Gas pipeline leakage detection method and device based on infrasonic wave and AI model

Abstract

The invention provides a gas pipeline leakage detection method and device based on infrasonic waves and an AI model, and relates to the technical field of gas pipeline detection. The method comprises the following steps: step 1, collecting leakage signals through continuous infrasonic waves; step 2, signal preprocessing; step 3, sound wave noise elimination is carried out and leakage signal characteristics are obtained; 4, processing and analyzing the signal features through an AI model; and step 5, obtaining a leakage identification condition. According to the method, the limitation that signals need to be monitored in real time in a traditional method is effectively solved, leakage study and judgment can be conducted through model matching recognition even if abrupt change signals are missed, and the method has the advantages of being high in accuracy, low in false alarm rate, low in missing report rate, measurable at any time and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of gas pipeline detection, and in particular to a gas pipeline leakage detection method and device based on infrasound and AI model. Background Art

[0002] Natural gas is an important source of fuel or raw materials in manufacturing and housing. Affected by factors such as pipeline construction quality not meeting specifications, corrosion and wear, human damage, and sudden natural disasters (earthquakes, mudslides), chemical gases in pipelines are prone to leakage under the action of pressure difference, causing destructive effects on human environment, production and life, and the safety of life and property. Therefore, leak detection and positioning of gas pipelines is not only of economic significance to the safe operation and management of gas pipelines, but also of great practical significance to the development of the national economy.

[0003] At present, the main methods for detecting gas pipeline leaks include helium mass spectrometry, negative pressure wave method, infrared detection method, infrasound method, etc.; helium mass spectrometry is suitable for sealed pipelines, and the test system is simple, but its detection cycle is long, and it needs repeated inflation and deflation and is limited by the volume of the detection structure; the negative pressure wave method has the advantages of fast deployment speed and rapid response, and can detect sudden pipeline leaks in time, but due to the poor sensitivity of the environment to slow leaks with small flow and small range, missed detection and misjudgment are prone to occur; infrared gas imaging detection can realize the visualization of the leakage source and realize non-contact imaging detection, but obstacles will affect the imaging and cause missed detection.

[0004] Therefore, there is an urgent need for a gas pipeline leak detection method and device based on infrasound and AI models that can improve detection accuracy and reduce the missed reporting rate. Summary of the invention

[0005] The purpose of the present invention is to provide a gas pipeline leakage detection method and device based on infrasound and AI model, which solves the technical problems of low detection accuracy and easy missed detection in the prior art. The many technical effects that can be produced by the preferred technical solution among the many technical solutions provided by the present invention are described in detail below.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention provides a gas pipeline leakage detection method based on infrasound and AI model, comprising the following steps:

[0008] Step 1: Collect leakage signals through continuous infrasound;

[0009] Step 2: signal preprocessing;

[0010] Step 3: De-noise the sound wave and obtain the leakage signal characteristics;

[0011] Step 4: Process and analyze the signal features through the AI ​​model;

[0012] Step 5: Obtain leakage identification information.

[0013] Preferably, an artificial leakage point is added in step one by means of a shock wave, and the leakage distance is calculated by the time interval between the signal of the artificial leakage point and the leakage signal in step five to determine the leakage position.

[0014] Preferably, in the step three, the denoised signal is integrated with the empirical mode decomposition, and the signal features are extracted from the decomposed signal by multi-scale permutation entropy.

[0015] Preferably, in step 4, the original signal and the signal features processed in step 3 are input into a neural network for learning to establish the AI ​​model, and whether a pipeline leaks is determined by the AI ​​model.

[0016] Preferably, it also includes:

[0017] Step 6: Evaluate the leakage identification accuracy of the proposed AI model.

[0018] A gas pipeline leakage detection device based on infrasound and AI model is applied to the above-mentioned gas pipeline leakage detection method based on infrasound and AI model, including an information collection pile and a camera, an edge computing instrument, a 485 collection server, a 4G network communicator, and an electronic control unit installed on the information collection pile from top to bottom, the information collection pile is fixedly connected to the wellhead edge of the valve well, and a plurality of pipe wall infrasound collectors and a plurality of gas infrasound collectors are installed on the natural gas pipeline in the valve well. A valve well mobile information collection box is detachably connected to the inner wall of the valve well, and the valve well mobile information collection box is electrically connected to the pipe wall infrasound collector and the gas infrasound collector.

[0019] Preferably, it also includes:

[0020] The battery panel and the backup battery, the battery panel is installed on the top of the information collection pile, the backup battery is installed in the information collection pile, and is electrically connected to the battery panel, the camera, the edge computing instrument, the 485 collection server, the 4G network communicator, and the electronic control unit.

[0021] Preferably, it also includes:

[0022] A circuit breaker, a protector and a charging controller, wherein the circuit breaker, the protector and the charging controller are electrically connected between the backup battery and the electronic control unit.

[0023] Preferably, it also includes:

[0024] A collection pile base is fixedly connected to the side wall of the valve well, and the information collection pile is fixedly connected to the wellhead of the valve well through the collection pile base.

[0025] Preferably, it also includes:

[0026] A lightning rod is connected above the solar panel.

[0027] In the technical solution provided by the present invention, infrasound technology is used in the application of continuous leakage signal detection in gas pipelines. This signal is different from the sudden change leakage signal, and its amplitude and frequency are relatively stable and its duration is relatively long. The traditional infrasound threshold method has poor accuracy and high false negative rate in judging the continuous signal. The present invention uses an AI model to put a section of infrasound signal data collected under the leakage condition into the AI ​​model for model matching and identification, so as to judge whether the gas pipeline is leaking. This effectively solves the limitation of the traditional method that the signal must be monitored in real time, and even if the sudden change signal is missed, the leakage can be judged through model matching and identification. It has the characteristics of high accuracy, low false alarm rate, low false negative rate, and can be measured at any time. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0029] Figure 1 It is a flow chart of the present invention;

[0030] Figure 2 It is a schematic diagram of the overall structure of the present invention;

[0031] Figure 3 It is a schematic diagram of the information collection pile and the collection pile base of the present invention.

[0032] In the figure, 1-battery panel; 2-camera; 3-edge computing device; 4-485 acquisition server; 5-4G network communicator; 6-backup battery; 7-circuit breaker and protector; 8-charging controller; 9-information acquisition pile; 10-collection pile base; 11-electronic control unit; 12-natural gas pipeline; 13-lightning rod; 14-pipe wall infrasound collector; 15-gas infrasound collector; 16-valve well mobile information acquisition box; 17-valve well. DETAILED DESCRIPTION

[0033] To make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0034] refer to Figure 1 The specific embodiment of the present invention provides a gas pipeline leakage detection method based on infrasound and AI model, comprising the following steps:

[0035] Step 1: Collect leakage signals through continuous infrasound;

[0036] Step 2: Signal preprocessing: by analyzing the amplitude of the original signal, smoothing it, and then analyzing the processed signal;

[0037] Step 3: De-noise the sound wave and obtain the leakage signal characteristics;

[0038] Step 4: Process and analyze signal features through AI models;

[0039] Step 5: Obtain leakage identification information.

[0040] At present, the main methods for detecting gas pipeline leaks include helium mass spectrometry, negative pressure wave method, infrared detection method, infrasound method, etc.; helium mass spectrometry is suitable for sealed pipelines, and the test system is simple, but its detection cycle is long, and it needs repeated inflation and deflation and is limited by the volume of the detection structure; the negative pressure wave method has the advantages of fast deployment speed and rapid response, and can detect sudden pipeline leaks in time, but due to the poor sensitivity of the environment to slow leaks with small flow and small range, missed detection and misjudgment are prone to occur; infrared gas imaging detection can realize the visualization of the leakage source and realize non-contact imaging detection, but obstacles will affect the imaging and cause missed detection. In the present application, infrasound technology is used to detect continuous leakage signals in gas pipelines. This signal is different from the sudden leakage signal, and its amplitude and frequency are relatively stable and its duration is relatively long. The traditional infrasound threshold method has poor accuracy and high false negative rate in judging the continuous signal. The present invention uses an AI model to put a section of infrasound signal data collected under the leakage condition into the AI ​​model for model matching and identification, so as to judge whether the gas pipeline is leaking. This effectively solves the limitation of the traditional method that the signal must be monitored in real time, and even if the sudden change signal is missed, the leakage can be judged through model matching and identification. It has the characteristics of high accuracy, low false alarm rate, low false negative rate, and can be measured at any time.

[0041] To further optimize the solution, an artificial leakage point is added in step one by means of shock wave, and the leakage distance is calculated by the time interval between the signal of the artificial leakage point and the leakage signal in step five to determine the leakage location.

[0042] Precise positioning is a positioning method based on shock wave detection. When a leak has already occurred, artificially adding a second leak will interfere with the original leak, thereby generating sound wave fluctuations. This method is called the shock wave method. In this way, a leak point can be artificially added to test the existing pipeline. When two peak points appear, it means that there is a leak point. The leakage distance can be calculated based on the time interval between the artificial signal and the second signal. If two shock wave signals are generated in a certain length, the distance between the two shock wave positions and the generated leakage distance can be calculated to determine the leakage direction and thus achieve leakage positioning.

[0043] To further optimize the solution, in step 3, the denoised signal is integrated with empirical mode decomposition, and the signal features of the decomposed signal are extracted through multi-scale permutation entropy; in step 4, the original signal and the signal features processed in step 3 are input into the neural network for learning. The features of the continuous leakage signal are different from those of the non-leakage signal, and this difference can be extracted and distinguished by signal processing. The specific means are to first perform wavelet processing, then eemd decomposition, and then perform multi-scale entropy dimensionality reduction processing. In this way, the signal features obtained are different in leakage and non-leakage situations. This data is used as a machine learning data set for training, thereby establishing an AI model, embedding the trained AI model into the software, and using the AI ​​model to determine whether the pipeline is leaking.

[0044] Further optimization plans also include:

[0045] Step 6: Evaluate the leakage identification accuracy of the proposed AI model, use the AI ​​model to test the pipeline signal to determine the leakage information, compare the judgment results with the expected results, and calculate the identification accuracy.

[0046] This application is based on infrasound and AI model calculations. It determines whether a leak occurs by extracting the leakage characteristics of the signal, and can remotely control and analyze it. It is accurate and efficient, and can be deployed over a wide area to form a grid-type monitoring system.

[0047] The leakage feature judgment process can use multi-batch judgment to improve accuracy, which means that after the collected data is divided into multiple groups, a comprehensive judgment is made on the multiple groups of data, and finally the results are uniformly output, which can improve the accuracy of leakage judgment. Due to the persistence of leakage events, leakage verification can be performed at any time and multiple times to ultimately determine whether a leakage has occurred. Suspected leakage is added on the basis of leakage judgment, and the spike signal generated at the beginning of the leakage and the leakage reaching a certain value in multi-batch processing are judged as suspected leakage; the threshold adjustment range of the spike signal detection and the threshold adjustment range of the suspected leakage are increased to achieve the effect of early warning and improving the judgment accuracy.

[0048] The signal processing means can be replaced, and the signal after wavelet processing can also be processed by blind source analysis method, which is suitable for the situation where multiple sensors observe leakage signals.

[0049] refer to Figure 2-3 The present application also provides a gas pipeline leakage detection device based on infrasound and AI model, which is applied to the above-mentioned gas pipeline leakage detection method based on infrasound and AI model, including an information collection pile 9 and a camera 2, an edge computing device 3, a 485 collection server 4, a 4G network communicator 5, and an electronic control unit 11 installed on the information collection pile 9 from top to bottom and electrically connected. The information collection pile 9 is fixedly connected to the wellhead edge of the valve well 17, and a plurality of pipe wall infrasound collectors 14 and a plurality of gas infrasound collectors 15 are installed on the natural gas pipeline 12 in the valve well 17. A valve well mobile information collection box 16 is detachably connected to the inner wall of the valve well 17, and the valve well mobile information collection box 16 is electrically connected to the pipe wall infrasound collector 14 and the gas infrasound collector 15.

[0050] By installing a gas infrasound collector 15 and a pipe wall infrasound collector 14 on the vent pipe in the valve well 17, the internal gas infrasound and pipe wall infrasound signals of the gas pipeline are collected; the pipe wall infrasound collector 14 and the gas infrasound collector 15 are collected by the edge computing device 3, and a preliminary analysis is performed, and the suspected leakage signal is reported to the server and the remote control end for further leakage judgment; the valve well 17 is monitored by the camera 2; the pipe wall infrasound collector 14 and the gas infrasound collector 15 transmit data to the 485 collection server through the 485 communication method, and the 485 collection server converts the data into network data through the TCP protocol and reports it to the MQTT server of the remote control center to store it in the database, or transmits it to the central analysis leakage judgment system and stores it in the database through the system. The central analysis leakage judgment system can read the alarm record data in the database for overall judgment.

[0051] Further optimization plans also include:

[0052] The battery panel 1 and the backup battery 6, the battery panel 1 is installed on the top of the information collection pile 9, the backup battery 6 is installed in the information collection pile 9, and is electrically connected to the battery panel 1, the camera 2, the edge computing device 3, the 485 collection server 4, the 4G network communicator 5, and the electronic control unit 11.

[0053] Due to the limitation of on-site working conditions, some valve wells 17 cannot be equipped with a solar power supply system. A mobile information collection box 16 can be installed in the valve well 17, and the pipeline infrasonic signal data can be detected and transmitted only by the backup battery 6. After the analysis is completed, the whole can be disassembled or the battery can be reinstalled. The user has less investment, less renovation construction, no land acquisition and coordination with relevant departments are required, and regular leakage detection of all gas pipelines can be achieved.

[0054] The MQTT protocol has the characteristics of low bandwidth consumption, publish / subscribe mode, persistent session, low power consumption, and security, and is suitable for the battery-powered and unstable field network working conditions of the present invention.

[0055] Further optimization plans also include:

[0056] The circuit breaker and protector 7 and the charging controller 8 are electrically connected between the backup battery 6 and the electronic control unit 11 .

[0057] Further optimization plans also include:

[0058] The collection pile foundation 10 is fixedly connected to the side wall of the valve well 17 , and the information collection pile 9 is fixedly connected to the wellhead of the valve well 17 through the collection pile foundation 10 .

[0059] Further optimization plans also include:

[0060] A lightning rod 13 is connected above the solar panel 1 .

[0061] To further optimize the solution, the valve well mobile information collection box 16 is fixedly connected to the inner wall of the valve well 17 by bolts.

[0062] Further optimization plans also include:

[0063] The cloud platform, the leakage data in the single valve well 17, the double valve well 17, and the several valve wells 17 are monitored through the cloud platform respectively.

[0064] The following table shows the experimental data and comparative data:

[0065]

[0066]

[0067] The number of equipment detections is the number of judgments made for each set of collected data, and the number of abnormal detections is the number of times a leak is not correctly judged or a normal condition is judged as a leak.

[0068] It should be noted that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", etc. used herein to indicate the orientation or position relationship is based on the orientation or position relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", "third", etc. are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0069] In the description of this article, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to the specific circumstances.

[0070] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A gas pipeline leakage detection method based on infrasound and AI model, characterized in that: The following steps are involved: Step 1: Collect leakage signals through continuous infrasound; Step 2: signal preprocessing; Step 3: De-noise the sound wave and obtain the leakage signal characteristics; Step 4: Process and analyze the signal features through the AI ​​model; Step 5: Obtain leakage identification information.

2. The gas pipeline leakage detection method based on infrasound and AI model according to claim 1 is characterized in that: An artificial leakage point is added in step one by means of a shock wave, and the leakage distance is calculated by the time interval between the signal of the artificial leakage point and the leakage signal in step five to determine the leakage position.

3. The gas pipeline leakage detection method based on infrasound and AI model according to claim 1 is characterized in that: In the step three, the denoised signal is integrated with the empirical mode decomposition, and the signal features are extracted from the decomposed signal through multi-scale permutation entropy.

4. The gas pipeline leakage detection method based on infrasound and AI model according to claim 3 is characterized in that: In the step 4, the original signal and the signal features processed in the step 3 are input into a neural network for learning to establish the AI ​​model, and whether the pipeline is leaking is determined by the AI ​​model.

5. The gas pipeline leakage detection method based on infrasound and AI model according to claim 1 is characterized in that: Also includes: Step 6: Evaluate the leakage identification accuracy of the proposed AI model.

6. A gas pipeline leakage detection device based on infrasound and AI model, characterized in that: A gas pipeline leakage detection method based on infrasound and AI model applied to any one of claims 1-5, comprising an information collection pile (9) and a camera (2), an edge computing device (3), a 485 collection server (4), a 4G network communicator (5), and an electronic control unit (11) installed on the information collection pile (9) from top to bottom, wherein the information collection pile (9) is fixedly connected to the wellhead edge of a valve well (17), and a plurality of pipe wall infrasound collectors (14) and a plurality of gas infrasound collectors (15) are installed on the natural gas pipeline (12) in the valve well (17), and a valve well mobile information collection box (16) is detachably connected to the inner wall of the valve well (17), and the valve well mobile information collection box (16) is electrically connected to the pipe wall infrasound collector (14) and the gas infrasound collector (15).

7. The gas pipeline leakage detection device based on infrasound and AI model according to claim 6 is characterized in that: Also includes: A battery panel (1) and a backup battery (6), wherein the battery panel (1) is installed on the top of the information collection pile (9), and the backup battery (6) is installed in the information collection pile (9) and is electrically connected to the battery panel (1), the camera (2), the edge computing device (3), the 485 collection server (4), the 4G network communication device (5), and the electronic control unit (11).

8. The gas pipeline leakage detection device based on infrasound and AI model according to claim 7 is characterized in that: Also includes: A circuit breaker and protector (7) and a charging controller (8), wherein the circuit breaker and protector (7) and the charging controller (8) are electrically connected between the backup battery (6) and the electronic control unit (11).

9. The gas pipeline leakage detection device based on infrasound and AI model according to claim 6 is characterized in that: Also includes: A collection pile base (10), wherein the collection pile base (10) is fixedly connected to the side wall of the valve well (17), and the information collection pile (9) is fixedly connected to the wellhead of the valve well (17) through the collection pile base (10).

10. The gas pipeline leakage detection device based on infrasound and AI model according to claim 7 is characterized in that: Also includes: A lightning rod (13), wherein the lightning rod (13) is connected above the solar panel (1).

Images