A gas leak perception, recognition and alarm method based on multi-sensor fusion

Abstract

The present invention belongs to the technical field of gas leakage monitoring, and in particular relates to a multi-sensor fusion gas leakage perception, identification and alarm method. The method first monitors the state of the gas pipeline in real time through integrated sensors, obtains multi-dimensional data and pre-processes it. Secondly, a pre-trained abnormal perception model is used to analyze sound waves and image data to evaluate the degree of gas abnormality. Thirdly, based on temperature, pressure and other data and abnormal monitoring scores, a discrimination rule is formulated to determine whether the gas pipeline is abnormal. Once an abnormal leakage is detected, the leakage point is quickly located through multi-position node sensors and a time positioning algorithm. Finally, based on the leakage location and the abnormal discrimination result, a gas leakage emergency strategy library and an alarm device are used to quickly generate a corresponding emergency treatment plan and quickly perform an abnormal fixed-point alarm to ensure gas safety.

Description

Technical Field

[0001] The present invention belongs to the technical field of gas leakage monitoring, and specifically is a multi-sensor fusion gas leakage perception, identification and alarm method. Background Art

[0002] With the acceleration of urbanization, gas, as one of the main energy sources, is increasingly used in daily life, but the resulting gas leakage problem has also become a major hidden danger to public safety. Traditional gas leak detection methods rely on a single sensor, which often has limitations in sensitivity, accuracy and response speed, and cannot meet safety requirements in complex environments.

[0003] For example, the patent with publication number CN 112542034 A discloses a method, system and device for identifying micro-flow leakage of a smart gas meter. The prior art focuses on monitoring micro-flow inside the smart gas meter and identifying leakage through pulse signal analysis. However, this single-point based monitoring method may not cover the complex situation of the entire gas pipeline network, and has limited ability to identify leakage points far away from the gas meter.

[0004] For example, the patent with publication number CN117746585A discloses a gas leakage safety monitoring method, system, equipment and medium based on the Internet of Things, including the following steps: obtaining an environmental image corresponding to a monitoring area; wherein the monitoring area is an area with a risk of gas leakage; identifying whether there are abnormal state features in the environmental image; if the abnormal state features are identified, receiving a power-on signal to power on a target gas alarm; although the image recognition technology of the prior art can effectively monitor a specific area, in a complex environment or in severe weather conditions, such as fog, rain, snow or strong light, the recognition accuracy of the environmental image may be affected, resulting in false alarms or missed alarms.

[0005] The above existing technologies have the following problems: 1) Most of the existing technologies focus on a single type of sensor or data source, and lack the deep fusion and complementarity of multi-sensor data; 2) A single technical means is difficult to comprehensively evaluate and quickly respond to various complex leakage situations, especially in the case of early minor leaks or atypical leaks. In order to solve the above problems, the present invention provides a multi-sensor fusion gas leak perception, identification and emergency response method. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes a multi-sensor fusion gas leak perception, identification and alarm method. The method first monitors the status of the gas pipeline in real time through integrated sensors, obtains multi-dimensional data and pre-processes it. Secondly, the pre-trained abnormal perception model is used to analyze the sound wave and image data to evaluate the degree of gas abnormality. Thirdly, based on the temperature, pressure and other data and the abnormal monitoring score, a discrimination rule is formulated to determine whether the gas pipeline is abnormal. Once an abnormal leak is detected, the leak point is quickly located through multi-location node sensors and time positioning algorithms. Finally, based on the leak location and the abnormal discrimination result, the gas leak emergency strategy library and the alarm device are used to quickly generate the corresponding emergency treatment plan and quickly perform abnormal fixed-point alarm to ensure gas safety.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A multi-sensor fusion gas leak perception, identification and alarm method, comprising:

[0009] S1. Configure integrated sensors to monitor the working status of indoor gas pipelines, obtain multi-dimensional status monitoring data, and use adaptive filtering algorithms to pre-process the collected multi-dimensional status monitoring data;

[0010] S2. Build and train an anomaly perception model, and input the sound wave and gas concentration data in the pre-processed multi-dimensional status monitoring data into the pre-trained anomaly perception model for analysis to obtain the gas anomaly monitoring scores of different sensor location nodes;

[0011] S3. Setting gas anomaly discrimination rules, inputting the temperature, pressure and gas anomaly monitoring score in the pre-processed multi-dimensional state monitoring data into the gas anomaly discrimination rules for discrimination analysis, and obtaining the comprehensive anomaly discrimination result of the gas pipeline;

[0012] S4. When an abnormal gas leak is detected according to S3, the multi-dimensional status monitoring data obtained by the multi-location node sensors is used to obtain the gas pipeline leakage location through a time-based positioning algorithm;

[0013] S5, inputting the comprehensive abnormality identification result obtained in S3 and the gas pipeline leakage location point obtained in S4 into the integrated sensor to use the built-in alarm device to perform a fixed-point abnormality alarm;

[0014] The steps of obtaining the comprehensive abnormality determination result of the gas pipeline and the leakage location point of the gas pipeline include:

[0015] S301. Setting an abnormality discrimination threshold interval according to the temperature, pressure and abnormality monitoring score in the multi-dimensional state monitoring data. The abnormality discrimination threshold interval specifically includes:

[0016] S3011. Set the abnormal score threshold to f h , when f t i >f h And f t i+1 >f h When , it is determined that an abnormal gas leakage occurs between the i-th integrated sensor and the i+1-th integrated sensor, and the distance from the leakage point to the i-th integrated sensor position is calculated using the acoustic wave positioning formula, and the calculated distance is used to obtain the position of the leakage point; f t i represents the abnormal monitoring score obtained at the i-th sensor location at time t;

[0017] When f t i >f h And f t i+1 ≤f h or t i ≤f h And f t i+1 >f h When the temperature and pressure obtained at the i-th or i+1-th integrated sensor position are used for secondary judgment, when it is judged that the gas pipeline is leaking abnormally, the comprehensive positioning formula is used to calculate the distance from the leakage point to the i-th or i+1-th integrated sensor position, and the specific leakage point is determined by the calculated distance.

[0018] Specifically, the steps for obtaining the gas anomaly monitoring score include:

[0019] S201, obtaining historical acoustic wave time domain data corresponding to gas leakage at the distances from N different leakage points to the corresponding i-th integrated sensor position and i+1-th integrated sensor position Abnormal concentration rate of gas per unit time Y t and the coordinate data of the i-th integrated sensor position and the i+1-th integrated sensor position relative to the historical abnormal leakage point (s i ,s i+1 ), and annotate the acquired sound wave data, where represents the historical acoustic wave time domain data sequence obtained at the i-th integrated sensor position at time t, represents the historical acoustic wave time domain data sequence obtained at the i+1th integrated sensor position at time t, s i represents the three-dimensional coordinate point of the i-th integrated sensor position, s i+1 Indicates the three-dimensional coordinate point of the i+1th integrated sensor position;

[0020] S202, performing time-frequency transformation on the acquired sound wave data by Fourier transform, extracting the frequency component characteristics and spectrum peaks corresponding to the sound wave, and extracting the sound intensity characteristics and sound wave entropy characteristics corresponding to the gas leakage at different distances by using a time domain analysis algorithm based on the acquired sound wave time domain data;

[0021] S203, aligning the frequency component characteristics, spectrum peaks, sound intensity characteristics and sound wave entropy characteristics corresponding to the acquired sound wave frequency domain and time domain and the abnormal concentration change rate of the monitored gas per unit time in the time dimension, and dividing the aligned sound wave feature data and abnormal concentration change rate according to the unit time length to obtain a multi-dimensional input sequence.

[0022] Specifically, the step of obtaining the gas anomaly monitoring score also includes:

[0023] S204, constructing an anomaly perception model through an integrated method based on a random forest algorithm and a multivariate regression algorithm, inputting the obtained multidimensional input sequence into the constructed anomaly perception model for training, and obtaining a trained anomaly perception model;

[0024] S205, deploy the acquired abnormal perception model to sensors at different positions of the gas pipeline, collect multi-dimensional status monitoring data of the gas pipeline in real time through sensors at different positions, and predict the probability of gas leakage at different position distances corresponding to each other through the random forest algorithm in the abnormal perception model [p1…p i …p n ] and the corresponding leaked gas type, p i represents the predicted probability of gas leakage obtained at the i-th sensor location, and n represents the total number of integrated sensors;

[0025] S206, input the predicted probability of gas leakage at different position distances and the multi-dimensional status monitoring data into the multivariate regression algorithm in the abnormal perception model to identify the abnormal monitoring scores corresponding to sensors at different positions [f1…f i …f n ], where f i Represents the abnormal monitoring score obtained at the i-th sensor location.

[0026] Specifically, the step of obtaining the comprehensive abnormality determination result of the gas pipeline in S3 also includes:

[0027] S3012, when f t i >f h And f t i+1 ≤f h When , the temperature and pressure obtained at the i-th integrated sensor position are used for secondary judgment. or Tt i >T h , it is determined that there is an abnormal gas leakage. If And T t i ≤T h When , it is determined that there is no abnormal gas leakage; among them, represents the gas pipeline pressure data at time t obtained by the i-th integrated sensor, p h Indicates the threshold value for determining abnormal gas pipeline pressure, T t i+1 represents the ambient temperature data of the gas pipeline at time t obtained by the i+1th integrated sensor, T h Indicates the threshold for determining abnormal gas pipeline temperature.

[0028] Specifically, the step of obtaining the comprehensive abnormality determination result of the gas pipeline in S3 also includes:

[0029] S3013, when f t i ≤f h And f t i+1 >f h When , the temperature and pressure obtained at the i+1th integrated sensor position are used for secondary judgment. or T t i+1 >T h , it is determined that there is an abnormal gas leakage. If And T t i+1 ≤T h When , it is determined that there is no abnormal gas leakage;

[0030] S3014, when f t i ≤f h And f t i+1 ≤f h When , it is determined that there is no abnormal gas leakage between the i-th integrated sensor and the i+1-th integrated sensor;

[0031] S302: Calculate the real-time abnormal monitoring scores [f1…f i …f n ], based on the acquired abnormal monitoring score and the temperature and pressure data in the real-time multi-dimensional status monitoring data, the comprehensive abnormality judgment result of the gas pipeline is obtained through the S3011-S3014 process.

[0032] Specifically, the specific formula for acoustic wave positioning is:

[0033]

[0034] Among them, D i represents the distance from the leak point to the i-th integrated sensor position, v represents the speed of sound in the propagation medium, D i,i+1 represents the distance between the position of the i-th integrated sensor and the i+1-th integrated sensor, ▽t i,i+1 A represents the time difference between the sound wave from the leakage point to the i-th integrated sensor position and the i+1-th integrated sensor position; f (v,f,D) represents the attenuation factor of the sound wave in the propagation medium, f represents the frequency characteristic of the sound wave, D represents the distance the sound wave is transmitted, a(f) represents the frequency attenuation factor of the sound wave during the transmission process, and k(T,H) represents the sound wave frequency correction coefficient.

[0035] Specifically, the specific steps of obtaining the gas pipeline leakage location point in S4 also include:

[0036] S402: When it is determined according to S3012 that the gas pipeline is leaking abnormally, the distance D″ from the leak point to the position of the i-th integrated sensor is calculated using the comprehensive positioning formula. i , the specific formula for comprehensive positioning is:

[0037]

[0038] Wherein, v1 represents the speed of gas transmission in the gas pipeline, v2 represents the propagation speed of the pressure wave in the gas when the gas leaks, and ▽t' represents the time difference between the i-th integrated sensor and the j-th integrated sensor receiving the pressure wave signal; D i represents the distance from the leakage point between the ith integrated sensor and the jth integrated sensor to the ith integrated sensor calculated by the formula in S401, D i,j represents the distance between the i-th integrated sensor and the j-th integrated sensor;

[0039] S403. When it is determined according to S3013 that the gas pipeline is leaking abnormally, the distance from the leak point to the i+1th integrated sensor is calculated by the distance formula in S402, and the position of the specific leak point is determined by the calculated distance.

[0040] Specifically, the specific steps of performing a fixed-point abnormality alarm in S5 include:

[0041] According to the obtained S301-S302 gas anomaly identification results and the distances of the S401 and S402 gas leakage points to the i-th integrated sensor, the following alarm settings are made:

[0042] When the gas anomaly discrimination result meets the conditions in S3011, the i-th integrated sensor is used to issue a fixed-point anomaly alarm through the alarm device built into the integrated sensor according to the location of the leakage point calculated in S401, otherwise no alarm is issued;

[0043] When the gas abnormality judgment result meets the secondary judgment in S3012 or T t i >T h Condition or secondary judgment in S3013 or T t i+1 >T h When the condition is met, the i-th integrated sensor is used to issue a fixed-point abnormal alarm through the alarm device built into the integrated sensor according to the position of the leakage point calculated in S402, otherwise no alarm is issued.

[0044] A computer-readable storage medium stores computer instructions, which, when executed, execute a multi-sensor fusion gas leak perception, identification and alarm method.

[0045] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements a multi-sensor fusion gas leak perception, identification and alarm method when executing the computer program.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] In view of the deficiencies in the prior art, the present invention integrates multiple sensors and adopts an adaptive filtering algorithm to pre-process multi-dimensional status monitoring data, thereby ensuring the accuracy and reliability of the data. The constructed and trained abnormal perception model can accurately analyze the sound wave and gas concentration data, so as to obtain the gas abnormality monitoring scores of different sensor location nodes, thereby improving the sensitivity of abnormality detection. The set gas abnormality discrimination rules combine multi-dimensional information such as temperature, pressure and monitoring scores for comprehensive analysis, thereby ensuring the accuracy and comprehensiveness of abnormality discrimination. When the system determines that there is a gas leak, the time-based positioning algorithm can be used to quickly and accurately determine the leak location. In addition, by using the built-in alarm device of the integrated sensor to perform a fixed-point abnormal alarm, relevant personnel can be notified in time to take emergency measures, which greatly shortens the emergency response time and effectively avoids safety accidents caused by gas leaks. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flow chart of a multi-sensor fusion gas leak perception, identification and alarm method according to Embodiment 1 of the present invention;

[0049] Figure 2A logical flow chart for obtaining the gas anomaly monitoring score of the present invention; DETAILED DESCRIPTION

[0050] Example 1

[0051] See also Figure 1 , an embodiment of the present invention provides: a multi-sensor fusion gas leak perception, identification and alarm method, comprising the following steps:

[0052] S1. Configure an integrated sensor to monitor the working status of the indoor gas pipeline, obtain multi-dimensional status monitoring data, and use an adaptive filtering algorithm to pre-process the collected multi-dimensional status monitoring data; further, according to the layout of the gas pipeline, deploy temperature sensors, pressure sensors, acoustic sensors and gas concentration sensors on the corresponding pipeline valve nodes to monitor and collect multi-dimensional data on the operating status of the gas pipeline; the corresponding adaptive filtering algorithms in this embodiment include Kalman filtering and adaptive median filtering;

[0053] S2. Build and train an anomaly perception model, and input the sound wave and gas concentration data in the pre-processed multi-dimensional status monitoring data into the pre-trained anomaly perception model for analysis to obtain the gas anomaly monitoring scores of different sensor location nodes;

[0054] S3. Setting gas anomaly discrimination rules, inputting the temperature, pressure and gas anomaly monitoring score in the pre-processed multi-dimensional state monitoring data into the gas anomaly discrimination rules for discrimination analysis, and obtaining the comprehensive anomaly discrimination result of the gas pipeline;

[0055] S4. When an abnormal gas leak is detected according to S3, the multi-dimensional status monitoring data obtained by the multi-location node sensors is used to obtain the gas pipeline leakage location through a time-based positioning algorithm;

[0056] S5, build a gas leak emergency strategy library, input the obtained gas pipeline leakage location and the gas pipeline comprehensive abnormality judgment result into the constructed gas leak emergency strategy library, obtain the corresponding gas emergency treatment strategy through the text generation algorithm, and input the comprehensive abnormality judgment result obtained in S3 and the gas pipeline leakage location obtained in S4 into the integrated sensor to use the built-in alarm device to perform fixed-point abnormality alarm, and transmit the fixed-point abnormality alarm information and gas emergency treatment strategy to the user or maintenance personnel through the mobile terminal for maintenance or emergency avoidance. Further, in this embodiment, through the abnormal judgment result and the gas pipeline leakage location, when the gas leaks, the alarm is issued through the built-in alarm device in the integrated sensor, and the specific location of the corresponding leakage point is sent to the user or maintenance personnel through the sensor.

[0057] For further information, see Figure 2,The steps of obtaining the gas anomaly monitoring score include:

[0058] S201, obtaining historical acoustic wave time domain data corresponding to gas leakage at the distances from N different leakage points to the corresponding i-th integrated sensor position and i+1-th integrated sensor position Abnormal concentration rate of gas per unit time Y t and the coordinate data of the i-th integrated sensor position and the i+1-th integrated sensor position relative to the historical abnormal leakage point (s i ,s i+1 ), and annotate the acquired sound wave data, where represents the historical acoustic wave time domain data sequence obtained at the i-th integrated sensor position at time t, represents the historical acoustic wave time domain data sequence obtained at the i+1th integrated sensor position at time t, s i represents the three-dimensional coordinate point of the i-th integrated sensor position, s i+1 Indicates the three-dimensional coordinate point of the i+1th integrated sensor position; further, the annotation here includes the gas type corresponding to the gas leakage sound wave data and the abnormal monitoring score obtained by monitoring the first sensor position and the second sensor position after the gas leakage; the i-th integrated sensor and the i+1-th integrated sensor here represent two adjacent integrated sensors in the gas pipeline; in order to determine the specific position coordinate points of each sensor, this embodiment uses the starting point position of the gas pipeline entering the house as the origin position of the spatial three-dimensional coordinate point, constructs a three-dimensional spatial structure model of the gas pipeline indoors, and annotates the coordinate points corresponding to each integrated sensor on the gas pipeline;

[0059] S202, performing time-frequency transformation on the acquired sound wave data by Fourier transform, extracting the frequency component characteristics and spectrum peaks corresponding to the sound wave, and extracting the sound intensity characteristics and sound wave entropy characteristics corresponding to the gas leakage at different distances by using the time domain analysis algorithm according to the acquired sound wave time domain data; the time domain analysis algorithm in this embodiment includes short-time energy analysis and waveform envelope analysis;

[0060] S203, aligning the frequency component characteristics, spectrum peaks, sound intensity characteristics and sound wave entropy characteristics corresponding to the acquired sound wave frequency domain and time domain and the abnormal concentration change rate of the monitored gas per unit time in the time dimension, and segmenting the aligned sound wave feature data and abnormal concentration change rate according to the unit time length to obtain a multidimensional input sequence;

[0061] S204, constructing an anomaly perception model through an integrated method based on a random forest algorithm and a multivariate regression algorithm, inputting the obtained multidimensional input sequence into the constructed anomaly perception model for training, and obtaining a trained anomaly perception model;

[0062] S205, deploy the acquired abnormal perception model to sensors at different positions of the gas pipeline, collect multi-dimensional status monitoring data of the gas pipeline in real time through sensors at different positions, and predict the probability of gas leakage at different position distances corresponding to each other through the random forest algorithm in the abnormal perception model [p1…p i …p n ] and the corresponding leaked gas type, p i represents the predicted probability of gas leakage obtained at the i-th sensor location, and n represents the total number of integrated sensors;

[0063] S206, inputting the acquired gas leakage probability and multi-dimensional status monitoring data at different position distances into the multivariate regression algorithm in the abnormality perception model to identify the abnormality monitoring scores corresponding to sensors at different positions [f1…f i …f n ], where f i Represents the abnormal monitoring score obtained at the i-th sensor location.

[0064] This process collects and analyzes historical acoustic time domain data, gas concentration change rate, and sensor location coordinate data, and combines random forest and multivariate regression algorithms to build an efficient anomaly perception model. The model can predict the probability and type of gas leaks in real time and give an anomaly monitoring score, thus achieving comprehensive monitoring and rapid response to gas pipelines.

[0065] Furthermore, in this embodiment, the step of obtaining the comprehensive abnormality determination result of the gas pipeline in S3 includes:

[0066] S301. Setting an abnormality discrimination threshold interval according to the temperature, pressure and abnormality monitoring score in the multi-dimensional state monitoring data. The abnormality discrimination threshold interval specifically includes:

[0067] S3011. Set the abnormal score threshold to f h , when f t i >f h And f t i+1 >f h When , it is determined that an abnormal gas leakage occurs between the i-th integrated sensor and the i+1-th integrated sensor;

[0068] S3012, when f t i >f h And f t i+1 ≤f h When , the temperature and pressure obtained at the i-th integrated sensor position are used for secondary judgment. or T ti >T h , it is determined that there is an abnormal gas leakage. If And T t i ≤T h When , it is determined that there is no abnormal gas leakage; among them, represents the gas pipeline pressure data at time t obtained by the i-th integrated sensor, p h Indicates the threshold value for determining abnormal gas pipeline pressure, T t i+1 represents the ambient temperature data of the gas pipeline at time t obtained by the i+1th integrated sensor, T h Indicates the threshold for determining abnormal gas pipeline temperature; f t i represents the abnormal monitoring score obtained at the i-th sensor location at time t;

[0069] S3013, when f t i ≤f h And f t i+1 >f h When , the temperature and pressure obtained at the i+1th integrated sensor position are used for secondary judgment. or T t i+1 >T h , it is determined that there is an abnormal gas leakage. If And T t i+1 ≤T h When , it is determined that there is no abnormal gas leakage;

[0070] S3014, when f t i ≤f h And f t i+1 ≤f h When , it is determined that there is no abnormal gas leakage between the i-th integrated sensor and the i+1-th integrated sensor;

[0071] S302: Calculate the real-time abnormal monitoring scores [f1…f i …f n ], based on the acquired abnormal monitoring score and the temperature and pressure data in the real-time multi-dimensional status monitoring data, the comprehensive abnormality judgment result of the gas pipeline is obtained through the S3011-S3014 process.

[0072] This process effectively improves the accuracy and reliability of gas pipeline anomaly detection by comprehensively considering the temperature, pressure and abnormal monitoring scores in the multi-dimensional status monitoring data and adopting a multi-level threshold judgment method. By setting a reasonable abnormal discrimination threshold interval and combining the multi-dimensional data obtained in real time, the system can quickly and accurately identify abnormal leaks in the gas pipeline. Especially when the abnormal score is at the critical value, the secondary judgment of temperature and pressure is used to further improve the detection accuracy and reduce the possibility of false alarms and missed alarms.

[0073] Furthermore, in this embodiment, the specific steps of obtaining the gas pipeline leakage location point in S4 include:

[0074] S401. When it is determined according to S3011 that the gas pipeline is leaking abnormally, it is known that the gas leak point is between the i-th integrated sensor position and the i+1-th integrated sensor. Then, the distance from the leak point to the i-th integrated sensor position is calculated using the acoustic wave positioning formula, and the calculated distance is used to obtain the position of the leak point. The specific formula for acoustic wave positioning is:

[0075]

[0076] Among them, D i represents the distance from the leak point to the i-th integrated sensor position, v represents the speed of sound in the propagation medium, D i,i+1 represents the distance between the position of the i-th integrated sensor and the i+1-th integrated sensor, ▽t i,i+1 A represents the time difference between the sound wave from the leakage point to the i-th integrated sensor position and the i+1-th integrated sensor position; f (v, f, D) represents the attenuation factor of the sound wave in the propagation medium, f represents the frequency characteristic of the sound wave, D represents the distance of sound wave transmission, a(f) represents the frequency attenuation factor of the sound wave during the transmission process, and k(T, H) represents the sound wave frequency correction coefficient; further, since D itself is an unknown quantity in this embodiment, for A f The calculation of (v, f, D) is pre-calibrated or estimated in real time by a person skilled in the art based on a theoretical model or experimental data; k(T, H) is a correction coefficient of the sound wave propagation frequency with respect to temperature and humidity, which is obtained by a person skilled in the art through an experimental verification process, and the optimal k(T, H) value is obtained by setting the change trend of the sound wave transmission frequency under M different temperature and humidity parameter pairs;

[0077] S402: When it is determined according to S3012 that the gas pipeline is leaking abnormally, the distance D″ from the leak point to the position of the i-th integrated sensor is calculated using the comprehensive positioning formula. i , the specific formula for comprehensive positioning is:

[0078]

[0079] Wherein, v1 represents the speed of gas transmission in the gas pipeline, v2 represents the propagation speed of the pressure wave in the gas when the gas leaks, and ▽t' represents the time difference between the i-th integrated sensor and the j-th integrated sensor receiving the pressure wave signal; D i represents the distance from the leakage point between the ith integrated sensor and the jth integrated sensor to the ith integrated sensor calculated by the formula in S401; D i,j represents the distance between the i-th integrated sensor and the j-th integrated sensor;

[0080] S403. When it is determined according to S3013 that the gas pipeline is leaking abnormally, the distance from the leak point to the i+1th integrated sensor is calculated by the distance formula in S402, and the position of the specific leak point is determined by the calculated distance.

[0081] When the system detects an abnormal gas leak, the process first calculates the distance from the leak point to the specific integrated sensor position based on the acoustic wave positioning formula. This method takes into account the attenuation factor and frequency correction coefficient during the acoustic wave propagation process, thereby improving the accuracy of positioning. When a secondary judgment is made through temperature and pressure data, a comprehensive positioning formula is used to combine the gas transmission speed and the pressure wave propagation speed, as well as the time difference between different sensors receiving the pressure wave signal, to further accurately calculate the position of the leak point. This comprehensive positioning method not only takes into account the characteristics of acoustic wave propagation, but also combines the characteristics of gas flow and pressure wave propagation, thereby achieving accurate and rapid positioning of the gas leak point.

[0082] Furthermore, in this embodiment, the specific steps of constructing the gas leak emergency strategy library include:

[0083] S501, collecting historical gas leak emergency plan text and video data, and preprocessing the collected data;

[0084] S502, input the preprocessed solution data into the keyword and relationship extraction model based on the Chinese pre-trained Bert to extract keywords and corresponding relationships, construct emergency solution triples, and input the constructed triple data into the graph database to obtain the emergency method knowledge base, and save the collected video emergency solution data into the emergency method knowledge base through the corresponding relationship attribute nodes;

[0085] S503, embedding the pre-trained reasoning model into the emergency method knowledge base to obtain a gas leak emergency strategy library with reasoning ability; further, the pre-trained reasoning models in this embodiment include RoBERTa, T5, YOLO series models and DeepSpeech;

[0086] Furthermore, in this embodiment, the specific steps of performing a fixed-point abnormality alarm in S5 include:

[0087] According to the obtained S301-S302 gas anomaly identification results and the distances of the S401 and S402 gas leakage points to the i-th integrated sensor, the following alarm settings are made:

[0088] When the gas anomaly discrimination result meets the conditions in S3011, the i-th integrated sensor is used to issue a fixed-point anomaly alarm through the alarm device built into the integrated sensor according to the location of the leakage point calculated in S401, otherwise no alarm is issued;

[0089] When the gas abnormality judgment result meets the secondary judgment in S3012 or T t i >T h Condition or secondary judgment in S3013 or T t i+1 >T h When the condition is met, the i-th integrated sensor is used to issue a fixed-point abnormal alarm through the alarm device built into the integrated sensor according to the position of the leakage point calculated in S402, otherwise no alarm is issued.

[0090] In this process, the system can intelligently determine whether to trigger an alarm by integrating the leaking gas type, abnormality identification results, and location information of the leak point obtained by the sensor, and can perform a fixed-point abnormality alarm through the built-in alarm device. This setting not only improves the accuracy and pertinence of the alarm, but also can quickly notify relevant personnel of the specific location of the leak, making it easier to take timely response measures and effectively avoid the occurrence of safety accidents.

[0091] Furthermore, in this embodiment, the specific steps of obtaining the corresponding gas emergency handling strategy include:

[0092] S504. According to the process of S301-S302, it is judged that there is a gas leak. When it is judged that there is a gas leak through S3011, the severity of the corresponding abnormal gas leak is judged by the abnormal monitoring score calculated by the process of S201-S206; the severity of the abnormal gas leak in this embodiment includes slight leakage, moderate leakage and severe leakage, wherein slight leakage, moderate leakage and severe leakage are specifically set and divided by technical personnel in this field according to the obtained abnormal monitoring score; wherein the emergency strategy corresponding to slight leakage is: closing the nearest valve, opening windows for ventilation, and on-site monitoring; the emergency strategy corresponding to moderate leakage is: immediately evacuating nearby personnel, closing valves in related areas, and notifying professional maintenance teams; the emergency strategy corresponding to severe leakage is: emergency evacuation of surrounding areas, blocking the scene, notifying fire and emergency rescue departments, and remotely closing the main valve;

[0093] S505, inputting the acquired severity of the abnormal gas leakage and the specific location of the gas leakage point acquired in S401 into the acquired gas leakage emergency strategy library, and generating the corresponding emergency strategy 1 through a text generation algorithm;

[0094] S504, when it is determined through S3012-S3013 that a gas leak occurs, the severity of the abnormal gas leak corresponding to the abnormal monitoring score calculated through the process of S201-S206;

[0095] S505. Input the acquired severity of abnormal gas leakage, the corresponding temperature data and pressure data, and the specific location of the gas leakage point acquired in S401 into the acquired gas leakage emergency strategy library, and generate the corresponding emergency strategy 2 through the text generation algorithm; further, the text generation algorithm of this embodiment includes the GPT series pre-training model, the Wenxinyiyan model and the Chinese pre-training model based on Bert.

[0096] Example 2

[0097] A computer-readable storage medium stores computer instructions, which, when executed, execute a multi-sensor fusion gas leak perception, identification and alarm method.

[0098] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements a multi-sensor fusion gas leak perception, identification and alarm method when executing the computer program.

[0099] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are within the protection of the present invention.

Claims

1. A multi-sensor fusion gas leak perception, identification and alarm method, characterized in that: include: S1. Configure integrated sensors to monitor the working status of indoor gas pipelines, obtain multi-dimensional status monitoring data, and use adaptive filtering algorithms to pre-process the collected multi-dimensional status monitoring data; S2. Build and train an anomaly perception model, and input the sound wave and gas concentration data in the pre-processed multi-dimensional status monitoring data into the pre-trained anomaly perception model for analysis to obtain the gas anomaly monitoring scores of different sensor location nodes; S3. Setting gas anomaly discrimination rules, inputting the temperature, pressure and gas anomaly monitoring score in the pre-processed multi-dimensional state monitoring data into the gas anomaly discrimination rules for discrimination analysis, and obtaining the comprehensive anomaly discrimination result of the gas pipeline; S4. When an abnormal gas leak is detected according to S3, the multi-dimensional status monitoring data obtained by the multi-location node sensors is used to obtain the gas pipeline leakage location through a time-based positioning algorithm; S5, inputting the comprehensive abnormality identification result obtained in S3 and the gas pipeline leakage location point obtained in S4 into the integrated sensor to use the built-in alarm device to perform a fixed-point abnormality alarm; The steps of obtaining the comprehensive abnormality determination result of the gas pipeline and the leakage location point of the gas pipeline include: S301. Setting an abnormality discrimination threshold interval according to the temperature, pressure and abnormality monitoring score in the multi-dimensional state monitoring data. The abnormality discrimination threshold interval specifically includes: S3011. Set the abnormal score threshold to f h , when f t i >f h And f t i+1 >f h When , it is determined that an abnormal gas leakage occurs between the i-th integrated sensor and the i+1-th integrated sensor, and the distance from the leakage point to the i-th integrated sensor position is calculated using the acoustic wave positioning formula, and the calculated distance is used to obtain the position of the leakage point; f t i represents the abnormal monitoring score obtained at the i-th sensor location at time t; When f t i >f h And f t i+1 ≤f h or t i ≤f h And f t i+1 >f h When the temperature and pressure obtained at the i-th or i+1-th integrated sensor position are used for secondary judgment, when it is judged that the gas pipeline is leaking abnormally, the comprehensive positioning formula is used to calculate the distance from the leakage point to the i-th or i+1-th integrated sensor position, and the specific leakage point is determined by the calculated distance.

2. A multi-sensor fusion gas leak perception, identification and alarm method as claimed in claim 1, characterized in that: The step of obtaining the gas anomaly monitoring score comprises: S201, obtaining historical acoustic wave time domain data corresponding to gas leakage at the distances from N different leakage points to the corresponding i-th integrated sensor position and i+1-th integrated sensor position The abnormal concentration rate of gas per unit time Yt and the coordinate data of the i-th integrated sensor position and the i+1-th integrated sensor position relative to the historical abnormal leakage point (s i ,s i+1 ), and annotate the acquired sound wave data, where represents the historical acoustic wave time domain data sequence obtained at the i-th integrated sensor position at time t, represents the historical acoustic wave time domain data sequence obtained at the i+1th integrated sensor position at time t, s i represents the three-dimensional coordinate point of the i-th integrated sensor position, s i+1 Indicates the three-dimensional coordinate point of the i+1th integrated sensor position; S202, performing time-frequency transformation on the acquired sound wave data by Fourier transform, extracting the frequency component characteristics and spectrum peaks corresponding to the sound wave, and extracting the sound intensity characteristics and sound wave entropy characteristics corresponding to the gas leakage at different distances by using a time domain analysis algorithm based on the acquired sound wave time domain data; S203, aligning the frequency component characteristics, spectrum peaks, sound intensity characteristics and sound wave entropy characteristics corresponding to the acquired sound wave frequency domain and time domain and the abnormal concentration change rate of the monitored gas per unit time in the time dimension, and dividing the aligned sound wave feature data and abnormal concentration change rate according to the unit time length to obtain a multi-dimensional input sequence.

3. A multi-sensor fusion gas leak perception, identification and alarm method as claimed in claim 2, characterized in that: The step of obtaining the gas anomaly monitoring score further includes: S204, constructing an anomaly perception model through an integrated method based on a random forest algorithm and a multivariate regression algorithm, inputting the obtained multidimensional input sequence into the constructed anomaly perception model for training, and obtaining a trained anomaly perception model; S205, deploy the acquired abnormal perception model to sensors at different positions of the gas pipeline, collect multi-dimensional status monitoring data of the gas pipeline in real time through sensors at different positions, and predict the probability of gas leakage at different position distances corresponding to each other through the random forest algorithm in the abnormal perception model [p1…p i …p n ] and the corresponding leaked gas type, p i represents the predicted probability of gas leakage obtained at the i-th sensor location, and n represents the total number of integrated sensors; S206, input the predicted probability of gas leakage at different position distances and the multi-dimensional status monitoring data into the multivariate regression algorithm in the abnormal perception model to identify the abnormal monitoring scores corresponding to sensors at different positions [f1…f i …f n ], where f i Represents the abnormal monitoring score obtained at the i-th sensor location.

4. A multi-sensor fusion gas leak perception, identification and alarm method as claimed in claim 3, characterized in that: The step of obtaining the comprehensive abnormality determination result of the gas pipeline in S3 also includes: S3012, when f t i >f h And f t i+1 ≤f h When , the temperature and pressure obtained at the i-th integrated sensor position are used for secondary judgment. or T t i >T h , it is determined that there is an abnormal gas leakage. If And T t i ≤T h When , it is determined that there is no abnormal gas leakage; represents the gas pipeline pressure data at time t obtained by the i-th integrated sensor, p h Indicates the threshold value for determining abnormal gas pipeline pressure, T t i+1 represents the ambient temperature data of the gas pipeline at time t obtained by the i+1th integrated sensor, T h Indicates the threshold for determining abnormal gas pipeline temperature.

5. A multi-sensor fusion gas leak perception, identification and alarm method as claimed in claim 4, characterized in that: The step of obtaining the comprehensive abnormality determination result of the gas pipeline in S3 also includes: S3013, when f t i ≤f h And f t i+1 >f h When , the temperature and pressure obtained at the i+1th integrated sensor position are used for secondary judgment. or T t i+1 >T h , it is determined that there is an abnormal gas leakage. If And T t i+1 ≤T h When , it is determined that there is no abnormal gas leakage; S3014, when f t i ≤f h And f t i+1 ≤f h When , it is determined that there is no abnormal gas leakage between the i-th integrated sensor and the i+1-th integrated sensor; S302: Calculate the real-time abnormal monitoring scores [f1…f i …f n ], based on the acquired abnormal monitoring score and the temperature and pressure data in the real-time multi-dimensional status monitoring data, the comprehensive abnormality judgment result of the gas pipeline is obtained through the S3011-S3014 process.

6. A multi-sensor fusion gas leak perception, identification and alarm method as claimed in claim 5, characterized in that: The specific formula for acoustic wave positioning is: A f (v,f,D)=10 a(f)×D / 10 , Among them, D i represents the distance from the leak point to the i-th integrated sensor position, v represents the speed of sound in the propagation medium, D i,i+1 represents the distance between the position of the i-th integrated sensor and the i+1-th integrated sensor, A represents the time difference between the sound wave from the leakage point to the i-th integrated sensor position and the i+1-th integrated sensor position; f (v,f,D) represents the attenuation factor of the sound wave in the propagation medium, f represents the frequency characteristic of the sound wave, D represents the distance the sound wave is transmitted, a(f) represents the frequency attenuation factor of the sound wave during the transmission process, and k(T,H) represents the sound wave frequency correction coefficient.

7. A multi-sensor fusion gas leak perception, identification and alarm method as claimed in claim 6, characterized in that: The specific step of obtaining the gas pipeline leakage location point in S4 also includes: S402: When it is determined according to S3012 that the gas pipeline is leaking abnormally, the distance D″ from the leak point to the position of the i-th integrated sensor is calculated using the comprehensive positioning formula. i , the specific formula for comprehensive positioning is: Among them, v1 represents the speed of gas transmission in the gas pipeline, v2 represents the propagation speed of the pressure wave in the gas when the gas leaks, represents the time difference between the i-th integrated sensor and the j-th integrated sensor receiving the pressure wave signal; D i represents the distance from the leakage point between the ith integrated sensor and the jth integrated sensor to the ith integrated sensor calculated by the formula in S401, D i,j represents the distance between the i-th integrated sensor and the j-th integrated sensor; S403. When it is determined according to S3013 that the gas pipeline is leaking abnormally, the distance from the leak point to the i+1th integrated sensor is calculated by the distance formula in S402, and the position of the specific leak point is determined by the calculated distance.

8. A multi-sensor fusion gas leak perception, identification and alarm method as claimed in claim 7, characterized in that: The specific steps of performing the fixed-point abnormality alarm in S5 include: According to the obtained S301-S302 gas anomaly identification results and the distances of the S401 and S402 gas leakage points to the i-th integrated sensor, the following alarm settings are made: When the gas anomaly discrimination result meets the conditions in S3011, the i-th integrated sensor is used to issue a fixed-point anomaly alarm through the alarm device built into the integrated sensor according to the location of the leakage point calculated in S401, otherwise no alarm is issued; When the gas abnormality judgment result meets the secondary judgment in S3012 or T t i >T h Condition or secondary judgment in S3013 or T t i+1 >T h When the condition is met, the i-th integrated sensor is used to issue a fixed-point abnormal alarm through the alarm device built into the integrated sensor according to the position of the leakage point calculated in S402, otherwise no alarm is issued.

9. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, the steps of a multi-sensor fusion gas leak perception, identification and alarm method as described in any one of claims 1-7 are executed.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the multi-sensor fusion gas leak perception, identification and alarm method described in any one of claims 1 to 7 is implemented.

Images