A method of detecting a leak in a pipe

Abstract

The application provides a pipeline leakage detection method, comprising: determining a pipeline to be detected; at least one set of sound collection arrays is arranged around the pipeline to be detected; different sound collection arrays correspond to different collection planes on the pipeline to be detected; collecting at least one sound wave in a detection plane, the detection plane being a collection plane corresponding to any sound collection array in the at least one set of sound collection arrays, and the at least one sound wave being collected based on any sound collection array; based on the at least one sound wave, performing offline calculation on the detection plane to obtain a beam space matrix, a kd tree and a transformation matrix corresponding to the detection plane; and based on online processing of the beam space matrix, the kd tree and the transformation matrix corresponding to the detection plane, determining a leakage position on the detection plane, so as to effectively and timely detect pipeline leakage and improve the accuracy of pipeline leakage position detection.

Description

Technical Field

[0001] This invention relates to the field of pipeline leak monitoring technology, and in particular to a method for detecting pipeline leaks. Background Technology

[0002] Pipelines are considered the safest and most economical method for energy transmission. However, with increasing service life and corrosion from various media, pipelines may age and break down, leading to leaks, especially when transporting gases. Pipeline inspection is an economical and effective way to protect pipeline safety. Detecting defects and repairing them promptly before serious problems occur can not only prevent accidents but also extend pipeline lifespan. Therefore, how to effectively and promptly detect pipeline leaks is a pressing issue that needs to be addressed.

[0003] Currently, acoustic sensors can be installed at various points along the pipeline, and combined with acoustic detection methods to detect leaks. However, installing acoustic sensors at various points along the pipeline increases detection and labor costs. Furthermore, once a leak is identified, this method struggles to pinpoint its exact location. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method for detecting pipeline leaks, which can effectively and timely detect pipeline leaks and improve the accuracy of detecting the location of pipeline leaks.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] This invention provides a method for detecting pipeline leaks, comprising: identifying the pipeline to be detected; wherein at least one set of sound acquisition arrays is arranged around the pipeline to be detected, and different sound acquisition arrays correspond to different acquisition planes on the pipeline to be detected; acquiring at least one beam of sound wave within the acquisition plane to be detected; wherein the acquisition plane to be detected is the acquisition plane corresponding to any one of the at least one set of sound acquisition arrays; the at least one beam of sound wave is obtained based on any one of the sound acquisition arrays; based on the at least one beam of sound wave, offline calculations are performed on the acquisition plane to be detected to obtain the beam space matrix, kd-tree, and transformation matrix corresponding to the acquisition plane to be detected; wherein the beam space matrix is ​​determined based on a first number of beams of sound waves from the at least one beam of sound waves; the kd-tree and transformation matrix are determined based on the sound waves corresponding to each grid cell in a second number of grid cells; the second number of grid cells are obtained by uniformly dividing the acquisition plane to be detected; one grid cell corresponds to one beam of sound wave, and different grid cells correspond to different beams of sound waves; and online processing is performed based on the beam space matrix, kd-tree, and transformation matrix corresponding to the acquisition plane to determine the leak location on the acquisition plane.

[0007] Based on the above technical solution, the present invention can be further improved as follows.

[0008] Furthermore, any sound acquisition array is a multi-arm helical acoustic array comprising a preset number of microphones. The distance between any sound acquisition array and the pipe to be tested is a first distance. The distance between any sound acquisition array and the ground is a second distance. The horizontal and vertical field of view angles of any sound acquisition array relative to the pipe to be tested are both preset field of view angles.

[0009] Furthermore, within at least one sound wave, a first type of beam and a second type of beam are identified. The sidelobes of the first type of beam are lower than those of the second type of beam. Four × 4 beams are selected from the first type of beam. Based on the selected 4 × 4 beams, the beam space matrix corresponding to the plane to be detected is obtained.

[0010] Furthermore, based on the selected 4×4 beams, a first matrix is ​​constructed, which is represented as follows:

[0011] .

[0012] in, Represents the first matrix, This represents the corresponding beam in the 4×4 beams, where n=1,2,…16.

[0013] Based on the first matrix, the beam space matrix is ​​obtained; the beam space matrix is ​​represented as:

[0014] ;

[0015] in, Represents the beam space matrix. for The conjugate transpose of .

[0016] Furthermore, based on at least one sound wave, the plane to be detected is divided into 41×41 grid cells. Based on the sound wave corresponding to each of the 41×41 grid cells, a weight matrix corresponding to the plane to be detected is determined. In the weight matrix, one column of symbols corresponds to one grid cell. The weight matrix is ​​transformed to a preset beam space to obtain a transformation matrix. Singular value decomposition and compression are performed on the transformation matrix to obtain the final transformation matrix. The transformation matrix is ​​expressed as:

[0017] ;

[0018] in, Represents the weight matrix. Represents the transformation matrix. for The conjugate transpose of; Represents the intermediate matrix; This represents the transformation matrix.

[0019] Furthermore, for any column vector included in the intermediate matrix, a normalization operation is performed on that column vector to obtain a normalized vector. Based on each normalized vector, a kd-tree is obtained.

[0020] Furthermore, the kd-tree includes an initial split axis, a left branch, and a right branch. The left branch includes at least one left sub-branch, and the right branch includes at least one right sub-branch. For any left sub-branch in the at least one left sub-branch, each left sub-branch includes at most one node. For any right sub-branch in the at least one right sub-branch, each right sub-branch includes at most one node. A node is used to store the variance of a normalized vector in the target dimension. The initial split axis is determined based on the target dimension. The target dimension is determined based on the variance of each dimension of the normalized vectors.

[0021] Furthermore, obtain the variance of each dimension of each normalized vector. The dimension corresponding to the maximum value of the variance of each dimension of each normalized vector is determined as the target dimension.

[0022] Further, based on the beam space matrix, kd-tree, and transformation matrix corresponding to the plane to be detected, the input signal is obtained. Standardization processing is performed on the input signal to obtain a standardized signal. Standardization processing includes at least one of denoising, trend removal, and fast Fourier transform. The standardized signal is then converted to a preset beam space to obtain the beam signal. The beam signal is represented as:

[0023] ;

[0024] in, Represents a standardized signal. Indicates beam signal;

[0025] calculate The first calculation result is obtained. Among them, express The conjugate matrix of the first calculation is normalized to obtain the second calculation result. Based on the kd-tree, a search is performed to find the conjugate matrix of the second calculation result. The value of q is the smallest. Let Z represent the normalized vector, Z represent the second calculation result, and q represent the column number of the column vector corresponding to the normalized vector in the intermediate matrix. This will make the equation... The location of the leak is determined by the position of the grid cell corresponding to the smallest value of q on the plane to be detected.

[0026] Furthermore, based on the second calculation result, a search is performed on the kd-tree to obtain the formula. The smallest value of q.

[0027] The beneficial effects of this invention are: compared with the traditional method of determining the location of pipeline leaks based on sound source localization, the pipeline leak detection method proposed in this invention is divided into offline and online parts, and most of the calculations can be completed in the offline part. Therefore, the calculation amount in the online part is small, which can complete the upstream task well. Combined with beamforming algorithms, leak detection methods, etc., it can realize real-time pipeline leak detection and equipment fault detection.

[0028] The offline part of the pipeline leakage detection method proposed in this invention not only calculates the beamforming matrix, but also calculates the beam space matrix to reduce the data dimensionality and compresses the beamforming matrix using the singular value decomposition method, which enables the subsequent online calculation part to be further accelerated.

[0029] The pipeline leakage detection method proposed in this invention uses the nearest neighbor search method of the kd-tree obtained from the offline calculation part to replace the traditional search method in the online calculation part, which further improves the speed of online calculation. Attached Figure Description

[0030] Figure 1 A schematic flowchart of a pipeline leakage detection method provided by the present invention;

[0031] Figure 2 A schematic diagram of the acquisition plane range provided by the present invention;

[0032] Figure 3 A flowchart illustrating an offline calculation process provided by the present invention;

[0033] Figure 4 This is a flowchart illustrating an online processing procedure provided by the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. In the description of this application, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in this application is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of this application, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Furthermore, to facilitate a clear description of the technical solutions in the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" are not necessarily different. Meanwhile, in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate that something is being used as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes.

[0035] Pipeline transportation of gases such as natural gas has unique advantages, and it has become the fifth largest mode of transportation after rail, road, waterway, and air transport. However, with the increase in pipeline usage time and corrosion from various media, pipelines may age and break, leading to leaks during transportation, especially when transporting gases, posing safety risks.

[0036] Pipeline inspection is an economical and effective method for protecting pipeline safety. The purpose of pipeline inspection technology is to detect the operational status of pipelines in a timely manner. Existing methods for detecting pipeline leaks can be divided into two categories based on their principles: one is to detect gases (such as methane) inside the pipeline, and the other is to inspect the pipeline itself.

[0037] Methods for detecting gas inside pipelines include fiber optic absorption, optical interferometry, and thermal conductivity measurement. These methods can directly detect the presence of natural gas in the air using optical and thermal means, thereby determining whether a pipeline leak has occurred. However, these methods are easily affected by environmental factors, increasing the false alarm rate. Furthermore, these methods can only detect whether a leak has occurred, but cannot pinpoint the exact location of the leak within the pipeline.

[0038] Methods for inspecting pipelines include acoustic detection and intelligent in-pipe detection. Intelligent in-pipe detection uses an intelligent crawler system equipped with various sensors to detect pressure, flow rate, temperature, and the integrity of the pipe wall, and determines whether a leak has occurred based on the detected data. However, intelligent in-pipe detection is only suitable for pipelines with few bends and connections, making it difficult to use in complex pipeline networks. Furthermore, operating the intelligent crawler system requires extensive experience, increasing labor costs. Acoustic detection involves installing acoustic sensors at various points along the pipeline to detect leaks. However, installing acoustic sensors at various points increases detection and labor costs. Moreover, even after determining that a leak has occurred, this method still struggles to pinpoint the exact location of the leak.

[0039] In summary, the main challenge of traditional gas pipeline leak detection methods lies in their difficulty in detecting leaks and accurately locating them within complex pipeline networks. Based on the characteristic that leak locations emit specific sound waves, the use of increasingly sophisticated microphone arrays combined with array signal processing algorithms can accurately locate the sound source and determine whether a leak has occurred based on the received signal. This approach has a lower false alarm rate and can accurately pinpoint leak locations even in complex pipeline networks. The key to leak detection using this method lies in the stability and reliability of the sound source localization algorithm. However, traditional sound source localization algorithms, while ensuring reliability, increase algorithm complexity, leading to a longer overall system response time. To achieve faster system response and real-time monitoring, it is necessary to improve traditional sound source localization algorithms to reduce computational load while maintaining accurate and reliable localization.

[0040] Therefore, this invention provides a method for detecting pipeline leaks, which involves setting up at least one fixed-position sound acquisition array (also referred to as an acoustic array in this embodiment) around the pipeline. The acoustic array is used to detect whether a natural gas leak has occurred in the pipeline within a fixed area that can be acquired. When a natural gas leak occurs in a pipeline, a stable sound wave with a specific amplitude spectrum is radiated from the leak location. After being received by the acoustic array, this sound wave can be used to locate the specific direction of the leak source using a sound source localization method. This achieves effective and timely detection of pipeline leaks and improves the accuracy of leak location detection.

[0041] Furthermore, the sound source localization method is divided into an offline part and an online part. The offline part calculates fixed intermediate results that can be stored based on the relative positional relationship between the acoustic array and the monitored target, including the beam space preprocessing matrix and some eigenvalue decomposition results. The online part performs partial calculations on the acoustic signal received by the array and the intermediate results in the offline part to obtain the localization result, without having to perform the calculations in the offline part again, thereby reducing the amount of computation.

[0042] See Figure 1 The present invention provides a method for detecting pipeline leakage, comprising the following steps S110-S140:

[0043] S110: Identify the pipeline to be inspected.

[0044] Among them, at least one set of sound acquisition arrays is set around the pipe to be tested, and different sound acquisition arrays correspond to different acquisition planes on the pipe to be tested.

[0045] In some embodiments, the sound acquisition array is a multi-arm helical acoustic array including a preset number of microphones. The distance between the sound acquisition array and the pipe to be detected is a first distance, and the distance between the sound acquisition array and the ground is a second distance. The horizontal and vertical field of view of the sound acquisition array relative to the pipe to be detected are both preset field of view angles.

[0046] It should be noted that those skilled in the art can set preset quantities, first distances, second distances, and preset field of view angles based on actual needs, and the embodiments of this application are not limited thereto.

[0047] For example, the sound acquisition array can contain 64 digital MEMS microphones. The structure of the sound acquisition array is as follows: eight microphones are distributed on an Archimedean spiral. Rotating the Archimedean spiral 45° clockwise eight times yields eight Archimedean spiral arms, each with eight microphones. The sound acquisition array is 10m away from the pipe and installed 1m above the ground. The horizontal and vertical field of view of the sound acquisition array are both 60°, and the array faces the center of the acquisition plane. Taking the plane of the sound acquisition array as the xOy plane and the vertical line through the center of the array as the z-axis, the beam direction u reaching the array when a sound source is present at any point on the acquisition plane corresponding to the pipe can be calculated. Furthermore, the range of the acquisition plane can be determined from the field of view of the sound acquisition array. Figure 2 As shown.

[0048] S120: Acquire at least one sound wave within the plane to be detected.

[0049] The plane to be detected is the acquisition plane corresponding to any one of the at least one set of sound acquisition arrays, and at least one beam of sound wave is acquired based on any one of the sound acquisition arrays.

[0050] S130: Based on at least one sound wave, perform offline calculations on the plane to be detected to obtain the beam space matrix, kd tree and transformation matrix corresponding to the plane to be detected.

[0051] The beam space matrix is ​​determined based on a first number of sound waves from at least one sound wave. The kd-tree and transformation matrix are determined based on the sound waves corresponding to each grid cell in a second number of grid cells. The second number of grid cells are obtained by uniformly dividing the plane to be detected. One grid cell corresponds to one sound wave, and different grid cells correspond to different sound waves.

[0052] The following is combined with Figure 3 The offline calculation process provided by this invention will be described in detail.

[0053] In some embodiments, a first type of beam and a second type of beam are determined from at least one sound wave beam. The sidelobes of the first type of beam are lower than those of the second type of beam. Four × 4 beams are selected from the first type of beam. Based on the selected 4 × 4 beams, a beam space matrix corresponding to the plane to be detected is obtained.

[0054] In some embodiments, a first matrix is ​​constructed based on the selected 4×4 beams. Based on the first matrix, a beam space matrix can be obtained.

[0055] The first matrix can be represented as:

[0056] ;

[0057] in, Describes the first matrix. This represents the corresponding beam in a 4×4 beam, where n=1,2,…16.

[0058] To satisfy the orthogonality of each beamspace matrix, the final beamspace matrix can be expressed as:

[0059] ;

[0060] in, Represents the beam space matrix; for The conjugate transpose of .

[0061] In some embodiments, the plane to be detected can be divided into 41×41 grid cells as acquisition locations based on at least one sound wave. Each grid cell corresponds to one sound wave, and the weight matrix corresponding to the plane to be detected can be determined based on the sound waves corresponding to each grid cell in the 41×41 grid cells.

[0062] In this system, each column of symbols in the weight matrix corresponds to a grid cell. Transforming the weight matrix to a preset beamspace yields the transformation matrix. Performing singular value decomposition and compression on the transformation matrix yields the final transformation matrix.

[0063] The transformation matrix can be represented as:

[0064] ;

[0065] in, Represents the weight matrix. Represents the transformation matrix. for The conjugate transpose of; Represents the intermediate matrix; This represents the transformation matrix.

[0066] In some embodiments, for any column vector included in the intermediate matrix, a normalization operation is performed on that column vector to obtain a normalized vector. Based on each normalized vector, a kd-tree can be obtained.

[0067] In some embodiments, the kd-tree includes an initial split axis, a left branch, and a right branch. The left branch includes at least one left sub-branch, and the right branch includes at least one right sub-branch. For any left sub-branch of the at least one left sub-branch, each left sub-branch includes at most one node. For any right sub-branch of the at least one right sub-branch, each right sub-branch includes at most one node. A node is used to store the variance of a normalized vector in the target dimension. The initial split axis is determined based on the target dimension. The target dimension is determined based on the variance of each dimension of the normalized vectors.

[0068] In some embodiments, the variance of each dimension of each normalized vector can be obtained. The dimension corresponding to the maximum value of the variance of each dimension of each normalized vector is determined as the target dimension.

[0069] Below is used Let represent the normalized vector, and q represent the column number of the column vector corresponding to the normalized vector. Describe in detail the construction process of the kd-tree T.

[0070] S131: Calculate all The variance of each dimension of the data is used to determine the target dimension r with the largest variance, and the target dimension r is used as the initial splitting axis.

[0071] S132: Retrieve all The variance in the r-dimensional dimension is selected by choosing the median of each variance. As a node.

[0072] S133: Divide into left and right branches, where the upper variance in the r-dimensional dimension is less than the median. As the left branch, it is not less than the median. As the right branch;

[0073] S134: Repeat steps S131-S133 for the left and right branches until all branches have no more than one node, thus obtaining the kd-tree T.

[0074] S140: Based on the beam space matrix, kd tree and transformation matrix corresponding to the plane to be detected, the leakage location on the plane to be detected is determined through online processing.

[0075] The following is combined with Figure 4 The online processing procedure provided by this invention will be described in detail below.

[0076] In some embodiments, the input signal can be obtained based on the beam space matrix, kd-tree, and transformation matrix corresponding to the plane to be detected. Standardization processing is then performed on the input signal to obtain a standardized signal.

[0077] The standardization process includes at least one of the following: noise reduction, trend removal, and fast Fourier transform.

[0078] By converting the standardized signal to a preset beam space, a beam signal can be obtained. The beam signal can be represented as:

[0079] ;

[0080] in, This represents the standardized signal. This refers to the beam signal.

[0081] Obtain beam signal After that, calculations can be performed. The first calculation result is obtained. Among them, express The conjugate matrix. Normalizing the first calculation result yields the second calculation result Z. Based on a kd-tree, a search can be performed to find the expression... The value of q is the smallest. Let represent the normalized vector, and q represent the column number of the column vector corresponding to the normalized vector in the intermediate matrix. This can be used to make the equation... The location of the leak is determined by the position of the grid cell corresponding to the smallest value of q on the plane to be detected.

[0082] In some embodiments, the kd-tree can be searched based on the second calculation result to obtain the expression The smallest value of q.

[0083] In some embodiments, the kd-tree can be searched based on the following steps S141-S144.

[0084] S141: Search based on the current node's dividing axis. If Z is less than the node in the r-dimensional dividing axis, search the left branch; otherwise, search the right branch.

[0085] S142: Repeat S141 to search for the lowest level node, mark it, calculate the distance and record it. If the current distance is less than the recorded distance, update the current nearest node and distance.

[0086] S143: Backtrack upwards. If backtracking reaches a marked point, continue backtracking to an unmarked point, then mark it, calculate the distance and compare it. At the same time, calculate the distance between Z and the dividing face corresponding to the node. If it is less than the current minimum distance between Z and the node, then another branch needs to be searched, and return to step S142. Otherwise, repeat S143 until backtracking is no longer possible.

[0087] S144: The nearest node to the final record is the search result.

[0088] As can be seen, compared with the traditional method of determining the location of pipeline leaks based on sound source localization, the pipeline leak detection method proposed in this invention is divided into offline and online parts, and most of the calculations can be completed in the offline part. Therefore, the computational load of the online part is small, which can complete the upstream task well. Combined with beamforming algorithms, leak detection methods, etc., it can realize real-time pipeline leak detection and equipment fault detection.

[0089] The offline part of the pipeline leakage detection method proposed in this invention not only calculates the beamforming matrix, but also calculates the beam space matrix to reduce the data dimensionality and compresses the beamforming matrix using the singular value decomposition method, which enables the subsequent online calculation part to be further accelerated.

[0090] The pipeline leakage detection method proposed in this invention uses the nearest neighbor search method of the kd-tree obtained from the offline calculation part to replace the traditional search method in the online calculation part, which further improves the speed of online calculation.

[0091] In some solutions, multiple embodiments of this application can be combined, and the combined solution can be implemented. Optionally, some operations in the processes of each method embodiment may be combined, and / or the order of some operations may be changed. Furthermore, the execution order between the steps of each process is merely exemplary and does not constitute a limitation on the execution order between steps; other execution orders are also possible. It is not intended to indicate that the execution order is the only possible order in which these operations can be performed. Those skilled in the art will conceive of various ways to reorder the operations described herein. In addition, it should be noted that the process details involved in one embodiment of this document are similarly applicable to other embodiments, or different embodiments may be combined.

[0092] Furthermore, some steps in the method embodiments can be equivalently replaced with other possible steps. Alternatively, some steps in the method embodiments may be optional and can be deleted in certain use cases. Or, other possible steps may be added to the method embodiments. Moreover, the various method embodiments can be implemented individually or in combination.

[0093] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above.

[0094] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.

[0095] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0096] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to it, or all or part of the technical solution, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0097] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for detecting pipeline leaks, characterized in that, include: The pipe to be inspected is identified; at least one set of sound acquisition arrays is arranged around the pipe to be inspected; different sound acquisition arrays correspond to different acquisition planes on the pipe to be inspected. At least one sound wave is acquired within the plane to be detected; the plane to be detected is the acquisition plane corresponding to any one of the at least one set of sound acquisition arrays. The at least one sound wave is acquired based on any of the sound acquisition arrays; Based on the at least one sound wave, offline calculations are performed on the plane to be detected to obtain the beam space matrix, kd-tree, and transformation matrix corresponding to the plane to be detected. The beam space matrix is ​​determined based on a first number of sound waves from the at least one sound wave. The kd-tree and the transformation matrix are determined based on the sound waves corresponding to each grid cell in a second number of grid cells. The second number of grid cells are obtained by uniformly dividing the plane to be detected. Each grid cell corresponds to one sound wave, and different grid cells correspond to different sound waves. Based on online processing of the beam space matrix, kd tree and transformation matrix corresponding to the plane to be detected, the leakage location on the plane to be detected is determined. The step of performing offline calculations on the plane to be detected based on the at least one sound wave to obtain the beam space matrix, kd-tree, and transformation matrix corresponding to the plane to be detected includes: In the at least one sound wave, a first type of beam and a second type of beam are identified; the sidelobes of the first type of beam are lower than the sidelobes of the second type of beam. Select 4×4 beams from the first type of beam; Based on the selected 4×4 beams, the beam space matrix corresponding to the plane to be detected is obtained; The step of obtaining the beam space matrix corresponding to the plane to be detected based on the selected 4×4 beams includes: Based on the selected 4×4 beams, a first matrix is ​​constructed; the first matrix is ​​represented as: ； in, Represents the first matrix, This represents the corresponding beam in the 4×4 beams, where n=1,2,…16; Based on the first matrix, the beam space matrix is ​​obtained; the beam space matrix is ​​represented as: ； in, Represents the beam space matrix; for The conjugate transpose of; The step of performing offline calculations on the plane to be detected based on the at least one sound wave to obtain the beam space matrix, kd-tree, and transformation matrix corresponding to the plane to be detected includes: Based on the at least one sound wave, the plane to be detected is divided into 41×41 grid cells; Based on the acoustic waves corresponding to each of the 41×41 grid cells, a weight matrix corresponding to the plane to be detected is determined; one column of symbols in the weight matrix corresponds to one of the grid cells. The weight matrix is ​​transformed to a preset beam space to obtain the transformation matrix; The transformation matrix is ​​obtained by performing singular value decomposition and compression on the transformation matrix; the transformation matrix is ​​expressed as: ； in, Represents the weight matrix, Represents the transformation matrix, for The conjugate transpose of; Represents the intermediate matrix; Represents the transformation matrix; The step of performing offline calculations on the plane to be detected based on the at least one sound wave to obtain the beam space matrix, kd-tree, and transformation matrix corresponding to the plane to be detected includes: For any column vector included in the intermediate matrix, a normalization operation is performed on the column vector to obtain a normalized vector; The kd-tree is obtained based on the normalized vectors described above.

2. The method according to claim 1, characterized in that, The sound acquisition array is a multi-arm helical sound array including a preset number of microphones; the distance between the sound acquisition array and the pipe to be tested is a first distance; the distance between the sound acquisition array and the ground is a second distance; the horizontal and vertical field of view of the sound acquisition array relative to the pipe to be tested are both preset field of view angles.

3. The method according to claim 1, characterized in that, The kd-tree includes an initial split axis, a left branch, and a right branch; the left branch includes at least one left sub-branch, and the right branch includes at least one right sub-branch. For any left sub-branch of the at least one left sub-branch, the left sub-branch includes at most one node; for any right sub-branch of the at least one right sub-branch, the right sub-branch includes at most one node; one of the nodes is used to store the variance of a normalized vector in the target dimension; the initial segmentation axis is determined based on the target dimension; the target dimension is determined based on the variance of each dimension of each of the normalized vectors.

4. The method according to claim 3, characterized in that, Also includes: Obtain the variance of each dimension of each of the normalized vectors; The dimension corresponding to the maximum value of the variance of each dimension of the normalized vector is determined as the target dimension.

5. The method according to claim 4, characterized in that, The method of determining the leakage location on the plane to be detected by online processing of the beam space matrix, kd-tree, and transformation matrix corresponding to the plane to be detected includes: The input signal is obtained based on the beam space matrix, kd tree, and transformation matrix corresponding to the plane to be detected; The input signal is subjected to standardization processing to obtain a standardized signal; the standardization processing includes at least one of denoising, trend removal, and fast Fourier transform; The standardized signal is converted to the preset beam space to obtain the beam signal; the beam signal is represented as: ； in, This represents the standardized signal. This represents the beam signal; calculate The first calculation result is obtained; among them, express The conjugate matrix; The first calculation result is normalized to obtain the second calculation result; Based on the kd-tree, the search expression is... The value of q is the smallest; where, Z represents the normalized vector; Z represents the second calculation result; q represents the column number of the column vector corresponding to the normalized vector in the intermediate matrix; The formula will be used The location of the leak is determined by the position of the grid cell corresponding to the smallest value of q on the plane to be detected.

6. The method according to claim 5, characterized in that, Based on the kd-tree, the search expression The smallest values ​​of q include: Based on the second calculation result, the kd-tree is searched to obtain the formula. The smallest value of q.

Images