As shown.

[0101] Comparative verification shows that the model trained by the method proposed in this invention has high robustness and still has high accuracy in gas classification tests on sensor data after drift.

[0102] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for VOCs multi-component oriented gas monitoring, characterized in that, The method is realized based on a VOCs multi-component oriented gas monitoring device, the device comprises a gas sensor array (1), a data sampling processing mainboard (2) and a gas sampling gas path (3), the gas sensor array (1) and the data sampling processing mainboard (2) are connected with each other, the gas sensor array (1) comprises a plurality of detection units, each detection unit comprises a semiconductor type gas sensor (6), each semiconductor type gas sensor (6) is connected to a matching resistor (7) and a filter capacitor, and each detection unit leads out a voltage signal source output; the voltage outputs of the detection units of the gas sensor array (1) and two input voltages respectively used for the heating electrode and the sensitive body electrode of the semiconductor type gas sensor (6) are connected to the data sampling processing mainboard (2) through a flexible wire; The gas sampling gas path (3) comprises a diaphragm air pump (4) and a sensor cavity (5), the diaphragm air pump (4) is connected with a sampling interface and the sensor cavity (5) respectively through silica gel hoses; The method comprises the following steps: S1, the voltage data output by the sensor array is subjected to the processing and calculation of a data acquisition module (8) and a main control module (9), and then a one-dimensional sequence of voltage values is output, the one-dimensional sequence of voltage values is processed, and the identification and preprocessing of real-time baseline voltage and response voltage are performed; specifically, the following steps are included: A1, after the device is preheated for a preset sensor preheating time through a built-in program, the voltage acquisition is started, and the initialization of the baseline voltage is performed first, the weighted average value of the initial voltage values is taken as the starting baseline voltage value; A2, after the initialization of the baseline voltage is completed, the subsequent sampling voltage and the baseline voltage value are subjected to sensor response value calculation based on a relationship, and the calculated sensor response value is judged, if the calculated sensor response value is within a preset threshold range, the sampling voltage is determined as the baseline voltage, and the existing baseline voltage value in the program is updated through a weighted moving average algorithm with dynamic adjustment of the weight value, and the expression is: A = L * (V in / V base ) If V in / V base >1, then: A = L * (V base / V in ); V base = (1 - A) * V base + A * V in In the formula, A represents a dynamic adjustment weight; V in represents a sampling voltage; V base represents a baseline voltage value; L represents a weight adjustment coefficient, and the smaller L is, the higher the baseline smoothness is. A3, if the calculated sensor response value exceeds the preset threshold range, the sampling voltage is determined as the response voltage, and the baseline voltage and the response voltage are output to the cache for subsequent calculation; S2, the sensor response values of each channel are calculated according to the baseline voltage and the response voltage, and the calculated sensor response values of each channel and the baseline voltage values at the corresponding time are output; specifically, the following steps are included: B1, the sensor response value is calculated according to the voltage value change of the sensor before and after the sensor contacts the target test gas, the circuit complication and the error caused by the measurement resistance are reduced, and the calculation expression of the sensor response value is: r = R air / R gas wherein r represents the defined sensor response value; R air represents the resistance value of the sensor in air, R gas represents 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 )) wherein R l represents the resistance value of the matching resistor (7) to which the sensor is connected; V s represents the voltage value output by the power module to the electrode end of the sensor array sensitive body S3, the gas type identification and the gas concentration calculation are performed according to the output response voltage and baseline voltage data, and specifically, the following contents are included: C1, in a specific gas testing system, the response values of each sensor of the sensor array are calibrated by using a standard gas sample set, wherein the standard gas sample set contains air with different humidity and single target gas and mixed target gas with different concentration, and the response value output and baseline voltage value output of each sensor in the sensor array corresponding to each standard gas sample are obtained in the calibration process; C2, the data of all sensor response values of the sensor array and corresponding gas sample types and concentrations are taken as input and output values, and the neural network is constructed and trained, and after the training reaches the target effect, the trained model is converted in format, and the network parameters are output as C code array; C3, a lightweight multilayer perceptron directly running on the device main control module (9) is built, including input layer, hidden layer, output layer and network parameters corresponding to each sensor array, and the network operation is directly performed by the main control module (9) to obtain the gas monitoring result in real time.

2. The method of claim 1, wherein the VOCs are selected from the group consisting of: The data sampling processing mainboard (2) includes integrated data acquisition module (8), main control module (9), multi-channel data transceiver module (10) and multi-channel power module (11); the gas sensor array (1) is connected to the data acquisition module (8), the data acquisition module (8) is connected with the main control module (9) through serial bus protocol, the main control module (9) is connected with the multi-channel data transceiver module (10) through serial data transceiver, and is connected to physical interface to communicate with external equipment, and the multi-channel power module (11) is connected with the data acquisition module (8), the main control module (9) and the multi-channel data transceiver module (10) respectively. ​ 3. The method of claim 2, wherein the VOCs are selected from the group consisting of: The data acquisition module (8) includes analog-to-digital conversion chip (12), voltage follower (13) and 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) through the voltage follower (13) to provide reference voltage. ​ 4. The method of claim 2, wherein the VOCs are selected from the group consisting of: The main control module (9) includes microcontroller (15) and gas pump control circuit (16), and the microcontroller (15) is connected with the analog-to-digital conversion chip (12) and the gas pump control circuit (16) respectively; ​ The microcontroller (15) has the following logical functions: 1) to initialize the microcontroller (15) itself and the serial interface connected with the outside; 2) to initialize, self-calibrate, select sampling rate and voltage value amplification multiple of the analog-to-digital conversion chip (12); 3) to read the register sampling value of the analog-to-digital conversion chip (12), and to process the original sampling value to obtain the collected voltage value data; 4) to perform gas detection calculation during the detection of target gas by the device; 5) to output the calculated result.

5. The method of claim 2, wherein the VOCs are selected from the group consisting of: 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 with the TTL-USB module (17), the TTL-RS485 module (18) and the TTL-CAN module (19) through serial data interfaces, the TTL-RS485 module (18) and the TTL-CAN module (19) are connected to the on-board serial port interface through double-pole switches, and are used for receiving data transmitted by external expansion devices on a serial bus or transmitting 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 to serve as a slave of a wireless gas monitoring device network or directly communicate with a data monitoring platform server. ​ 6. The method of claim 2, wherein the VOCs are selected from the group consisting of: The multi-channel power supply module (11) comprises a power switch power supply chip, a low-dropout linear voltage stabilizer, and adjustable voltages output by the switch power supply chip are connected to the gas sensor array (1) through zero-ohm resistors in two paths, and the voltage output by the linear voltage stabilizer is connected to each chip on the data sampling processing mainboard (2) to provide working voltage. ​

Images