A multi-parameter gas sensor, monitoring method and monitoring system thereof

Abstract

The application discloses a kind of multi-parameter gas sensor, monitoring method and its monitoring system, belong to gas sensing detection technical field;Method includes: determining target gas response enhancement core dimension, collecting characteristic parameters and being analyzed by correlation degree and redundancy, constructs multi-target gas core benchmark parameter system;Matching core parameter and sensor acquisition channel and signal threshold, collect original signal and complete directional interception and differentiating gain enhancement;Interference feature library is constructed, multivariate linear decoupling model is trained to analyze interference amplitude, stripping interference and verifying pure signal, finally complete gas identification;System contains parameter construction, signal acquisition enhancement, interference processing identification three big modules;Sensor is integrated by sensing probe, signal acquisition circuit and data processing unit and is constituted;The application can optimize parameter structure, improve signal quality and anti-interference ability, effectively improve the stability and identification precision of multi-component gas monitoring, adapt to the gas detection demand of complex scene.

Description

Technical Field

[0001] This invention belongs to the field of gas sensing and detection technology, specifically relating to a multi-parameter gas sensor, monitoring method, and monitoring system. Background Technology

[0002] Currently, multi-component gas monitoring equipment is widely used in industrial and environmental monitoring fields, and the requirements for monitoring stability and identification accuracy are constantly increasing. Traditional monitoring methods still have shortcomings in parameter selection, signal processing, and interference suppression, and the overall effectiveness needs to be improved. Therefore, this invention urgently needs to solve the following technical problems: The lack of a reasonable screening process for monitoring relevant characteristic parameters easily leads to a large amount of invalid and duplicate information, making it difficult to form a stable and reliable monitoring basis and increasing the system's operating load. The signal preprocessing method is relatively simple, and strong and weak signals are not differentiated. Some weak signals are easily affected, and the signal quality and recognition are difficult to meet the usage requirements. Interference signals are difficult to separate effectively in complex operating environments, it is difficult to obtain pure signals, the anti-interference capability of the equipment is limited, and the stability of gas identification results is insufficient. To address this, we propose a multi-parameter gas sensor, monitoring method, and monitoring system. Summary of the Invention

[0003] The purpose of this invention is to provide a multi-parameter gas sensor, monitoring method, and monitoring system to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a monitoring method for a multi-parameter gas sensor, comprising: Step 1: Determine the core research dimensions for response enhancement of each target gas, collect all characteristic monitoring parameters and form a set; based on the response evaluation index of the target gas, conduct correlation and redundancy analysis, screen out the multi-component correlation core benchmark parameters specific to the target gas, and construct a multi-target gas response enhancement core benchmark parameter system. Step 2: Match the sensor hardware acquisition channels and signal characteristic thresholds of each core reference parameter of the target gas, acquire the original response signal of the target environment scene and extract the signal segments that meet the requirements to form a preliminary response signal set; perform differentiated gain enhancement on each preliminary response signal segment to form an enhanced preliminary response signal segment set. Step 3: Construct a multi-target gas interference feature library and match target gas interference templates to identify and mark interference signal segments; train a multivariate linear decoupling model and analyze the interference amplitude to complete signal interference stripping and verification of the pure response signal set; output the pure response signal set of each target gas and complete gas identification.

[0005] Preferably, the specific process of conducting correlation analysis based on the target gas's corresponding response evaluation index is as follows: For each target gas, the corresponding core research dimension for gas response enhancement is determined, and full feature monitoring parameters under each dimension are collected and formed into parameter subsets. The total set of full feature monitoring parameters for the target gas is then obtained. To obtain the gas response evaluation index of the target gas, the Pearson correlation coefficient method is used to calculate the linear correlation between each characteristic monitoring parameter and the response evaluation index within each core research dimension. Obtain the absolute value of the Pearson correlation between each feature monitoring parameter and the response evaluation index, preset the correlation threshold, filter feature monitoring parameters whose absolute correlation value is greater than or equal to the preset threshold, and compile a subset of strongly correlated feature monitoring parameters for the current core research dimension.

[0006] Preferably, the specific process of screening out the multi-component correlation core benchmark parameters specific to the target gas and constructing a multi-target gas response enhancement core benchmark parameter system is as follows: The mutual information method is used to calculate the information redundancy of any two parameters in the strongly correlated feature monitoring parameter subset. A redundancy threshold is preset, and parameters with redundancy exceeding the threshold and low correlation with the response evaluation index are removed until there are no highly redundant parameters in the subset, thus obtaining the core benchmark parameter set of the current core research dimension. Organize the core benchmark parameter sets of each dimension to obtain the total set of core benchmark parameters related to the multi-components specific to the target gas; By compiling a comprehensive set of dedicated parameters for all target gases, a core reference parameter system for enhancing the multi-target gas response of multi-parameter gas sensors is formed.

[0007] Preferably, the specific process of acquiring the original response signal of the target environment scene and extracting signal segments that meet the requirements to form a preliminary response signal set is as follows: For each target gas to be monitored, retrieve the set of multi-component related core reference parameters, clarify the core research dimensions and sensor hardware signal channels corresponding to each core reference parameter, and make the acquisition channels correspond one-to-one with the core reference parameters. The signal characteristic thresholds of each core benchmark parameter are obtained, the target environment scene is continuously monitored and the original response signal is collected; the original response signal is selectively extracted based on the corresponding channel and characteristic threshold, the matching channel data and the signal content that meets the threshold range are retained, and irrelevant signals and out-of-threshold noise are removed to obtain the preliminary response signal fragments; all fragments are integrated to form the preliminary response signal set of the target gas.

[0008] Preferably, the specific process of performing differentiated gain enhancement on each preliminary response signal segment to form an enhanced set of preliminary response signal segments is as follows: The weak signal judgment threshold of the core reference parameter corresponding to the target gas is retrieved, the amplitude of the initial response signal segment is compared with the threshold, and the gain coefficient of the strong response signal segment is set to 1; the low amplitude deviation value of the weak response signal segment is calculated, and the gain coefficient is matched according to the preset deviation range. The signal segments are multiplied according to the gain coefficient to obtain the enhanced segments, and all the enhanced segments are integrated to form a set of enhanced preliminary response signal segments.

[0009] Preferably, the specific process of constructing a multi-target gas interference feature library, matching target gas interference templates, and identifying and marking interference signal segments is as follows: Construct a multi-target gas interference feature library containing cross-interference feature templates and environmental interference feature templates, and match the corresponding interference templates according to the type of target gas. The cosine similarity algorithm is used to calculate the feature similarity between the enhanced signal segment and the two types of templates. Based on the preset similarity threshold, the cross-interference signal segments and environmental noise signal segments are labeled, and the enhanced response signal segment set with the interference location and interference type labeled is output.

[0010] Preferably, the specific process of training a multivariate linear decoupling model and analyzing the disturbance amplitude is as follows: According to the current target gas type, obtain the sensor response signal sample in the corresponding multi-component mixed gas environment; the sample includes the complete response signal with the interference location and interference type marked, as well as the amplitude of the corresponding cross-interference component and the amplitude of the environmental noise component. The samples are divided into training set and validation set according to a preset ratio. The response signal labeled with the location and type of interference is used as the model input, and the amplitude of the corresponding cross-interference component and the amplitude of the environmental noise component are used as the model output. The gradient descent algorithm is used for iterative training and validation is performed on the validation set to obtain a multivariate linear decoupling model that is suitable for the current target gas. Input the set of enhanced response signal segments labeled with interference location and interference type, train the multivariate linear decoupling model, analyze each segment and output the amplitude of the cross-interference component and the amplitude of the environmental noise component in each signal segment that are only for the current target gas.

[0011] Preferably, the specific process of completing signal interference removal and pure response signal set verification, outputting the pure response signal set of each target gas, and completing gas identification is as follows: An interference stripping operation is performed on the enhanced preliminary response signal fragments to obtain the pure response signal fragments corresponding to each core reference parameter of the current target gas, and these fragments are integrated to form a pure response signal set for the target gas. The standard response signal data is retrieved to calculate the feature similarity between the pure response signal set and the standard response signal. The validity of the signal is determined according to the preset verification threshold. If the threshold is not met, the interference removal is deemed insufficient. The cosine similarity threshold for interference identification is adjusted, or the multivariate linear decoupling model is re-iterated and trained to optimize its model parameters, or the original response signal of the target environment is re-acquired. The interference identification and interference removal steps are repeated until a qualified pure response signal set is obtained. All target gases to be monitored are processed sequentially, and a set of pure response signals corresponding to each target gas is output. The set of pure response signals is then sent to the gas identification and processing terminal, and the target gas identification is completed based on the pure response signals.

[0012] Compared with the prior art, the beneficial effects of the present invention are: (1) The multi-parameter gas sensor, monitoring method and monitoring system of this invention, by conducting correlation analysis and redundancy elimination on the monitoring characteristic parameters, constructs a core reference parameter system adapted to multiple target gases, which can effectively screen out invalid and duplicate data, optimize the parameter structure, reduce the system operating load, provide a stable and reliable parameter basis for the overall monitoring process, and improve the standardization and rationality of the monitoring system from the source.

[0013] (2) The multi-parameter gas sensor, monitoring method and monitoring system can significantly improve the problem of weak signals being easily overwhelmed by interference by combining hardware acquisition channels with core parameters, signal threshold directional interception and strong and weak signal differential gain enhancement, improve the overall signal recognition and feature expression, optimize signal preprocessing effect, and lay a high-quality data foundation for subsequent interference processing and gas identification.

[0014] (3) The multi-parameter gas sensor, monitoring method and monitoring system can effectively separate cross interference and environmental noise in complex environments by constructing an interference feature library, training a multivariable linear decoupling model and completing interference stripping and signal verification, stably obtain pure response signals, greatly improve the anti-interference capability of the equipment, improve the stability and accuracy of gas identification results, and optimize the overall monitoring performance. Attached Figure Description

[0015] Figure 1 This is a flowchart of the present invention; Figure 2 This is a device diagram of the multi-parameter gas sensor of the present invention. Figure 3 This is a schematic diagram of the internal wiring structure of the multi-parameter gas sensor of the present invention; Figure 4 This is a schematic diagram of the overall test status of the multi-parameter gas sensor of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1; Please see Figure 1-4 This invention provides a monitoring method for a multi-parameter gas sensor, comprising: Step 1: Determine the core research dimensions for response enhancement of each target gas, collect all characteristic monitoring parameters and form a summary; based on the corresponding response evaluation indicators of the target gas, conduct correlation and redundancy analysis, screen out the multi-component correlation core benchmark parameters specific to the target gas, and construct a multi-target gas response enhancement core benchmark parameter system. The specific process is as follows: Identify all target gases for the electrochemical multi-parameter gas sensor, and for each target gas, determine its own core research dimensions for enhancing gas response, including but not limited to: target gas response characteristics, electrode catalytic activity response to target gas, cross-interference between target gas and other gases, stability of target gas response signal, and the impact of environmental interference on target gas response. For each target gas, collect all characteristic monitoring parameters related to the target gas's response sensitivity, response speed, and selectivity under each core research dimension, forming a subset of characteristic monitoring parameters for each core research dimension. Then, organize all the subsets of characteristic monitoring parameters for each core research dimension to form a total set of all characteristic monitoring parameters for the target gas. The core research dimensions of the target gas are denoted as: The parameter subset corresponding to the core research dimension of the target gas is Where i is the index of the core research dimension under the target gas, n is the total number of monitoring parameters corresponding to the core research dimension, and m is the total number of core research dimensions corresponding to the target gas. Obtain gas response evaluation indicators for the target gas (e.g., response amplitude, response rate, selectivity coefficient, signal-to-noise ratio, etc.). Using these corresponding evaluation indicators as the core, conduct dual data analysis—multi-component gas correlation analysis and parameter redundancy analysis—on the entire set of monitoring parameters for the target gas. Based on the analysis results, select core benchmark parameters for multi-component correlation specific to the target gas. Specifically: For each core research dimension of the target gas, the Pearson correlation coefficient method is used to calculate the linear correlation between each characteristic monitoring parameter and the response evaluation index within the parameter subset of the core research dimension (the parameters are normalized and dimensionless during calculation). The calculation formula is as follows: , in, For the target gas, the first Under the core research dimension, the Pearson correlation coefficient between the j-th feature monitoring parameter and the corresponding response evaluation index; For the target gas, the first The k-th sampled value of the j-th feature monitoring parameter under the core research dimension, where j is the index of the feature monitoring parameter under the core research dimension; For the target gas, the first The mean of the kth sampled value of the jth feature monitoring parameter under the core research dimension; For the target gas, the first Target gas response evaluation indicators under the core research dimensions; For the target gas, the first Under the core research dimension, the target gas response evaluation index corresponding to the kth sampling value; For the target gas, the first The mean values ​​of the target gas response evaluation indicators under the core research dimensions; K is the total number of samples; Obtain the Pearson correlation values ​​of each feature monitoring parameter and response evaluation index under the current core research dimension, and take their absolute values; preset the correlation threshold, filter the feature monitoring parameters in the core research dimension whose absolute correlation value is greater than or equal to the preset correlation threshold, and sort them to obtain a subset of strongly correlated feature monitoring parameters in the current core research dimension. The mutual information method is used to calculate the information redundancy between any two feature monitoring parameters within the subset of strongly correlated feature monitoring parameters of the current core research dimension, quantifying the degree of information overlap and cross-interference between different feature monitoring parameters. The calculation formula is as follows: , in, For the target gas, the mutual information value (i.e., redundancy value) between the j-th and l-th feature monitoring parameters within the subset of strongly correlated feature monitoring parameters in the i-th core research dimension. For the target gas, under the i-th core research dimension, the joint probability distribution of the j-th feature monitoring parameter and the l-th feature monitoring parameter within the strongly correlated feature monitoring parameter subset (the j-th parameter under the i-th core research dimension). With the l-th parameter The historical sampled values ​​are divided into several value intervals. The frequency of common occurrence of the two sets of parameters under each interval combination is counted. The discrete joint probability distribution is obtained by dividing the frequency by the total number of samples. For the target gas, under the i-th core research dimension, the marginal probability distribution of the j-th feature monitoring parameter within the subset of strongly correlated feature monitoring parameters (based on the joint probability distribution of the discretized bins). Summing the joint distribution over all values ​​of the l-th parameter (i.e., summing the row / column dimensions of the joint distribution) yields the corresponding j-th parameter. Discrete marginal probability distribution); For the target gas, under the i-th core research dimension, the marginal probability distribution of the l-th feature monitoring parameter within the strongly correlated feature monitoring parameter subset (based on the joint probability distribution of the discretized bins). Summing the joint distribution over all values ​​of the j-th parameter (i.e., summing the row / column dimensions of the joint distribution) yields the corresponding l-th parameter. Discrete marginal probability distribution); l is the index of the other feature monitoring parameters in the strongly correlated feature monitoring parameter subset, excluding the j-th feature monitoring parameter; A redundancy threshold is preset, and the redundancy value between any two parameters in the strongly correlated feature monitoring parameter subset of the current core research dimension is judged one by one. If the redundancy value between two parameters is greater than or equal to the preset threshold, it is judged as a high redundancy parameter combination, and feature monitoring parameters with relatively low correlation with the target gas response evaluation index are removed. This process is repeated until there are no high redundancy parameter combinations in the strongly correlated feature monitoring parameter subset, and finally the core benchmark parameter set of the current core research dimension is obtained. By compiling the core benchmark parameter sets for each core research dimension corresponding to the current target gas, a total set of multi-component correlation core benchmark parameters specific to the current target gas is obtained. By compiling a complete set of dedicated multi-component correlation core reference parameters for all target gases, a multi-target gas response enhancement core reference parameter system for multi-parameter gas sensors is obtained.

[0018] It should be noted that by identifying the core research dimensions of response enhancement for each target gas and collecting a full range of characteristic monitoring parameters, key dimensions such as gas response, catalytic activity, cross-interference, signal stability, and environmental impact can be fully covered, providing a comprehensive and complete data foundation for subsequent monitoring processes. Using the Pearson correlation coefficient method to screen the correlation of characteristic parameters can accurately retain effective parameters that are highly correlated with gas response evaluation indicators, weaken and eliminate weakly correlated terms, and improve the effectiveness and representativeness of core benchmark parameters. By using the mutual information method to remove highly redundant information within parameters, the computational overhead and identification errors caused by duplicate data can be reduced. The constructed multi-target gas response enhancement core benchmark parameter system can ensure the accuracy and stability of the entire process of signal acquisition, enhancement, interference removal and gas identification from the source.

[0019] Step Two: Match the sensor hardware acquisition channels and signal characteristic thresholds of each core reference parameter of the target gas, acquire the raw response signal of the target environment scene, and extract signal segments that meet the requirements to form a preliminary response signal set; perform differentiated gain enhancement on each preliminary response signal segment to form an enhanced preliminary response signal segment set. The specific process is as follows: The target environment is monitored by an electrochemical multi-parameter gas sensor. For each target gas to be monitored in the target environment, the set of core reference parameters related to the multi-components of the target gas is retrieved. The core research dimension to which each core reference parameter belongs and the corresponding sensor hardware signal channel are identified one by one, so that the signal acquisition channel corresponds to the core reference parameter (e.g., "response amplitude" and "response rise slope" belong to the gas response characteristic dimension and correspond to sensor current acquisition channel 1; "steady-state current value" belongs to the signal stability dimension and corresponds to sensor current acquisition channel 1). By using the sensor operation logs, obtain the signal characteristic thresholds of each core reference parameter (e.g., the effective rise time threshold for "response rise slope" is 0.5-2s, the threshold for "steady-state current value" is 2-10s, and the threshold for "response amplitude" is 0.05-1.0mA). The electrochemical multi-parameter gas sensor is activated to continuously monitor the target environment and collect the raw response signal directly output by the sensor (this signal is a time-varying current / potential time-series electrical signal, which includes the response components of all target gases, the cross-interference components caused by the overlap of core reference parameters, and the ambient temperature, humidity and noise components, and the acquisition time, signal amplitude and corresponding channel information are recorded simultaneously). For each core benchmark parameter of the current target gas, based on its corresponding signal acquisition channel and feature threshold, the original response signal is selectively extracted to obtain a preliminary response signal segment corresponding to the core benchmark parameter, specifically: Only retain channel data that matches the core reference parameter, and remove irrelevant signals from other channels; Extract signal time periods and amplitude data that conform to the characteristic threshold range of the core benchmark parameter, remove noise exceeding the threshold, and obtain the preliminary response signal segment corresponding to the core benchmark parameter; The preliminary response signal segments corresponding to all core reference parameters of the current target gas are integrated to form a preliminary response signal set for the target gas. For each segment of the initial response signal in the current set of initial response signals for the target gas, perform the following enhancement operations: Retrieve the weak signal determination threshold of the core reference parameter corresponding to the target gas from the sensor operation log, and compare the amplitude of the initial response signal segment with the corresponding weak signal determination threshold. If the amplitude of the initial response signal segment is greater than or equal to the corresponding weak signal determination threshold, it is determined to be a strong response signal segment, and the gain coefficient of the strong response signal segment is directly set to 1. If the amplitude of the initial response signal segment is less than the corresponding weak signal determination threshold, it is determined to be a weak response signal segment. The deviation between the amplitude of the segment and the weak signal determination threshold is calculated and recorded as the low amplitude deviation value. Several low-amplitude deviation value ranges are preset, and a corresponding gain coefficient is pre-configured for each deviation value range. The rule is: the larger the upper and lower limits of the low-amplitude deviation value range, the larger the corresponding gain coefficient value. The low-amplitude deviation value of the initial response signal segment is matched with all low-amplitude deviation value intervals, and the corresponding gain coefficient is output. The enhanced initial response signal segment is obtained by multiplying the initial response signal segment based on the gain coefficient of the initial response signal segment (i.e., enhanced signal segment = initial response signal segment × specific gain coefficient). The enhanced preliminary response signal segments corresponding to all core reference parameters of the current target gas are integrated to form an enhanced preliminary response signal segment set.

[0020] It should be noted that by matching the sensor hardware acquisition channels with the core reference parameters and selectively extracting the original response signal based on the signal characteristic threshold, it is possible to extract the effective signal components related to the target gas, reduce irrelevant channel data, over-threshold clutter and invalid timing information, reduce the interference of redundant data on subsequent processing, and improve the effectiveness and relevance of the initial response signal set. A differentiated gain enhancement method is used to partition strong and weak signal segments. Strong signal segments maintain their original output to ensure stability, while weak signal segments are reasonably amplified based on the corresponding gain coefficient matched according to the low amplitude deviation. This improves the problem of low identification of weak response components in the original signal and easy coverage by noise, thereby enhancing the overall uniformity and usability of the signal. Through systematic processing of channel matching, targeted interception, and differentiated enhancement, signal quality can be optimized and signal characteristic performance can be strengthened, providing more stable data support for subsequent interference identification, interference stripping, and gas identification, thereby improving the reliability and identification accuracy of the overall monitoring process from the signal processing stage.

[0021] Step 3: Construct a multi-target gas interference feature library and match target gas interference templates to identify and label interference signal segments; train a multivariate linear decoupling model and analyze the interference amplitude to complete signal interference stripping and verification of the pure response signal set; output the pure response signal set of each target gas and complete gas identification. The specific process is as follows: A multi-target gas interference feature library is constructed. For each target gas, a corresponding cross-interference feature template and an environmental interference feature template are preset in the interference feature library. Among them, the cross-interference feature template is the typical feature of each non-target gas response signal (including amplitude range and time series variation pattern). The environmental interference feature template is the typical characteristics of noise signals caused by temperature and humidity changes in the target environment (including frequency and amplitude fluctuation patterns). Based on the type of target gas to be processed, the cross-interference feature template and environmental interference feature template corresponding to the target gas are matched from the multi-target gas interference feature library; The enhanced preliminary response signal fragment set is matched segment by segment with the obtained cross-interference feature template and environmental interference feature template, respectively, to output an enhanced response signal fragment set labeled with interference location and interference type. The specific matching judgment is as follows: The cosine similarity algorithm is used to calculate the feature similarity values ​​between the enhanced preliminary response signal segment and the cross-interference feature template and the environmental interference feature template, respectively. If the feature similarity value between a signal segment and the cross-interference feature template is greater than a preset similarity threshold, then the signal segment is determined to be highly consistent with the typical features of the cross-interference feature template, and the signal segment is marked as a cross-interference signal segment. If the feature similarity value between a signal segment and the environmental interference feature template is greater than the preset similarity threshold, then the signal segment is determined to be highly consistent with the typical features of the environmental interference feature template, and the signal segment is marked as an environmental noise signal segment. Construct a multivariable linear decoupling model: Based on the current target gas type, obtain sensor response signal samples in a multi-component mixed gas environment; the samples include complete response signals labeled with interference location and interference type, as well as the amplitude of the corresponding labeled cross-interference component and the amplitude of the environmental noise component; The samples are divided into training set and validation set according to a preset ratio. The response signal labeled with the location and type of interference is used as the model input, and the amplitude of the corresponding cross-interference component and the amplitude of the environmental noise component are used as the model output. The gradient descent algorithm is used for iterative training and validation is performed on the validation set to obtain a multivariate linear decoupling model that is suitable for the current target gas. Input the set of enhanced response signal segments labeled with interference location and interference type, train the multivariate linear decoupling model, analyze each segment and output the amplitude of the cross-interference component and the amplitude of the environmental noise component in each signal segment that are only for the current target gas. For each signal segment in the enhanced preliminary response signal segment set, an interference stripping operation is performed, i.e.: pure response signal segment = enhanced preliminary response signal segment − amplitude of the cross-interference component corresponding to the signal segment − amplitude of the ambient noise component corresponding to the signal segment; After completing the interference stripping of all signal segments, the pure response signal segments corresponding to each core reference parameter of the current target gas are obtained; The pure response signal segments corresponding to all core benchmark parameters of the current target gas are integrated to form a pure response signal set for the target gas. The standard response signal data of the target gas is retrieved from the sensor operation log, and the cosine similarity algorithm is used to calculate the feature similarity value between the pure response signal set and the standard response signal. If the similarity value is greater than the preset verification threshold, the set of clean response signals is determined to be free of interference and meets the requirements; if the similarity value is less than or equal to the preset verification threshold, the interference removal is determined to be insufficient. The cosine similarity threshold for interference identification is adjusted, or the multivariate linear decoupling model is retrained to optimize its model parameters (the model parameters are generated by the gradient descent algorithm through iterative training, including the cross-interference component analysis weight and the environmental noise component analysis weight, which are used to analyze the amplitude of the cross-interference component and the amplitude of the environmental noise component in the signal), or the original response signal of the target environment is re-acquired, and the interference identification and interference removal steps are repeated until a qualified set of clean response signals is obtained. For all target gases to be monitored in the target environment, the following operations are performed sequentially: preliminary response signal capture, signal enhancement, interference identification, model decoupling, interference removal and signal verification, to obtain a set of pure response signals for each target gas to be monitored. The set of pure response signals corresponding to each target gas to be monitored is sent to the gas identification and processing terminal. The processing terminal quickly completes the identification of each target gas based on the pure and interference-free response signals, thereby improving the response performance of the target gas and optimizing the identification accuracy.

[0022] It should be noted that by constructing a multi-target gas interference feature library and completing interference template matching, it is possible to distinguish between cross-interference and environmental noise components in the signal, reduce the impact of various interferences on gas identification results, and improve the anti-interference capability of the monitoring process. By analyzing and stripping the interference amplitude using a multivariable linear decoupling model, the effective response of the target gas and the interference components can be effectively separated, resulting in a pure response signal that is closer to the real state and improving the identification bias caused by signal distortion and interference superposition. By combining the standard response signal to complete the verification of the pure signal set, the reliability of the output signal is further guaranteed. Gas identification based on the pure response signal can improve the stability of the identification results and optimize the overall response performance and monitoring accuracy of the sensor.

[0023] A monitoring system for a multi-parameter gas sensor includes: Parameter construction module: Determine the core research dimensions for response enhancement of each target gas, collect all characteristic monitoring parameters and form a set; Based on the response evaluation index of the target gas, conduct correlation and redundancy analysis, screen out the multi-component correlation core benchmark parameters specific to the target gas, and construct a multi-target gas response enhancement core benchmark parameter system; Signal acquisition and enhancement module: Matches the sensor hardware acquisition channels and signal characteristic thresholds of each core reference parameter of the target gas, acquires the original response signal of the target environment scene and extracts signal segments that meet the requirements to form a preliminary response signal set; performs differentiated gain enhancement on each preliminary response signal segment to form an enhanced preliminary response signal segment set; Interference processing and identification module: Constructs a multi-target gas interference feature library and matches target gas interference templates to identify and mark interference signal segments; trains a multivariate linear decoupling model and analyzes the interference amplitude to complete signal interference stripping and verification of the pure response signal set; outputs the pure response signal set of each target gas and completes gas identification.

[0024] A multi-parameter gas sensor includes: a sensing probe, a signal acquisition circuit, and a data processing unit; the sensing probe is connected to the signal acquisition circuit, and the signal acquisition circuit is connected to the data processing unit; the data processing unit integrates a parameter construction module, a signal acquisition enhancement module, and an interference processing and identification module, and is used to execute the monitoring system of the multi-parameter gas sensor to realize parameter construction, signal acquisition enhancement, interference processing, and gas identification of the target gas.

[0025] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A monitoring method using a multi-parameter gas sensor, characterized in that, include: Step 1: Determine the core research dimensions for response enhancement of each target gas, collect all characteristic monitoring parameters and form a set; based on the response evaluation index of the target gas, conduct correlation and redundancy analysis, screen out the multi-component correlation core benchmark parameters specific to the target gas, and construct a multi-target gas response enhancement core benchmark parameter system. Step 2: For each target gas to be monitored, match the sensor hardware acquisition channels and signal characteristic thresholds of each core reference parameter of the target gas, acquire the original response signal of the target environment scene and extract the signal segments that meet the requirements to form a preliminary response signal set; perform differentiated gain enhancement on each preliminary response signal segment to form an enhanced preliminary response signal segment set. Step 3: Construct a multi-target gas interference feature library. For each target gas, the multi-target gas interference feature library has a pre-set corresponding cross-interference feature template and environmental interference feature template. Based on the cross-interference feature template and environmental interference feature template corresponding to the current target gas, the cosine similarity algorithm is used to calculate the feature similarity between the enhanced preliminary response signal segment and the cross-interference feature template and the environmental interference feature template, respectively. Based on the comparison results of the feature similarity with the preset similarity threshold, the interference position and interference type in each enhanced preliminary response signal segment are marked, and a set of enhanced preliminary response signal segments with interference position and interference type markings is obtained. Based on the current target gas type, obtain sensor response signal samples in the corresponding multi-component mixed gas environment. The samples include complete response signals with interference location and interference type labels, as well as the amplitudes of the corresponding cross-interference components and environmental noise components. Using the complete response signals with interference location and interference type labels as model inputs and the amplitudes of the corresponding cross-interference components and environmental noise components as model outputs, a multivariate linear decoupling model adapted to the current target gas is trained. The enhanced preliminary response signal segments labeled with interference location and type are input into the trained multivariate linear decoupling model. The amplitude of the cross-interference component and the amplitude of the ambient noise component for the current target gas in each enhanced preliminary response signal segment are analyzed one by one. The corresponding amplitude of the cross-interference component and the amplitude of the ambient noise component are subtracted from each enhanced preliminary response signal segment to obtain the pure response signal segments corresponding to each core benchmark parameter of the current target gas. These segments are then integrated to form the pure response signal set of the current target gas. Calculate the feature similarity between the set of pure response signals and the standard response signals of the current target gas; When the feature similarity is less than or equal to the preset verification threshold, adjust the cosine similarity threshold used for interference identification, re-train the multivariate linear decoupling model, or re-acquire the original response signal of the target environment scene, and repeat the interference identification and interference stripping until a set of pure response signals with feature similarity greater than the preset verification threshold is obtained; output the set of pure response signals corresponding to each target gas and complete the target gas identification.

2. The monitoring method of a multi-parameter gas sensor according to claim 1, characterized in that, The specific process of conducting correlation analysis based on the target gas response evaluation index is as follows: For each target gas, the corresponding core research dimension for gas response enhancement is determined, and full feature monitoring parameters under each dimension are collected and formed into parameter subsets. The total set of full feature monitoring parameters for the target gas is then obtained. To obtain the gas response evaluation index of the target gas, the Pearson correlation coefficient method is used to calculate the linear correlation between each characteristic monitoring parameter and the response evaluation index within each core research dimension. Obtain the absolute value of the Pearson correlation between each feature monitoring parameter and the response evaluation index, preset the correlation threshold, filter feature monitoring parameters whose absolute correlation value is greater than or equal to the preset threshold, and compile a subset of strongly correlated feature monitoring parameters for the current core research dimension.

3. The monitoring method of a multi-parameter gas sensor according to claim 2, characterized in that, The specific process of screening out the multi-component correlation core benchmark parameters specific to the target gas and constructing a multi-target gas response enhancement core benchmark parameter system is as follows: The mutual information method is used to calculate the information redundancy of any two parameters in the strongly correlated feature monitoring parameter subset. A redundancy threshold is preset, and parameters with redundancy exceeding the threshold and low correlation with the response evaluation index are removed until there are no highly redundant parameters in the subset, thus obtaining the core benchmark parameter set of the current core research dimension. Organize the core benchmark parameter sets of each dimension to obtain the total set of core benchmark parameters related to the multi-components specific to the target gas; By compiling a comprehensive set of dedicated parameters for all target gases, a core reference parameter system for enhancing the multi-target gas response of multi-parameter gas sensors is formed.

4. The monitoring method of a multi-parameter gas sensor according to claim 3, characterized in that, The specific process of collecting the raw response signals of the target environment and extracting signal segments that meet the requirements to form a preliminary set of response signals is as follows: For each target gas to be monitored, retrieve the set of multi-component related core reference parameters, clarify the core research dimensions and sensor hardware signal channels corresponding to each core reference parameter, and make the acquisition channels correspond one-to-one with the core reference parameters. The signal characteristic thresholds of each core benchmark parameter are obtained, the target environment scene is continuously monitored and the original response signal is collected; the original response signal is selectively extracted based on the corresponding channel and characteristic threshold, the matching channel data and the signal content that meets the threshold range are retained, and irrelevant signals and out-of-threshold noise are removed to obtain the preliminary response signal fragments; all fragments are integrated to form the preliminary response signal set of the target gas.

5. The monitoring method of a multi-parameter gas sensor according to claim 4, characterized in that, The specific process of performing differentiated gain enhancement on each preliminary response signal segment to form an enhanced set of preliminary response signal segments is as follows: Retrieve the weak signal determination threshold of the core reference parameters corresponding to each target gas in the sensor operation log, and compare the amplitude of each preliminary response signal segment with the corresponding weak signal determination threshold. If the amplitude of the initial response signal segment is greater than or equal to the corresponding weak signal determination threshold, it is determined to be a strong response signal segment, and the gain coefficient of the strong response signal segment is directly set to 1. If the amplitude of the initial response signal segment is less than the corresponding weak signal determination threshold, it is determined to be a weak response signal segment. The deviation between the amplitude of the segment and the weak signal determination threshold is calculated, and the corresponding gain coefficient is matched according to the deviation. The enhanced initial response signal segment is obtained by multiplying the initial response signal segment based on the gain coefficient of the initial response signal segment. The enhanced preliminary response signal segments corresponding to all core reference parameters of the current target gas are integrated to form an enhanced preliminary response signal segment set.

6. The monitoring method of a multi-parameter gas sensor according to claim 5, characterized in that, The process of labeling the interference location and type in each enhanced preliminary response signal segment based on the comparison results of feature similarity and preset similarity threshold, and obtaining a set of enhanced preliminary response signal segments with interference location and type labels, includes: When the feature similarity between the enhanced preliminary response signal fragment and the cross-interference feature template is greater than the preset similarity threshold, the enhanced preliminary response signal fragment is marked as a cross-interference signal fragment. When the feature similarity between the enhanced preliminary response signal segment and the environmental interference feature template is greater than the preset similarity threshold, the enhanced preliminary response signal segment is labeled as an environmental noise signal segment. By integrating the annotated enhanced preliminary response signal fragments, a set of enhanced preliminary response signal fragments with annotations of interference location and interference type is obtained.

7. A monitoring system for a multi-parameter gas sensor, applied to the monitoring method of a multi-parameter gas sensor as described in any one of claims 1-6, characterized in that, include: Parameter construction module: Determine the core research dimensions for response enhancement of each target gas, collect all characteristic monitoring parameters and form a set; Based on the response evaluation index of the target gas, conduct correlation and redundancy analysis, screen out the multi-component correlation core benchmark parameters specific to the target gas, and construct a multi-target gas response enhancement core benchmark parameter system; Signal acquisition and enhancement module: For each target gas to be monitored, the sensor hardware acquisition channel and signal characteristic threshold of each core reference parameter of the target gas are matched, the original response signal of the target environment scene is acquired and the signal segments that meet the requirements are extracted to form a preliminary response signal set; the preliminary response signal segments are enhanced with differentiated gain to form an enhanced preliminary response signal segment set. Interference processing and identification module: Constructs a multi-target gas interference feature library. For each target gas, the multi-target gas interference feature library has a preset corresponding cross-interference feature template and environmental interference feature template. Based on the cross-interference feature template and environmental interference feature template corresponding to the current target gas, the cosine similarity algorithm is used to calculate the feature similarity between the enhanced preliminary response signal segment and the cross-interference feature template and the environmental interference feature template, respectively. Based on the comparison results of the feature similarity with the preset similarity threshold, the interference position and interference type in each enhanced preliminary response signal segment are marked, and a set of enhanced preliminary response signal segments with interference position and interference type markings is obtained. Based on the current target gas type, obtain sensor response signal samples in the corresponding multi-component mixed gas environment. The samples include complete response signals with interference location and interference type labels, as well as the amplitudes of the corresponding cross-interference components and environmental noise components. Using the complete response signals with interference location and interference type labels as model inputs and the amplitudes of the corresponding cross-interference components and environmental noise components as model outputs, a multivariate linear decoupling model adapted to the current target gas is trained. The enhanced preliminary response signal segments labeled with interference location and type are input into the trained multivariate linear decoupling model. The amplitude of the cross-interference component and the amplitude of the ambient noise component for the current target gas in each enhanced preliminary response signal segment are analyzed one by one. The corresponding amplitude of the cross-interference component and the amplitude of the ambient noise component are subtracted from each enhanced preliminary response signal segment to obtain the pure response signal segments corresponding to each core benchmark parameter of the current target gas. These segments are then integrated to form the pure response signal set of the current target gas. Calculate the feature similarity between the set of pure response signals and the standard response signals of the current target gas; When the feature similarity is less than or equal to the preset verification threshold, adjust the cosine similarity threshold used for interference identification, re-train the multivariate linear decoupling model, or re-acquire the original response signal of the target environment scene, and repeat the interference identification and interference stripping until a set of pure response signals with feature similarity greater than the preset verification threshold is obtained; output the set of pure response signals corresponding to each target gas and complete the target gas identification.

8. A multi-parameter gas sensor, characterized in that, include: The system comprises a sensing probe, a signal acquisition circuit, and a data processing unit; the sensing probe is connected to the signal acquisition circuit, and the signal acquisition circuit is connected to the data processing unit; the data processing unit integrates a parameter construction module, a signal acquisition enhancement module, and an interference processing and identification module, and is used to execute the monitoring system described in claim 7 to achieve the complete operation of the target gas monitoring process.

Images