Self-calibration method of natural gas detection equipment

Abstract

The invention discloses a self-calibration method for natural gas detection equipment. The self-calibration method comprises the following steps: constructing a calibration correction model based on detection data in historical natural gas data and calibration data difference; acquiring a first detection result of the target position based on a quantum sensing technology, and acquiring a second detection result according to the first detection result and the calibration correction model; obtaining the dissipation similarity difference between the dissipation condition and the first detection result and the dissipation similarity difference between the dissipation condition and the second detection result based on the dissipation waveform similarity; environment similarity differences are obtained according to the environment air data, the first detection result and the second detection result; obtaining a sample similarity difference according to the sample waveform, the first detection result and the second detection result; and outputting a natural gas leakage calibration result according to a comparison result of the dissipation similarity difference, the environment similarity difference, the sample similarity difference and a preset difference threshold. The method has the beneficial effects that the accuracy of calibration of the natural gas equipment is improved, and additional acquisition equipment is not needed.

Description

Technical Field

[0001] The present invention relates to the technical field of natural gas detection, and particularly to a self-calibration method for natural gas detection equipment. Background Art

[0002] Natural gas detection technology plays a crucial role in fields such as industry, residential life, and environmental protection, especially in monitoring and preventing natural gas leaks. However, detecting natural gas leaks at low concentrations faces many challenges, mainly reflected in weak signals, susceptibility to interference, and insufficient reliability.

[0003] At low concentrations, the signals generated by natural gas leaks are often very weak, making it extremely difficult for detection equipment to capture these signals. At the same time, environmental noise and interference from other gases further exacerbate the complexity of detection. For example, other combustible gases or volatile organic compounds in the environment may produce signals similar to those of natural gas leaks, misleading the detection equipment and causing false alarms. Such false alarms not only cause unnecessary panic and waste of resources but may also disrupt normal production and living order.

[0004] The patent "Natural Gas Leak Detection Method and Multimodal Natural Gas Leak Detection System", publication number: CN116379359A, publication date: July 4, 2023, specifically discloses that the method includes: obtaining at least two consecutive frames of infrared video images of the area to be detected; after transforming the obtained two consecutive frames of infrared video images to a preset size, inputting them into a trained natural gas detection model for detection to obtain corresponding detection results respectively; the natural gas detection model is mainly constructed with the yolov5 model framework, replaces the C3 network structure with the C2f network structure, and uses the SoftPool operation to replace the convolution operation for 2-fold downsampling for feature downsampling; based on the detection results corresponding to the two consecutive frames of infrared video images, determine whether it is a false result, otherwise it is considered that a natural gas leak is detected and the detection result is output. This solution judges whether there is a natural gas leak through the detection results of two consecutive frames of infrared video images, but can only exclude detection errors caused by accidental factors, and cannot exclude detection errors caused by continuous factors.

[0005] Patent "A Multimodal Natural Gas Leak Detection System and Method", Publication Number: CN116295788A, Publication Date: June 23, 2023, specifically discloses that the system includes an optical fiber acoustic wave monitoring subsystem, an optoelectronic signal monitoring subsystem, a movable optical image detection subsystem, and a multimodal analysis platform; the optical fiber acoustic wave monitoring subsystem includes an optical fiber monitoring module; each group of optical fiber monitoring modules corresponds to a section of natural gas pipeline, including two sections of auxiliary optical fibers, at least one section of monitoring optical fiber, and corresponding optical fiber pulse emitters and optical fiber sensors; the optoelectronic signal monitoring subsystem includes an optoelectronic monitoring module; each group of optoelectronic monitoring modules corresponds to a pipeline valve, including a laser emitter and an optoelectronic sensor respectively arranged on both sides of the pipeline valve; the movable optical image detection subsystem includes a mobile detection module; each group of mobile detection modules includes a mobile device, an optical gas camera, a locator, and a signal transceiver. This solution improves the accuracy of natural gas leak detection by using a multimodal mutual inspection method. However, it requires setting up multiple detection units, resulting in a high cost, and the error probability of multimodal data acquisition is higher. Summary of the Invention

[0006] In view of the problems of low accuracy and high cost of multimodal detection in the existing technology for natural gas detection, the present invention provides a self-calibration method for natural gas detection equipment, which determines whether there is a natural gas leak problem currently through the bidirectional comparison between a quantum detection sensor and a calibration correction model, and improves the accuracy of the final natural gas leak calibration by excluding environmental interference based on real-time detection data and historical data prediction values, without the need to set up multiple detection units, reducing the detection cost while improving the detection accuracy.

[0007] To achieve the above technical objectives, a technical solution provided by the present invention is a self-calibration method for natural gas detection equipment, including the following steps: S1: Using a machine learning algorithm to construct a calibration correction model based on the difference between the detection data and the calibration data in historical natural gas data; S2: Obtaining a first detection result of the target position based on quantum sensing technology, and obtaining a second detection result based on the first detection result and the calibration correction model; S3: Obtaining the difference in the escape similarity between the escape situation and the first detection result and the second detection result based on the similarity of the escape waveform; S4: Obtaining environmental air data, and obtaining the difference in environmental similarity based on the environmental air data, the first detection result, and the second detection result; S5: Adding natural gas according to a preset concentration to obtain a sample waveform, and obtaining the difference in sample similarity based on the sample waveform, the first detection result, and the second detection result; S6: Outputting a natural gas leak calibration result based on the comparison results of the escape similarity difference, the environmental similarity difference, the sample similarity difference, and a preset difference threshold.

[0008] Further, S1 includes: obtaining historical natural gas data, calculating calibration differences based on detection data and calibration data for each time series; constructing an influence relationship between features and calibration differences based on the time series characteristics and spatial characteristics of natural gas data for each time series and the calibration differences; and constructing a calibration correction model based on the influence relationship between features and calibration differences.

[0009] Further, S2 includes: obtaining gas waveform information of the target location based on quantum sensing technology as the first detection result; using the first detection result, current time series information, and spatial information as the input of the calibration correction model, and using the output of the calibration correction model as the second detection result.

[0010] Further, S3 includes: in response to the output of the second detection result, opening the sealed cavity, and obtaining the escape waveform according to the natural gas escape situation; obtaining the first escape similarity difference based on the escape waveform and the first detection result, and obtaining the second escape similarity difference based on the escape waveform and the second detection result.

[0011] Further, S5 includes: retrieving the corresponding matching type of natural gas according to the similarity interval between the first detection result and the second detection result, matching the corresponding preset concentration according to the natural gas leakage safety threshold, and obtaining a sample waveform according to the corresponding matching type of natural gas and the preset concentration; obtaining the sample similarity difference based on the sample waveform, the first detection result, and the second detection result.

[0012] Further, the matching of the corresponding preset concentration according to the natural gas leakage safety threshold includes: obtaining the maximum possible leakage concentration according to the first detection result and the second detection result; obtaining the corresponding preset concentration according to the difference between the maximum possible leakage concentration and the natural gas leakage safety threshold.

[0013] Further, S6 includes: calculating the weighted sum of differences of the escape similarity difference, environment similarity difference, and sample similarity difference based on preset weight coefficients, comparing the weighted sum of differences with a preset difference threshold. If the weighted sum of differences is less than or equal to the preset difference threshold, the natural gas leakage calibration result is natural gas leakage; if the weighted sum of differences is greater than the preset difference threshold, the natural gas leakage calibration result is no natural gas leakage.

[0014] Further, the escape similarity difference, environment similarity difference, and sample similarity difference at least include difference data points.

[0015] Further, S6 includes: correcting the escape similarity difference with the environment similarity difference and the sample similarity difference, and outputting the natural gas leakage calibration result according to the comparison result between the corrected escape similarity difference and the preset difference threshold.

[0016] Furthermore, the correction of the scattered similarity difference using the environmental similarity difference and the sample similarity difference includes: comparing the first environmental similarity difference with the first scattered similarity difference, retaining the same data points of the first environmental similarity difference and the first scattered similarity difference, and obtaining a first corrected difference; comparing the second environmental similarity difference with the second scattered similarity difference, retaining the same data points of the second environmental similarity difference and the second scattered similarity difference, and obtaining a second corrected difference; obtaining a first correction weight and a second correction weight according to the first sample similarity and the second sample similarity; and obtaining a scattered similarity difference correction value using the first corrected difference, the second corrected difference, the first correction weight, and the second correction weight.

[0017] The beneficial effects of the present invention are as follows: (1) by utilizing a machine learning algorithm to construct a calibration correction model, by learning the characteristics of the detection data and calibration data in the historical natural gas data and the difference between the calibration result and the detection result caused by the characteristic difference, a corrected second detection result is obtained according to the calibration correction model based on the actual detection result, i.e., the first detection result. Then, by comparing the similarity of the escape waveform, the ambient air data, and the waveform of the sample after adding natural gas of known concentration, interference factors are eliminated from multiple dimensions, the error caused by a single factor is reduced, and the reliability of the detection result is improved.

[0018] (2) Using quantum sensing principles, such as quantum dot fluorescence enhancement technology, to improve the ability to detect weak absorption signals. Quantum states are extremely sensitive to small changes and can provide clearer signal characteristics at extremely low concentrations. The timing information includes at least the detection time and the data acquisition time interval, and the spatial information includes at least the sensor position, the spatial layout of the target position, and the environmental data. Through the calibration correction model, the detection deviation that may exist in the current timing and space is obtained according to the matching of the time and space characteristics, and the second detection result after the time and space factors are compensated is output. Since the aging of the equipment is not only related to the timing, but also to the spatial layout, such as the target location is relatively humid and the space is small, it may cause the equipment to age faster and the detection accuracy to decrease faster. Therefore, detection compensation is performed according to the time and space characteristics to further improve the detection accuracy.

[0019] 3. With the sealed chamber open, observe the waveform characteristics of the escaping gas. Compare the similarity of the escaping waveform with the first detection result and the second detection result. Quantify the leakage possibility through the waveform difference, and reflect the possible situation of the leakage source through the separate comparison of the first detection result and the second detection result. If the difference in the escaping similarity between the first detection result and the second detection result is small, it indicates a high probability of natural gas leakage. If the difference in the escaping similarity between the first detection result and the second detection result is large, it indicates a low probability of natural gas leakage. If the difference in the escaping similarity of the first detection result is inconsistent with the difference in the escaping similarity of the second detection result, and the difference in the escaping similarity of any detection result is small, it is considered that there is a probability of natural gas leakage, and there is an error in the sensor or the calibration correction model, and the sensor or the calibration correction model needs to be repaired. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic flowchart of a self - calibration method for a natural gas detection device of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only the best embodiments of the present invention, which are only used to explain the present invention and do not limit the protection scope of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0022] As Figure 1 shown, a self - calibration method for a natural gas detection device includes the following steps: S1: Use a machine learning algorithm to construct a calibration correction model based on the detection data and the difference in calibration data in historical natural gas data; S2: Obtain the first detection result of the target position based on quantum sensing technology, and obtain the second detection result based on the first detection result and the calibration correction model; S3: Obtain the difference in the escaping similarity between the escaping situation and the first detection result and the second detection result based on the escaping waveform similarity; S4: Obtain the environmental air data, and obtain the difference in environmental similarity based on the environmental air data, the first detection result, and the second detection result; S5: Add natural gas according to a preset concentration to obtain a sample waveform, and obtain the difference in sample similarity based on the sample waveform, the first detection result, and the second detection result; S6: Output the natural gas leakage calibration result according to the comparison results of the difference in escaping similarity, the difference in environmental similarity, the difference in sample similarity, and a preset difference threshold.

[0023] In this embodiment, by using a machine learning algorithm to construct a calibration correction model, the characteristics of the detection data and the calibration data in the historical natural gas data and the differences between the calibration results and the detection results caused by the characteristic differences are learned, so that the corrected second detection result is obtained according to the actual detection result, that is, the first detection result, based on the calibration correction model. Furthermore, by comparing the similarity of the escape waveforms, comparing the ambient air data, and comparing the sample waveforms after adding natural gas with a known concentration, interference factors are excluded from multiple dimensions, the error caused by a single factor is reduced, and the reliability of the detection result is improved.

[0024] Specifically, using a machine learning algorithm to construct a calibration correction model based on the detection data and the calibration data differences in the historical natural gas data includes: Obtain historical natural gas data, and calculate the calibration difference based on the detection data and the calibration data at each time series; Based on the time series characteristics and the spatial characteristics of the natural gas data at each time series, construct the influence relationship between the characteristics and the calibration difference based on the calibration difference; Construct a calibration correction model based on the influence relationship between the characteristics and the calibration difference.

[0025] The historical natural gas data at least includes the detection data and the calibration data of each working condition at each historical time series. By using a machine learning algorithm to analyze the historical natural gas data, the influence of the changes in the time series characteristics and the spatial characteristics in the historical state on the difference is obtained, the correction of the detection data is realized, and the detection accuracy is improved. The historical natural gas data contains the detection and calibration conditions of the equipment under different working conditions and environments. The machine learning algorithm mines the law of the difference between the detection data and the calibration data, and establishes a model that can reflect the performance change of the equipment. For example, if the equipment ages over time and causes a detection deviation, the model learns this trend and predicts the correction value, and adjusts the data according to the influence relationship between the characteristics and the calibration difference during subsequent detection to improve the detection accuracy.

[0026] Step S2 includes: Obtain the gas waveform information of the target position as the first detection result based on the quantum sensing technology; Use the first detection result, the current time series information, and the spatial information as the input of the calibration correction model, and use the output of the calibration correction model as the second detection result.

[0027] Using quantum sensing principles, such as quantum dot fluorescence enhancement technology, to improve the detection ability of weak absorption signals. Quantum states are extremely sensitive to tiny changes and can provide clearer signal characteristics at extremely low concentrations. The timing information includes at least the detection time and the data acquisition time interval, and the spatial information includes at least the sensor position, the spatial layout of the target position, and the environmental data. The calibration correction model obtains the possible detection deviation at the current time sequence and space according to the matching of the spatio-temporal characteristics, and then outputs the second detection result after compensating for the spatio-temporal factors. Since the aging condition of the device is related not only to the time sequence but also to the spatial layout, such as the target position being prone to moisture and the space being narrow, etc., it may cause the device to age faster and the detection accuracy to decline faster. Therefore, detecting compensation is carried out according to the spatio-temporal characteristics to further improve the detection accuracy. Machine learning algorithms can adopt algorithms such as support vector machine regression algorithm and neural network algorithm.

[0028] Step S3 includes: In response to the output of the second detection result, open the sealed cavity and obtain the escape waveform according to the natural gas escape situation; Obtain the first escape similarity difference with the escape waveform and the first detection result, and obtain the second escape similarity difference with the escape waveform and the second detection result.

[0029] When the sealed cavity is opened, observe the waveform characteristics of the escaped gas, compare the escape waveform with the first detection result and the second detection result for similarity, quantify the leakage possibility through the waveform difference, and reflect the possible situation of the leakage source through the separate comparison of the first detection result and the second detection result. If the escape similarity differences between the first detection result and the second detection result are both small, it indicates a high probability of natural gas leakage. If the escape similarity differences between the first detection result and the second detection result are large, it indicates a low probability of natural gas leakage. If the escape similarity difference of the first detection result is inconsistent with the escape similarity difference of the second detection result, and the escape similarity difference of any detection result is small, it is considered that there is a probability of natural gas leakage and there is an error in the sensor or the calibration correction model, and the sensor or the calibration correction model needs to be repaired.

[0030] Step S4 includes: Obtain the ambient air data and obtain the ambient detection waveform according to the ambient air data; Obtain the first ambient similarity difference with the ambient detection waveform and the first detection result, and obtain the second ambient similarity difference with the ambient detection waveform and the second detection result.

[0031] By comparing the waveform similarity of the ambient detection waveform with the first detection result and the second detection result, the gas interference is excluded, and the accuracy of the final natural gas leakage judgment result is improved.

[0032] Step S5 includes: Retrieve natural gas of the corresponding matching type according to the similarity interval of the first detection result and the second detection result, match the corresponding preset concentration according to the natural gas leakage safety threshold, and obtain the sample waveform based on the natural gas of the corresponding matching type and the preset concentration; obtain the sample similarity difference based on the sample waveform, the first detection result, and the second detection result.

[0033] Match the type of natural gas according to the similarity interval of the natural gas content in the first detection result and the second detection result. At this time, the first detection result and the second detection result at least include gas components. Match the type of natural gas according to the gas components included in both the first detection result and the second detection result to improve the adaptability of the added natural gas and more accurately simulate the real leakage situation.

[0034] Matching the corresponding preset concentration according to the natural gas leakage safety threshold includes: Obtain the maximum possible leakage concentration based on the first detection result and the second detection result; Obtain the corresponding preset concentration according to the difference between the maximum possible leakage concentration and the natural gas leakage safety threshold.

[0035] The natural gas leakage safety threshold can be set according to expert experience. For example, it can be set to 20% - 30% of the lower explosion limit of natural gas, so as to ensure that even if natural gas leaks, the addition of sample natural gas will not pose a safety risk. Obtain the corresponding preset concentration based on the difference between the maximum possible leakage concentration in the first detection result and the second detection result and the natural gas leakage safety threshold. Under the premise of ensuring safety, ensure the adaptability of the added sample natural gas as much as possible, thereby further improving the accuracy of natural gas leakage detection.

[0036] Inject natural gas of the corresponding matching type into the test cavity at the preset concentration, record the waveform change in the test cavity after injecting the natural gas, and obtain the sample waveform.

[0037] Since in order to avoid the same leakage risk in the test cavity, the corresponding preset concentration is obtained based on the difference between the maximum possible leakage concentration and the natural gas safety threshold. At this time, there is a difference between the injection concentration of the sample natural gas and the leakage concentration of the natural gas. Therefore, construct the correlation between concentration and waveform based on historical natural gas concentration and waveform data. At this time, obtaining the sample similarity difference based on the sample waveform, the first detection result, and the second detection result includes: Compensate and correct the concentration of the sample waveform according to the correlation between concentration and waveform to obtain the corrected sample waveform; Obtain the first sample similarity difference based on the corrected sample waveform and the first detection result; Obtain the second sample similarity difference based on the corrected sample waveform and the second detection result.

[0038] In this embodiment, the similarity is calculated through a similarity algorithm, such as the correlation coefficient, Euclidean distance, etc. For example, the dispersion similarity, environmental similarity, and sample similarity are calculated according to the Euclidean distance. The Euclidean distance measures the straight-line distance between two waveforms in a multi-dimensional space. The smaller the Euclidean distance, the closer the two waveforms are: where x i and y i are respectively the i-th data points of the two waveforms being compared, and I is the total number of waveform data points. When calculating the first dispersion similarity difference, the Euclidean distance is calculated based on the dispersion waveform and the detected waveform in the first detection result. Taking the i-th data point of the dispersion waveform as x i,s , and taking the i-th data point of the detected waveform in the first detection result as y i,s1 . At this time, the first dispersion similarity difference is: where d s1 is the first dispersion similarity.

[0039] At this time, step S6 includes: Calculating the weighted sum of differences of the dispersion similarity difference, environmental similarity difference, and sample similarity difference based on a preset weight coefficient, and comparing the weighted sum of differences with a preset difference threshold. If the weighted sum of differences is less than or equal to the preset difference threshold, the natural gas leakage calibration result is natural gas leakage. If the weighted sum of differences is greater than the preset difference threshold, the natural gas leakage calibration result is no natural gas leakage.

[0040] The weighted sum of differences of the dispersion similarity difference, environmental similarity difference, and sample similarity difference calculated based on a preset weight coefficient is: P = αA + βB + γC; α + β + γ = 1; α > 0, β < 0, γ > 0.

[0041] where P is the weighted sum of differences, α is the weight coefficient of the dispersion similarity difference, β is the weight coefficient of the environmental similarity difference, γ is the weight coefficient of the sample similarity difference, A is the dispersion similarity difference, B is the environmental similarity difference, C is the sample similarity difference, d s1 is the first dispersion similarity difference, d s2 is the second dispersion similarity difference, d h1 is the first environmental similarity difference, d h2 is the second environmental similarity difference, d b1 is the first sample similarity difference, d b2It is the second sample similarity difference.

[0042] By making the sensor data and the model data approximate each other, the deviation value of the difference is reduced, thereby improving the calibration accuracy. At the same time, through the preset weight coefficients, the weighted sum calculation is carried out among the similarity differences, approaching a more realistic natural gas leakage situation, and the final difference value is obtained.

[0043] Specifically, the historical natural gas data is divided into model training data and weight training data. The calibration correction model is trained with the model training data, and the preset weight coefficients are trained with the weight training data. In this embodiment, step S1 further includes: dividing the historical natural gas data into model training data and weight training data based on the dichotomy method; The detection data and calibration data are obtained from the model training data.

[0044] Step S6 further includes: The fugitive similarity difference, environmental similarity difference, and sample similarity difference are obtained from the weight training data. According to the preset difference threshold, the correlation between the weight coefficients of the fugitive similarity difference, environmental similarity difference, and sample similarity difference and the environment is calculated; the current environmental data is obtained, and the preset weight coefficients are obtained according to the current environmental data and the correlation between the weight coefficients and the environment.

[0045] In this embodiment, the historical natural gas data is interleaved and divided according to the time sequence. The continuous historical natural gas data is divided into model training data and weight training data according to the front and back time sequences respectively, so as to further adaptively correct the proportion of each difference on the basis of calibration correction, so as to improve the accuracy of the final natural gas calibration. In other cases, the fugitive similarity difference, environmental similarity difference, and sample similarity difference at least include difference data points. At this time, step S6 includes: The fugitive similarity difference is corrected with the environmental similarity difference and the sample similarity difference, and the natural gas leakage calibration result is output according to the comparison result between the corrected fugitive similarity difference and the preset difference threshold.

[0046] The fugitive similarity difference is corrected through the environmental similarity difference and the sample similarity difference, thereby excluding the differences brought by environmental gases and natural gas type concentrations, and further improving the accuracy of the natural gas leakage calibration result.

[0047] Specifically, correcting the fugitive similarity difference with the environmental similarity difference and the sample similarity difference includes: Comparing the first environmental similarity difference with the first fugitive similarity difference, retaining the same data points of the first environmental similarity difference and the first fugitive similarity difference, and obtaining the first corrected difference; Compare the second environmental similarity difference with the second dissipation similarity difference, retain the same data points of the second environmental similarity difference and the second dissipation similarity difference, and obtain the second correction difference; Obtain a first correction weight and a second correction weight according to the first sample similarity difference and the second sample similarity difference; Obtain a dissipation similarity difference correction value based on the first correction difference, the second correction difference, the first correction weight, and the second correction weight.

[0048] Obtain a first correction difference and a second correction difference respectively according to the comparison results of the first environmental similarity difference and the first dissipation similarity difference, and the second environmental similarity difference and the second dissipation similarity difference. The environmental similarity difference represents the difference between the current environmental gas and the detection result, and the dissipation similarity difference represents the difference between the dissipated gas and the detection result. By screening out the same data points of the two, the part where there are differences between the dissipated gas and the environmental gas is obtained, so as to further judge whether there is a possibility of dissipation in this part, and judge the accurate probability of the result of the sensor and the calibration correction model according to the comparison of the first sample similarity difference and the second sample similarity difference, that is, the known concentration is obtained according to the maximum leakage possible concentration of the first detection result and the second detection result. If the sample waveform is more similar to the first detection result, the first detection result has a higher first correction weight. According to the first correction weight, the second correction weight, the first correction difference, and the second correction difference, the dissipation difference after excluding environmental interference, incorporating the sensor, and considering the accuracy of the calibration correction model is obtained, so as to judge whether there is leaked natural gas, thereby improving the accuracy of the calibration of natural gas equipment.

[0049] Judge whether natural gas leaks according to the comparison result of the dissipation similarity difference correction value and the preset difference threshold. The preset difference threshold can be set according to the actual safety threshold. In some other cases, after step S2, judge whether the first detection result and the second detection result are greater than the safety threshold. If so, it is considered that natural gas leaks. If not, execute steps S3-S6. If it is judged that natural gas leaks, immediately start the corresponding leakage response procedures, such as closing the valve, evacuating personnel, etc., and judge the accuracy of the sensor and the calibration correction model according to the difference between the judgment result and the first detection result and the second detection result, and prompt the operator to perform maintenance.

[0050] The above specific implementation manners are the preferred implementation manners of a self-calibration method for a natural gas detection device of the present invention, and do not limit the specific scope of the present invention. The scope of the present invention includes but is not limited to this specific implementation manner. All equivalent changes made according to the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A self-calibration method for a natural gas detection device, characterized in that: It includes the following steps: S1: Construct a calibration correction model based on the detection data and calibration data differences in historical natural gas data using a machine learning algorithm; S2: Obtain the first detection result of the target location based on quantum sensing technology, and obtain the second detection result using the first detection result and the calibration correction model; S3: Obtain the difference in the fugitive similarity between the fugitive situation and the first detection result and the second detection result based on the similarity of the fugitive waveforms; S4: Obtain environmental air data, and obtain the difference in environmental similarity based on the environmental air data, the first detection result, and the second detection result; S5: Add natural gas according to a preset concentration to obtain a sample waveform, and obtain the difference in sample similarity based on the sample waveform, the first detection result, and the second detection result; S6: Output the calibration result of natural gas leakage according to the comparison results of the fugitive similarity difference, the environmental similarity difference, the sample similarity difference, and the preset difference threshold.

2. The self-calibration method of a natural gas detection device according to claim 1, wherein: The S1 includes: Obtain historical natural gas data, and calculate the calibration difference based on the detection data and calibration data at each time series; Based on the time series characteristics and spatial characteristics of the natural gas data at each time series, construct the influence relationship between the characteristics and the calibration difference; Construct a calibration correction model based on the influence relationship between the characteristics and the calibration difference.

3. The self-calibration method of a natural gas detection device according to claim 1, wherein: The S2 includes: Obtain the gas waveform information of the target location based on quantum sensing technology as the first detection result; Use the first detection result, the current time series information, and the spatial information as the input of the calibration correction model, and use the output of the calibration correction model as the second detection result.

4. The self-calibration method of a natural gas detection device according to claim 1, wherein: The S3 includes: In response to the output of the second detection result, open the sealed chamber, and obtain the fugitive waveform according to the natural gas fugitive situation; Obtain the first fugitive similarity difference using the fugitive waveform and the first detection result, and obtain the second fugitive similarity difference using the fugitive waveform and the second detection result.

5. The self-calibration method of a natural gas detection device according to claim 1, wherein: The S5 includes: Retrieve the corresponding matched type of natural gas according to the similarity interval of the first detection result and the second detection result, match the corresponding preset concentration according to the natural gas leakage safety threshold, and obtain the sample waveform according to the corresponding matched type of natural gas and the preset concentration; obtain the difference in sample similarity based on the sample waveform, the first detection result, and the second detection result.

6. The self-calibration method of a natural gas detection device according to claim 5, wherein: The matching of the corresponding preset concentration according to the natural gas leakage safety threshold includes: Obtain the maximum possible leakage concentration based on the first detection result and the second detection result; Obtain the corresponding preset concentration according to the difference between the maximum possible leakage concentration and the natural gas leakage safety threshold.

7. The self-calibration method of a natural gas detection device according to claim 1, wherein: The S6 includes: Calculate the weighted sum of differences of the fugitive similarity difference, environmental similarity difference, and sample similarity difference based on the preset weight coefficients. Compare the weighted sum of differences with the preset difference threshold. If the weighted sum of differences is less than or equal to the preset difference threshold, the natural gas leakage calibration result is natural gas leakage. If the weighted sum of differences is greater than the preset difference threshold, the natural gas leakage calibration result is no natural gas leakage.

8. The self-calibration method of a natural gas detection device according to claim 1, wherein: The fugitive similarity difference, environmental similarity difference, and sample similarity difference at least include difference data points.

9. The self-calibration method of a natural gas detection device according to claim 8, wherein: The S6 includes: Correct the fugitive similarity difference with the environmental similarity difference and the sample similarity difference, and output the natural gas leakage calibration result according to the comparison result between the corrected fugitive similarity difference and the preset difference threshold.

10. The self-calibration method of a natural gas detection device according to claim 9, wherein: The correcting the fugitive similarity difference with the environmental similarity difference and the sample similarity difference includes: Compare the first environmental similarity difference with the first fugitive similarity difference, retain the same data points of the first environmental similarity difference and the first fugitive similarity difference, and obtain the first corrected difference; Compare the second environmental similarity difference with the second fugitive similarity difference, retain the same data points of the second environmental similarity difference and the second fugitive similarity difference, and obtain the second corrected difference; Obtain the first correction weight and the second correction weight according to the first sample similarity and the second sample similarity; Obtain the corrected value of the fugitive similarity difference with the first corrected difference, the second corrected difference, the first correction weight, and the second correction weight.

Images