Portable gas intelligent measuring system based on TDALS

Abstract

The application relates to the technical field of gas detection, and provides a portable intelligent gas detection system based on TDALS, which is characterized in that parallel sensors are arranged through an air inlet and in parallel with the air duct direction and located on the same horizontal plane, the inhaled gas is monitored in real time, multi-component gas concentration data of the to-be-detected gas are acquired, the influence of environmental pressure fluctuation on the measurement accuracy of the gas sensor is effectively overcome through a multi-parameter self-adaptive intelligent compensation algorithm, the detection stability and data reliability of the system under complex working conditions are significantly improved, and the system has the functions of environmental multi-parameter intelligent detection and inspection, so that the measurement accuracy of the to-be-detected gas can be effectively improved in an industrial or accident site with complex gas components and multiple components.

Description

Technical Field

[0001] This application relates to the field of gas detection technology, and in particular to a portable intelligent gas measurement system based on TDALS. Background Technology

[0002] With the widespread application of portable gas analyzers in both daily life and industry, they play a vital role as a flexible and efficient gas detection tool in various scenarios. For example, in industrial settings or during emergency response, quickly determining the composition and concentration of gases is crucial for ensuring personnel safety and preventing secondary disasters. Failure to detect changes in the composition and concentration of gases in the environment in a timely manner may escalate dangerous situations, resulting in casualties and significant economic losses. Summary of the Invention

[0003] The purpose of this application is to provide a portable intelligent gas measurement system based on TDALS to solve or alleviate the problems existing in the prior art.

[0004] To achieve the above objectives, this application provides the following technical solution: This application provides a portable intelligent gas measurement system based on TDALS, comprising: multiple parallel gas sensors of different types arranged on the same horizontal plane with air inlets parallel to the direction of the air duct; wherein the multiple parallel gas sensors of different types monitor the inhaled gas to be measured based on the principle of tunable semiconductor laser absorption spectroscopy to obtain multi-component gas concentration data of the gas to be measured; according to the ambient pressure of the gas to be measured, the multi-component gas concentration data is compensated by a multi-parameter adaptive intelligent compensation algorithm, and the compensated data is inversely normalized to obtain the actual concentration value of the gas to be measured.

[0005] Preferably, the gas sensor includes a carbon monoxide sensor for acquiring the concentration of carbon monoxide in the gas to be measured, and the ambient pressure of the gas to be measured is acquired by a pressure sensor built into the carbon monoxide sensor.

[0006] Preferably, a hybrid strategy of particle swarm optimization and cooperative optimization is used to globally optimize the model parameters of the multi-parameter adaptive intelligent compensation algorithm, and the globally optimal model parameters are used as the initial parameters of the multi-parameter adaptive intelligent compensation algorithm. The multi-parameter adaptive intelligent compensation algorithm with initial parameters is locally fine-tuned, and the standardized residual prediction value of the gas sensor is output based on the fine-tuned multi-parameter adaptive intelligent compensation algorithm. Furthermore, the standardized residual prediction value of the gas sensor The actual residual value of the gas sensor is obtained by performing inverse normalization. ; Pressure compensation was performed on the ambient pressure of the gas to be measured through a gas calibration experiment to obtain the pressure correction value of the gas sensor. ; Based on the actual residual value of the gas sensor Pressure correction value Raw concentration measurements from the gas sensor Compensation is performed to obtain the gas replenishment concentration of the gas sensor. .

[0007] Preferably, the individual fitness of network group parameters is achieved through a multi-parameter adaptive intelligent compensation algorithm. Optimal individual fit for updating network group parameters And select the individual fit from all network group parameters. The optimal individual fitness of the minimum network group parameters is the group fitness of the multi-parameter adaptive intelligent compensation algorithm. ; Based on the parameter quality of network group parameters, calculate the collaborative optimization collision speed of network group parameters. Furthermore, the particle swarm optimization inertia weights are determined by linear decay of the network group parameters. And combined with the individual adaptability of network group parameters Optimal single-unit fit Group fit Calculate the fusion speed of global optimization of network group parameters. ; By fusion speed Individual fit of network group parameters Perform iterative updates, and correct the individual fitness of the network group parameters that are ranked at the bottom 30% in the individual fitness after the iterative updates. The network group parameters corresponding to the smallest individual fitness among all group network parameters are used as the initial parameters for the multi-parameter adaptive intelligent compensation algorithm.

[0008] Preferably, multi-component gas concentration sample data of the gas to be tested under different environmental pressures are obtained and divided into training set and validation set according to pressure, so as to train and validate the multi-parameter adaptive intelligent compensation algorithm respectively; According to the single-unit fit model: Calculate the first Individual fit of group network parameters In the formula, The multi-parameter adaptive intelligent compensation algorithm adopts the first The first validation set output when grouping network group parameters Predicted concentration values ​​for each sample data point. To verify the first set The true concentration value of each sample data point. The multi-parameter adaptive intelligent compensation algorithm adopts the first... When grouping network parameters, the weights from the input layer to the hidden layer and the weights from the hidden layer to the output layer are used. To verify the quantity of the sample data in the set.

[0009] Preferably, based on the collaborative optimization mechanism and random pairing of network group parameters, collision simulation is performed on the network group parameters, and according to the parameter quality model: Calculate the first Collision speed of group network parameters ; In the formula, They are randomly paired for the first time. Group, No. The quality of group parameters in the group network. These are the maximum and minimum values ​​of individual fitness among all network group parameters, respectively; It is a very small positive number; The first Group, No. Optimal individual fit of group network parameters; The first Group, No. Pre-collision velocity of the group parameters of the network. These are the collaborative optimization recovery coefficients for network group parameters determined through linear decay.

[0010] Preferably, the global optimization fusion model is adopted: Calculate the first Fusion speed of global optimization of group parameters in group network ; In the formula, Particle swarm optimization inertia weights for network swarm parameters determined by linear decay; For the first Collision speed optimization of group parameters in the network For the first Optimal individual fit of group parameters in the network For group fit, For the first Individual fit of group parameters in the group network All A random number within a given interval.

[0011] Preferably, the model is updated and corrected according to the fit: For the first Individual fit of group network parameters Perform iterative updates and individual component adaptation corrections; In the formula, For the first Individual fit before updating group parameters in the network. For the first Individual fit after updating group parameters of the network; For the first Individual fitness after group parameter correction in the group network For group fit.

[0012] Preferably, multi-component gas concentration sample data of the gas to be tested under different environmental pressures are obtained and divided into a training set and a validation set according to the pressure; wherein, the training set includes Group; pass Set up a training set and fine-tune the model according to the parameters: Local fine-tuning of a multi-parameter adaptive intelligent compensation algorithm with initial parameters; In the formula, To improve the multi-parameter adaptive intelligent compensation algorithm The model parameters to be fine-tuned in the next step; To fine-tune the model parameters The model parameters obtained after local fine-tuning; For the model parameters to be fine-tuned gradient, gradients First-order moment, second-order moment, First-order moments Second-order torch Correction value; First-order moments Second-order torch The attenuation rate, For the model parameters to be fine-tuned The weighted average of historical gradients, For the model parameters to be fine-tuned The weighted average of the squared historical gradients.

[0013] Preferably, the gas to be tested is collected under different environmental pressures through a gas calibration experiment. The multi-component gas concentration sample data was divided into training and validation sets according to pressure to train and validate the multi-parameter adaptive intelligent compensation algorithm, respectively; and according to the concentration compensation model: Raw concentration measurements from the gas sensor Compensation is performed to obtain the gas replenishment concentration of the gas sensor. ; In the formula, For gas sensors under ambient pressure The pressure correction value below, The gas to be tested collected for the gas calibration experiment at ambient pressure A collection of multi-component gas concentration sample data. The first parameter in the validation set output by the multi-parameter adaptive intelligent compensation algorithm. Predicted concentration values ​​for each sample data point. To verify the first set The true concentration value of each sample data; For gas sensors under ambient pressure The actual residual.

[0014] Beneficial effects: The portable environmental multi-parameter intelligent measurement system based on TDLAS provided in this application embodiment has multiple parallel gas sensors of different types arranged on the same horizontal plane with air inlets parallel to the air duct direction. The multiple parallel gas sensors of different types monitor the inhaled gas to be measured based on the principle of tunable semiconductor laser absorption spectroscopy, and obtain multi-component gas concentration data of the gas to be measured. Then, according to the environmental pressure of the gas to be measured, the multi-component gas concentration data is compensated by a multi-parameter adaptive intelligent compensation algorithm, and the compensated data is normalized to obtain the actual concentration value of the gas to be measured.

[0015] By using a parallel sensor layout with the air inlet and duct parallel to each other and on the same horizontal plane, and by monitoring the gas in real time through the suction method, the gas sensor is placed under the same pressure condition as the gas to be measured to obtain multi-component gas concentration data of the gas to be measured. Through a multi-parameter adaptive intelligent compensation algorithm, the influence of environmental parameter changes on sensor accuracy is effectively eliminated. It has environmental adaptive multi-parameter intelligent measurement and inspection, which can effectively improve the measurement accuracy of the gas to be measured in industrial or accident sites with complex and diverse gas components. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. Wherein: Figure 1 This is a schematic diagram of a portable intelligent environmental multi-parameter measurement system based on TDLAS, according to some embodiments of this application. Figure 2 for Figure 1 The diagram shows the internal structure of a portable intelligent environmental multi-parameter measurement system based on TDLAS. Figure 3 This is a schematic diagram of the layout of a portable intelligent environmental multi-parameter measurement system based on TDLAS according to some embodiments of this application; Figure 4 Standard gas concentrations provided according to some embodiments of this application Figures showing test results before and after pressure compensation; Figure 5 Standard gas concentrations provided according to some embodiments of this application and A schematic diagram of concentration compensation at that time; Figure 6 For an embodiment of this application, the gas to be measured before and after compensation is provided. Comparison diagram; Figure 7 For an embodiment of this application, the gas to be measured before and after compensation is provided. Comparison diagram; Figure 8 This is a schematic diagram of a test for 2% carbon dioxide and 5% methane according to another embodiment of this application; Figure 9 Standard gas concentration provided according to another embodiment of this application A schematic diagram of a carbon monoxide concentration test. Detailed Implementation

[0017] The present application will now be described in detail with reference to the accompanying drawings and embodiments. Various examples are provided by way of explanation and not by way of limitation. In fact, those skilled in the art will understand that modifications and variations can be made to the present application without departing from the scope or spirit of the present application. For example, a feature shown or described as part of one embodiment may be used in another embodiment to produce yet another embodiment. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention should fall within the scope of protection of the embodiments of the present invention.

[0018] Developing portable gas analyzers to enable rapid and accurate detection in complex gas environments is of great significance for industrial safety and environmental protection. Most existing portable gas analyzers are designed for detecting specific single gases, making them ineffective in handling the complex and diverse gas compositions at industrial or accident sites. Furthermore, detecting only a single gas's concentration cannot comprehensively reflect the overall gas distribution and dynamic changes in the current environment, resulting in users not obtaining complete gas analysis data. This, in turn, affects the scientific validity of on-site emergency decision-making and the accuracy of operations.

[0019] To address the limitations of current portable gas analyzers, which detect only a single gas or employ electrochemical sensors that suffer from short lifespans and susceptibility to cross-interference from other gases or environmental disturbances, thus failing to effectively handle complex and diverse gas compositions in industrial or accident scenarios, this embodiment provides a portable environmental multi-parameter intelligent measurement system based on TDLAS. Employing a parallel sensor layout and intelligent risk control mechanism, it features environmentally adaptive intelligent measurement and inspection of multiple environmental parameters. Through an environmental parameter compensation algorithm adapted to the instrument, it achieves long lifespan, multi-parameter detection, no cross-interference, and strong resistance to environmental interference. It possesses gas detection capabilities including rapid response, high selectivity, strong anti-interference, portability, and support for intelligent inspection and data fusion early warning.

[0020] like Figures 1 to 9 As shown, this portable intelligent environmental multi-parameter measurement system based on TDLAS mainly includes a housing, a gas acquisition and distribution module, a multi-component gas sensor array, a central processing and control motherboard, a human-machine interface module, a power supply module, and an intelligent inspection auxiliary module. The housing adopts a lightweight and explosion-proof design, and its internal multi-layer modular layout results in higher overall system integration and a more compact size, facilitating portability and on-site movement. A miniature silent fan and air vents are located at the bottom of the housing, forming a unidirectional airflow channel to effectively ensure that ambient gases can enter the detection chamber quickly and uniformly.

[0021] The gas acquisition and distribution module includes a miniature silent fan and an air duct to replace the vacuum pump method for drawing in the gas to be measured. The miniature silent fan draws in ambient gas (the gas to be measured) in a low-power, low-noise manner, and distributes the gas to the inlets of various parallel-arranged gas sensors of different types. This ensures that the gas sensors and the gas to be measured are under the same pressure, effectively reducing measurement errors caused by pressure differences, avoiding zero-point drift caused by pressure interference, improving the accuracy of sensor calibration, and effectively enhancing gas measurement precision. Here, the inlets of different types of parallel gas sensors are located on the same horizontal plane, below which is the gas chamber to be measured. The gas chamber to be measured is refreshed by the miniature silent fan and the outlet. The lower edge of the gas sensor inlet is parallel to the air duct formed by the miniature silent fan and the outlet, effectively ensuring the synchronization and consistency of gas sample detection by each gas sensor.

[0022] The multi-component gas sensor array (parallel gas sensors of different types) includes oxygen sensors, methane sensors, carbon monoxide sensors, carbon dioxide sensors, etc. Multiple different types of gas sensors are connected in parallel and fixed to the bracket inside the housing through a standard interface. The gas sensors use the principle of tunable semiconductor laser absorption spectroscopy to select the characteristic absorption lines of specific gases. By using laser wavelength tuning and harmonic detection, they can achieve high selectivity and high sensitivity detection of inhaled target gases and obtain multi-component gas concentration data of the gas to be measured. This fundamentally and effectively solves the problem of cross-interference between different gases, and the lifespan of the (gas) sensors is significantly longer than that of traditional electrochemical sensors.

[0023] The central processing and control motherboard is integrated and installed above the sensor array. It integrates multiple functional units such as a central processing module, data storage module, data analysis module, communication module, and temperature and humidity sensor module. Through modular design and high integration, it effectively ensures the system's operational stability and ease of maintenance, while also providing support for expanding the system with more functions, enabling it to adapt to complex and ever-changing industrial application needs.

[0024] The control motherboard is embedded with a device that offers excellent measurement accuracy and stability, with a temperature error within ± Within [a certain range], the humidity error does not exceed [a certain value]. Temperature and humidity sensors are used to sense the temperature and humidity parameters of the surrounding environment in real time.

[0025] The central processing module includes a signal converter and a processor, which work together to realize the entire process of data acquisition and processing. The signal converter is electrically connected to a miniature silent fan and multiple different types of gas sensors. It receives raw electrical signals (including analog signals or other forms of signals) transmitted from the gas sensors and converts these signals into standard digital signals suitable for subsequent processing. The processor uses built-in algorithms and logic modules to correct, filter, and analyze the received signals, ultimately generating gas concentration values ​​and trend data for subsequent display, storage, or further analysis, ensuring that the entire system's data processing is efficient, stable, and accurate.

[0026] The data storage module stores monitoring data and system status information in real time. The data analysis module, connected to the data storage module, acquires and analyzes data transmitted from multiple different gas sensors within the data storage module. The data analysis module incorporates a multi-parameter adaptive intelligent compensation algorithm. Based on the ambient pressure of the gas being measured, it compensates for the acquired multi-component gas concentration data and performs inverse normalization on the compensated data to obtain the actual concentration value of the gas being measured, effectively eliminating the impact of environmental parameter changes on sensor accuracy.

[0027] The communication module supports wireless transmission, sending data to an external monitoring center or mobile terminal in real time. The human-machine interface (HMI) module includes a touchscreen display on the front of the housing, connected to the motherboard via a data interface. As the core component of HMI, it features high-sensitivity capacitive touch, supporting multi-touch and gesture operations. Its main functions include: real-time display of gas concentration data, trend curves, and environmental parameters; providing system status indicators and warning information; and receiving user commands such as mode switching, parameter settings, data querying, and exporting. Thus, the display shows the analysis results processed by the data analysis module in real time, including gas concentration values, trend curves, and other information, providing a clear understanding of the detection data. Simultaneously, it receives user-input commands such as parameter settings, curve zooming, mode switching, and data storage commands.

[0028] The touch display uses high-sensitivity capacitive touch technology, supporting multi-touch and gesture operation, effectively ensuring fast and efficient interactive operation in different environments, thereby improving the overall user experience and convenience of the device; at the same time, the display adopts an explosion-proof design, adapting to various hazardous industrial environments.

[0029] The power module uses a high-performance explosion-proof lithium iron phosphate battery, which, together with the motherboard's built-in high-efficiency voltage regulation power management circuit, provides a stable and reliable power supply for the entire system, ensuring that the system can work stably and continuously without an external power source.

[0030] The measurement system in this embodiment is preheated after ensuring that the outer casing and the plane are undamaged. Place the measurement system in the test environment and click the start button on the screen to activate the miniature silent fan and begin the test. The carbon monoxide gas sensor, oxygen gas sensor, and carbon dioxide / methane dual-mode sensor monitor gas concentrations to obtain multi-component gas concentration data. The temperature and humidity sensors on the motherboard, as well as the pressure sensor built into the carbon monoxide gas sensor, obtain environmental data (ambient temperature, ambient humidity, and ambient pressure). These data are then converted from electrical signals to digital signals by a signal converter, further processed to form data, and then compensated using a multi-parameter adaptive intelligent compensation algorithm.

[0031] Among them, the multi-parameter adaptive intelligent compensation algorithm uses a hybrid strategy of particle swarm optimization and cooperative optimization to globally optimize the model parameters, and uses the globally optimal model parameters as the initial parameters of the multi-parameter adaptive intelligent compensation algorithm. First, the individual fitness of the network group parameters in the multi-parameter adaptive intelligent compensation algorithm is used. Specifically, multi-component gas concentration sample data of the gas to be tested under different environmental pressures are obtained and divided into training and validation sets according to pressure, so as to train and validate the multi-parameter adaptive intelligent compensation algorithm respectively; according to the single-component adaptation model: Calculate the first Individual fit of group network parameters In the formula, The multi-parameter adaptive intelligent compensation algorithm adopts the first The first validation set output when grouping network group parameters Predicted concentration values ​​for each sample data point. To verify the first set The true concentration value of each sample data point. The multi-parameter adaptive intelligent compensation algorithm adopts the first... When grouping network parameters, the weights from the input layer to the hidden layer and the weights from the hidden layer to the output layer are used. To verify the quantity of the sample data in the set.

[0032] Then, individual fitness is determined through network group parameters. Update the individual optimal fitness of the corresponding group network parameters. In this process, the fitness of each network group parameter is compared with its historical best fitness value, and the smaller value is taken as the optimal fitness of that network group parameter; and the individual fitness is selected from all network group parameters. The optimal individual fitness of the minimum network group parameters is the group fitness of the multi-parameter adaptive intelligent compensation algorithm. .

[0033] Next, based on the collaborative optimization mechanism, collision simulations are performed on the network group parameters through random pairing, bringing the network group parameters closer to the group fitness. Specifically, based on the collaborative optimization mechanism and random pairing of network group parameters, collision simulations are performed on the network group parameters, according to the parameter quality model: Calculate the first Collision speed of group network parameters In the formula, They are randomly paired for the first time. Group, No. The quality of group parameters in the group network. These are the maximum and minimum values ​​of individual fitness among all network group parameters, respectively; It is a very small positive number; The first Group, No. Optimal individual fit of group network parameters; The first Group, No. Pre-collision velocity of the group parameters of the network. These are the collaborative optimization recovery coefficients for network group parameters determined through linear decay.

[0034] Furthermore, the particle swarm optimization inertia weights are determined by linear decay of the network group parameters. And combined with the individual adaptability of network group parameters Optimal single-unit fit Group fit and collaborative optimization of collision speed According to the global optimization fusion model: Calculate the first Fusion speed of global optimization of group parameters in group network In the formula, Particle swarm optimization inertia weights for network swarm parameters determined by linear decay; For the first Collision speed optimization of group parameters in the network For the first Optimal individual fit of group parameters in the network For group fit, For the first Individual fit of group parameters in the group network All A random number within a given interval.

[0035] In this embodiment, linear decay measurement is used to determine the particle swarm optimization inertia weight. and co-optimized recovery coefficient Specifically, according to the attenuation measurement model: Determine the inertia weights for particle swarm optimization and co-optimized recovery coefficient In the formula, This represents the current iteration number. This represents the maximum number of iterations.

[0036] After determining the fusion speed of global optimization of network group parameters Then, through fusion speed Individual fit of network group parameters Iterative updates are performed, and the individual fitness of the bottom 30% of network group parameters in the updated individual fitness ranking is corrected. Specifically, the fitness update and correction model is implemented according to the following: For the first Individual fit of group network parameters Perform iterative updates and individual unit fit corrections; where, For the first Individual fit before updating group parameters in the network. For the first Individual fit after updating group parameters of the network; For the first Individual fitness after group parameter correction in the group network For group fit.

[0037] Finally, when the change in individual fitness after iterative correction of network group parameters is less than a threshold... or reaching the maximum number of iterations. When the minimum individual fitness among all group network parameters is used, the network group parameter corresponding to the minimum individual fitness is used as the initial parameter of the multi-parameter adaptive intelligent compensation algorithm.

[0038] The globally optimal model parameters (initial parameters) of the multi-parameter adaptive intelligent compensation algorithm obtained through a hybrid strategy of particle swarm optimization and collaborative optimization are not locally optimal. Therefore, in this embodiment, multi-component gas concentration sample data of the gas to be measured under different environmental pressures are obtained and divided according to pressure. A training set and a validation set, and through The training set is used to locally fine-tune the multi-parameter adaptive intelligent compensation algorithm with initial parameters.

[0039] Specifically, fine-tune the model according to the parameters: Local fine-tuning is performed on the multi-parameter adaptive intelligent compensation algorithm with initial parameters; where, To improve the multi-parameter adaptive intelligent compensation algorithm The model parameters to be fine-tuned in the next step; To fine-tune the model parameters The model parameters obtained after local fine-tuning; For the model parameters to be fine-tuned gradient, gradients First-order moment, second-order moment, First-order moments Second-order torch Correction value; First-order moments Second-order torch The attenuation rate, For the model parameters to be fine-tuned The weighted average of historical gradients, For the model parameters to be fine-tuned The weighted average of the squared historical gradients.

[0040] Furthermore, the finely tuned multi-parameter adaptive intelligent compensation algorithm outputs the standardized residual prediction value of the gas sensor through the validation set. Furthermore, the standardized residual prediction value of the gas sensor The actual residual value of the gas sensor is obtained by performing inverse normalization. Specifically, according to the sensor residual model: The standardized residual prediction value output by the multi-parameter adaptive intelligent compensation algorithm after local parameter fine-tuning. Inverse normalization to actual residual value In the formula, These are the actual residual values. The standard deviation and mean.

[0041] In this embodiment, the gas to be tested is collected under different environmental pressures through a gas calibration experiment. Multi-component gas concentration sample data were used, and the data was divided into training and validation sets according to pressure to train and validate the multi-parameter adaptive intelligent compensation algorithm, respectively; and the algorithm was tested under different environmental pressures. The actual residual value of the gas sensor is determined by a multi-parameter adaptive intelligent compensation algorithm after local parameter fine-tuning. Therefore, according to the concentration compensation model: Raw concentration measurements from the gas sensor Compensation is performed to obtain the gas replenishment concentration of the gas sensor. In the formula, For gas sensors under ambient pressure The pressure correction value below, The gas to be tested collected for the gas calibration experiment at ambient pressure A collection of multi-component gas concentration sample data. The first parameter in the validation set output by the multi-parameter adaptive intelligent compensation algorithm. Predicted concentration values ​​for each sample data point. To verify the first set The true concentration value of each sample data; For gas sensors under ambient pressure The actual residual.

[0042] Therefore, the multi-parameter adaptive intelligent compensation algorithm effectively overcomes the impact of environmental pressure fluctuations on the measurement accuracy of gas sensors, significantly improving the detection stability and data reliability of the system under complex on-site conditions. It has environmentally adaptive multi-parameter intelligent measurement and inspection capabilities, which can effectively improve the measurement accuracy of the gas to be measured in industrial or accident sites with complex and diverse gas compositions.

[0043] In a specific application scenario, the accuracy of the sensor measurement values ​​is significantly improved after compensation, as shown in Table 1: Table 1. Comparison of concentration compensation under different environmental pressures In this application scenario, the standard gas concentration is Under the same conditions, the error between the compensated concentration and the true concentration is significantly reduced. (Based on pressure...) For example, The value decreased from 4.819 to 1.136, representing an improvement rate of [missing information]. ; The value decreased from 4.625 to 0.836, representing an improvement rate of [missing information]. Under pressure Under the same conditions, the error reduction exceeded This indicates that the model possesses good generalization ability under different pressure environments. The results show that after compensation... and All decreased significantly, with some pressure points below. The improvement rate exceeded 70%. With an improvement rate exceeding 80%, the adaptive capability and compensation accuracy in complex environments are fully verified.

[0044] In this embodiment, a miniature silent fan is used to extract gas and ensure stable airflow. Multiple gas sensors (such as oxygen, methane, carbon monoxide, and carbon dioxide sensors) detect gas components and generate accurate concentration data. The sensor signals are acquired, processed, corrected, and analyzed by a signal converter and processor to ensure the accuracy and stability of the data. A hybrid strategy of particle swarm optimization and collaborative optimization is used to globally optimize the model parameters of the multi-parameter adaptive intelligent compensation algorithm, and the global optimization multi-parameter adaptive intelligent compensation algorithm is locally fine-tuned to compensate for sensor errors and effectively eliminate gas sensor errors caused by pressure.

[0045] The data analysis module analyzes gas data and transmits data to external devices via the communication module; the touch screen provides real-time data display and convenient user interaction functions, supporting multi-touch and gesture operation to enhance the user experience. This measurement system integrates gas detection, data processing, and human-machine interaction, enabling efficient and accurate real-time monitoring and analysis of various gases in industrial environments.

[0046] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and are not intended to 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.

[0047] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "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, an electrical connection, or a connection that allows communication between them; 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, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0048] In this invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0049] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A portable intelligent environmental multi-parameter measurement system based on TDLAS, characterized in that, include: Multiple parallel gas sensors of different types are arranged on the same horizontal plane with their air inlets parallel to the direction of the air duct. Multiple parallel gas sensors of different types monitor the inhaled gas to be tested based on the principle of tunable semiconductor laser absorption spectroscopy, and obtain multi-component gas concentration data of the gas to be tested. Based on the ambient pressure of the gas to be measured, a multi-parameter adaptive intelligent compensation algorithm is used to compensate for the concentration data of the multi-component gas, and the compensated data is then inversely normalized to obtain the actual concentration value of the gas to be measured.

2. The system according to claim 1, characterized in that, The gas sensor includes a carbon monoxide sensor for acquiring the concentration of carbon monoxide in the gas to be measured, and the ambient pressure of the gas to be measured is acquired through a pressure sensor built into the carbon monoxide sensor.

3. The system according to claim 1, characterized in that, The model parameters of the multi-parameter adaptive intelligent compensation algorithm are globally optimized by a hybrid strategy of particle swarm optimization and cooperative optimization, and the globally optimal model parameters are used as the initial parameters of the multi-parameter adaptive intelligent compensation algorithm. The multi-parameter adaptive intelligent compensation algorithm with initial parameters is locally fine-tuned, and the standardized residual prediction value of the gas sensor is output based on the fine-tuned multi-parameter adaptive intelligent compensation algorithm. Furthermore, the standardized residual prediction value of the gas sensor The actual residual value of the gas sensor is obtained by performing inverse normalization. ; Pressure compensation was performed on the ambient pressure of the gas to be measured through a gas calibration experiment to obtain the pressure correction value of the gas sensor. ; Based on the actual residual value of the gas sensor Pressure correction value Raw concentration measurements from the gas sensor Compensation is performed to obtain the gas replenishment concentration of the gas sensor. .

4. The system according to claim 3, characterized in that, Individual fitness of network group parameters through multi-parameter adaptive intelligent compensation algorithm Optimal individual fit for updating network group parameters And select the individual fit from all network group parameters. The optimal individual fitness of the minimum network group parameters is the group fitness of the multi-parameter adaptive intelligent compensation algorithm. ; Based on the parameter quality of network group parameters, calculate the collaborative optimization collision speed of network group parameters. Furthermore, the particle swarm optimization inertia weights are determined by linear decay of the network group parameters. And combined with the individual adaptability of network group parameters Optimal single-unit fit Group fit Calculate the fusion speed of global optimization of network group parameters. ; By fusion speed Individual fit of network group parameters Perform iterative updates, and correct the individual fitness of the network group parameters that are ranked at the bottom 30% in the individual fitness after the iterative updates. The network group parameters corresponding to the smallest individual fitness among all group network parameters are used as the initial parameters for the multi-parameter adaptive intelligent compensation algorithm.

5. The system according to claim 4, characterized in that, Multi-component gas concentration sample data of the gas to be tested under different environmental pressures were obtained and divided into training set and validation set according to pressure, so as to train and validate the multi-parameter adaptive intelligent compensation algorithm respectively. According to the single-unit fit model: Calculate the first Individual fit of group network parameters In the formula, The multi-parameter adaptive intelligent compensation algorithm adopts the first The first validation set output when grouping network group parameters Predicted concentration values ​​for each sample data point. To verify the first set The true concentration value of each sample data point. The multi-parameter adaptive intelligent compensation algorithm adopts the first... When grouping network parameters, the weights from the input layer to the hidden layer and the weights from the hidden layer to the output layer are used. To verify the quantity of the sample data in the set.

6. The system according to claim 4, characterized in that, Based on the collaborative optimization mechanism and random pairing of network group parameters, collision simulations are performed on the network group parameters, and the parameter quality model is followed: Calculate the first Collision speed of group network parameters ; In the formula, They are randomly paired for the first time. Group, No. The quality of group parameters in the group network. These are the maximum and minimum values ​​of individual fitness among all network group parameters, respectively; It is a very small positive number; The first Group, No. Optimal individual fit of group parameters in the network; The first Group, No. Pre-collision velocity of the group parameters of the network. These are the collaborative optimization recovery coefficients for network group parameters determined through linear decay.

7. The system according to claim 4, characterized in that, According to the global optimization fusion model: Calculate the first Fusion speed of global optimization of group parameters in group network ; In the formula, Particle swarm optimization inertia weights for network swarm parameters determined by linear decay; For the first Collision speed optimization of group parameters in the network For the first Optimal individual fit of group parameters in the network For group fit, For the first Individual fit of group parameters in the group network All A random number within a given interval.

8. The system according to claim 4, characterized in that, Update and correct the model based on its fit: For the Individual fit of group parameters in group network Perform iterative updates and individual component adaptation corrections; In the formula, For the first Individual fit before updating group parameters in the network. For the first Individual fit after updating group parameters of the network; For the first Individual fitness after group parameter correction in the group network For group fit.

9. The system according to claim 3, characterized in that, Multi-component gas concentration sample data of the gas to be tested under different environmental pressures were acquired and divided into training set and validation set according to pressure; the training set includes... Group; pass Set up a training set and fine-tune the model according to the parameters: Local fine-tuning of a multi-parameter adaptive intelligent compensation algorithm with initial parameters; In the formula, To improve the multi-parameter adaptive intelligent compensation algorithm The model parameters to be fine-tuned in the next step; To fine-tune the model parameters The model parameters obtained after local fine-tuning; For the model parameters to be fine-tuned gradient, gradients First-order moment, second-order moment, First-order moments Second-order torch Correction value; First-order moments Second-order torch The attenuation rate, For the model parameters to be fine-tuned The weighted average of historical gradients, For the model parameters to be fine-tuned The weighted average of the squared historical gradients.

10. The system according to claim 3, characterized in that, The test gas was collected under different ambient pressures through a gas calibration experiment. The multi-component gas concentration sample data were divided into training and validation sets according to pressure, so as to train and validate the multi-parameter adaptive intelligent compensation algorithm respectively. And according to the concentration compensation model: Raw concentration measurements from the gas sensor Compensation is performed to obtain the gas replenishment concentration of the gas sensor. ; In the formula, For gas sensors under ambient pressure The pressure correction value below, The gas to be tested collected for the gas calibration experiment at ambient pressure A collection of multi-component gas concentration sample data. The first parameter in the validation set output by the multi-parameter adaptive intelligent compensation algorithm. Predicted concentration values ​​for each sample data point. To verify the first set The true concentration value of each sample data; For gas sensors under ambient pressure The actual residual.

Images