Low-pressure gas pipeline small leakage detection method and system

Abstract

The invention discloses a low-pressure gas pipeline small leakage detection method and system, and relates to the field of low-pressure gas pipeline small leakage detection. Small-aperture leakage signals are collected and segmented, and signal segments are generated; carrying out preprocessing and denoising operation on the segmented signal segments to obtain reconstructed signals; carrying out convolution on the reconstructed signal and the wavelet basis, carrying out modulus taking, and carrying out smoothing processing on a result to obtain a first-order scattering coefficient; transforming the first-order scattering coefficient, generating logarithmic scattering transformation of a reconstructed signal, and taking an average value obtained along a time dimension as input of a classification model, and training the classification model; a pipeline inspection robot based on a sound source localization algorithm detects a signal at a leakage position and inputs the signal to the trained classification model to obtain a pipeline leakage aperture recognition result; according to the invention, accurate pipeline leakage aperture identification and pipeline leakage point positioning are effectively realized, fusion of heterogeneous sensor data is realized, and an automatic inspection task of an industrial pipeline is completed.

Description

Technical Field

[0001] The present invention relates to the field of small leak detection in low-pressure gas pipelines, and in particular to a small leak detection method and system for low-pressure gas pipelines based on a Beidou positioning and navigation inspection robot. Background Art

[0002] With the continuous development of industry, industrial transportation, as an important part of industrial production, has become a key issue for the safety inspection of transportation facilities. Pipelines, as one of the important modes of industrial transportation and the main carrier of oil and gas transportation, often experience varying degrees of leakage due to factors such as corrosion, construction legacy, aging, or geological deformation, posing potential hazards to the environment and human safety. Therefore, it is very necessary to locate the leak point and determine the leakage extent of small-diameter pipeline leaks.

[0003] There are four main types of existing technical solutions:

[0004] 1. Use magnetic flux leakage technology to detect pipeline leakage, and use target detection to classify targets in the magnetic flux leakage image to determine the degree of leakage.

[0005] 2. Use optimized convolutional neural network to identify leakage aperture.

[0006] 3. Using traditional acoustic features, such as energy, amplitude, ASL, RMS, frequency and other characteristic parameters, the features extracted from the collected signals are input into the classifier to identify the pipeline leakage aperture.

[0007] 4. Use harmonic analysis technology to analyze the vibration of the pipeline and build a leakage model to identify the leakage status.

[0008] Existing technologies, due to the long pipeline distances and the relatively short acquisition time at a single point, result in a sharp drop in the number of samples. Some models and algorithms are unable to adapt to small sample sizes, resulting in low recognition accuracy and failure to meet the requirements of industrial pipeline inspection. Furthermore, due to the complex environment in which industrial transport pipelines operate, with some pipelines deeply buried, many inspection technologies require sensors to come into contact with the pipeline, making many inspection technologies unsuitable for pipeline leak detection. Existing sensors are therefore not suitable for the inspection environment. Summary of the Invention

[0009] The purpose of the present invention is to provide a method and system for detecting small leaks in low-pressure gas pipelines based on a Beidou positioning and navigation inspection robot, which can effectively realize accurate pipeline leakage aperture identification and accurate pipeline leakage point positioning, and complete the automatic inspection task of industrial pipelines.

[0010] According to a first aspect of an embodiment of the present disclosure, a method for detecting leakage in a low-pressure gas pipeline is provided, comprising the following steps:

[0011] Collect small aperture leakage signals and segment them to generate signal segments;

[0012] The segmented signal segments are preprocessed and denoised to obtain reconstructed signals;

[0013] The reconstructed signal is convolved with the wavelet basis, modulo-ed, and the result is smoothed to obtain the first-order scattering coefficient;

[0014] The first-order scattering coefficient is transformed to generate the logarithmic scattering transform of the reconstructed signal, and the average value obtained along the time dimension is used as the input of the classification model to train the classification model;

[0015] The pipeline inspection robot based on the sound source localization algorithm detects the signal at the leakage location, inputs it into the trained classification model, and obtains the pipeline leakage aperture identification result.

[0016] In one embodiment, the pipeline inspection robot based on the sound source localization algorithm includes:

[0017] The mobile chassis includes a chassis and wheels, and the chassis is equipped with a ROS host module, a sound source positioning module, a Beidou navigation module, a lidar module and a motor module.

[0018] The ROS host module includes a PC, in which the ROS system is installed;

[0019] The sound source positioning module, Beidou navigation module, and lidar module signals are connected to the PC of the ROS host module;

[0020] The motor module includes a chassis drive chip, a motor driver, and a motor. The PC sends a control signal to the chassis drive chip, which controls the motor through the motor driver. The motor output shaft is connected to the wheels of the mobile chassis.

[0021] In one embodiment, a pipeline inspection robot based on a sound source localization algorithm includes the following steps:

[0022] The sound source localization module uses the sound source localization algorithm to obtain the location of the sound source signal and sends the sound source coordinate data to the ROS host module;

[0023] The Beidou navigation module receives Beidou satellite signals for precise positioning and sends real-time location data to the ROS host module;

[0024] Scan surrounding objects through the LiDAR module and send the data of surrounding objects to the ROS host module;

[0025] After receiving data from the sound source positioning module, Beidou navigation module, and lidar module, the ROS host module plans a path based on the data and sends control signals to the motor module.

[0026] After receiving the control signal, the motor module controls the mobile chassis to move, so that the pipeline inspection robot approaches the sound source;

[0027] The pipeline inspection robot collects sound source information through microphones.

[0028] In one embodiment, the sound source localization algorithm comprises the following steps:

[0029] Q1. Place three microphones at the three endpoints of an isosceles right triangle on a plane.

[0030] Q2. Select two microphones and obtain the sound emitted by the sound source received by the two microphones;

[0031] Q3. Using the improved and optimized GCC-PHAT algorithm, we suppress reverberation and noise from the sound received by the two microphones and obtain the time delay between the sound reaching the two microphones.

[0032] Repeat Q2-Q3 to obtain the time delay between the sound reaching each two microphones.

[0033] Q5. Based on the time delay between the sound reaching the microphone, the exact location of the sound source is obtained through the relative position algorithm.

[0034] In one embodiment, the detailed steps of the improved and optimized GCC-PHAT algorithm are as follows:

[0035] The sound value and cross-spectrum after Fourier transformation are obtained according to the sound emitted by the sound source received by the microphone. In order to solve the phase distortion of the cross-spectrum caused by reverberation, an improved recursive smoothing method is introduced to estimate the cross-spectrum:

[0036]

[0037] In the above formula, λ is the smoothing factor, and the cross-spectral estimation of the historical frame and the current frame is fused by the smoothing factor λ. Y1(f,k) represents the short-time Fourier transform coefficient of the k-th frame signal of the first sensor at frequency f. is the complex conjugate of the short-time Fourier transform coefficient of the k-th frame signal of the second sensor at frequency f;

[0038] The cross spectrum is weighted by the PHAT weighting function to suppress noise and reverberation interference and obtain the cross power spectrum;

[0039] The generalized cross-correlation function is obtained by Fourier transform, cross-spectrum and cross-power spectrum of the simultaneous signals;

[0040] The improved PHAT weighting function is obtained, and the phase spectrum is obtained by combining the PHAT weighting function and the cross power spectrum. The improved PHAT weighting function is:

[0041]

[0042] In the above formula, g mod (f) is the PHAT weighting function, is the cross-spectrum, α is the regularization parameter;

[0043] In order to solve the problem of poor adaptability of a single weighting function, the ROTH weighting is combined with the regularized improved PHAT weighting:

[0044]

[0045] In the above formula is the power spectrum, β is the control mixing ratio, β∈[0,1];

[0046] The phase spectrum is:

[0047]

[0048] In the above formula is the phase spectrum, τ is the relative time delay between the signal arriving at the two microphones;

[0049] The weighted generalized cross-correlation function obtained by combining the generalized cross-correlation function, the ROTH weighted and regularized improved PHAT weighted function and the phase spectrum is:

[0050]

[0051] In the above formula is the weighted generalized cross-correlation function, p is the peak value of the weighted generalized cross-correlation function, and τ is the time delay between the signals reaching the two microphones;

[0052] The time delay τ between the signals reaching the two microphones can be determined by finding the peak position of the weighted generalized cross-correlation function.

[0053] In one embodiment, the relative position algorithm is implemented as follows:

[0054] Establish a plane coordinate system, set the sound source position to S(x, y), microphone M1(0, 0), microphone M2(d, 0), microphone M3(0, d);

[0055] According to the position relationship on the plane coordinate system, the TDOA equation is obtained;

[0056] The TDOA equation is:

[0057]

[0058] In the above equation, c is the speed of sound, τ 12is the time delay between the sound reaching microphone M1 and microphone M2, τ 23 is the time delay between the sound reaching microphone M2 and microphone M3.

[0059] Assuming the distance from the sound source to microphone M1 is r1, the distance from the sound source to microphone M2 is r2, and the distance from the sound source to microphone M3 is r3, the distance equation can be obtained as follows:

[0060] r1=r2+cτ 12

[0061] r2=r3+cτ 23

[0062]

[0063] By solving the TDOA equation and the distance equation simultaneously, we can obtain:

[0064]

[0065] By calculating the values ​​of x and y, the sound source coordinates S(x, y) can be obtained.

[0066] According to a second aspect of an embodiment of the present disclosure, a low-pressure gas pipeline small leak detection system is provided, comprising:

[0067] Control processing module: This module is mainly composed of a laptop and an STM32 microcontroller;

[0068] Leak detection module: collects acoustic leakage data and feeds it back to the host computer to identify the leakage aperture;

[0069] Sound source localization module: This module consists of three microphone sensors in an array. It mainly collects sound source localization data and feeds it back to the control processing module to realize the coordinate positioning of the leakage sound source;

[0070] Mapping module: This module uses lidar to map the working environment and provide a map for subsequent positioning and navigation;

[0071] Positioning and navigation module: This module mainly uses the Beidou positioning and navigation module to realize the robot's positioning and path planning in the saved map to ensure that the robot travels along the inspection pipeline;

[0072] Power module: This module provides power to sensors, motors, etc. that require power;

[0073] Drive motor: drives the wheels by receiving instructions from the host computer.

[0074] According to the third aspect of the embodiment of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored and running on the memory. When the processor executes the program, it implements the method for detecting small leaks in low-pressure gas pipelines based on the Beidou positioning and navigation inspection robot.

[0075] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the method for detecting small leaks in a low-pressure gas pipeline based on a Beidou positioning and navigation inspection robot is implemented.

[0076] The above technical solutions adopted by the present invention offer the following advantages over existing technologies: A method for detecting small leaks in low-pressure gas pipelines based on a Beidou positioning and navigation inspection robot is proposed. Accurate pipeline leak aperture identification is achieved through the use of a wavelet scattering transform algorithm; the use of a microphone array and optimization of positioning algorithm parameters allows for relatively accurate location of pipeline leaks; and the integration of heterogeneous sensor data, such as Beidou, lidar, and microphones, based on the ROS communication framework, enables automated inspection of industrial pipelines. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The drawings in the specification, which constitute a part of this application, are used to provide further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute improper limitations on this application.

[0078] Figure 1 This is the overall framework diagram of a low-pressure gas pipeline small leak detection system based on a Beidou positioning and navigation inspection robot. DETAILED DESCRIPTION

[0079] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

[0080] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present application belongs.

[0081] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0082] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the methods and systems according to the various embodiments of the present disclosure. It should be noted that each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code can include one or more executable instructions for implementing the logical functions specified in the various embodiments. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, or they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the flowchart and / or block diagram, and the combination of the boxes in the flowchart and / or block diagram, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or can be implemented using a combination of dedicated hardware and computer instructions.

[0083] Example 1:

[0084] This embodiment provides a method for detecting small leaks in low-pressure gas pipelines using a Beidou positioning and navigation inspection robot. Based on the GB50028 Urban Gas Design Specification, the detection target is a steel gas pipeline with an internal pressure of less than 0.4 MPa and a leak diameter of less than 2 mm. The method includes the following steps:

[0085] S1. Collect leakage signals for four leak apertures: 0.5 mm, 1 mm, 1.5 mm, and 2 mm. Each aperture is sampled three times, each time for 30 seconds. Use a 1-second non-overlapping rectangular window to segment the recorded measurement values ​​every 30 seconds, generating 30 signal segments.

[0086] f k (t) = f(t + k·T w )·ω(t)

[0087] In the above formula, f k (t) is the segmented signal; k is the index (k = 0, 1, 2....30); t is the local time variable; ω(t) is the rectangular window function, ensuring that t∈(0, T w ) is 1; T w Set the window duration to 1s.

[0088] S2. Select the segmented signal segment for windowing operation and use it as the input signal:

[0089] x m =[x(m*M),x(m*M+1)...x(m*M+N-1)]

[0090] x w(m,n)=x(m*M+n)*w(n)

[0091] In the above formula, x m represents the fragment of the mth frame; x w (m,n) is the signal after weighting the window; m is the frame index.

[0092]

[0093] In the above formula, w(n) is the windowing function (Hanning window); N is the frame length (N=512); and M is the frame shift (M=0.5*N).

[0094] Perform FFT on each frame of the windowed signal to obtain the complex spectrum X(m,k):

[0095]

[0096] Noise estimation, obtain the noise amplitude spectrum:

[0097]

[0098] In the above formula, D(k) is the noise amplitude spectrum; X t (k) is the spectral coefficient of the t-th frame.

[0099] Perform amplitude spectrum denoising and phase spectrum recovery operations:

[0100]

[0101] Reconstruct the signal by inverse short-time Fourier transform:

[0102]

[0103] In the above formula, This is the reconstructed signal after denoising.

[0104] S3. Extract the first-order scattering coefficient

[0105] (1) Select the wavelet basis and determine the scaling function

[0106] This embodiment uses the db2 wavelet basis and sets the minimum scale function to j min =1, the minimum scale function is j max = J = log2N, the scale sequence is [j min ,j min +1,......,J].

[0107] (2) Calculate the convolution of the reconstructed signal and the wavelet basis selected above and take the modulus:

[0108] |U1f(x,j)|=|f*ψ j(x)|

[0109] (3) Smoothing the modular operation result can obtain the first-order scattering coefficient:

[0110] S1f(x,j)=|U1f(x,j)|*φ J (x)

[0111] In the above formula, φ J (x) is a Gaussian window of scale J.

[0112] S4. Training the model: transform the extracted first-order scattering coefficient through the coefficient obtained by the natural logarithm pair, generate a series of logarithmic scattering transforms of the signal and take the average value along the time dimension as the input of the classification model. The classification model is trained, and the zero-order scattering coefficient is not considered as an input feature.

[0113] S1f(x,j)=log(ε+S[p]x(t))

[0114]

[0115] In the above formula, S1f(x,j) is the first-order scattering coefficient; p is the scattering path; ε is the regularization constant and is set to 10 -6 ; T is the signal duration; μ p is the average value of the logarithmic scattering coefficient.

[0116] S5. Identify the leak diameter

[0117] The pipeline inspection robot based on the sound source localization algorithm detects the signal at the leakage location, inputs it into the trained classification model, and obtains the pipeline leakage aperture identification result.

[0118] Specifically, pipeline inspection robots based on sound source localization algorithms include:

[0119] The mobile chassis includes a chassis and wheels, and the chassis is equipped with a ROS host module, a sound source positioning module, a Beidou navigation module, a lidar module and a motor module.

[0120] The ROS host module includes a PC, in which the ROS system is installed;

[0121] The sound source positioning module, Beidou navigation module, and lidar module signals are connected to the PC of the ROS host module;

[0122] The motor module includes a chassis drive chip, a motor driver, and a motor. The PC sends a control signal to the chassis drive chip, which controls the motor through the motor driver. The motor output shaft is connected to the wheels of the mobile chassis.

[0123] Specifically, the implementation method of the pipeline inspection robot based on the sound source localization algorithm includes the following steps:

[0124] The sound source localization module uses the sound source localization algorithm to obtain the location of the sound source signal and sends the sound source coordinate data to the ROS host module;

[0125] The Beidou navigation module receives Beidou satellite signals for precise positioning and sends real-time location data to the ROS host module;

[0126] Scan surrounding objects through the LiDAR module and send the data of surrounding objects to the ROS host module;

[0127] After receiving data from the sound source positioning module, Beidou navigation module, and lidar module, the ROS host module plans a path based on the data and sends control signals to the motor module.

[0128] After receiving the control signal, the motor module controls the mobile chassis to move, so that the pipeline inspection robot approaches the sound source;

[0129] The pipeline inspection robot collects sound source information through microphones.

[0130] Specifically, the implementation method of the pipeline inspection robot sound source localization algorithm includes the following steps:

[0131] Q1. Place three microphones at the three endpoints of an isosceles right triangle on a plane.

[0132] Q2. Select two microphones and obtain the sound emitted by the sound source received by the two microphones;

[0133] Q3. Use the GCC-PHAT algorithm to suppress reverberation and noise from the sound received by the two microphones and obtain the time delay between the sound reaching the two microphones.

[0134] Repeat Q2-Q3 to obtain the time delay between the sound reaching each two microphones.

[0135] Q5. Based on the time delay between the sound reaching the microphone, the exact location of the sound source is obtained through the relative position algorithm.

[0136] Specifically, the GCC-PHAT algorithm of the sound source localization algorithm is implemented as follows:

[0137] Acquire a Fourier-transformed sound value and a cross-spectrum according to the sound emitted by the sound source received by the microphone;

[0138] The sound value after Fourier transform is:

[0139]

[0140] In the above formula, Y n (f) is the sound value after Fourier transform, y n (k) is the sound value received by the nth microphone at discrete time point k, j is the imaginary number symbol, π is pi, and f is the frequency of the sound;

[0141] In order to solve the cross-spectrum phase distortion caused by reverberation, an improved recursive smoothing method is introduced to estimate the cross-spectrum:

[0142]

[0143] In the above formula, λ is the smoothing factor. The working scene is set to outdoor, which is regarded as a reverberation scene. λ is set to 0.9. The cross-spectral estimation of the historical frame and the current frame is fused by the smoothing factor λ. Y1(f,k) represents the short-time Fourier transform coefficient of the k-th frame signal of the first sensor at frequency f. is the complex conjugate of the short-time Fourier transform coefficient of the k-th frame signal of the second sensor at frequency f;

[0144] The cross spectrum is weighted by the PHAT weighting function to suppress noise and reverberation interference, and the cross power spectrum is obtained:

[0145]

[0146] In the above formula, is the cross power spectrum, g(f) is the PHAT weighting function, φ smooth (f, k) is the recursively smoothed estimated cross-spectrum;

[0147] The Fourier transform, cross-spectrum and cross-power spectrum of the simultaneous signals are used to obtain the generalized cross-correlation function, which is:

[0148]

[0149] In the above formula is the generalized cross-correlation function, To perform Fourier transform, is the cross power spectrum, φ smooth (f, k) is the recursively smoothed estimated cross-spectrum, and p is the peak value of the generalized cross-correlation function;

[0150] In the low signal-to-noise ratio (SNR) frequency band, the cross-spectrum value is small, resulting in an excessively large weighting function, which amplifies the noise. The above weighting function is regularized and improved to obtain an improved PHAT weighting function. The phase spectrum is obtained by combining the PHAT weighting function and the cross-power spectrum. The improved PHAT weighting function is:

[0151]

[0152] In the above formula, gmod (f) is the PHAT weighting function, is the cross-spectrum, α is the regularization parameter, and the initial setting range of α is [0.1, 0.5] according to the characteristics of pipeline leakage acoustic signal. In order to avoid numerical instability, α is finally set to 0.2;

[0153] In order to solve the problem of poor adaptability of a single weighting function, the ROTH weighting is combined with the regularized improved PHAT weighting:

[0154]

[0155] In the above formula is the power spectrum, β is the control mixing ratio, β∈[0,1]. According to ISO 18436-5:2021 "Machinery Vibration Monitoring", in industrial environments dominated by impact noise, the mixing weight coefficient β is recommended to be 0.65-0.75 to balance delay resolution and noise robustness. In this embodiment, β is set to 0.7 to suppress strong noise while sacrificing a small amount of delay resolution.

[0156] The phase spectrum is:

[0157]

[0158] In the above formula is the phase spectrum, τ is the relative time delay between the signal arriving at the two microphones;

[0159] The weighted generalized cross-correlation function obtained by combining the generalized cross-correlation function, the ROTH weighted and regularized improved PHAT weighted function and the phase spectrum is:

[0160]

[0161] In the above formula, is the weighted generalized cross-correlation function, p is the peak value of the weighted generalized cross-correlation function, and τ is the time delay between the signals reaching the two microphones;

[0162] The time delay τ between the signals reaching the two microphones can be determined by finding the peak position of the weighted generalized cross-correlation function.

[0163] Specifically, the relative position algorithm of the sound source localization algorithm is implemented as follows:

[0164] Establish a plane coordinate system, set the sound source position to S(x,y), microphone M1(0,0), microphone M2(d,0), microphone M3(0,d);

[0165] According to the position relationship on the plane coordinate system, the TDOA equation is obtained;

[0166] The TDOA equation is:

[0167]

[0168] In the above equation, c is the speed of sound, τ 12 is the time delay between the sound reaching microphone M1 and microphone M2, τ 23 is the time delay between the sound reaching microphone M2 and microphone M3,

[0169] Assuming the distance from the sound source to microphone M1 is r1, the distance from the sound source to microphone M2 is r2, and the distance from the sound source to microphone M3 is r3, the distance equation can be obtained as follows:

[0170] r1=r2+cτ 12

[0171] r2=r3+cτ 23

[0172]

[0173] By solving the TDOA equation and the distance equation simultaneously, we can obtain:

[0174]

[0175] By calculating the values ​​of x and y, the sound source coordinates S(x, y) can be obtained.

[0176] Example 2:

[0177] This embodiment provides a low-pressure gas pipeline small leak detection system based on a Beidou positioning and navigation inspection robot, including:

[0178] Control processing module, which is mainly composed of a laptop and an STM32 microcontroller;

[0179] The sound source localization module consists of three microphone sensors in an array. It mainly collects sound source localization data and feeds it back to the control processing module to realize the coordinate positioning of the leakage sound source;

[0180] Leak detection module, which collects acoustic leakage data and feeds it back to the host computer to identify the leakage aperture;

[0181] Mapping module, which uses LiDAR to map the working environment and provide a map for subsequent positioning and navigation;

[0182] Positioning and navigation module: This module mainly uses the Beidou positioning and navigation module to realize the robot's positioning and path planning in the saved map to ensure that the robot travels along the inspection pipeline;

[0183] Power module, which provides power to sensors, motors, etc. that require power;

[0184] The drive motor drives the wheels by receiving instructions from the host computer.

[0185] Example 3:

[0186] An electronic device includes a memory, a processor, and a computer program stored and running on the memory, wherein the processor, when executing the program, implements the above-mentioned method for detecting small leaks in a low-pressure gas pipeline based on a Beidou positioning and navigation inspection robot, including:

[0187] Collect small aperture leakage signals and segment them to generate signal segments;

[0188] The segmented signal segments are preprocessed and denoised to obtain reconstructed signals;

[0189] The reconstructed signal is convolved with the wavelet basis, modulo-ed, and the result is smoothed to obtain the first-order scattering coefficient;

[0190] The first-order scattering coefficient is transformed to generate the logarithmic scattering transform of the reconstructed signal, and the average value obtained along the time dimension is used as the input of the classification model to train the classification model;

[0191] The pipeline inspection robot based on the sound source localization algorithm detects the signal at the leakage location, inputs it into the trained classification model, and obtains the pipeline leakage aperture identification result.

[0192] Example 4:

[0193] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for detecting small leaks in low-pressure gas pipelines based on a Beidou positioning and navigation inspection robot, comprising:

[0194] Collect small aperture leakage signals and segment them to generate signal segments;

[0195] The segmented signal segments are preprocessed and denoised to obtain reconstructed signals;

[0196] The reconstructed signal is convolved with the wavelet basis, modulo-ed, and the result is smoothed to obtain the first-order scattering coefficient;

[0197] The first-order scattering coefficient is transformed to generate the logarithmic scattering transform of the reconstructed signal, and the average value obtained along the time dimension is used as the input of the classification model to train the classification model;

[0198] The pipeline inspection robot based on the sound source localization algorithm detects the signal at the leakage location, inputs it into the trained classification model, and obtains the pipeline leakage aperture identification result.

[0199] Those skilled in the art will appreciate that the modules or steps of the present disclosure described above can be implemented using a general-purpose computer device. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. Alternatively, they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. The present disclosure is not limited to any specific combination of hardware and software.

[0200] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

[0201] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Those skilled in the art should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A method for detecting small leaks in a low-pressure gas pipeline, characterized in that: The following steps are involved: Collect small aperture leakage signals and segment them to generate signal segments; The segmented signal segments are preprocessed and denoised to obtain reconstructed signals; The reconstructed signal is convolved with the wavelet basis, modulo-ed, and the result is smoothed to obtain the first-order scattering coefficient; The first-order scattering coefficient is transformed to generate the logarithmic scattering transform of the reconstructed signal, and the average value obtained along the time dimension is used as the input of the classification model to train the classification model; The pipeline inspection robot based on the sound source localization algorithm detects the signal at the leakage location, inputs it into the trained classification model, and obtains the pipeline leakage aperture identification result.

2. A method for detecting small leaks in a low-pressure gas pipeline according to claim 1, characterized in that: The pipeline inspection robot based on the sound source localization algorithm includes: The mobile chassis consists of a chassis and wheels. The chassis is equipped with a ROS host module, a sound source positioning module, a Beidou navigation module, a lidar module and a motor module. The ROS host module includes a PC, in which the ROS system is installed; The sound source positioning module, Beidou navigation module, and lidar module signals are connected to the PC of the ROS host module; The motor module includes a chassis drive chip, a motor driver, and a motor. The PC sends a control signal to the chassis drive chip, which controls the motor through the motor driver. The motor output shaft is connected to the wheels of the mobile chassis.

3. A method for detecting small leaks in a low-pressure gas pipeline according to claim 1, characterized in that: The detection method of the pipeline inspection robot based on the sound source localization algorithm at the leakage location includes the following steps: The sound source localization module uses the sound source localization algorithm to obtain the location of the sound source signal and sends the sound source coordinate data to the ROS host module; The Beidou navigation module receives Beidou satellite signals for precise positioning and sends real-time location data to the ROS host module; Scan surrounding objects through the LiDAR module and send the data of surrounding objects to the ROS host module; After receiving data from the sound source positioning module, Beidou navigation module, and lidar module, the ROS host module plans a path based on the data and sends control signals to the motor module. After receiving the control signal, the motor module controls the mobile chassis to move, so that the pipeline inspection robot approaches the sound source; The pipeline inspection robot collects sound source information through microphones.

4. A method for detecting small leaks in a low-pressure gas pipeline according to claim 1, characterized in that: The implementation method of the pipeline inspection robot sound source localization algorithm includes the following steps: S1. Place three microphones at the three endpoints of an isosceles right triangle on a plane. S2. Select two microphones and obtain the sound emitted by the sound source received by the two microphones; S3. Use the optimized GCC-PHAT algorithm to suppress reverberation and noise from the sound received by the two microphones and obtain the time delay of the sound reaching the two microphones; S4. Repeat S2-S3 to obtain the time delay between the sound reaching each two microphones; S5. According to the time delay between the sound reaching the microphone, the accurate position of the sound source is obtained through the relative position algorithm.

5. A method for detecting small leaks in a low-pressure gas pipeline according to claim 1, characterized in that: The GCC-PHAT algorithm implementation of the sound source localization algorithm is as follows: The sound value and cross-spectrum after Fourier transformation are obtained according to the sound emitted by the sound source received by the microphone. In order to solve the phase distortion of the cross-spectrum caused by reverberation, an improved recursive smoothing method is introduced to estimate the cross-spectrum: In the above formula, λ is the smoothing factor, and the cross-spectral estimation of the historical frame and the current frame is fused by the smoothing factor λ. Y1(f, k) represents the short-time Fourier transform coefficient of the k-th frame signal of the first sensor at frequency f. is the complex conjugate of the short-time Fourier transform coefficient of the k-th frame signal of the second sensor at frequency f; The cross spectrum is weighted by the PHAT weighting function to suppress noise and reverberation interference and obtain the cross power spectrum; The generalized cross-correlation function is obtained by Fourier transform, cross-spectrum and cross-power spectrum of the simultaneous signals; The improved PHAT weighting function is obtained, and the phase spectrum is obtained by combining the PHAT weighting function and the cross power spectrum. The improved PHAT weighting function is: In the above formula, g mod (f) is the PHAT weighting function, is the cross-spectrum, α is the regularization parameter; In order to solve the problem of poor adaptability of a single weighting function, the ROTH weighting is combined with the regularized improved PHAT weighting: In the above formula is the power spectrum, β is the control mixing ratio, β∈[0,1]; The phase spectrum is: In the above formula is the phase spectrum, τ is the relative time delay between the signal arriving at the two microphones; The weighted generalized cross-correlation function obtained by combining the generalized cross-correlation function, the ROTH weighted and regularized improved PHAT weighted function and the phase spectrum is: In the above formula is the weighted generalized cross-correlation function, and p is the weighted generalized cross-correlation function The peak value of τ is the time delay of the signal reaching the two microphones; The time delay τ between the signals reaching the two microphones can be determined by finding the peak position of the weighted generalized cross-correlation function.

6. A method for detecting small leaks in a low-pressure gas pipeline according to claim 1, characterized in that: The relative position algorithm of the sound source localization algorithm is implemented as follows: Establish a plane coordinate system, set the sound source position to S(x,y), microphone M1(0,0), microphone M2(d,0), microphone M3(0,d); According to the position relationship on the plane coordinate system, the TDOA equation is obtained; The TDOA equation is: In the above equation, c is the speed of sound, τ 12 is the time delay between the sound reaching microphone M1 and microphone M2, τ 23 is the time delay between the sound reaching microphone M2 and microphone M3, Assuming the distance from the sound source to microphone M1 is r1, the distance from the sound source to microphone M2 is r2, and the distance from the sound source to microphone M3 is r3, the distance equation can be obtained as follows: r1=r2+cτ 12 r2=r3+cτ 23 By solving the TDOA equation and the distance equation simultaneously, we can obtain: By calculating the values ​​of x and y, the sound source coordinates S(x, y) can be obtained.

7. Low-pressure gas pipeline small leak detection system, characterized by: include: Control processing module: This module is mainly composed of a laptop and an STM32 microcontroller; Leak detection module: collects acoustic leakage data and feeds it back to the host computer to identify the leakage aperture; Sound source localization module: This module consists of three microphone sensors in an array. It mainly collects sound source localization data and feeds it back to the control processing module to realize the coordinate positioning of the leakage sound source; Mapping module: This module uses lidar to map the working environment and provide a map for subsequent positioning and navigation; Positioning and navigation module: This module mainly uses the Beidou positioning and navigation module to realize the robot's positioning and path planning in the saved map to ensure that the robot travels along the inspection pipeline; Power module: This module provides power to sensors, motors, etc. that require power; Drive motor: drives the wheels by receiving instructions from the host computer.

8. An electronic device comprising a memory, a processor, and a computer program stored and running on the memory, characterized in that: When the processor executes the program, the low-pressure gas pipeline small leak detection method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for detecting small leaks in a low-pressure gas pipeline according to any one of claims 1 to 6 is implemented.

Images