GIS equipment defect three-dimensional visual positioning method based on acoustic imaging

Abstract

The invention discloses a GIS equipment defect three-dimensional visual positioning method based on acoustic imaging, and relates to the related technical field of GIS equipment, and the method comprises the steps: constructing a sound intensity detection dual system; sound source signals of the GIS equipment are synchronously acquired, and near-field acoustic signals and far-field acoustic signals are output; performing frequency band separation through a self-adaptive band-pass filter, and identifying near acoustic abnormal features and far acoustic abnormal features of the separated frequency band; carrying out spatial positioning fusion and outputting a positioning point; and importing a GIS equipment three-dimensional model according to the sound source fusion positioning point, and outputting a GIS equipment visual positioning result. The technical problems that in the prior art, the GIS equipment defect detection range is limited, the positioning precision is insufficient, the defect space position is difficult to visually present, the equipment defect positioning is not rapid and accurate enough, and the equipment operation and maintenance efficiency and reliability are affected are solved. The technical effects of quickly and accurately positioning the equipment defects and effectively improving the operation and maintenance efficiency and reliability of the equipment are achieved.

Description

Technical Field

[0001] The present application relates to the technical field related to GIS equipment, and in particular to a three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging. Background Art

[0002] During long-term operation, GIS equipment (gas-insulated switchgear) may develop internal defects such as partial discharge, abnormal mechanical vibration, and gas leakage. If these defects are not discovered and addressed in a timely manner, they will affect the safe and stable operation of the power grid. Existing GIS equipment defect detection methods, such as partial discharge detection and infrared thermal imaging detection, have limited detection range, insufficient positioning accuracy, and difficulty in visually presenting the spatial location of defects. For example, partial discharge detection focuses on capturing electrical signals and has difficulty accurately locating non-electrical mechanical faults. Infrared thermal imaging detection is significantly affected by factors such as ambient temperature and equipment surface condition, making it impossible to perform efficient and accurate GIS equipment status monitoring and fault diagnosis, affecting equipment operation and maintenance efficiency and reliability.

[0003] At present, relevant technologies have technical problems such as limited detection range of GIS equipment defects, insufficient positioning accuracy, and difficulty in intuitively presenting the spatial location of defects, resulting in insufficient speed and accuracy in equipment defect positioning, affecting the efficiency and reliability of equipment operation and maintenance. Summary of the Invention

[0004] This application provides a three-dimensional visual positioning method for GIS equipment defects based on acoustic imaging, which solves the technical problems in the existing technology such as limited GIS equipment defect detection range, insufficient positioning accuracy, and difficulty in intuitively presenting the spatial location of defects, resulting in insufficiently fast and accurate positioning of equipment defects, and affecting the efficiency and reliability of equipment operation and maintenance. It achieves the technical effect of realizing rapid and accurate positioning of equipment defects and effectively improving the efficiency and reliability of equipment operation and maintenance.

[0005] The present application provides a three-dimensional visual positioning method for GIS equipment defects based on acoustic imaging, which includes: constructing a dual sound intensity detection system, wherein the dual sound intensity detection system includes a near-field sound intensity detection system and a far-field sound intensity detection system; starting the dual sound intensity detection system to synchronously collect sound source signals of the GIS equipment, and output near-field acoustic signals and far-field acoustic signals; performing frequency band separation on the near-field acoustic signals and the far-field acoustic signals through an adaptive band-pass filter, and identifying near-acoustic anomaly features and far-acoustic anomaly features of the separated frequency bands; performing spatial positioning fusion according to the near-acoustic anomaly features and the far-acoustic anomaly features, and outputting a sound source fusion positioning point; importing a three-dimensional model of the GIS equipment according to the sound source fusion positioning point, and outputting a visual positioning result of the GIS equipment.

[0006] In a possible implementation, the three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging also performs the following processing: performing short-time Fourier transform and continuous wavelet transform on the near-field acoustic signal and the far-field acoustic signal, respectively, and outputting a near-field-time-frequency feature group and a far-field-time-frequency feature group, wherein each time-frequency feature group includes a time-frequency matrix and a spectrum matrix; the adaptive bandpass filter determines adaptive filtering parameters according to the near-field-time-frequency feature group and the far-field-time-frequency feature group, and the adaptive filtering parameters include a center frequency and a bandwidth; and the near-field acoustic signal and the far-field acoustic signal are separated by the adaptive filtering parameters to obtain a near-field main acoustic frequency band and a far-field main acoustic frequency band.

[0007] In a possible implementation, the three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging also performs the following processing: identifying the key feature vectors corresponding to the near-field main acoustic frequency band and the far-field main acoustic frequency band, including sound intensity, frequency peak, harmonics and time-frequency changes; using the pre-built near-field-abnormal acoustic feature vector library and far-field-abnormal acoustic feature vector library to perform anomaly detection on the key feature vectors, and output near-field acoustic anomaly features and far-field acoustic anomaly features.

[0008] In a possible implementation, the three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging also performs the following processing: performing conflict segment detection based on the near acoustic anomaly features and the far acoustic anomaly features, and outputting conflict segments and non-conflict segments, wherein the conflict segments satisfy the frequency range intersection and the distance between the sound source positioning coordinates is less than a preset threshold; constructing a collaborative fusion positioning channel to perform spatial positioning fusion inversion on the conflict segments, and outputting a first type of sound source fusion positioning point; constructing an independent fusion positioning channel to perform spatial positioning fusion on the non-conflict segments, and outputting a second type of sound source fusion positioning point; fusing the first type of sound source fusion positioning point and the second type of sound source fusion positioning point, and outputting a sound source fusion positioning point.

[0009] In a possible implementation, the three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging also performs the following processing: the collaborative fusion positioning channel includes a collaborative mapping relationship between near and far features; the near acoustic abnormality features and the far acoustic abnormality features of the conflict section are frequency cross-compared to extract the acoustic distribution differences of the conflict section, wherein the acoustic distribution differences include the sound source reception time difference, the frequency main peak offset and the time-frequency feature difference; the acoustic distribution differences are jointly inverted and analyzed using the collaborative mapping relationship between near and far features, the spatial intersection points corresponding to the near and far sound sources in the conflict section are calculated, and the first type of sound source fusion positioning points are output.

[0010] In a possible implementation, the three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging also performs the following processing: the independent fusion positioning channel includes a confidence evaluation model; the near acoustic anomaly features and the far acoustic anomaly features of the non-conflicting section are spatially positioned and analyzed respectively to construct a near-field sound source positioning initial value set and a far-field sound source positioning initial value set; the confidence evaluation model is used to perform confidence evaluation on the near-field sound source positioning initial value set and the far-field sound source positioning initial value set, and a confidence evaluation index set is output; the optimal solution is selected based on the confidence evaluation index set, and a second-category sound source fusion positioning point is output.

[0011] In a possible implementation, the three-dimensional visual positioning method for GIS equipment defects based on acoustic imaging further performs the following processing: performing time-frequency consistency verification on the second-category sound source fusion positioning points, and obtaining time-frequency consistency verification results, including verification pass and verification fail; screening the second-category sound source fusion positioning points according to the time-frequency consistency verification results, and outputting updated second-category sound source fusion positioning.

[0012] In a possible implementation, the three-dimensional visual positioning method for GIS equipment defects based on acoustic imaging also performs the following processing: constructing a three-dimensional coordinate system of the sound source space; converting the three-dimensional model of the GIS equipment according to the three-dimensional coordinate system of the sound source space, and outputting the converted three-dimensional model of the GIS equipment; importing the sound source fusion positioning point into the converted three-dimensional model of the GIS equipment for visual positioning, and outputting the visual positioning result of the GIS equipment.

[0013] The proposed three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging will be used to construct a dual sound intensity detection system. The system will synchronously collect sound source signals from GIS equipment and output near-field and far-field acoustic signals. Adaptive bandpass filters will be used to separate frequency bands and identify near- and far-field acoustic anomaly characteristics in the separated frequency bands. Spatial positioning fusion will be performed to output sound source fusion positioning points. The three-dimensional model of the GIS equipment will be imported according to the sound source fusion positioning points to output the visualization positioning results of the GIS equipment. This solves the technical problems in the prior art of limited GIS equipment defect detection range, insufficient positioning accuracy, and difficulty in visually presenting the spatial location of defects, resulting in insufficiently fast and accurate positioning of equipment defects and impacting equipment operation and maintenance efficiency and reliability. The system achieves the technical effect of rapidly and accurately positioning equipment defects and effectively improving equipment operation and maintenance efficiency and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention are briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in precise order. Instead, various steps may be processed in reverse order or simultaneously as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0015] Figure 1 A schematic diagram of a process for a three-dimensional visual location method for GIS equipment defects based on acoustic imaging provided in an embodiment of the present application; Figure 2 A schematic diagram of the flow of frequency band separation in the three-dimensional visual positioning method for GIS equipment defects based on acoustic imaging provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0017] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0018] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms “first\second” involved are merely to distinguish similar objects and do not represent a specific ordering of the objects. The terms “including” and “having” and any variations are intended to cover non-exclusive inclusions. For example, a process, method, product, or server comprising a series of steps is not necessarily limited to those steps clearly listed, but may include other steps that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used herein are for the purpose of describing the embodiments of this application only.

[0019] The present application provides a method for three-dimensional visualization of GIS equipment defects based on acoustic imaging. Figure 1 As shown, the method includes: Step S100 : constructing a sound intensity detection dual system, wherein the sound intensity detection dual system includes a near-field sound intensity detection system and a far-field sound intensity detection system.

[0020] Preferably, a dual sound intensity detection system with synchronous acquisition and complementary positioning is constructed, including a near-field sound intensity detection system and a far-field sound intensity detection system. The near-field sound intensity detection system is responsible for the fine positioning and feature identification of defects (such as determining the specific location and frequency characteristics of the discharge point), and the far-field system is responsible for the global screening and trend analysis of defects (such as determining the equipment interval or area where the defect is located). Specifically, the near-field sound intensity detection system refers to a subsystem that collects acoustic signals at close range (close to the surface of the GIS equipment or a local area), usually consisting of a high-density microphone array, a high-precision synchronous acquisition module, a near-field acoustic positioning algorithm module, etc. The high-density microphone array uses high-sensitivity microphones and is deployed in a dense array (such as a planar array, a circular array, etc.) on or near the surface of the GIS equipment to capture subtle changes in the close-range sound field; the synchronous acquisition module ensures the time synchronization accuracy of each microphone signal to avoid positioning deviation due to delay error; the near-field acoustic positioning algorithm module is based on near-field acoustic theory (such as beamforming algorithm, acoustic holography technology, etc.) to perform high-precision positioning of close-range sound sources. Close-range acquisition can capture high-frequency detail signals (such as ultrasonic waves generated by partial discharge and high-frequency noise from mechanical vibration), with positioning accuracy reaching centimeters or even millimeters. It is suitable for detecting tiny defects on the surface or inside of equipment at close range, and for performing detailed scanning of suspected defective areas to determine the specific location and type of defects, such as detecting insulator cracks and partial discharge caused by poor contact.

[0021] A far-field sound intensity detection system is a subsystem that collects acoustic signals at long distances (several to tens of meters). It typically consists of a sparse microphone array, a wide-angle acoustic sensor, and a far-field sound source localization algorithm module. The sparse microphone array uses low-sensitivity microphones deployed at large intervals around the GIS equipment to capture sound waves propagating over long distances. The wide-angle sensor supports broadband signal acquisition (such as audible sound and infrasound), covering the entire sound field radiated by the equipment. The far-field sound source localization algorithm module uses far-field acoustic theory (such as sonar time difference localization) to globally locate distant sound sources. It can capture low-frequency signals radiated by the entire GIS equipment (such as audible sound from gas leaks and overall equipment vibration noise), with a coverage range of up to tens of meters. It is suitable for quickly scanning large areas, screening for potential defects, and determining the approximate location of defects, such as near a certain interval or busbar section.

[0022] Step S200: Start the dual sound intensity detection system to synchronously collect sound source signals from the GIS equipment and output near-field acoustic signals and far-field acoustic signals.

[0023] Preferably, combined with the online monitoring data of GIS equipment (such as partial discharge warning signals, vibration sensor threshold alarms, etc.), when an abnormality is detected in the GIS equipment, the sound intensity detection is automatically started. The dual system synchronously collects the sound source signal to ensure that the near-field and far-field signals are strictly aligned in the time dimension, and realize real-time response to defects. Specifically, the high-precision microphone array is usually deployed on the surface of the GIS equipment shell (within 0.1~1 meters of the equipment) by attachment, magnetic attraction or bracket fixing, covering key parts prone to defects (such as casing connections, circuit breaker units, disconnector interfaces, etc.), and the acquisition parameters of the high-precision microphone array are configured, such as sampling rate ≥100 kHz (to meet the needs of ultrasonic signal sampling), the number of channels ≥ 32 channels (high-density array to ensure spatial resolution) and acquisition time requirements (single acquisition lasts 1 to 5 seconds to capture transient signals); then the microphone array is driven by a robotic arm or mobile platform to scan each area to ensure that there are no blind spots in the detection, and then the near-field acoustic signal is output, which mainly includes ultrasonic waves (above 20kHz) and high-frequency audible sounds (20Hz~20kHz), corresponding to the sound waves generated by fast transient defects such as partial discharge, insulator breakdown, and collision of mechanical parts. Among them, there are obvious sound pressure level gradients and phase differences in the close-range sound field, and the array signal processing algorithm can be used to achieve high-precision positioning of defects.

[0024] Preferably, a sparse microphone array is deployed around the interval where the GIS equipment is located (5 to 50 meters away from the equipment), forming a circular or fan-shaped monitoring area, covering the acoustic radiation range of the entire equipment group, and configuring the acquisition parameters of the high-precision microphone array, such as a sampling rate ≥ 48 kHz (to meet the sampling requirements of audible sound signals), a number of channels ≥ 8 channels (the sparse array achieves wide-angle coverage), and an acquisition time (lasting several minutes to several hours). In addition, several monitoring points are selected in the substation (such as above the equipment, on both sides of the channel, etc.), and fixed sensors are used to monitor the background noise and the overall acoustic characteristics of the equipment for a long time, thereby forming a far-field acoustic signal, which mainly includes audible sound (20 Hz to 20 kHz) and infrasound (<20 Hz), corresponding to turbulent noise generated by gas leakage, overall mechanical vibration of the equipment, and low-frequency discharge caused by insulation aging. Among them, the long-distance sound field tends to be a plane wave with a small sound pressure level gradient, but the sound source direction can be estimated through time difference of air (TDOA) or phase difference of air (FDOA) analysis of multiple sensors.

[0025] Step S300 : performing frequency band separation on the near-field acoustic signal and the far-field acoustic signal through an adaptive bandpass filter, and identifying near-field acoustic abnormality features and far-field acoustic abnormality features in the separated frequency bands.

[0026] Further, such as Figure 2As shown, step S300 also includes step S310, performing short-time Fourier transform and continuous wavelet transform on the near-field acoustic signal and the far-field acoustic signal, respectively, and outputting a near-field-time-frequency feature group and a far-field-time-frequency feature group, wherein each time-frequency feature group includes a time-frequency matrix and a spectrum matrix; step S320, the adaptive bandpass filter determines adaptive filtering parameters according to the near-field-time-frequency feature group and the far-field-time-frequency feature group, and the adaptive filtering parameters include a center frequency and a bandwidth; step S330, separating the near-field acoustic signal and the far-field acoustic signal using the adaptive filtering parameters to obtain a near-field main acoustic frequency band and a far-field main acoustic frequency band.

[0027] Preferably, an adaptive bandpass filter is used to separate the frequency bands of near-field acoustic signals and far-field acoustic signals. Specifically, near-field acoustic signals are mostly transient high-frequency signals (such as pulse sounds of local discharges and burst noises of mechanical collisions), and their frequency components change rapidly with time. Then, a short-time Fourier transform is performed on the near-field acoustic signal, that is, the near-field signal is framed by a sliding window function, and each frame signal is Fourier transformed to generate a time-frequency matrix (the horizontal axis is time, the vertical axis is frequency, and the amplitude is color / grayscale value); at the same time, the maximum value of the spectrum of each frame signal is extracted to generate a spectrum matrix to reflect the changing trend of the energy distribution of each frequency band over time; finally, the near-field-time-frequency feature group is output according to the time-frequency matrix and the spectrum matrix.

[0028] Preferably, the far-field acoustic signals are mostly low-frequency slowly varying signals (such as turbulent noise from gas leakage and periodic noise from overall vibration of the equipment), and their frequency components have multi-scale characteristics. The far-field acoustic signals are then subjected to continuous wavelet transform, that is, wavelet basis functions of different scales are convolved with the far-field signals to generate a time-frequency matrix (energy distribution on the scale-time plane, where the scale is inversely proportional to the frequency); at the same time, the energy at each scale is counted to generate a spectrum matrix (reflecting the energy distribution in different frequency ranges); finally, the time-frequency matrix and the spectrum matrix are fused to output a far-field-time-frequency feature group.

[0029] Preferably, feature analysis and extraction are performed on the near-field acoustic signal and the far-field acoustic signal respectively. Specifically, abnormal high-frequency components (such as the ultrasonic characteristic frequency of partial discharge 20~100kHz) are extracted from the near-field-time-frequency feature group, and environmental noise (such as the low-frequency vibration noise of the fan) is excluded. The frequency interval corresponding to the energy peak in the spectrum matrix is ​​the characteristic frequency band of the near-field signal, such as a narrowband signal of 30kHz±5kHz; low-frequency abnormal components (such as 10~500Hz turbulent noise of gas leakage and 100Hz frequency doubling of mechanical resonance) are extracted from the far-field-time-frequency feature group. The energy area that persists in the spectrum matrix is ​​the characteristic frequency band of the far-field signal, such as a broadband signal of 200Hz±100Hz.

[0030] Preferably, the adaptive bandpass filter calculates and determines the adaptive filtering parameters, including the center frequency and bandwidth, based on the extracted features. For near-field acoustic signals, the geometric center of the characteristic frequency band (such as 30kHz) is taken as the center frequency, and the bandwidth is determined according to the frequency range of the characteristic frequency band (such as ±5kHz, bandwidth 10kHz); for far-field acoustic signals, the energy-weighted center of the characteristic frequency band (such as calculating the weighted average frequency by the energy proportion of each frequency component of the spectrum matrix) is taken as the center frequency, and the bandwidth is determined according to the wide-band characteristics (such as ±100kHz, bandwidth 200kHz); in addition, if there are multiple energy peaks in the time-frequency feature group (such as multi-source defects), the filter can be automatically split into multiple bandpass channels to track different characteristic frequency bands respectively.

[0031] Preferably, the near-field acoustic signal and the far-field acoustic signal are finally separated using adaptive filtering parameters determined according to the near-field acoustic signal and the far-field acoustic signal, that is, a band-pass filter with a center frequency and bandwidth corresponding to the near-field acoustic signal is used to filter out low-frequency environmental noise and high-frequency interference from the near-field original signal, and retain the 20-100kHz high-frequency signal directly related to local discharge and mechanical defects, that is, the near-field main acoustic frequency band; a band-pass filter with a center frequency and bandwidth corresponding to the far-field acoustic signal is used to filter out high-frequency transient noise and DC components from the far-field original signal, and retain the 10-500Hz low-frequency signal related to gas leakage and structural resonance, that is, the far-field main acoustic frequency band.

[0032] Furthermore, step S330 further includes step S331: Preferably, the key eigenvectors corresponding to the near-field main acoustic frequency band and the far-field main acoustic frequency band are identified, including sound intensity, frequency peak, harmonics and time-frequency changes, where sound intensity refers to the sound power per unit area, reflecting the energy of the sound source; the frequency peak refers to the frequency point or frequency band with the highest energy in the spectrum, corresponding to the characteristic frequency of the defect. For example, the frequency peak of corona discharge is mostly in the range of 300MHz-3GHz (ultra-high frequency band), and the frequency peak of mechanical vibration is related to the natural frequency of the component (such as the vibration of circuit breaker opening and closing is often in the range of 100Hz-1kHz); harmonics refer to frequency components at integer multiples of the fundamental frequency, reflecting nonlinear acoustic effects; time-frequency changes refer to the change pattern of characteristic parameters over time, which are used to distinguish steady-state from transient defects. For example, the time-frequency changes of partial discharge are manifested as intermittent pulses (time domain sparsity), and the time-frequency changes of gas leakage are manifested as continuous low-frequency noise (time domain continuity); the eigenvectors corresponding to the near-field acoustic signal and the far-field acoustic signal are shown in Table 1: Table 1 Differences between near-field eigenvectors and far-field eigenvectors

[0033] Preferably, in a laboratory environment, different types of defects (such as needle tip discharge, suspended potential discharge, mechanical loosening, etc.) are imposed on the GIS equipment, and near-field and far-field acoustic signals are collected synchronously. Key feature vectors are extracted from the signals of known defect types, and a feature-defect mapping relationship is constructed. For example, during needle tip discharge, the feature vector is manifested as a high-frequency pulse (frequency peak at 300MHz), the sound intensity increases exponentially with increasing voltage, and the harmonic content is low; during mechanical loosening, the feature vector is manifested as a low-frequency vibration (frequency peak at 200Hz), the time-frequency variation is periodic, and the harmonic distribution is odd-dominated; then, a machine learning algorithm (such as K-means, DBSCAN) is used to cluster the feature vectors to form feature clusters of different defect types, and the corresponding defect types are labeled, and finally a near-field-abnormal acoustic feature vector library and a far-field-abnormal acoustic feature vector library are constructed respectively.

[0034] Preferably, the near-field-abnormal acoustic feature vector library and the far-field-abnormal acoustic feature vector library are used to perform abnormality detection on the key feature vectors respectively, that is, the similarity (such as cosine similarity, Euclidean distance) between the feature vector extracted in real time and the feature vector pre-stored in the feature vector library is calculated. When the similarity exceeds a threshold, it is determined to be an abnormality of the corresponding type, and then the near-acoustic abnormality features and far-acoustic abnormality features of the separated frequency bands are output, including identifying frequency components or time domain pulse signals with energy significantly higher than the background noise from the near-field main acoustic frequency band, such as ultra-high frequency pulses generated by partial discharge (manifested as narrow pulses in the time domain and high-frequency energy peaks in the frequency domain), corona discharge noise (continuous high-frequency noise) before insulator breakdown; and non-periodic or energy mutation signal features from the far-field main acoustic frequency band, such as impact vibration caused by loose mechanical parts (manifested as sudden spikes in the time domain and broadband energy distribution in the frequency domain), and turbulent noise caused by gas leakage (continuous medium and low frequency noise, with energy concentrated at a specific frequency).

[0035] Step S400: performing spatial positioning fusion according to the near acoustic anomaly features and the far acoustic anomaly features, and outputting a sound source fusion positioning point.

[0036] Step S400 further includes step S410, performing conflict segment detection based on the near acoustic anomaly features and the far acoustic anomaly features, and outputting conflict segments and non-conflict segments, wherein the conflict segments satisfy the frequency range intersection and the distance between the sound source positioning coordinates is less than a preset threshold; step S420, constructing a collaborative fusion positioning channel to perform spatial positioning fusion inversion on the conflict segments, and outputting a first type of sound source fusion positioning point; step S430, constructing an independent fusion positioning channel to perform spatial positioning fusion on the non-conflict segments, and outputting a second type of sound source fusion positioning point; step S440, fusing the first type of sound source fusion positioning point and the second type of sound source fusion positioning point, and outputting a sound source fusion positioning point.

[0037] Preferably, conflict segment detection is performed based on the near-field acoustic abnormality features and the far-field acoustic abnormality features. Specifically, the intersection of the near-field characteristic frequency interval and the far-field characteristic frequency interval is calculated to obtain the frequency range intersection; the three-dimensional Euclidean distance between the near-field sound source positioning coordinates and the far-field sound source positioning coordinates is calculated. If the distance is less than a preset threshold (such as 1 meter), it is determined to be spatially adjacent; and then the conflict segment and the non-conflict segment are output, wherein the conflict segment refers to the area where the near-field and far-field detection results have an intersection, satisfying the frequency range intersection (the frequency interval of the near-field abnormality feature overlaps with the frequency interval of the far-field abnormality feature, such as a 200Hz~1kHz vibration signal detected in the near field and a 500Hz~2kHz noise detected in the far field) and positioning coordinate proximity (the spatial distance between the near-field positioning result and the far-field positioning result is less than a preset threshold); the non-conflict segment refers to the area where there is no frequency intersection or a relatively far spatial distance between the near-field and far-field detection results, indicating that the two may be independent defects or background interference.

[0038] Preferably, a collaborative fusion positioning channel is constructed to perform spatial positioning fusion inversion on the conflicting section. Specifically, when there is a conflict (correlation) between the near-field and far-field detection results, the positioning accuracy needs to be improved through information complementarity, that is, the high-frequency detail signals of the near field (such as the ultra-high frequency characteristics of local discharge) are used to achieve centimeter-level precise positioning, and the global sound field coverage of the far field (such as the low-frequency propagation characteristics of mechanical vibration) is used to verify the rationality of the positioning results; then the signals of the near-field high-density array and the far-field sparse array are jointly processed to construct a cross-scale beamforming matrix to achieve focused positioning of the conflicting section, and then the high-frequency update data of the near-field positioning and the low-frequency trend data of the far-field positioning are combined, and the positioning coordinates are iteratively corrected through Kalman filtering to suppress random noise interference, and then the first type of sound source fusion positioning point is output. The positioning accuracy after fusion can reach the centimeter level, and is accompanied by frequency cross-validation results.

[0039] Preferably, an independent fusion positioning channel is constructed to perform spatial positioning fusion on non-conflicting sections. Specifically, when the near-field and far-field detection results are non-conflicting (independent), the two types of signals need to be processed separately to avoid cross-source interference. The near-field signal is processed independently, that is, local defects are accurately positioned based on high-frequency features. The far-field signal is processed independently, and global defects are roughly positioned based on low-frequency features. Then, acoustic holography technology or near-field beamforming is used to perform near-field independent fusion, and only near-field microphone array data is used for positioning, and local positioning points with an output accuracy of centimeters are output. At the same time, time difference positioning or wave direction estimation is used, and only far-field microphone array data is used for positioning, and global positioning points with an output accuracy of meters are output. Finally, the second type of sound source fusion positioning points are output, including near-field independent positioning points and far-field independent positioning points.

[0040] Preferably, the first-category sound source fusion positioning points and the second-category sound source fusion positioning points are finally fused. Specifically, the first-category positioning points (collaborative fusion results) have higher priority than the second-category positioning points because they have cross-validation of near-field and far-field data, and the conflicting sections are fused first; for the independent positioning points in the second-category positioning points whose spatial distance is less than the preset threshold, it is further determined whether there is a potential association (such as whether they belong to the same equipment unit). If so, spatial proximity merging is performed to generate a composite positioning point; finally, all the first-category positioning points are retained, and the second-category positioning points that have no spatial conflict with the first-category positioning points are retained as the final sound source fusion positioning point output, thereby ensuring positioning reliability and operation and maintenance efficiency in complex scenarios.

[0041] Furthermore, step S420 also includes step S421, wherein the collaborative fusion positioning channel includes a collaborative mapping relationship between near and far features; step S422, performing frequency cross-comparison on the near acoustic anomaly features and the far acoustic anomaly features of the conflict section, and extracting the acoustic distribution difference of the conflict section, wherein the acoustic distribution difference includes the sound source reception time difference, the frequency main peak offset and the time-frequency feature difference; step S423, utilizing the collaborative mapping relationship of the near and far features to jointly invert and analyze the acoustic distribution difference, calculate the spatial intersection point corresponding to the near and far sound sources in the conflict section, and output the first type of sound source fusion positioning point.

[0042] Preferably, the collaborative fusion positioning channel includes a collaborative mapping relationship between near-field and far-field features, which is used to process conflicting sections, that is, to integrate near-field and far-field detection data, complement the advantages of the two types of signals, and finally output high-precision positioning results. Among them, the collaborative mapping relationship between near-field and far-field features is the association rule between near-field features and far-field features established through historical data and theoretical models. Specifically, based on acoustic propagation theory (such as the attenuation law and refraction and reflection characteristics of sound waves in GIS equipment), a mathematical model of the propagation path of near-field high-frequency signals and far-field low-frequency signals is established, and then a large number of near-field and far-field acoustic data of known defect types (such as partial discharge and mechanical vibration) are used. Through machine learning (such as neural networks and decision trees), the correlation between the two types of signals in frequency, time and space dimensions is mined to form a mapping rule library, and then a correlation mapping of high-frequency pulses and low-frequency vibrations is established.

[0043] Preferably, the near-field acoustic anomaly features and far-field acoustic anomaly features of the conflicting section are frequency-cross-compared to extract the acoustic distribution differences of the conflicting section, including the sound source reception time difference, the frequency main peak offset, and the time-frequency feature difference. Specifically, the sound waves generated by the same defect have different reception times at the near-field and far-field sensors. Since the near-field sensor is closer to the sound source, the reception time is earlier, while the far-field sensor is relatively delayed. By calculating the time difference between multiple sensors receiving the same signal and combining it with the sound wave propagation speed, the sound source position can be inferred. During the propagation of the sound wave, due to factors such as medium attenuation and reflection, the main frequency peak (the frequency point with the strongest energy) of the signal detected in the near field and far field may shift. For example, high-frequency signals attenuate faster during long-distance propagation, causing the main frequency peak detected in the far field to move toward low frequency. The sound waves generated by different types of defects have unique patterns in the time-frequency domain, and the time-frequency features (such as pulse width and spectrum shape) detected in the near field and far field are different. By comparing the time-frequency matrix with the spectrum matrix, the defect characteristics can be identified.

[0044] Preferably, the collaborative mapping relationship between near and far features is utilized, and the extracted acoustic distribution differences are used as input to infer the spatial position of the sound source. Specifically, the acoustic distribution differences (such as time difference, frequency offset) of the conflicting sections are matched with the rules in the collaborative mapping relationship library of near and far features to find the most consistent association pattern; then, based on the matched association pattern, a group of equations containing near-field and far-field positioning parameters is established, and the group of equations is solved by the least squares method, iterative algorithm, etc., to calculate the intersection of the near-field sound source and the far-field sound source in three-dimensional space, and finally the calculated spatial intersection is used as the first-class sound source fusion positioning point.

[0045] Furthermore, step S430 also includes step S431, wherein the independent fusion positioning channel includes a confidence evaluation model; step S432, performing spatial positioning analysis on the near acoustic anomaly features and the far acoustic anomaly features of the non-conflicting section respectively, and constructing a near-field sound source positioning initial value set and a far-field sound source positioning initial value set; step S433, using the confidence evaluation model to perform confidence evaluation on the near-field sound source positioning initial value set and the far-field sound source positioning initial value set, and outputting a confidence evaluation index set; step S434, selecting the best solution based on the confidence evaluation index set, and outputting the second type of sound source fusion positioning point.

[0046] Preferably, the independent fusion positioning channel processes the acoustic anomaly characteristics of the near field and far field separately for non-conflicting sections (i.e., areas where there is no frequency overlap or the spatial distance between the near-field and far-field detection results is far), and performs reliability evaluation on each positioning result through a confidence evaluation model to avoid the uncertainty of single system positioning and ensure the credibility of the output positioning point. The confidence evaluation model is constructed based on historical data, sensor characteristics and signal processing algorithms to quantify the reliability of the positioning results. Specifically, a large amount of near-field and far-field acoustic data of known defect locations is collected, and the error distribution of positioning results under different conditions (such as environmental noise, signal strength, and sensor layout) is analyzed to establish a correlation model between error and influencing factors; then the confidence is comprehensively calculated from multiple dimensions such as signal quality (signal-to-noise ratio, stability), positioning algorithm accuracy, and sensor layout rationality.

[0047] Preferably, in the non-conflict section, spatial positioning analysis is performed on the acoustic anomaly characteristics of the near field and far field respectively to obtain preliminary positioning results, that is, the high-frequency acoustic signals collected by the near-field microphone array are used to calculate multiple possible sound source positions through beam forming, acoustic holography, etc., to form an initial value set for near-field sound source positioning; based on the low-frequency acoustic signals collected by the far-field microphone array, TDOA (time difference positioning) and DOA (direction of arrival estimation) are used to calculate the approximate position of the far-field sound source, to form an initial value set for far-field sound source positioning. Then, each positioning point in the near-field sound source positioning initial value set and the far-field sound source positioning initial value set is input into the confidence evaluation model for evaluation, and the corresponding confidence score is calculated to form a confidence evaluation index set, and the final confidence value is obtained by weighted calculation; finally, the positioning point with the highest confidence is selected from the confidence evaluation index set as the optimal solution. If there are multiple high-confidence positioning points, they are further screened based on conditions such as spatial clustering and signal feature similarity, and the optimal positioning point finally selected is used as the second-type sound source fusion positioning point, representing the most reliable defect position in the near-field and far-field independent positioning in the non-conflicting section.

[0048] Furthermore, step S434 also includes step a, performing time-frequency consistency verification on the second-category sound source fusion positioning point, and obtaining a time-frequency consistency verification result, including verification pass and verification fail; step b, screening the second-category sound source fusion positioning point according to the time-frequency consistency verification result, and outputting an updated second-category sound source fusion positioning.

[0049] Preferably, in the non-conflicting section, the near-field and far-field acoustic signals usually belong to independent sound sources (such as local discharge and mechanical vibration at different locations), and the time-frequency characteristics of their signals should be independent. If the near-field and far-field signals of a certain positioning point are abnormally consistent in the time-frequency domain (such as the same frequency components and synchronized time evolution laws), it indicates that the two actually come from the same sound source. It may be a misjudgment to regard them as independent sound sources in the non-conflicting section. The time-frequency consistency of the second-type sound source fusion positioning points is verified to eliminate "pseudo-independent" positioning points caused by signal coupling, propagation path overlap and other factors in the non-conflicting section, to ensure that the final output second-type sound source positioning points truly correspond to independent and unrelated sound sources, and to avoid misjudging the near-field and far-field signals of the same sound source as two independent sound sources.

[0050] Preferably, time-frequency domain features are extracted for the near-field acoustic anomaly features and the far-field acoustic anomaly features, including time-frequency diagrams, frequency peak sequences, time-frequency energy distribution, and phase consistency, and then the time-frequency similarity of the near-field and far-field signals is evaluated using quantitative indicators, such as calculating the Pearson correlation coefficient of the near-field and far-field time-frequency diagrams. If the absolute value of the coefficient is close to 1, it indicates that the time-frequency features are highly similar; the alignment similarity of the two time-frequency sequences on the time axis is measured; the overlap ratio of the frequency peaks of the near-field and far-field signals is counted, and if the overlap ratio exceeds a threshold (such as 60%), the frequency components are considered to be highly consistent; finally, a time-frequency consistency verification threshold is set according to the device type, operating environment, etc., and a judgment is made to obtain a time-frequency consistency verification result, including verification pass and verification fail. If the time-frequency consistency index of the near-field and far-field signals is lower than the time-frequency consistency verification threshold, it indicates that the time-frequency features of the two are significantly different and belong to independent sound sources, and the verification passes; if the consistency index is higher than the time-frequency consistency verification threshold, it indicates that the time-frequency features of the two are highly correlated, and they may be near-field and far-field signals of the same sound source, and the original positioning result has been misjudged, and the verification fails.

[0051] Preferably, the second-class sound source fusion positioning points are screened according to the time-frequency consistency verification results. Specifically, for the positioning points that pass the verification, the second-class sound source fusion positioning points are retained and confirmed as valid independent sound sources; for the positioning points that fail the verification, they are no longer regarded as independent sound sources in the non-conflicting section to avoid repeated positioning. If it is judged that the near-field and far-field signals actually come from the same sound source (such as the near-field and far-field propagation of the same discharge point), the frequency band and spatial information corresponding to the positioning point are transferred to the collaborative fusion positioning channel for reprocessing, and the unique sound source position is determined by collaborative inversion of near-field and far-field features; finally, the updated second-class sound source fusion positioning is output to ensure the positioning accuracy of the defect positioning point.

[0052] Step S500: importing the GIS device three-dimensional model according to the sound source fusion positioning point, and outputting the GIS device visual positioning result.

[0053] Step S500 further includes step S510, constructing a three-dimensional coordinate system of the sound source space; step S520, converting the three-dimensional model of the GIS device according to the three-dimensional coordinate system of the sound source space, and outputting the converted three-dimensional model of the GIS device; step S530, importing the sound source fusion positioning point into the converted three-dimensional model of the GIS device for visual positioning, and outputting the visual positioning result of the GIS device.

[0054] Preferably, a three-dimensional coordinate system for the sound source space is constructed with a physical point of the GIS device as the origin (such as the center point of the main transformer or the geographic coordinate zero point of the substation), the X-axis is along the device arrangement direction (such as from phase A to phase C), the Y-axis is perpendicular to the device arrangement direction (such as from the busbar room to the circuit breaker room), and the Z-axis is in the vertical direction (such as the ground is Z=0, and the device height direction is the positive direction of the Z-axis), to ensure that the near-field and far-field positioning results are expressed in the same coordinate frame; then, a precise three-dimensional model of the GIS device is constructed using CAD, laser scanning, and three-dimensional modeling software (such as SolidWorks), including key components such as the shell, bushing, circuit breaker, and disconnector, and the three-dimensional model of the GIS device is converted from its original coordinate system (such as the design coordinate system) to the three-dimensional coordinate system of the sound source space, including moving the model origin to the origin of the sound source coordinate system (such as the center point of the device base); adjusting the model posture so that its coordinate axis is aligned with the sound source coordinate system (such as the device length direction corresponds to the X-axis); ensuring that the model size is consistent with the actual physical size (such as a 1:1 ratio); and outputting the converted three-dimensional model of the GIS device.

[0055] Preferably, the three-dimensional coordinates of the sound source fusion positioning points (including the first and second types of positioning points) are directly marked in the converted three-dimensional model of the GIS equipment for visual positioning, and the visual positioning results of the GIS equipment are output. Specifically, the first type of positioning points (conflict section fusion results) are marked with red cubes, marking precise positioning points; the second type of positioning points (independent positioning in non-conflict sections) are marked with blue spheres, marking coarse positioning points; and the marked positioning points are associated with attributes. Clicking the mark can display detailed information of the positioning point, such as defect type (partial discharge / mechanical vibration), characteristic parameters (sound intensity, frequency, confidence), detection time and historical trend data, so as to achieve rapid and accurate positioning of GIS equipment defects.

[0056] The above specific embodiments do not constitute a limitation to the scope of protection of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of this application should be included in the scope of protection of this application. In some cases, the actions or steps recorded in this application can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging, characterized by: The method comprises: Constructing a dual sound intensity detection system, wherein the dual sound intensity detection system includes a near-field sound intensity detection system and a far-field sound intensity detection system; The dual sound intensity detection system is activated to synchronously collect sound source signals from the GIS equipment and output near-field acoustic signals and far-field acoustic signals; Performing frequency band separation on the near-field acoustic signal and the far-field acoustic signal through an adaptive bandpass filter, and identifying near-field acoustic abnormality features and far-field acoustic abnormality features in the separated frequency bands; Perform spatial positioning fusion according to the near acoustic anomaly features and the far acoustic anomaly features, and output a sound source fusion positioning point; The three-dimensional model of the GIS device is imported according to the sound source fusion positioning point, and the visual positioning result of the GIS device is output.

2. The three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging according to claim 1 is characterized in that: Performing frequency band separation on the near-field acoustic signal and the far-field acoustic signal by using an adaptive bandpass filter, the method further comprising: Performing short-time Fourier transform and continuous wavelet transform on the near-field acoustic signal and the far-field acoustic signal, respectively, to output a near-field-time-frequency feature group and a far-field-time-frequency feature group, wherein each time-frequency feature group includes a time-frequency matrix and a spectrum matrix; The adaptive bandpass filter determines adaptive filtering parameters according to the near-field-time-frequency feature group and the far-field-time-frequency feature group, wherein the adaptive filtering parameters include a center frequency and a bandwidth; The near-field acoustic signal and the far-field acoustic signal are separated by using the adaptive filtering parameters to obtain a near-field main acoustic frequency band and a far-field main acoustic frequency band.

3. The three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging according to claim 2 is characterized in that: After obtaining the near-field main acoustic frequency band and the far-field main acoustic frequency band, the method includes: Identify the key eigenvectors corresponding to the near-field main acoustic frequency band and the far-field main acoustic frequency band, including sound intensity, frequency peaks, harmonics and time-frequency changes; The pre-built near-field-abnormal acoustic feature vector library and far-field-abnormal acoustic feature vector library are used to perform anomaly detection on key feature vectors respectively, and output near-field acoustic anomaly features and far-field acoustic anomaly features.

4. The three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging according to claim 2 is characterized in that: Performing spatial positioning fusion according to the near acoustic anomaly features and the far acoustic anomaly features, the method comprising: Detecting conflicting sections based on the near acoustic anomaly features and the far acoustic anomaly features, and outputting conflicting sections and non-conflicting sections, wherein the conflicting sections satisfy frequency range intersection and the distance between the sound source positioning coordinates is less than a preset threshold; Constructing a collaborative fusion positioning channel to perform spatial positioning fusion inversion on the conflicting section and output a first-class sound source fusion positioning point; Construct an independent fusion positioning channel to perform spatial positioning fusion on the non-conflicting segments and output the second type of sound source fusion positioning points; The first type of sound source fusion positioning point and the second type of sound source fusion positioning point are fused to output a sound source fusion positioning point.

5. The three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging according to claim 4 is characterized in that: Constructing a collaborative fusion positioning channel to perform spatial positioning fusion inversion on the conflicting section and outputting a first-class sound source fusion positioning point, including: The collaborative fusion positioning channel includes a collaborative mapping relationship between near and far features; Performing frequency cross-comparison on the near acoustic anomaly features and the far acoustic anomaly features of the conflicting section to extract the acoustic distribution difference of the conflicting section, wherein the acoustic distribution difference includes the sound source reception time difference, the frequency main peak offset, and the time-frequency feature difference; The acoustic distribution difference is jointly inverted and analyzed by utilizing the collaborative mapping relationship of the near and far features, the spatial intersection points corresponding to the near and far sound sources in the conflict section are calculated, and the first type of sound source fusion positioning points are output.

6. The three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging according to claim 4 is characterized in that: Construct an independent fusion positioning channel to perform spatial positioning fusion on the non-conflicting segments and output the second type of sound source fusion positioning points. The method includes: The independent fusion positioning channel includes a confidence evaluation model; Performing spatial positioning analysis on the near acoustic anomaly features and the far acoustic anomaly features of the non-conflicting section respectively, and constructing a near-field sound source localization initial value set and a far-field sound source localization initial value set; Performing confidence evaluation on the near-field sound source localization initial value set and the far-field sound source localization initial value set using the confidence evaluation model, and outputting a confidence evaluation index set; An optimal solution is selected based on the confidence evaluation indicator set, and a second type of sound source fusion positioning point is output.

7. The three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging according to claim 6 is characterized in that: After performing spatial positioning fusion based on the confidence evaluation indicator set and outputting a second type of sound source fusion positioning point, the method includes: Performing time-frequency consistency verification on the second type of sound source fusion positioning points to obtain a time-frequency consistency verification result, including a verification pass or a verification fail; The second type of sound source fusion positioning points are screened according to the time-frequency consistency verification result, and an updated second type of sound source fusion positioning is output.

8. The three-dimensional visualization positioning method for GIS equipment defects based on acoustic imaging according to claim 1 is characterized in that: Importing the GIS device three-dimensional model according to the sound source fusion positioning point and outputting the GIS device visual positioning result, the method includes: Construct a three-dimensional coordinate system of the sound source space; Converting the GIS device three-dimensional model according to the sound source space three-dimensional coordinate system and outputting the converted GIS device three-dimensional model; The sound source fusion positioning point is imported into the converted GIS device three-dimensional model for visual positioning, and the GIS device visual positioning result is output.

Images