Sensor integrated system for detecting various harmful gases

Abstract

The invention discloses a sensor integrated system for detecting various harmful gases, which relates to the field of gas detection and comprises a power supply module, a multifunctional sensor array, a signal processing module and a main control module, the multifunctional sensor array is used for acquiring a gas response electric signal of gas flowing through the multifunctional sensor array; the signal processing module is used for processing the gas response electric signal to obtain a gas response digital signal; the main control module is also used for carrying out preprocessing, feature extraction and recognition on the gas response digital signal to obtain a gas recognition result; wherein the gas identification result comprises no harmful gas or the type and concentration of the harmful gas; wherein a gas identification convolutional neural network model for qualitative analysis and a gas quantitative analysis model for quantitative analysis are preset in the main control module. According to the invention, the accuracy of system detection is improved, and the processing efficiency is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of harmful gas detection, and in particular to a sensor integrated system for detecting multiple harmful gases. Background Art

[0002] In the field of harmful gas detection and monitoring, traditional detection and monitoring equipment often uses single gas sensors. However, traditional single gas sensors are usually limited to detecting one or a few specific gases and generally respond not only to a single gas but also to other gases, which significantly limits their application in complex environments. In addition, due to interference from environmental factors and the characteristics of the sensor itself, the detection results of gas sensors often have certain errors, which to a certain extent affects the accuracy and reliability of detection.

[0003] To overcome these limitations, multiple gas sensors have begun to be integrated into a unified detection and monitoring system to achieve comprehensive detection of multiple harmful gases. However, this integration brings a lot of data and calculations. Therefore, a sensor integration system is needed that can efficiently process and analyze a large amount of data from different sensors and extract accurate harmful gas detection results from them. Summary of the invention

[0004] In order to solve the technical problems existing in the background technology, the present invention proposes a sensor integration system for detecting multiple harmful gases.

[0005] The present invention proposes a sensor integrated system for detecting multiple harmful gases, comprising: a power module, a gas flow rate control module, a multifunctional sensor array, a signal processing module, an environmental parameter acquisition module and a main control module;

[0006] The multifunctional sensor array is electrically connected to the signal processing module, and the power module and the signal processing module are electrically connected to the main control module respectively;

[0007] The multifunctional sensor array is used to obtain a gas response electrical signal of the gas flowing through it;

[0008] The signal processing module is used to process the gas response electrical signal to obtain a gas response digital signal;

[0009] The main control module is also used to pre-process, extract features and identify the gas response digital signal to obtain a gas identification result; wherein the gas identification result includes the absence of harmful gas or the type and concentration of harmful gas;

[0010] Among them, the main control module is preset with a gas identification convolutional neural network model for qualitative analysis and a gas quantitative analysis model for quantitative analysis, wherein the gas quantitative analysis model includes a kernel principal component analysis model and a gradient boosting decision tree model.

[0011] Preferably, during the identification process, the main control module first uses a gas identification convolutional neural network model to identify the extracted features to obtain a first identification result; wherein the first identification result includes no harmful gas or types of harmful gas;

[0012] When the first identification result is that there is no harmful gas, the first identification result is used as the gas identification result;

[0013] When the first recognition result is the type of harmful gas, the main control module uses the kernel principal component analysis model to reduce the dimension of the extracted features, and then inputs the reduced features into the gradient boosting decision tree model for recognition to obtain a second recognition result; wherein the second recognition result includes the type and concentration of the harmful gas; at this time, the second recognition result is used as the gas recognition result.

[0014] Preferably, the main control module is also used to retrain the gas identification convolutional neural network model and the gas quantitative analysis model through online learning or regularly according to the stored gas identification results and corresponding features.

[0015] Preferably, the multifunctional sensor array includes a nitric oxide gas sensor, an alcohol gas sensor, a carbon monoxide gas sensor, an ammonia gas sensor, a hydrogen sulfide gas sensor and a hydrogen chloride gas sensor.

[0016] Preferably, it also includes a gas flow rate control module and an environmental parameter acquisition module, and the gas flow rate control module and the environmental parameter acquisition module are electrically connected to the main control module respectively;

[0017] The environmental parameter acquisition module is used to acquire environmental parameters; wherein the environmental parameters include temperature, humidity and wind speed and direction data;

[0018] The gas flow rate control module is used to make the gas flow evenly through the functional sensor array at a preset flow rate and flow direction in a closed environment;

[0019] The main control module is used to determine environmental conditions according to environmental parameters; wherein the environmental conditions include ordinary open environments and closed environments;

[0020] Among them, when the environmental conditions are ordinary open environments, the main control module is used to control the gas flow rate control module to close, and the gas in the environment can directly enter the multi-functional sensor array for detection; when the environmental conditions are closed environments, the main control module is used to control the gas flow rate control module to open, so that the gas in the environment flows evenly through the functional sensor array according to the preset flow rate and flow direction.

[0021] Preferably, the main control module is used to perform feedback compensation on the gas response digital signal according to environmental parameters.

[0022] Preferably, it also includes a storage module, which is electrically connected to the main control module.

[0023] Preferably, it also includes a touch screen, which is electrically connected to the main control module.

[0024] Preferably, the power module includes a solar power supply submodule and a lithium battery.

[0025] Preferably, the lithium battery is also equipped with a lithium battery power monitoring mechanism, and the lithium battery power monitoring mechanism is electrically connected to the main control module.

[0026] Preferably, it also includes a wireless communication module, which is electrically connected to the main control module and is used to connect to a background management platform or a user terminal.

[0027] Preferably, a water immersion sensor is further included, and the water immersion sensor is electrically connected to the main control module.

[0028] Preferably, it also includes a GPS / Beidou positioning module, and the GPS / Beidou positioning module is electrically connected to the main control module.

[0029] Preferably, a vibration sensor is also included, and the vibration sensor is electrically connected to the main control module.

[0030] In the present invention, the proposed sensor integrated system for detecting a variety of harmful gases uses a gas recognition convolutional neural network and a combination of a kernel principal component analysis model and a gradient boosting decision tree in the recognition process. It can handle nonlinear features and automatically learn the optimal feature representation, thereby improving the performance of the system in complex environments, which not only improves the accuracy of system detection, but also greatly improves processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A block diagram of a sensor integration system for detecting multiple harmful gases in one embodiment of the present invention. DETAILED DESCRIPTION

[0032] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0033] Reference Figure 1 , a sensor integrated system for detecting multiple harmful gases proposed by the present invention includes: a power module, a gas flow rate control module, a multifunctional sensor array, a signal processing module, an environmental parameter acquisition module and a main control module;

[0034] The multifunctional sensor array is electrically connected to the signal processing module, and the power module and the signal processing module are electrically connected to the main control module respectively;

[0035] The multifunctional sensor array is used to obtain a gas response electrical signal of the gas flowing through it;

[0036] The signal processing module is used to process the gas response electrical signal to obtain a gas response digital signal;

[0037] The main control module is also used to pre-process, extract features and identify the gas response digital signal to obtain the gas identification result;

[0038] Among them, the main control module is preset with a gas recognition convolutional neural network (GRCNN) model for qualitative analysis and a gas quantitative analysis model for quantitative analysis, wherein the gas quantitative analysis model includes a kernel principal component analysis (KPCA) model and a gradient boosting decision tree (GBDT) model.

[0039] Traditional gas detection methods often rely on linear models and manually designed features. The present invention uses a combination of GRCNN and KPCA+GBDT in the recognition process, which can handle nonlinear features and automatically learn the optimal feature representation, thereby improving the performance of the system in complex environments. It not only improves the accuracy of system detection, but also greatly improves processing efficiency.

[0040] The gas identification result includes the absence of harmful gases or the type and concentration of harmful gases.

[0041] In the recognition process, the main control module first uses the gas recognition convolutional neural network (GRCNN) model to recognize the extracted features to obtain a first recognition result; wherein the first recognition result includes no harmful gas or the type of harmful gas;

[0042] When the first identification result is that there is no harmful gas, the first identification result is used as the gas identification result;

[0043] When the first recognition result is the type of harmful gas, the main control module uses the kernel principal component analysis (KPCA) model to reduce the dimension of the extracted features, and then inputs the reduced features into the gradient boosting decision tree (GBDT) model for recognition to obtain a second recognition result; wherein the second recognition result includes the type and concentration of the harmful gas; at this time, the second recognition result is used as the gas recognition result.

[0044] This embodiment uses a combination of GRCNN and KPCA+GBDT to perform qualitative analysis first and then quantitative analysis, which effectively improves processing efficiency. Among them, in the qualitative analysis, this embodiment uses a gas recognition convolutional neural network (GRCNN) to identify the type of gas. GRCNN can determine the type of gas and classify it by analyzing the patterns in the sensor data. In the quantitative analysis, this embodiment uses kernel principal component analysis (KPCA) to reduce the dimension of the data, extract the most representative features, and reduce data redundancy; then, the reduced-dimensional features are input into the gradient boosting decision tree (GBDT) model, and the GBDT model further analyzes and predicts the concentration of the gas.

[0045] It can be seen that this embodiment can process nonlinear features and automatically learn the optimal feature representation through the combination of GRCNN and KPCA+GBDT, which improves the performance of the system in complex environments, not only improves the accuracy of the system, but also greatly improves the processing efficiency.

[0046] Of course, the gas recognition convolutional neural network (GRCNN) model and the gas quantitative analysis model used for quantitative analysis in this embodiment have been pre-trained using a harmful gas training set to improve the accuracy of model prediction.

[0047] The harmful gas training set is from the University of California's dynamic mixed gas sensor array dataset. The dataset not only includes experimental data of multiple gases, but also contains diverse data under different environments and different temperature and humidity conditions, which helps to improve the generalization ability of the model and avoid overfitting.

[0048] In a further embodiment, the main control module is also used to retrain the gas recognition convolutional neural network (GRCNN) model and the gas quantitative analysis model through online learning or regularly based on the stored gas identification results and corresponding features, so as to continuously optimize the existing model, adapt to environmental changes and improve accuracy.

[0049] In this embodiment, it also includes a gas flow rate control module and an environmental parameter acquisition module, and the gas flow rate control module and the environmental parameter acquisition module are electrically connected to the main control module respectively;

[0050] The environmental parameter acquisition module is used to acquire environmental parameters; wherein the environmental parameters include temperature, humidity and wind speed and direction data;

[0051] The main control module is used to determine environmental conditions according to environmental parameters; wherein the environmental conditions include ordinary open environments and closed environments;

[0052] The gas flow rate control module is used to make the gas flow evenly through the functional sensor array at a preset flow rate and flow direction in a closed environment;

[0053] Among them, when the environmental conditions are ordinary open environments, the main control module is used to control the gas flow rate control module to close, and the gas in the environment can directly enter the multi-functional sensor array for detection; when the environmental conditions are closed environments, the main control module is used to control the gas flow rate control module to open, so that the gas in the environment flows evenly through the functional sensor array according to the preset flow rate and flow direction.

[0054] This embodiment can dynamically adjust the working state of the gas flow rate control module according to environmental conditions. In a normal open environment, the gas flow rate control module is turned off, and the gas directly enters the sensor array for detection; in a closed environment, the gas flow rate control module is enabled, and the uniformity of the sample gas is ensured by adjusting the gas flow rate and distribution, thereby further improving the accuracy of detection and reducing power consumption.

[0055] Of course, in a normal open environment, if the gas flow rate is too slow, the main control module can also start the gas flow rate control module.

[0056] In a further embodiment, the main control module is used to perform feedback compensation on the gas response digital signal according to environmental parameters to avoid the influence of temperature and humidity in the environment on the gas response electrical signal obtained by the multifunctional sensor array.

[0057] In this embodiment, the preprocessing includes denoising, normalization or standardization processing, and missing value processing.

[0058] In this embodiment, the interference and noise from the environment in the gas response digital signal are removed by a signal filter, i.e., denoising; then, the gas response digital signal is normalized or standardized so that the data of different sensors can be compared and analyzed at the same scale; then, the gas response digital signal processes the missing values ​​to ensure the integrity of the data. Specifically, in the process of processing the missing values, the missing data is usually filled by an interpolation method.

[0059] Among them, the harmful gases in this embodiment include nitric oxide (NO), alcohol, carbon monoxide (CO), ammonia (NH3), hydrogen sulfide (H2S) and hydrogen chloride (HCl).

[0060] Therefore, the multifunctional sensor array includes a nitric oxide (NO) gas sensor, an alcohol gas sensor, a carbon monoxide (CO) gas sensor, an ammonia (NH3) gas sensor, a hydrogen sulfide (H2S) gas sensor, and a hydrogen chloride (HCl) gas sensor.

[0061] The signal processing module is used to amplify, filter and perform analog-to-digital conversion (AD conversion) on the gas response electrical signal, eliminate noise and interference, obtain the gas response digital signal, and transmit the gas response digital signal to the main control module. Specifically, the signal processing module includes 6 analog signal processing circuits, and the 6 analog signal processing circuits are connected one by one to the 6 sensors in the multifunctional sensor array.

[0062] It should be understood that in the feature extraction process, the extracted features include time domain features (such as mean, standard deviation, maximum value, minimum value, etc.) and frequency domain features (such as various frequency components in spectrum analysis), which are conducive to extracting the part that best reflects the gas characteristics, thereby reducing the computational complexity of subsequent identification and analysis.

[0063] In this embodiment, the main control module is also used to control the entire system to enter a low-power standby mode after obtaining the gas identification result, so as to reduce energy consumption.

[0064] In a further embodiment, the main control module supports user-defined wake-up cycles and automatically starts executing a new round of detection tasks at a predetermined time.

[0065] In this embodiment, a touch screen is also included. The touch screen is electrically connected to the main control module. The touch screen is used to display the gas identification result and can perform human-computer interaction.

[0066] In this embodiment, a storage module is also included. The storage module is electrically connected to the main control module. The storage module is used to store gas identification results and corresponding environmental parameters and gas response digital signals for subsequent analysis.

[0067] In this embodiment, the power module includes a solar power supply submodule and a lithium battery.

[0068] The solar power supply module in this embodiment provides a stable 12V output voltage through the maximum power point tracking (MPPT) technology, which can directly supply power and charge the lithium battery. The lithium battery is used as a backup power source. When the light is insufficient or the solar power supply is interrupted, the lithium battery is automatically switched to power supply, ensuring that the system continues to operate stably under any power supply conditions.

[0069] In a further embodiment, the lithium battery is further configured with a lithium battery power monitoring mechanism, and the lithium battery power monitoring mechanism is electrically connected to the main control module.

[0070] With this configuration, the main control module can monitor the lithium battery power in real time, dynamically adjust the system operating status, and ensure efficient use of power supply.

[0071] In this embodiment, a water immersion sensor is also included. The water immersion sensor is electrically connected to the main control module and is used to monitor in real time whether the system has water ingress;

[0072] When flooding is detected, the main control module immediately triggers an alarm, records the abnormality and sends relevant information to the background management platform or user terminal for timely disposal.

[0073] In this embodiment, a wireless communication module is also included. The wireless communication module is electrically connected to the main control module, and the wireless communication module is used to connect to a background management platform or a user terminal.

[0074] Specifically, the wireless communication module supports Bluetooth, WiFi and 4G networks, realizes the interaction between short-range devices and the system through Bluetooth and WiFi, and realizes remote data transmission and management through 4G networks, making it convenient for users to check the device status and detection data at any time.

[0075] In a further embodiment, a GPS / Beidou positioning module is also included. The GPS / Beidou positioning module is electrically connected to the main control module, and the GPS / Beidou positioning module is used to obtain the geographic location of the device in real time.

[0076] The main control module of this embodiment can obtain the device's geographic location in real time through the GPS / Beidou positioning module and upload it to the background management system for easy positioning and management.

[0077] In a further embodiment, a vibration sensor is further included, the vibration sensor is electrically connected to the main control module, and the vibration sensor is used to obtain the operating status of the system in real time;

[0078] When the system detects that the device has fallen, vibrated, or moved abnormally, the main control module will record the abnormality and send an alarm in time to prevent damage or loss of the device due to environmental influences or external interference.

[0079] In specific implementation, the sensor integrated system for detecting multiple harmful gases is started, and the main control module first performs a self-check on each functional module to ensure that the module status is normal;

[0080] After the self-check is completed, the main control module obtains the environmental parameters through the environmental parameter acquisition module and determines the environmental conditions according to the environmental parameters;

[0081] When the environmental condition is a normal open environment, the main control module is used to control the gas flow rate control module to close, and the gas in the environment can directly enter the multifunctional sensor array for detection;

[0082] When the environmental condition is a closed environment, the main control module is used to control the gas flow rate control module to open, so that the gas in the environment flows evenly through the multifunctional sensor array according to the preset flow rate and flow direction;

[0083] Next, the main control module obtains the gas response electrical signal of the harmful gas under environmental conditions through the multifunctional sensor array;

[0084] The gas response electrical signal is processed by a signal processing module to obtain a gas response digital signal;

[0085] Next, the main control module performs feedback compensation on the gas response digital signal according to the environmental parameters;

[0086] Next, the main control module preprocesses, extracts features and identifies the gas response digital signal;

[0087] During the recognition process, the main control module first uses the gas recognition convolutional neural network (GRCNN) model to recognize the extracted features and obtain a first recognition result; wherein the first recognition result includes no harmful gas or the type of harmful gas;

[0088] When the first identification result is that there is no harmful gas, the first identification result is used as the gas identification result;

[0089] When the first recognition result is the type of harmful gas, the main control module uses the kernel principal component analysis (KPCA) model to reduce the dimension of the extracted features, and then inputs the reduced-dimensional features into the gradient boosting decision tree (GBDT) model for recognition to obtain a second recognition result; wherein the second recognition result includes the type and concentration of the harmful gas; at this time, the second recognition result is used as the gas recognition result;

[0090] When the concentration of harmful gases exceeds the preset threshold of the type, the main control module immediately alarms and sends the abnormal information to the remote device or background system through the wireless communication module.

[0091] The above description is only a preferred specific implementation manner 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 scheme and inventive concept 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 sensor integrated system for detecting multiple harmful gases, characterized in that: include: Power module, multifunctional sensor array, signal processing module and main control module; The multifunctional sensor array is electrically connected to the signal processing module, and the power module and the signal processing module are electrically connected to the main control module respectively; The multifunctional sensor array is used to obtain a gas response electrical signal of the gas flowing through it; The signal processing module is used to process the gas response electrical signal to obtain a gas response digital signal; The main control module is also used to pre-process, extract features and identify the gas response digital signal to obtain a gas identification result; wherein the gas identification result includes the absence of harmful gas or the type and concentration of harmful gas; Among them, the main control module is preset with a gas identification convolutional neural network model for qualitative analysis and a gas quantitative analysis model for quantitative analysis, wherein the gas quantitative analysis model includes a kernel principal component analysis model and a gradient boosting decision tree model.

2. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: During the recognition process, the main control module first uses the gas recognition convolutional neural network model to recognize the extracted features to obtain a first recognition result; wherein the first recognition result includes no harmful gas or the type of harmful gas; When the first identification result is that there is no harmful gas, the first identification result is used as the gas identification result; When the first recognition result is the type of harmful gas, the main control module uses the kernel principal component analysis model to reduce the dimension of the extracted features, and then inputs the reduced features into the gradient boosting decision tree model for recognition to obtain a second recognition result; wherein the second recognition result includes the type and concentration of the harmful gas; at this time, the second recognition result is used as the gas recognition result. Preferably, the main control module is also used to retrain the gas identification convolutional neural network model and the gas quantitative analysis model through online learning or regularly according to the stored gas identification results and corresponding features.

3. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: It also includes a gas flow rate control module and an environmental parameter acquisition module, which are electrically connected to the main control module respectively; The environmental parameter acquisition module is used to acquire environmental parameters; wherein the environmental parameters include temperature, humidity and wind speed and direction data; The gas flow rate control module is used to make the gas flow evenly through the functional sensor array at a preset flow rate and flow direction in a closed environment; The main control module is used to determine environmental conditions according to environmental parameters; wherein the environmental conditions include ordinary open environments and closed environments; Among them, when the environmental condition is a normal open environment, the main control module is used to control the gas flow rate control module to be closed, and the gas in the environment can directly enter the multifunctional sensor array for detection; when the environmental condition is a closed environment, the main control module is used to control the gas flow rate control module to be opened, so that the gas in the environment flows evenly through the functional sensor array according to the preset flow rate and flow direction; Preferably, the main control module is used to perform feedback compensation on the gas response digital signal according to environmental parameters.

4. The sensor integrated system for detecting multiple harmful gases according to claim 3, characterized in that: The multifunctional sensor array includes a nitric oxide gas sensor, an alcohol gas sensor, a carbon monoxide gas sensor, an ammonia gas sensor, a hydrogen sulfide gas sensor, and a hydrogen chloride gas sensor.

5. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: It also includes a storage module, which is electrically connected to the main control module; Preferably, it also includes a touch screen, which is electrically connected to the main control module.

6. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: The power module includes a solar power supply module and a lithium battery; Preferably, the lithium battery is also equipped with a lithium battery power monitoring mechanism, and the lithium battery power monitoring mechanism is electrically connected to the main control module.

7. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: It also includes a wireless communication module, which is electrically connected to the main control module and is used to connect to a background management platform or a user terminal.

8. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: It also includes a water immersion sensor, which is electrically connected to the main control module.

9. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: It also includes a GPS / Beidou positioning module, which is electrically connected to the main control module.

10. The sensor integrated system for detecting multiple harmful gases according to claim 1, characterized in that: It also includes a vibration sensor, which is electrically connected to the main control module.

Images