Partial discharge positioning method, partial discharge positioning equipment, partial discharge positioning device and readable storage medium

Abstract

The invention relates to a partial discharge positioning method, equipment and device and a readable storage medium. The method comprises the following steps: collecting a visible light image and an infrared image of a to-be-detected area containing a partial discharge phenomenon, fusing the visible light image and the infrared image to generate a double-spectrum positioning map, collecting a sound wave signal of the to-be-detected area through a spiral microphone array, carrying out fusion processing on the double-spectrum positioning map and a sound source thermodynamic diagram of the sound wave signal, and obtaining a sound source thermodynamic diagram of the sound wave signal. Generating a comprehensive discharge positioning map, adopting a clustering algorithm to identify a plurality of features of the visible light image, the infrared image and the sound wave signal, determining the discharge position and the discharge type of the partial discharge phenomenon in the comprehensive discharge positioning map based on the plurality of features, and outputting the comprehensive discharge positioning map. Marking the discharge position and the discharge type of the partial discharge phenomenon in the comprehensive discharge positioning map; the optical, thermal and acoustic multi-mode information of the discharge is integrated, the positioning precision of the partial discharge is greatly improved, the fault condition can be conveniently and quickly judged, and the quick inspection requirement is met.

Description

Technical Field

[0001] The present application relates to the technical field of partial discharge locating technology, and in particular to a partial discharge locating method, device, apparatus and readable storage medium. Background Art

[0002] In rail transit systems, the health of power supply and transformation equipment is directly related to the safety and stability of train operation. Partial discharge is an important manifestation of insulation defects in power supply and transformation equipment. Its accurate detection and positioning are of key significance for equipment failure prevention and maintenance.

[0003] In related technologies, the detection methods of partial discharge have obvious limitations, such as insufficient positioning accuracy, cumbersome operation, and single detection means, which cannot effectively cope with the complex environment of rail transit equipment, such as high noise and high electromagnetic interference. Summary of the Invention

[0004] Based on this, it is necessary to provide a partial discharge positioning method, equipment, device, computer equipment, computer-readable storage medium and computer program product that can accurately locate the partial discharge position of rail transit equipment in complex environments to address the above technical problems.

[0005] In a first aspect, the present application provides a method for locating a local discharge, comprising: collecting a visible light image and an infrared image of a region to be measured containing a local discharge phenomenon, fusing the visible light image and the infrared image to generate a dual-spectrum positioning map; collecting an acoustic wave signal of the region to be measured through a spiral microphone array; fusing the dual-spectrum positioning map with a sound source thermal map of the acoustic wave signal to generate a comprehensive discharge positioning map; using a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image, and the acoustic wave signal, and determining the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map based on the multiple features; outputting the comprehensive discharge positioning map, and marking the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map.

[0006] In one embodiment, after collecting the sound wave signal of the area to be measured through the spiral microphone array, it includes: processing the sound wave signal through a beamforming algorithm, and determining the three-dimensional coordinates of the sound source of the sound wave signal in combination with a three-dimensional positioning error model; using a frequency-weighted signal-to-noise ratio to remove electromagnetic interference in the specific frequency band of rail transit in the sound wave signal.

[0007] In one embodiment, the dual-spectrum localization map and the sound source heat map of the sound wave signal are fused to generate a comprehensive discharge localization map, including: determining the sound source point of the sound wave signal and assigning a weight to each sound source point based on the energy intensity and / or sound pressure level of each sound source point; generating a sound source heat map using a Gaussian blur algorithm according to the weight of each sound source point; and superimposing the sound source heat map on the dual-spectrum localization map with a preset transparency coefficient through an image processing algorithm for fusion processing to generate the comprehensive discharge localization map.

[0008] In one embodiment, the clustering algorithm is used to respectively identify multiple features of the visible light image, the infrared image and the acoustic wave signal, and the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map are determined based on the multiple features, including: identifying the phase features and amplitude features in the phase-resolved local discharge spectrum of the acoustic wave signal; performing time-frequency analysis on the acoustic wave signal to obtain wavelet transform coefficients, and determining the frequency domain characteristics of the transient discharge of the local discharge phenomenon; extracting the grayscale co-occurrence matrix texture of the visible light discharge image and the thermal distribution gradient of the infrared image; and processing the phase features, the amplitude features, the wavelet transform coefficients, the frequency domain characteristics, the grayscale co-occurrence matrix texture and the thermal distribution gradient through the clustering algorithm to obtain the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map.

[0009] In one embodiment, the phase characteristics, the amplitude characteristics, the wavelet transform coefficients, the frequency domain characteristics, the gray-level co-occurrence matrix texture and the thermal distribution gradient are processed by the clustering algorithm to determine the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map, including: determining the discharge type of the local discharge phenomenon based on the phase characteristics, the amplitude characteristics and the frequency domain characteristics of the acoustic wave signal in the clustering result obtained by the clustering algorithm, combined with the thermal distribution gradient; determining the discharge type of the local discharge phenomenon based on the correlation between the gray-level co-occurrence matrix texture and the acoustic transient characteristics of the acoustic wave signal; and using the cluster center coordinates as the discharge position of the local discharge phenomenon in the comprehensive discharge positioning map.

[0010] In one embodiment, outputting the comprehensive discharge location map and marking the discharge position and the discharge type of the partial discharge phenomenon in the comprehensive discharge location map includes:

[0011] The comprehensive discharge map is output, and the discharge position and discharge type of the local discharge phenomenon are marked with rainbow dots in the comprehensive discharge map; different colors of the rainbow dots represent different discharge types, the positions of the rainbow dots represent the discharge positions, and the sizes of the rainbow dots are positively correlated with the energy intensity of the sound source point.

[0012] In the second aspect, the present application also provides a local discharge locating device, including: an infrared camera for collecting infrared images of the area to be measured containing local discharge phenomena; a visible light camera for collecting visible light images of the area to be measured containing local discharge phenomena; a spiral microphone array for collecting acoustic wave signals of the area to be measured; a processor for fusing the visible light image and the infrared image to generate a dual-spectrum locating map; and fusing the dual-spectrum locating map with the sound source thermal map of the acoustic wave signal to generate a comprehensive discharge locating map; using a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image and the acoustic wave signal, and determining the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge locating map based on the multiple features; a visual display screen for outputting the comprehensive discharge locating map, and marking the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge locating map.

[0013] In a third aspect, the present application further provides a partial discharge locating device, comprising:

[0014] An image acquisition module is provided for acquiring visible light images and infrared images of the area to be measured containing partial discharge phenomena, and fusing the visible light images and the infrared images to generate a dual-spectrum positioning map; a signal acquisition module is provided for acquiring acoustic wave signals of the area to be measured through a spiral microphone array; an image fusion module is provided for fusing the dual-spectrum positioning map with the sound source thermal map of the acoustic wave signal to generate a comprehensive discharge positioning map; a discharge recognition module is provided for respectively identifying multiple features of the visible light image, the infrared image and the acoustic wave signal using a clustering algorithm, and determining the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge positioning map based on the multiple features; an image output module is provided for outputting the comprehensive discharge positioning map, and marking the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge positioning map.

[0015] In a fourth aspect, the present application also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program: collecting visible light images and infrared images of the area to be measured containing local discharge phenomena, and fusing the visible light image and the infrared image to generate a dual-spectrum positioning map; collecting the acoustic wave signal of the area to be measured through a spiral microphone array; fusing the dual-spectrum positioning map with the sound source thermal map of the acoustic wave signal to generate a comprehensive discharge positioning map; using a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image and the acoustic wave signal, and determining the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map based on the multiple features; outputting the comprehensive discharge positioning map, and marking the discharge position and the discharge type of the local discharge phenomenon in the comprehensive discharge positioning map.

[0016] In a fifth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, which implements the following steps when executed by a processor: collecting visible light images and infrared images of the area to be measured containing local discharge phenomena, and fusing the visible light image and the infrared image to generate a dual-spectrum positioning map; collecting the acoustic wave signal of the area to be measured through a spiral microphone array; fusing the dual-spectrum positioning map with the sound source thermal map of the acoustic wave signal to generate a comprehensive discharge positioning map; using a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image and the acoustic wave signal, and determining the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map based on the multiple features; outputting the comprehensive discharge positioning map, and marking the discharge position and the discharge type of the local discharge phenomenon in the comprehensive discharge positioning map.

[0017] In a sixth aspect, the present application also provides a computer program product, comprising a computer program, which, when executed by a processor, implements the following steps: collecting visible light images and infrared images of the area to be measured containing local discharge phenomena, and fusing the visible light image and the infrared image to generate a dual-spectrum positioning map; collecting the acoustic wave signal of the area to be measured through a spiral microphone array; fusing the dual-spectrum positioning map with the sound source thermal map of the acoustic wave signal to generate a comprehensive discharge positioning map; using a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image and the acoustic wave signal, and determining the discharge position and discharge type of the local discharge phenomenon in the comprehensive discharge positioning map based on the multiple features; outputting the comprehensive discharge positioning map, and marking the discharge position and the discharge type of the local discharge phenomenon in the comprehensive discharge positioning map.

[0018] The above-mentioned partial discharge positioning method, equipment, device, computer equipment, computer-readable storage medium and computer program product realize the complementarity of optical features by collecting visible light images and infrared images and fusing them to generate a dual-spectrum positioning map. The fusion of the two effectively solves the problem of single infrared technology missing detection of weak discharges and being unable to distinguish between normal temperature rise and discharge heat sources, while improving the anti-interference ability of discharge signals in complex environments; the use of a spiral microphone array to collect sound wave signals can effectively filter out noise interference in the complex environment of rail transit, solving the problems of poor anti-interference and limited range of traditional ultrasonic detection; the dual-spectrum positioning map and the sound source heat are combined to form a dual-spectrum positioning map. The fusion processing of the image aligns and superimposes the optical and acoustic features in time and space, integrating the multi-modal information of light, heat and sound of the discharge, greatly improving the positioning accuracy of partial discharge, and especially solving the problem of inaccurate positioning of existing technologies in complex equipment, hidden parts and multi-point discharge scenarios; using clustering algorithms to analyze multiple features of visible light, infrared and acoustic wave signals, it can accurately identify different types of discharges such as corona discharge, suspended discharge, and tip discharge; outputting a comprehensive discharge positioning map marked with the discharge position and type, and intuitively presenting the detection results in a visual interface, which is convenient for on-site inspection personnel to quickly judge the fault situation and meet the needs of rapid inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 is a schematic diagram of a partial discharge locating device according to one embodiment;

[0021] Figure 2 is a schematic flow chart of a partial discharge locating method according to one embodiment;

[0022] Figure 3 1 is a flow chart of a method for generating a comprehensive discharge positioning map in one embodiment;

[0023] Figure 4 is a schematic flow chart of a partial discharge locating method according to another embodiment;

[0024] Figure 5 is a structural block diagram of a partial discharge locating device in one embodiment;

[0025] Figure 6 FIG. 4 is a diagram showing the internal structure of a partial discharge locating device in one embodiment. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0027] At present, the detection methods for partial discharge locations mainly include ultrasonic detection technology, infrared imaging technology, electrical signal method, etc., but these technologies have obvious limitations in the complex environment of rail transit.

[0028] Ultrasonic detection technology uses ultrasonic signals generated by partial discharge (PD) to detect the source of PD using ultrasonic sensors. However, the complex and often noisy environment of rail transit operations makes ultrasonic signals susceptible to interference, resulting in weak detection signals. Furthermore, ultrasonic detection has a limited range, making it difficult to accurately identify the source of PD in complex equipment.

[0029] Infrared imaging technology utilizes the minute thermal effects of partial discharge (PD) for detection using an infrared thermal imager. However, this technology is insensitive to the thermal effects of weak PD, making it prone to missed detections. Furthermore, single thermal imaging information cannot distinguish other heat sources (such as normal temperature rises caused by equipment operation). Furthermore, infrared detection often only detects anomalies late in the development of equipment problems, making it difficult to provide early warning of faults.

[0030] Electrical signal methods, such as ultra-high frequency and pulse current methods, determine the discharge location by monitoring the high-frequency electromagnetic waves or pulse current signals generated by partial discharge. However, the complex electromagnetic environment of rail transit power supply and transformation systems makes detection equipment susceptible to electromagnetic interference, resulting in a high false alarm rate. Furthermore, their positioning accuracy is insufficient, making them difficult to adapt to the complex environment of multiple discharge signals.

[0031] In summary, existing detection methods have problems such as a single detection method, an inability to effectively cope with the complex environment of rail transit equipment (such as high noise and high electromagnetic interference), insufficient positioning accuracy (difficulty in achieving accurate diagnosis in hidden locations or scenarios with multiple discharge sources), and large equipment size or cumbersome operation (not conducive to the rapid inspection needs in the rail transit field). There is an urgent need for a more efficient and accurate partial discharge positioning technology.

[0032] In order to solve the above technical problems, the present invention provides a method for locating partial discharge, which can be applied to Figure 1The illustrated partial discharge locating device 100 includes an infrared camera 102, a visible light camera 104, a spiral microphone array 106, a processor, and a visual display screen 110. The infrared camera 102 is configured to capture an infrared image of the area to be measured containing partial discharge phenomena; the visible light camera 104 is configured to capture a visible light image of the area to be measured containing partial discharge phenomena; the spiral microphone array 106 is configured to capture acoustic signals from the area to be measured; the processor is configured to fuse a dual-spectrum locating map with a sound source thermogram of the acoustic signal to generate a comprehensive discharge locating map, employing a clustering algorithm to identify multiple features of the visible light image, infrared image, and acoustic signal to determine the discharge location and type of the partial discharge phenomenon in the comprehensive discharge locating map; and the visual display screen 110 is configured to output the comprehensive discharge locating map and annotate the discharge location and type of the partial discharge phenomenon in the comprehensive discharge locating map. In one embodiment, the infrared camera 102, visible light camera 104, and spiral microphone array 106 can be located on the back of the partial discharge locating device 100, while the visual display screen 110 can be located on the front of the partial discharge locating device 100. In one embodiment, the visible light camera 104 may be disposed in the middle of the spiral microphone array 106 .

[0033] In an exemplary embodiment, Figure 2 As shown, a partial discharge location method is provided, which is applied to Figure 1 The partial discharge location device in the embodiment is taken as an example to illustrate, including the following steps 201 to 205. Among them:

[0034] Step 201 : collecting a visible light image and an infrared image of a region to be measured containing a partial discharge phenomenon, and fusing the visible light image and the infrared image to generate a dual-spectrum positioning map.

[0035] Partial discharge (PD) refers to discharges caused by insulation defects in power supply and transformation equipment and is a key indicator of equipment failure. Examples include corona discharge and spike discharge caused by aging insulation within the equipment. The test area refers to specific locations within rail transit power supply and transformation equipment where PD testing is required, such as transformer windings and internal wiring nodes in switchgear. Visible light images, captured by visible light imaging sensors or cameras, clearly demonstrate the corona light signatures of PD. For example, a faint light spot or halo can be observed at the PD point in a visible light image. Infrared images, captured by infrared imaging sensors or cameras, reveal the subtle thermal effects of PD. For example, a PD point appears as a hotspot with a temperature difference from the surrounding area in an infrared image. Dual-spectral localization maps are generated by algorithmically fusing visible light and infrared images, integrating information from both. For example, the light spot at the discharge point in the visible light image is overlaid with the corresponding hotspot in the infrared image, more clearly indicating the possible location of the discharge source.

[0036] In one embodiment, using rail transit switchgear as an example, the test area (the busbar connection inside the switchgear) is inspected. First, a visible light camera captures a visible light image of the area, capturing the faint corona spot generated by partial discharge at the busbar connection. Simultaneously, an infrared camera captures an infrared image of the area, detecting hotspots formed by the minute thermal effect of partial discharge at that location. The two images are then fused using a fusion algorithm within the processor's data processing unit to generate a dual-spectrum localization map. This dual-spectrum localization map retains the location information of the corona spot in the visible light image and overlays the temperature information of the hotspot in the infrared image, more accurately indicating the possible location of partial discharge sources at the switchgear busbar connection.

[0037] In the above embodiment, by collecting visible light images and infrared images of the area to be tested and fusing them to generate a dual-spectrum positioning map, the visible light image is used to capture the corona light characteristics of local discharge to enhance the recognition ability, and the infrared image is used to capture weak thermal effects to improve the detection effect of weak signals in hidden environments. The fusion of the two realizes the comprehensive detection of optical and thermal effect characteristics, effectively making up for the defects of single infrared technology in insufficient sensitivity to weak local discharge, easy missed detection and difficulty in distinguishing normal temperature rise from discharge heat source, and at the same time enhances the anti-interference ability of discharge signals in complex environments, laying a reliable optical foundation for subsequent precise positioning of local discharge positions and identification of discharge types.

[0038] Step 202: Collect the sound wave signal of the area to be measured by using a spiral microphone array.

[0039] A spiral microphone array is a multi-microphone system that arranges multiple microphones evenly along a spiral path. In one embodiment, the spiral microphone array can employ a microphone distribution design based on an annular array. It can consist of 128 highly sensitive microphones arranged in a specific spiral geometry. Its layout is designed with 16 spiral arms and 8 ring layers. The number of microphones in each ring layer increases from 9 in the first ring, 11 in the second ring, 13 in the third ring, and so on, to 23 in the eighth ring. The maximum radius of the spiral microphone array is 0.06 meters, and the minimum radius is 0.012 meters. For example, when inspecting transformers in rail transit, the array can accurately capture the acoustic signals generated by partial discharges within the transformer. Its unique structure optimizes sound source localization performance.

[0040] Acoustic wave signals are the acoustic characteristic signals generated by partial discharge phenomena and are an important basis for locating the source of partial discharge. For example, when partial discharge occurs in power supply and transformation equipment due to insulation defects, sound waves with specific frequencies and amplitudes are generated. The spiral microphone array can capture these sound wave signals for subsequent analysis.

[0041] In one exemplary embodiment, using the inspection of gas-insulated switchgear (GIS) in rail transit as an example, a spiral microphone array is aimed at the GIS test area (e.g., busbar connections). The spiral microphone array, with 16 spiral arms and 8 ring layers, features 128 microphones working together to collect acoustic signals generated by partial discharge (PD) in the test area in real time. The optimized distribution and number of microphones in different ring layers effectively captures the acoustic waves generated by discharges at different locations. Even in the complex noisy environments in which GIS equipment operates, the acoustic characteristics of partial discharges can be accurately captured, providing reliable data for subsequent sound source location.

[0042] In the above embodiment, the sound wave signal of the area to be tested is collected by a spiral microphone array. Its special spiral geometric structure has high-resolution acoustic imaging capabilities and can effectively filter out noise interference in the complex environment of rail transit. It solves the problems of traditional ultrasonic detection being easily affected by environmental noise and having a limited detection range. At the same time, it provides key acoustic feature information for hidden discharge sources that are difficult to cover by optical detection, enhances the ability to capture local discharge phenomena, and lays a solid acoustic data foundation for subsequent fusion with dual-spectrum positioning maps to improve the accuracy of local discharge positioning.

[0043] Step 203 : Fusion processing is performed on the dual-spectrum positioning map and the sound source heat map of the acoustic wave signal to generate a comprehensive discharge positioning map.

[0044] The sound source heat map is a visualization generated based on the acoustic signals collected by a spiral microphone array. It reflects the intensity and distribution of the sound source through a color gradient (e.g., from blue to red). The closer the color is to red, the stronger the sound source signal. For example, when inspecting switchgear, areas with strong acoustic signals generated by partial discharge appear as red spots in the heat map, visually indicating the approximate location of the discharge source. The comprehensive discharge localization map is a fusion of the dual-spectrum localization map and the sound source heat map. It integrates the optical, thermal, and acoustic multimodal characteristics of partial discharge and can simultaneously display the optical characteristics, thermal effects, and acoustic distribution of the discharge. For example, the comprehensive discharge localization map shows both the corona spots in the visible light image and the hot spots in the infrared image, as well as the red strong signal areas in the sound source heat map, comprehensively reflecting the location and characteristics of the partial discharge source.

[0045] In one embodiment, using the inspection of GIS equipment in rail transit as an example, the partial discharge location device acquires a dual-spectral location map (containing visible optical corona spots and infrared hotspots indicating partial discharge within the GIS) and an acoustic source heat map (showing strong red signal areas of the discharge acoustic waves). The device then performs a fusion process: The two images are first spatially and temporally aligned to ensure that the same discharge source corresponds in both images. A fusion algorithm then overlays the optical and thermal features of the dual-spectral location map with the acoustic features of the sound source heat map to generate a comprehensive discharge location map. This comprehensive discharge location map clearly displays the locations of the discharge point's corona light and hotspots, while also enhancing its acoustic signal characteristics through the red heat map. This allows inspectors to intuitively and comprehensively understand the specific location and multi-dimensional characteristics of the discharge source within the GIS.

[0046] In the above embodiment, a comprehensive discharge positioning map is generated by fusing the dual-spectrum positioning map with the sound source thermal map, which integrates the light, heat, and sound multimodal information of the local discharge. This makes up for the shortcomings of single optical detection in identifying hidden discharge sources and the defect of single acoustic detection lacking intuitive optical features. It effectively improves the positioning accuracy and information integrity of the local discharge position in complex environments, and at the same time provides a comprehensive image basis for the subsequent identification of the discharge position and type through clustering algorithms, which helps detection personnel to judge the local discharge situation more efficiently and accurately.

[0047] In step 204 , a clustering algorithm is used to respectively identify multiple features of the visible light image, the infrared image, and the acoustic wave signal, and the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge location map are determined based on the multiple features.

[0048] The discharge location refers to the specific spatial coordinates of the partial discharge phenomenon in the comprehensive discharge location map. This precise location is determined by fusing the features of visible light images, infrared images, and acoustic wave signals and analyzing them through a clustering algorithm. For example, in the comprehensive discharge location map of a rail transit transformer, the discharge location of a certain partial discharge source is marked as (0.3m, 0.5m, 0.8m), clarifying its three-dimensional spatial orientation inside the transformer.

[0049] Discharge types are categorized based on the characteristics of local discharge (such as corona light intensity, thermal effect distribution, and acoustic wave frequency). Common types include corona discharge, suspension discharge, and tip discharge. For example, by analyzing the scattered weak corona light in visible light images, the uniform low thermal effect area in infrared images, and the specific frequency characteristics of acoustic wave signals, the local discharge can be identified as a corona discharge.

[0050] In one embodiment, taking the inspection of rail transit switchgear as an example, first, features such as the corona light intensity and shape of the discharge point are extracted from the visible light image, features such as the hotspot temperature and range are extracted from the infrared image, and features such as the frequency and amplitude are extracted from the acoustic wave signal. Then, a clustering algorithm (such as K-Means clustering) is used to analyze these multi-dimensional features, grouping signals with similar features into one category to distinguish different partial discharge sources. Finally, two discharge locations are determined in the comprehensive discharge positioning map: the upper busbar connection inside the switchgear (0.2m, 0.1m, 0.4m) and the lower terminal block (0.15m, 0.6m, 0.3m). The former is identified as a tip discharge (characterized by a concentrated strong light spot, local high temperature, and high-frequency acoustic waves), while the latter is a suspended discharge (characterized by dispersed weak light, uniform low heat, and acoustic waves in a specific frequency band), achieving accurate identification of the discharge location and type.

[0051] In the above embodiment, the clustering algorithm is used to analyze multiple features of visible light images, infrared images and acoustic wave signals, which can accurately identify the discharge location and discharge type of the partial discharge phenomenon, effectively solving the problem of difficulty in distinguishing multiple point discharges and inaccurately judging the discharge type in the existing technology. At the same time, combined with the comprehensive analysis of multimodal data features, the accuracy of partial discharge location positioning and the reliability of discharge type identification in complex environments are improved, providing an accurate basis for targeted maintenance of subsequent equipment failures, and further improving the efficiency of fault detection and diagnosis of rail transit power supply and transformation equipment.

[0052] Step 205: output a comprehensive discharge location map, and mark the discharge location and discharge type of the partial discharge phenomenon in the comprehensive discharge location map.

[0053] In one embodiment, taking the inspection of high-voltage switchgear in a rail transit substation as an example, after the generation of a comprehensive discharge location map and the identification of the discharge location and type are completed, step 205 is executed: the partial discharge location device outputs a comprehensive discharge location map that integrates visible light images, infrared images, and acoustic signal features. This comprehensive discharge location map clearly displays the structural details within the switchgear. In the comprehensive discharge location map, the discharge locations can be marked with rainbow dots of different colors. For example, a red dot (representing a tip discharge) is marked at the busbar connection on the left side of the switchgear, with coordinates corresponding to (0.4m, 0.2m, 0.6m); a blue triangle dot (representing a suspended discharge) is marked at the right terminal, with coordinates (0.3m, 0.5m, 0.6m). Each annotation is accompanied by a text description, such as "Tip discharge (location: 0.4m, 0.2m, 0.6m)" and "Suspended discharge (location: 0.3m, 0.5m, 0.6m)." Inspectors can intuitively view this information through the human-computer interface of the partial discharge location device, quickly understanding the specific situation of partial discharge within the switchgear.

[0054] In the above embodiment, by outputting a comprehensive discharge positioning map and marking the discharge position and discharge type, the multimodal information of partial discharge is presented in an intuitive and visual manner, which solves the problem of the existing technology that the discharge position and discharge type cannot be intuitively displayed. It is convenient for detection personnel to quickly and clearly grasp the specific situation of partial discharge, improves the efficiency of on-site fault diagnosis, and provides a clear reference basis for subsequent equipment maintenance, which helps to take targeted measures to deal with faults in a timely manner and ensure the safe and stable operation of rail transit power supply and transformation equipment.

[0055] In an exemplary embodiment, after collecting the sound wave signal of the area to be measured through the spiral microphone array, it can also include: processing the sound wave signal through a beamforming algorithm, and determining the three-dimensional coordinates of the sound source of the sound wave signal in combination with a three-dimensional positioning error model; and using a frequency-weighted signal-to-noise ratio to remove electromagnetic interference in the specific frequency band of rail transit in the sound wave signal.

[0056] The beamforming algorithm spatially filters and enhances the acoustic signals collected by a spiral microphone array. By focusing on the sound source signal in a specific direction, it suppresses noise interference from other directions and improves the resolution of sound source localization. For example, when detecting partial discharge in a switchgear cabinet, the algorithm can focus on the direction of the partial discharge point, enhancing the acoustic signal there and weakening irrelevant noise such as equipment vibration. The three-dimensional positioning error model comprehensively calculates the deviation of the sound source positioning in three-dimensional space (x, y, and z directions). By taking the square root of the sum of the squares of the deviations from multiple detections and averaging them, it comprehensively reflects the spatial deviation between the positioning result and the actual fault point. For example, it can calculate the comprehensive error for a positioning error of 0.02 meters on the x-axis, 0.01 meters on the y-axis, and 0.03 meters on the z-axis, thus optimizing positioning accuracy. The three-dimensional coordinates of the sound source refer to the specific location information of the partial discharge source in three-dimensional space, and can accurately describe the spatial location of the partial discharge source inside the equipment. For example, the three-dimensional coordinates of the partial discharge source inside a transformer are determined to be (0.5m, 0.3m, 1.2m), which clarifies the specific location of the partial discharge location in the equipment.

[0057] For example, the three-dimensional coordinates of the sound source, that is, the three-dimensional positioning coordinates of the sound wave signal source. The calculation formula of the three-dimensional positioning error of the sound wave signal may include:

[0058]

[0059] in, 、 、 is the three-dimensional coordinate of the detected sound source, 、 、 is the actual three-dimensional coordinate of the sound source, and n is the number of microphones i in the spiral microphone array.

[0060] In one embodiment, when performing partial discharge detection on a transformer in a rail transit substation, after a spiral microphone array collects a signal containing discharge sound waves, it first uses frequency-weighted signal-to-noise ratio processing. To address the 50 Hz power frequency electromagnetic interference present in the transformer's operating environment and the specific high-frequency interference caused by passing trains, the array assigns high weights to signals in the characteristic frequency bands of partial discharge (e.g., 80-150 kHz) and low weights to signals in the interference frequency band, effectively filtering out electromagnetic interference in these specific frequency bands. Subsequently, the denoised sound wave signal is processed using a beamforming algorithm, focusing on the possible direction of the partial discharge source and enhancing the sound wave signal in that direction. Combined with a three-dimensional positioning error model, the positioning deviations in the x, y, and z directions are comprehensively calculated and optimized, ultimately determining the three-dimensional coordinates of the partial discharge source inside the transformer as (0.8m, 0.4m, 1.5m), completing the precise processing of the sound wave signal and sound source location.

[0061] Frequency-weighted signal-to-noise ratio (SNR) is a calculation method that considers the weighting of signals at different frequencies. By setting weights based on the impact of noise and signals at different frequencies in the rail transit environment, it can more accurately measure signal quality. For example, signals with characteristic frequencies of partial discharge (PD) (e.g., 100kHz) are given a high weight, while signals with 50Hz power frequency interference generated by train operation are given a low weight, highlighting valid signals. Electromagnetic interference (EMI) refers to electromagnetic signals in specific frequency bands generated by train operation and equipment in the rail transit environment, which can interfere with the detection of PD acoustic signals. For example, electromagnetic signals in specific frequency bands generated by traction motor operation can be mixed with PD acoustic signals, affecting detection accuracy.

[0062] For example, the calculation formula of the frequency-weighted signal-to-noise ratio may include:

[0063]

[0064] in, is the weight corresponding to different frequency signals, m is the number of detection frequency bands (from the tail frequency band to the starting frequency band) with a difference of 100HZ, j is the starting frequency band number of the detection frequency band, and the signal power j is the signal power of the acoustic signal j, the noise power j is the signal power of the noise signal j.

[0065] Through this frequency-weighted signal-to-noise ratio calculation and processing, the effective partial discharge acoustic wave signal is highlighted and the electromagnetic interference signal in a specific frequency band is suppressed, so that the subsequent analysis of the acoustic wave signal is more focused on the real partial discharge information and the impact of electromagnetic interference on the detection results is reduced.

[0066] In an exemplary embodiment, Figure 3FIG. 1 is a flow chart of a method for generating a comprehensive discharge positioning map in an embodiment. Figure 3 The method for generating the comprehensive discharge positioning map shown is specifically as follows: Figure 2 The sub-steps of step 203 in the embodiment shown involve a specific implementation method of fusing the dual-spectrum localization map with the sound source heat map of the acoustic wave signal to generate a comprehensive discharge localization map. Figure 2 Based on the embodiment shown, for details not described in this embodiment, please refer to Figure 2 This method is applied to Figure 1 Taking the partial discharge locating device in the embodiment as an example, the method includes the following steps 2031 to 2033:

[0067] Step 2031 : determining the sound source points of the sound wave signal and assigning a weight to each sound source point based on the energy intensity and / or sound pressure level of each sound source point.

[0068] Among them, the sound source point refers to the specific point in space of the local discharge source that generates the sound wave signal. It is the source position of the sound wave signal and can be determined from the signal collected by the spiral microphone array through acoustic imaging algorithm or positioning algorithm; for example, when detecting partial discharge of a transformer, algorithm analysis can determine that there are two sound source points inside the equipment, corresponding to discharge phenomena at different positions.

[0069] The energy of the sound source point refers to the energy carried by the sound waves generated by local discharge, and the sound pressure level is a physical quantity that measures the strength of the sound wave pressure. Both can reflect the discharge intensity of the sound source point; for example, a higher energy or higher sound pressure level at a sound source point indicates that the local discharge phenomenon at that location is more intense.

[0070] In one embodiment, taking the inspection of rail transit switch cabinets as an example, after the spiral microphone array collects the sound wave signal in the switch cabinet, it is analyzed through an acoustic imaging algorithm to determine that there are two sound source points in the switch cabinet, located at the busbar connection and the bolt fixing point respectively; then the energy and / or sound pressure level of each sound source point is calculated, and it is found that the energy of the sound source point at the busbar connection is 50 microjoules (μJ) and the sound pressure level is 65 decibels (dB), and the energy of the sound source point at the bolt fixing point is 30 μJ and the sound pressure level is 55 dB; then, a weight is assigned to each sound source point according to the energy and / or sound pressure level. The greater the energy and / or sound pressure level, the higher the weight. Therefore, a weight of 0.6 is assigned to the sound source point at the busbar connection, and a weight of 0.4 is assigned to the sound source point at the bolt fixing point, so that the strong discharge signal area can be highlighted when the sound source heat map is subsequently drawn.

[0071] In the above embodiment, by determining the sound source point and the energy or sound pressure level of each sound source point and assigning weights, an accurate basis is provided for the subsequent drawing of the sound source heat map, so that the heat map can more accurately reflect the intensity differences of different discharge sources and highlight the strong discharge signal area.

[0072] Step 2032: Generate a sound source heat map using a Gaussian fuzzy algorithm based on the weight of each sound source point.

[0073] Among them, the Gaussian blur algorithm is an algorithm that smoothes images through a Gaussian function. It can make the pixel values ​​in the image show a normally distributed attenuation centered on a certain point, forming a gradual blur effect. It is often used to enhance the diffusion performance of feature areas in the image; for example, in the generation of sound source heat maps, the algorithm can allow the signals of high-weight sound source points to diffuse smoothly to the surrounding areas, forming a gradual transition from strong signals in the center to weak signals at the edge, which is more in line with the propagation characteristics of sound waves in space.

[0074] In one embodiment, using GIS equipment inspection at a rail transit substation as an example, after determining the presence of three sound source points within the equipment and assigning them weights (0.7, 0.5, and 0.3, respectively), step 2032 is executed. A Gaussian blur algorithm is applied to each sound source point based on its weight. A sound source point with a weight of 0.7 (corresponding to a stronger discharge) is processed by the Gaussian function, forming a large diffusion area centered on its location, with a dark center (e.g., red) and gradually lighter edges (e.g., orange). A sound source point with a weight of 0.5 forms a medium-sized diffusion area with an orange center and yellow edges. A sound source point with a weight of 0.3 forms a smaller diffusion area with a yellow center and blue edges. Finally, the signals from these Gaussian-blurred sound source points are superimposed and fused to generate a sound source heat map that intuitively reflects the intensity and spatial distribution of each discharge source. Darker and larger areas represent more intense discharges.

[0075] In the above embodiment, a Gaussian blur algorithm is used to generate a sound source heat map based on the weight of the sound source point, so that the signal of the sound source point is diffused in a gradual manner, which not only highlights the position and intensity of the high-weight (strong discharge) sound source point, but also intuitively reflects the propagation characteristics of the sound wave in space through the smooth transition of color changes, solving the problem that a single labeled sound source point cannot reflect the distribution of discharge intensity.

[0076] In step 2033, the sound source thermal map is superimposed on the dual-spectrum positioning map with a preset transparency coefficient through an image processing algorithm for fusion processing to generate a comprehensive discharge positioning map.

[0077] The preset transparency coefficient refers to the transparency parameter set for the sound source heat map during image fusion. It is used to balance the display effects of the sound source heat map and the bispectral localization map, preventing information from one from obscuring the other. The transparency coefficient is typically between 0 and 1. The closer the coefficient is to 0, the more transparent the heat map and the clearer the bispectral localization map; the closer the coefficient is to 1, the less transparent the heat map and the more prominent the acoustic features. For example, when inspecting complex switchgear, the transparency coefficient can be preset to 0.5, ensuring that the corona spots and hot spots in the bispectral localization map are clearly visible while also clearly displaying strong signal areas in the sound source heat map.

[0078] In one embodiment, taking the inspection of transformers in rail transit as an example, after obtaining a dual-spectrum positioning map (including visible optical corona and infrared hotspots of partial discharge inside the transformer) and an acoustic source thermogram (displaying the strength distribution of discharge sound waves), a fusion process is performed: first, a preset transparency coefficient is set to 0.4. This coefficient is determined through preliminary testing, which allows both the optical and thermal features in the dual-spectrum positioning map to be clearly visible and the acoustic features of the acoustic source thermogram to be effectively superimposed; then, using an image processing algorithm, the acoustic source thermogram is superimposed on the dual-spectrum positioning map at a transparency coefficient of 0.4. After superposition, the corona spots (white bright spots) and infrared hotspots (red areas) in the dual-spectrum positioning map remain clear, while the orange-red areas representing strong discharge in the acoustic source thermogram are covered in a semi-transparent state at the corresponding positions, and the information of the two does not obstruct each other; finally, a comprehensive discharge positioning map is generated, which simultaneously presents the optical, thermal, and acoustic characteristics of partial discharge, allowing inspection personnel to intuitively see the specific location and multi-dimensional characteristics of the discharge source in the transformer.

[0079] In the above embodiment, the sound source thermogram is superimposed on the dual-spectrum positioning map by presetting the transparency coefficient, which not only ensures the clear presentation of the light and heat characteristics in the dual-spectrum positioning map, but also effectively superimposes the acoustic characteristics of the sound source thermogram, avoids the masking of single image information, realizes the organic fusion of multimodal information, and solves the problem of information conflict or missing during image fusion. The generated comprehensive discharge positioning map can comprehensively and intuitively reflect the light, heat and sound characteristics of local discharge, providing a clear basis for detection personnel to quickly grasp the location and characteristics of the discharge source, and further improving the efficiency and accuracy of fault diagnosis.

[0080] In an exemplary embodiment, Figure 4 FIG. 1 is a flow chart of a partial discharge location method according to another embodiment. Figure 4 The partial discharge location method shown is specifically Figure 2 The sub-steps of step 204 in the embodiment shown involve using a clustering algorithm to respectively identify multiple features of the visible light image, infrared image and acoustic wave signal to obtain the discharge location and discharge type of the partial discharge phenomenon. Figure 2Based on the embodiment shown, for details not described in this embodiment, please refer to Figure 2 This method is applied to Figure 1 Taking the partial discharge locating device in the embodiment as an example, the method includes the following steps 2041 to 2044:

[0081] Step 2041 : Identify the phase features and amplitude features in the phase-resolved partial discharge spectrum of the acoustic wave signal.

[0082] In one embodiment, when performing partial discharge detection on GIS equipment in a rail transit power supply and transformation system, a spiral microphone array collects acoustic signals and processes the data to generate a phase-resolved partial discharge (PRPD) pattern. The horizontal axis of this pattern represents the phase (0-720 degrees, corresponding to two power frequency cycles), and the vertical axis represents the normalized acoustic signal amplitude.

[0083] The phase characteristics of this spectrum include: the local discharge acoustic wave signals are mainly concentrated in the phase intervals of 100-150 degrees and 460-510 degrees, showing a clear double-cluster distribution characteristic, and the phase widths of these two intervals are relatively narrow, indicating that the discharge is more active in these specific phases. The amplitude characteristics of this spectrum include: the normalized signal amplitudes in the two phase intervals are mostly within the range of 0.6-0.8, with the maximum amplitude reaching 0.85 in the 100-150 degree interval and 0.78 in the 460-510 degree interval. The overall amplitude distribution is relatively concentrated, indicating that the discharge intensity is relatively stable and has a certain regularity. The identification of phase and amplitude characteristics provides a key basis for the subsequent determination of local discharge types (such as corona discharge and suspension discharge).

[0084] Step 2042 : Perform time-frequency analysis on the acoustic wave signal to obtain wavelet transform coefficients, and determine the frequency domain characteristics of transient discharge of the partial discharge phenomenon.

[0085] For example, assume that a piece of sound wave data containing a local discharge signal is collected through a spiral microphone array (duration: 1 second, sampling frequency: 1 megahertz (MHz)). The signal contains both transient sound waves generated by local discharge and background environmental noise.

[0086] First, a time-frequency analysis is performed on the acoustic signal: an appropriate wavelet basis function (such as the Daubechies4 wavelet) is selected, the number of decomposition layers is set to 8, and a wavelet transform is performed on the acoustic signal. The resulting wavelet transform coefficients reveal the distribution of signal energy at different time points and frequency bands. The coefficients corresponding to transient partial discharges appear as pulsed energy concentrations within a short time dimension and cover a specific frequency range in the frequency dimension.

[0087] Subsequently, by extracting the wavelet transform coefficients corresponding to these pulse energy concentration areas, their frequency domain characteristics were analyzed: for example, it was found that the energy of the transient discharge was mainly distributed in the frequency band of 50kHz~200kHz, and an energy peak appeared near 150kHz. This frequency domain characteristic is consistent with the known transient acoustic wave frequency domain characteristics of the air gap discharge inside the cable insulation. Therefore, it can be determined that the transient discharge frequency domain characteristics of this local discharge phenomenon are "wide-band pulse characteristics centered at 150kHz and mainly distributed in the range of 50kHz~200kHz."

[0088] Step 2043 : extracting the grayscale co-occurrence matrix texture of the visible light discharge image and the thermal distribution gradient of the infrared image.

[0089] For example, when local discharge occurs in a high-voltage cable, a high-definition visible light camera is used to capture a visible light image of the discharge area (with a resolution of 1920×1080 pixels), and an infrared camera is used to synchronously capture an infrared image of the area (temperature measurement range -20°C~150°C, spatial resolution 640×512 pixels).

[0090] For visible light discharge images, preprocessing is first performed: the color image is converted to a grayscale image, and irrelevant background areas at the image edge are removed, retaining the discharge spot and surrounding area. Subsequently, the grayscale co-occurrence matrix texture features are extracted. The pixel distance is set to 1, and the orientation is set to 0 degrees (°), 45 degrees, 90 degrees, and 135 degrees. The grayscale co-occurrence matrix is ​​calculated, and texture parameters such as contrast (reflecting the intensity of grayscale changes in the image; the contrast between the discharge spot and the surrounding area is generally high), energy (reflecting the uniformity of the image's grayscale distribution; the energy value in the discharge area will exhibit a specific value due to the concentration of the spot), and entropy (indicating the complexity of the image information; the entropy value of the discharge area is different from that of normal areas) are extracted from the matrix. These parameters can quantitatively describe the visual texture characteristics of the discharge area.

[0091] For infrared images, noise reduction and temperature calibration are first performed to ensure the accuracy of the thermal distribution data. The thermal distribution gradient is then extracted: Taking each pixel in the image as the center, the temperature difference between it and its eight neighboring pixels is calculated. The thermal gradient amplitude and direction at that point are then calculated using a gradient operator. For example, in the infrared image region corresponding to a partial discharge point, a distinct area of ​​elevated temperature appears. The thermal distribution gradient exhibits a gradient shift from the high-temperature center to the surrounding low-temperature region. Areas with larger gradient amplitudes are concentrated at the edges of the discharge point, and the gradient direction exhibits a radial distribution. These characteristics intuitively reflect the heat diffusion pattern generated by partial discharge.

[0092] Step 2044 , using a clustering algorithm to process phase features, amplitude features, wavelet transform coefficients, gray-level co-occurrence matrix textures, and thermal distribution gradients, to obtain the discharge location and discharge type of the partial discharge phenomenon.

[0093] For example, assuming that the characteristic data related to the partial discharge to be analyzed have been obtained, it may include phase characteristics (such as the corresponding relationship parameters between the partial discharge pulse and the voltage phase), amplitude characteristics (such as the peak value and average value of the discharge pulse), wavelet transform coefficients (reflecting the frequency domain characteristics of transient discharge, the energy proportion of a total of 8 frequency bands), gray-level co-occurrence matrix texture (three parameters: contrast, energy, and entropy), and thermal distribution gradient (two parameters: gradient amplitude mean and directional concentration), totaling 1+2+8+3+2=16 dimensions of features, and each sample corresponds to the coordinates of a detection point on the cable.

[0094] First, considering the principle of different importance of features in different dimensions, the weight of each feature is calculated by the information gain method: for example, the information gain of the energy proportion of the 150kHz frequency band in the wavelet transform coefficients is the highest (weight 0.15), followed by the contrast of the gray-level co-occurrence matrix (weight 0.12), the phase feature weight is 0.10, the peak weight in the amplitude feature is 0.09, and the mean weight of the thermal distribution gradient amplitude is 0.08. The remaining features are assigned weights of 0.03-0.07 according to their importance, and the sum of the weights of all dimensions is 1.

[0095] Next, select the K-means clustering algorithm to process these features:

[0096] Determine the number of clusters: Based on the possible discharge locations (such as joints and insulation defects) and discharge types (such as air gap discharge and surface discharge) of the cable, the number of clusters is initially set to 3.

[0097] Calculate the improved distance: For sample i and sample k in the same cluster, calculate the distance of the j-th dimension feature (such as Euclidean distance), multiply it by the weight of the dimension, and then sum it to get the improved average distance within the same cluster ; Similarly, calculate the average distance between sample i and the nearest sample in other clusters , where w j The weight given to each dimension feature, m is the total number of feature dimensions, is the distance within the same cluster calculated based on the j-th dimension feature, The distance from sample i to the nearest sample of other clusters calculated based on the j-th dimension feature, and then according to the improved silhouette coefficient formula , calculate the silhouette coefficient for each sample , where N is the total number of samples. The average silhouette coefficient of all samples is maximized by adjusting the number of clusters (the optimal number of clusters is ultimately determined to be 3).

[0098] Perform clustering: Through iterative calculations, samples with similar features are clustered into one category.

[0099] In one embodiment, the clustering results are analyzed as follows:

[0100] Cluster 1: The wavelet transform coefficients of the samples are concentrated in the 50-200kHz frequency band, the grayscale co-occurrence matrix has high contrast (corresponding to obvious light spots), and the thermal distribution gradient is radial (high temperature is concentrated in the center). Combined with the energy proportion characteristics of the 150kHz frequency band with a higher weight, it is judged to be air gap discharge. The detection point coordinates corresponding to this type of samples are concentrated in a certain area in the middle section of the cable, and the discharge location is determined to be the air gap in the insulation layer in the middle section of the cable.

[0101] Cluster 2: Phase characteristics show that the discharge pulses are concentrated near the voltage peak, the amplitude characteristics fluctuate greatly, and the thermal distribution gradient extends along the cable surface, indicating surface discharge. The sample coordinates are concentrated at the cable joint, confirming that the discharge location is the sealing surface of the cable joint.

[0102] Cluster 3: The values ​​of all features are small, and the silhouette coefficient shows that it is far away from other clusters. Combined with the background noise characteristics, it is judged to be a normal area with no obvious discharge.

[0103] Finally, after processing with the clustering algorithm, the discharge location (the air gap in the insulation layer of the middle section of the cable, the sealing surface of the cable joint) and the discharge type (air gap discharge, surface discharge) of the partial discharge phenomenon are obtained.

[0104] In one embodiment, the phase, amplitude, and frequency characteristics of the acoustic signal in the clustering results obtained by the clustering algorithm, combined with the thermal distribution gradient, can be used to determine whether the discharge type is a concealed discharge. For example, cluster 1 has a stable phase distribution, minimal amplitude fluctuations, and a gentle thermal distribution gradient (heat does not diffuse easily), which meets the characteristics of a concealed discharge: "energy is concentrated internally, and external thermal characteristics are not significant." Therefore, the discharge type of cluster 1 is determined to be a concealed discharge.

[0105] In one embodiment, surface discharge can be determined based on the correlation between the gray-level co-occurrence matrix texture and the acoustic transient characteristics of the acoustic signal. For example, for cluster 2, the correlation coefficient between the gray-level co-occurrence matrix texture and the acoustic transient characteristics of the acoustic signal is high, indicating a strong correlation. Combined with the surface discharge characteristics of "rough texture and persistent acoustic transient signals," the discharge type in cluster 2 is determined to be surface discharge.

[0106] In one embodiment, the coordinate position of the cluster center may be used as the discharge position of the partial discharge phenomenon in the comprehensive discharge location map.

[0107] In the above embodiment, the clustering algorithm is used to analyze multiple features of visible light images, infrared images and acoustic wave signals, which can accurately identify the discharge location and discharge type of the partial discharge phenomenon, effectively solving the problem of difficulty in distinguishing multiple point discharges and inaccurately judging the discharge type in the existing technology. At the same time, combined with the comprehensive analysis of multimodal data features, the accuracy of partial discharge location positioning and the reliability of discharge type identification in complex environments are improved, providing an accurate basis for targeted maintenance of subsequent equipment failures, and further improving the efficiency of fault detection and diagnosis of rail transit power supply and transformation equipment.

[0108] In an exemplary embodiment, a comprehensive discharge map is output, and the discharge location and discharge type of the partial discharge phenomenon are marked with rainbow dots in the comprehensive discharge map. Different colors of the rainbow dots represent different discharge types, and the location of the rainbow dots represents the discharge location. The size of the rainbow dots is positively correlated with the energy intensity and / or sound pressure level of the sound source. For example, a blue rainbow dot represents a hidden discharge, a red rainbow dot represents a surface discharge, and the center of the rainbow dot represents the discharge location. The larger the size of the rainbow dot, the higher the energy intensity and sound pressure level of the sound source.

[0109] The aforementioned partial discharge localization method captures visible light and infrared images and fuses them to generate a dual-spectrum localization map, achieving complementary optical features. This fusion effectively addresses the issues of single-infrared technology missing weak discharges and failing to distinguish between normal temperature rises and discharge heat sources, while also improving the anti-interference capability of discharge signals in complex environments. The use of a spiral microphone array to collect acoustic signals effectively filters out noise interference in the complex environment of rail transit, addressing the poor anti-interference and limited range issues of traditional ultrasonic detection. The fusion of the dual-spectrum localization map and the acoustic source heat map aligns and superimposes optical and acoustic features in time and space, integrating the multimodal information of light, heat, and sound from the discharge, significantly improving the localization accuracy of partial discharges. This particularly addresses the inaccurate localization issues of existing technologies in complex equipment, hidden locations, and multi-point discharge scenarios. A clustering algorithm is used to analyze multiple features of visible light, infrared, and acoustic signals, accurately identifying different types of discharges, such as corona discharge, suspended discharge, and tip discharge. The method outputs a comprehensive discharge localization map labeled with discharge location and type, presenting the detection results intuitively on a visual interface, enabling on-site inspectors to quickly determine fault conditions and meet the needs of rapid inspections.

[0110] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily performed in sequence in the order indicated by the arrows. Unless clearly stated herein, the execution of these steps is not strictly limited in order, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above may include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of the steps or stages in other steps or other steps. It is understandable that the various steps in different embodiments can be freely combined as needed, and the various non-contradictory schemes formed by the combination all fall within the scope of protection of this application.

[0111] Based on the same inventive concept, embodiments of the present application further provide a partial discharge locating device for implementing the aforementioned partial discharge locating method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more partial discharge locating device embodiments provided below can be found in the aforementioned limitations of the partial discharge locating method and are not further elaborated here.

[0112] In an exemplary embodiment, Figure 5 As shown, a partial discharge locating device 500 is provided, comprising: an image acquisition module 510, a signal acquisition module 520, an image fusion module 530, a discharge identification module 540, and an image output module 550, wherein: the image acquisition module 510 is used to acquire a visible light image and an infrared image of a test area containing a partial discharge phenomenon, and fuse the visible light image and the infrared image to generate a dual-spectrum positioning map; the signal acquisition module 520 is used to acquire an acoustic wave signal of the test area through a spiral microphone array; the image fusion module 530 is used to fuse the dual-spectrum positioning map with a sound source heat map of the acoustic wave signal to generate a comprehensive discharge positioning map; the discharge identification module 540 is used to use a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image, and the acoustic wave signal, and determine the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge positioning map based on the multiple features; and the image output module 550 is used to output the comprehensive discharge positioning map and mark the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge positioning map.

[0113] In the above-mentioned partial discharge positioning device, the image acquisition module 510 realizes the complementarity of optical features by collecting visible light images and infrared images and fusing them to generate a dual-spectrum positioning map. The fusion of the two effectively solves the problem that a single infrared technology misses detection of weak discharges and cannot distinguish between normal temperature rise and discharge heat sources, while improving the anti-interference ability of discharge signals in complex environments; the signal acquisition module 520 uses a spiral microphone array to collect sound wave signals, which can effectively filter out noise interference in the complex environment of rail transit, and solves the problems of poor anti-interference and limited range of traditional ultrasonic detection; the image fusion module 530 fuses the dual-spectrum positioning map with the sound source heat map to obtain the optical The features are aligned and superimposed in time and space with the acoustic features, integrating the multimodal information of light, heat and sound of the discharge, greatly improving the positioning accuracy of local discharge, and especially solving the problem of inaccurate positioning of existing technologies in complex equipment, hidden parts and multi-point discharge scenarios; the discharge recognition module 540 uses a clustering algorithm to analyze multiple features of visible light, infrared and acoustic wave signals, and can accurately identify different types such as corona discharge, suspended discharge, and tip discharge; the image output module 550 outputs a comprehensive discharge positioning map marked with the discharge position and type, and intuitively presents the detection results in a visual interface, which is convenient for on-site inspection personnel to quickly judge the fault situation and meet the needs of rapid inspection.

[0114] In one embodiment, the signal acquisition module 520 is also used to process the acoustic wave signal through a beamforming algorithm, and determine the three-dimensional coordinates of the sound source of the acoustic wave signal in combination with a three-dimensional positioning error model; and use a frequency-weighted signal-to-noise ratio to remove electromagnetic interference in a specific frequency band of rail transit in the acoustic wave signal.

[0115] In one embodiment, the image fusion module 530 is also used to determine the sound source point of the sound wave signal and assign a weight to each sound source point based on the energy intensity and / or sound pressure level of each sound source point; according to the weight of each sound source point, a Gaussian blur algorithm is used to generate a sound source heat map; and the sound source heat map is superimposed on the dual-spectrum positioning map with a preset transparency coefficient through an image processing algorithm for fusion processing to generate the comprehensive discharge positioning map.

[0116] In one embodiment, the discharge identification module 540 is also used to identify the phase characteristics and amplitude characteristics in the phase-resolved local discharge map of the acoustic wave signal; perform time-frequency analysis on the acoustic wave signal to obtain wavelet transform coefficients; the wavelet transform coefficients characterize the frequency domain characteristics of the transient discharge of the local discharge phenomenon; extract the grayscale co-occurrence matrix texture of the visible light discharge image and the thermal distribution gradient of the infrared image; and process the phase characteristics, the amplitude characteristics, the wavelet transform coefficients, the grayscale co-occurrence matrix texture and the thermal distribution gradient through the clustering algorithm to obtain the discharge position and discharge type of the local discharge phenomenon.

[0117] In one embodiment, the discharge identification module 540 is also used to determine the discharge type of the local discharge phenomenon based on the phase characteristics, the amplitude characteristics and the frequency domain characteristics of the acoustic wave signal in the clustering results obtained by the clustering algorithm, combined with the thermal distribution gradient; determine the discharge type of the local discharge phenomenon based on the correlation between the gray level co-occurrence matrix texture and the acoustic transient characteristics of the acoustic wave signal; and use the cluster center coordinates as the discharge position of the local discharge phenomenon in the comprehensive discharge positioning map.

[0118] In one embodiment, the image output module 550 is further used to output the comprehensive discharge map, and mark the discharge position and discharge type of the local discharge phenomenon with rainbow dots in the comprehensive discharge map; different colors of the rainbow dots represent different discharge types, the position of the rainbow dots represents the discharge position, and the size of the rainbow dots is positively correlated with the energy intensity and / or sound pressure level of the sound source point.

[0119] Each module in the aforementioned partial discharge locating device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a computer device memory in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0120] In an exemplary embodiment, a partial discharge locating device is provided. The internal structure of the partial discharge locating device can be as follows: Figure 6As shown. The partial discharge locating device includes a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. The processor of the partial discharge locating device provides computing and control capabilities. The memory of the partial discharge locating device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the partial discharge locating device is used to exchange information between the processor and external devices. The communication interface of the partial discharge locating device is used to communicate with external terminals via wired or wireless communication, where the wireless communication can be achieved via Wi-Fi, a mobile cellular network, near field communication (NFC), or other technologies. When executed by the processor, the computer program implements a partial discharge locating method. The display unit of the partial discharge locating device is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the partial discharge locating device can be a touch layer covered on the display screen, or a button, trackball or touchpad provided on the housing of the partial discharge locating device.

[0121] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0122] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0123] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0124] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0125] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), quantum computing-based data processing logic devices, artificial intelligence (AI) processors, and the like.

[0126] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0127] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for locating partial discharge, characterized in that: The method comprises: Collecting a visible light image and an infrared image of a region to be measured containing a partial discharge phenomenon, and fusing the visible light image and the infrared image to generate a dual-spectrum positioning map; Collecting the sound wave signal of the area to be measured by a spiral microphone array; Fusing the dual-spectrum localization map with the sound source thermal map of the acoustic wave signal to generate a comprehensive discharge localization map; using a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image, and the acoustic wave signal, and determining a discharge location and a discharge type of the partial discharge phenomenon in the comprehensive discharge location map based on the multiple features; The comprehensive discharge location map is output, and the discharge position and the discharge type of the partial discharge phenomenon are marked in the comprehensive discharge location map.

2. The method according to claim 1, characterized in that After collecting the sound wave signal of the test area by the spiral microphone array, the method includes: Processing the acoustic wave signal using a beamforming algorithm and determining the three-dimensional coordinates of the sound source of the acoustic wave signal in combination with a three-dimensional positioning error model; The frequency-weighted signal-to-noise ratio is used to remove electromagnetic interference in the specific frequency band of rail transit in the sound wave signal.

3. The method according to claim 1, characterized in that The step of fusing the dual-spectrum positioning map with the sound source heat map of the acoustic wave signal to generate a comprehensive discharge positioning map includes: Determining the sound source points of the sound wave signal and assigning a weight to each of the sound source points based on the energy intensity and / or sound pressure level of each of the sound source points; According to the weight of each sound source point, a Gaussian fuzzy algorithm is used to generate a sound source heat map; The sound source thermal map is superimposed on the dual-spectrum positioning map with a preset transparency coefficient through an image processing algorithm for fusion processing to generate the comprehensive discharge positioning map.

4. The method according to claim 1, wherein The adopting of a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image, and the acoustic wave signal, and determining the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge location map based on the multiple features, includes: Identifying phase features and amplitude features in a phase-resolved partial discharge spectrum of the acoustic wave signal; Performing time-frequency analysis on the acoustic wave signal to obtain wavelet transform coefficients; the wavelet transform coefficients represent the frequency domain characteristics of the transient discharge of the partial discharge phenomenon; extracting a grayscale co-occurrence matrix texture of the visible light discharge image and a thermal distribution gradient of the infrared image; The phase feature, the amplitude feature, the wavelet transform coefficient, the gray-level co-occurrence matrix texture and the thermal distribution gradient are processed by the clustering algorithm to determine the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge localization map.

5. The method according to claim 4, characterized in that The step of processing the phase feature, the amplitude feature, the wavelet transform coefficient, the gray-level co-occurrence matrix texture, and the thermal distribution gradient by the clustering algorithm to determine the discharge position and discharge type of the partial discharge phenomenon in the comprehensive discharge location map includes: Determining the discharge type of the partial discharge phenomenon based on the phase characteristics, the amplitude characteristics, and the frequency domain characteristics of the acoustic wave signal in the clustering result obtained by the clustering algorithm, combined with the heat distribution gradient; determining the discharge type of the partial discharge phenomenon according to the correlation between the gray level co-occurrence matrix texture and the acoustic transient characteristics of the acoustic wave signal; The cluster center coordinates are used as the discharge positions of the local discharge phenomenon in the comprehensive discharge location map.

6. The method according to claim 1, characterized in that The outputting the comprehensive discharge location map and marking the discharge position and the discharge type of the partial discharge phenomenon in the comprehensive discharge location map includes: Output the comprehensive discharge map, and mark the discharge position and discharge type of the local discharge phenomenon with rainbow dots in the comprehensive discharge map; different colors of the rainbow dots represent different discharge types, the positions of the rainbow dots represent the discharge positions, and the sizes of the rainbow dots are positively correlated with the energy intensity and / or sound pressure level of the sound source point.

7. A partial discharge locating device, characterized in that: The device comprises: Infrared camera, used to collect infrared images of the area to be tested containing partial discharge phenomena; A visible light camera, used to collect visible light images of the area to be tested containing partial discharge phenomena; A spiral microphone array, used to collect sound wave signals from the area to be measured; a processor configured to fuse the visible light image and the infrared image to generate a dual-spectrum positioning map; and fuse the dual-spectrum positioning map with a sound source thermal map of the acoustic wave signal to generate a comprehensive discharge positioning map; and employ a clustering algorithm to respectively identify multiple features of the visible light image, the infrared image, and the acoustic wave signal, and determine a discharge location and discharge type of the partial discharge phenomenon in the comprehensive discharge positioning map based on the multiple features; A visual display screen is used to output the comprehensive discharge location map and mark the discharge position and the discharge type of the partial discharge phenomenon in the comprehensive discharge location map.

8. A partial discharge locating device, characterized in that: The device comprises: An image acquisition module is used to acquire a visible light image and an infrared image of the area to be measured containing the partial discharge phenomenon, and fuse the visible light image and the infrared image to generate a dual-spectrum positioning map; A signal acquisition module, configured to acquire the sound wave signal of the area to be measured through a spiral microphone array; An image fusion module is used to fuse the dual-spectrum positioning map with the sound source heat map of the acoustic wave signal to generate a comprehensive discharge positioning map; a discharge identification module, configured to respectively identify multiple features of the visible light image, the infrared image, and the acoustic wave signal using a clustering algorithm, and determine a discharge location and a discharge type of the partial discharge phenomenon in the comprehensive discharge location map based on the multiple features; An image output module is used to output the comprehensive discharge location map and mark the discharge position and the discharge type of the partial discharge phenomenon in the comprehensive discharge location map.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images