Intelligent partial discharge detection method and system

Abstract

This invention discloses an intelligent method and system for partial discharge detection. The method first simultaneously acquires multiple ultrasonic signals and surface infrared image sequences from the electrical equipment under test. Pulse detection, cross-channel clustering based on a sliding time window, and time-difference localization are performed on the ultrasonic signals to obtain preliminary discharge points. Based on the active periods of the ultrasonic pulses, infrared images from multiple time points are intelligently selected and fused to generate infrared images of the target surface highlighting abnormal temperature rise areas. The ultrasonic localization points and abnormal infrared regions are spatially aggregated to form suspected discharge sources with acoustic-thermal correlation. Waveform features and temperature features are extracted from the aggregated ultrasonic pulse set and infrared regions, respectively, fused, and input into a dual-channel neural network model. Finally, the model outputs the discharge type identification result and discharge intensity assessment value of the suspected discharge source. This invention significantly improves the localization accuracy and identification accuracy of partial discharge detection through deep spatiotemporal fusion and feature-level fusion.

Description

Technical Field

[0001] This invention belongs to the field of partial discharge detection technology, and particularly relates to an intelligent method and system for partial discharge detection. Background Technology

[0002] Partial discharge is an important early sign of insulation degradation in power equipment. Effective detection and identification of partial discharge is crucial for ensuring the safe operation of the power grid and preventing sudden equipment failures. Currently, the mainstream partial discharge detection technologies include pulsed current method, ultra-high frequency method, and ultrasonic method. Among these, the ultrasonic method is widely used due to its strong resistance to electromagnetic interference and ease of localization. However, relying solely on ultrasonic signals has some inherent limitations: firstly, ultrasonic signals are easily affected by the internal structure of the equipment, the propagation medium, and background noise, resulting in insufficient sensitivity for detecting minute or deep discharges; secondly, single ultrasonic localization results have certain errors and are difficult to distinguish discharge types (such as internal discharge, surface discharge, corona discharge, etc.).

[0003] Infrared thermal imaging technology provides direct morphological evidence of partial discharge by detecting abnormal temperature rises on the surface of equipment caused by discharge. However, infrared methods are usually slow to react, making it difficult to capture instantaneous discharge pulses. Furthermore, the temperature rise is easily affected by non-discharge factors such as ambient temperature, load current, and sunlight, and using them alone can easily lead to misjudgments.

[0004] While existing technologies have attempted to combine ultrasound and infrared methods, these methods are mostly simple information juxtapositions or manual comparisons, lacking in-depth spatiotemporal correlation and feature-level fusion. For example, existing technologies have failed to address the core issue of how to actively and efficiently acquire high-quality infrared image evidence in partial discharge detection, and extract features that reliably characterize the thermal effects of the discharge and are strongly correlated with the ultrasonic event. Summary of the Invention

[0005] This invention aims to overcome the above-mentioned defects and proposes an intelligent detection method and system for partial discharge. It adopts an intelligent infrared image processing strategy guided by ultrasonic signals, focusing on the active discharge period, and fusing information from multiple moments, so as to significantly improve the efficiency and value of infrared technology in the joint detection of partial discharge.

[0006] In a first aspect, the present invention provides a smart method for detecting partial discharge, comprising:

[0007] Multiple ultrasonic signals from the electrical equipment under test within a preset time period are acquired, wherein the multiple ultrasonic signals are respectively acquired by multiple sensors;

[0008] Pulse analysis is performed on the multiple ultrasonic signals to obtain ultrasonic pulse signal sequences corresponding to the multiple ultrasonic signals. Based on a preset sliding window, the ultrasonic pulse signal sequences are slid across each ultrasonic pulse signal sequence to cluster the ultrasonic pulse signals that meet the time synchronization condition, thereby obtaining at least one ultrasonic pulse signal set. In this set, an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal.

[0009] Determine the initial discharge point corresponding to the at least one set of ultrasonic pulse signals;

[0010] Based on a preset image processing strategy, the infrared image of the target surface of the power equipment under test is determined within a preset time period. Based on at least one abnormal temperature rise area in the infrared image of the target surface, each preliminary discharge point is aggregated to obtain at least one set of preliminary discharge points. Among them, one set of preliminary discharge points corresponds to one suspected discharge source, and one set of preliminary discharge points is associated with one abnormal temperature rise area.

[0011] Extract a waveform feature vector corresponding to a certain initial discharge point set and a temperature feature vector of a certain abnormal temperature rise region, and fuse the waveform feature vector and the temperature feature vector to obtain a fused feature vector;

[0012] The fused feature vector is input into a preset partial discharge identification model, and the partial discharge identification model outputs the discharge type identification result and discharge intensity evaluation value of a suspected discharge source.

[0013] In a second aspect, the present invention provides a partial discharge intelligent detection system, comprising:

[0014] The acquisition module is configured to acquire multiple ultrasonic signals from the power equipment under test within a preset time period, wherein the multiple ultrasonic signals are respectively acquired by multiple sensors;

[0015] The analysis module is configured to perform pulse analysis on the multiple ultrasonic signals according to preset signal detection rules, obtain ultrasonic pulse signal sequences corresponding to the multiple ultrasonic signals, and slide on each ultrasonic pulse signal sequence based on a preset sliding window to cluster the ultrasonic pulse signals that meet the time synchronization condition to obtain at least one ultrasonic pulse signal set, wherein an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal.

[0016] The determination module is configured to determine the initial discharge point corresponding to the at least one set of ultrasonic pulse signals;

[0017] The aggregation module is configured to determine the infrared image of the target surface of the power equipment under test within a preset time period based on a preset image processing strategy, and aggregate each preliminary discharge point based on at least one abnormal temperature rise area in the infrared image of the target surface to obtain at least one set of preliminary discharge points, wherein one set of preliminary discharge points corresponds to one suspected discharge source, and one set of preliminary discharge points is associated with one abnormal temperature rise area.

[0018] The extraction module is configured to extract a waveform feature vector corresponding to a certain set of initial discharge points and a temperature feature vector of a certain abnormal temperature rise region, and to fuse the waveform feature vector and the temperature feature vector to obtain a fused feature vector.

[0019] The output module is configured to input a certain fused feature vector into a preset partial discharge identification model, and the partial discharge identification model outputs the discharge type identification result and discharge intensity evaluation value of a certain suspected discharge source.

[0020] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the partial discharge intelligent detection method of any embodiment of the present invention.

[0021] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the steps of the partial discharge intelligent detection method of any embodiment of the present invention.

[0022] This application presents a partial discharge intelligent detection method and system. Through strict alignment of multi-channel synchronous ultrasound and infrared data, combined with a sliding window clustering algorithm based on a physical propagation model, it achieves accurate extraction and cross-sensor correlation of discharge events. Subsequently, it intelligently guides the screening of infrared keyframes during ultrasound active periods and generates enhanced diagnostic images through regional spatiotemporal fusion, effectively locating thermal anomalies. Furthermore, it utilizes a spatial aggregation mechanism of acoustic-thermal dual-modal evidence chains to cross-validate and fuse ultrasound positioning points and infrared hot zones, filtering out false signals. Based on this, it extracts multi-dimensional features from the aggregated pulse sequences and hot zones, constructing a unified fusion feature vector characterizing the physical essence of the discharge. Finally, the partial discharge identification model analyzes this vector, simultaneously outputting high-precision discharge type identification results and quantitative intensity assessments. This method significantly improves the accuracy, reliability, and intelligence level of partial discharge detection. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart of a partial discharge intelligent detection method provided in an embodiment of the present invention;

[0025] Figure 2 A structural block diagram of a partial discharge intelligent detection system provided in one embodiment of the present invention;

[0026] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Please see Figure 1 The diagram shows a flowchart of a partial discharge intelligent detection method according to this application.

[0029] like Figure 1 As shown, the intelligent partial discharge detection method specifically includes the following steps:

[0030] Step S101: Acquire multiple ultrasonic signals from the power equipment under test within a preset time period, wherein the multiple ultrasonic signals are acquired by multiple sensors.

[0031] In this step, a broadband or resonant ultrasonic sensor suitable for partial discharge detection in power equipment is selected. Its effective frequency range typically covers 40kHz to 300kHz to capture the characteristic ultrasonic signals generated by different discharge types. The sensor should possess high sensitivity, a good signal-to-noise ratio, and a certain degree of directivity.

[0032] Based on the structure, dimensions, and key monitoring areas of the power equipment under test (such as transformers, GIS, and switchgear), multiple sensors are pre-planned and installed. The deployment principle is to ensure that the sensors form a certain geometric distribution in space (such as triangles, quadrilaterals, or around key parts of the equipment) to facilitate subsequent time-of-flight positioning. The three-dimensional spatial coordinates (X, Y, Z) of each sensor need to be accurately calibrated and recorded through measurement or design drawings.

[0033] Within a preset time period, an integrated multi-channel synchronous data acquisition system is used to sample ultrasonic signals from various sensors at high speed and convert them into digital signals. Specifically, the integrated multi-channel synchronous data acquisition system includes: a synchronous acquisition card with an independent analog-to-digital converter (ADC) and sample-and-hold circuit, where each channel shares the same high-stability crystal oscillator clock source to ensure strict synchronization of the sampling clocks across all channels, with a synchronization error of less than 10 nanoseconds; a signal conditioning unit with a built-in programmable gain amplifier (PGA) and anti-aliasing filter. For each channel signal, the gain is uniformly set to 60dB, and a bandpass filter with a bandwidth of 20kHz-400kHz is configured; and a timing and triggering module integrating a GPS / BeiDou dual-mode timing unit, providing accurate UTC time synchronization signals (PPS) and absolute time information.

[0034] Step S102: Perform pulse analysis on the plurality of ultrasonic signals to obtain ultrasonic pulse signal sequences corresponding to the plurality of ultrasonic signals, and slide on each ultrasonic pulse signal sequence based on a preset sliding window to cluster the ultrasonic pulse signals that meet the time synchronization condition to obtain at least one ultrasonic pulse signal set, wherein an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal.

[0035] Specifically, the implementation of step S102 includes the following sub-steps executed sequentially:

[0036] Step S1021, Parallel preprocessing of multi-channel ultrasonic signals:

[0037] For each raw digital sequence of ultrasonic signal channel obtained in step S101, preprocessing is performed in parallel:

[0038] Bandpass filtering: A zero-phase-distortion digital bandpass filter (such as an FIR filter) is used to limit the signal of each channel to the ultrasonic frequency band of partial discharge characteristics from 40kHz to 300kHz, so as to suppress low-frequency mechanical vibration and high-frequency electromagnetic noise to the maximum extent.

[0039] Noise baseline estimation: For each filtered channel signal, the moving median method is used to estimate its dynamic noise baseline level. With a window length of 1 millisecond, the median absolute value of the signal amplitude within the window is calculated as the estimated noise level N(t) for that period.

[0040] Preliminary noise reduction: Optional, a noise reduction algorithm based on wavelet threshold can be applied to further improve the signal-to-noise ratio.

[0041] Step S1022, Pulse event detection based on adaptive threshold:

[0042] Pulse detection is performed independently for each preprocessed channel signal:

[0043] Dynamic threshold calculation: The impulse detection threshold Th(t) is dynamically calculated based on the noise baseline, for example: Where K is a preset signal-to-noise ratio coefficient (typically 3 to 5). This mechanism ensures that constant detection sensitivity is maintained even when noise fluctuates.

[0044] Threshold-crossing point identification and pulse segmentation: Identify all time points where the signal amplitude exceeds Th(t). Combine the signal amplitudes corresponding to consecutive threshold-crossing time points into a single ultrasonic pulse signal, and determine the peak occurrence time of this ultrasonic pulse signal, where the peak occurrence time is the time corresponding to the maximum signal amplitude in the ultrasonic pulse signal.

[0045] For a single channel, all the ultrasound pulse signals detected within a preset time period are arranged in ascending order of timestamps to form the ultrasound pulse signal sequence corresponding to that channel.

[0046] Step S1023, Multi-channel pulse clustering based on sliding time window:

[0047] The goal of this step is to correlate pulses generated by the same spatial discharge power source in different channel sequences.

[0048] Initialize the sliding window: Set a sliding time window W, the length of which is determined based on the maximum propagation delay of the ultrasonic wave in the device medium and the maximum sensor spacing, for example, 50 microseconds. The window slides on the global time axis in fixed steps.

[0049] Intra-window pulse collection and correlation: For each window position, execute:

[0050] Collection: From the ultrasound pulse signal sequences of all channels, identify all ultrasound pulses whose timestamps fall within the current window.

[0051] Channel deduplication: If there are multiple pulses in the same channel within the current window, usually only the one with the largest amplitude is retained to avoid multi-pulse interference within a single channel.

[0052] Forming a candidate set: The collected ultrasound pulse signals from different channels are considered as a candidate set of ultrasound pulse signals. This set is initially considered to potentially correspond to an independent discharge event.

[0053] Set validity verification and refinement: Verify each candidate set:

[0054] Quantity verification: The set must contain ultrasonic pulses from at least two sensors at different spatial locations. Sets with fewer than two ultrasonic pulses are discarded.

[0055] Time synchronization verification (core condition): Calculate the arrival time difference between all pairs of ultrasonic pulses within the set. Theoretically, ultrasonic pulses generated from the same source should have an arrival time difference less than the maximum theoretical time difference Tmax required for the pulse to propagate from the source to the two farthest sensors. If the time differences of all pulse pairs within the set satisfy this physical constraint, then it is determined that the "time synchronization condition" is met.

[0056] Output: After the above sliding window traversal and condition verification, at least one set of ultrasonic pulse signals is obtained.

[0057] Step S103: Determine the initial discharge point corresponding to the at least one set of ultrasonic pulse signals.

[0058] In this step, it is determined whether the number of ultrasonic pulse signals in a certain set of ultrasonic pulse signals is greater than a preset number threshold;

[0059] If the number exceeds a preset threshold, then select ultrasonic pulse signals from three sensors at different spatial locations from a set of ultrasonic pulse signals.

[0060] The timestamps and peak amplitudes of ultrasonic pulse signals from three sensors at different spatial positions are obtained, and the arrival time difference between any two ultrasonic pulse signals is calculated based on the pre-calibrated three-dimensional spatial coordinates of the three sensors at different spatial positions.

[0061] Based on the known propagation speed of ultrasound in the insulating medium of the equipment, the three-dimensional spatial coordinates of three sensors at different spatial positions, and the arrival time difference between any two ultrasonic pulse signals, a set of hyperbolic positioning equations is established.

[0062] The least squares estimation algorithm is used to solve the hyperbola localization equations to obtain a three-dimensional spatial coordinate corresponding to a certain set of ultrasonic pulse signals, which is used as a certain initial discharge point.

[0063] If the number is not greater than a preset threshold, a target sensor is selected from the sensors corresponding to each ultrasonic pulse signal, and the three-dimensional spatial coordinates of the target sensor are used as reference coordinates to obtain a preliminary discharge point. The target sensor is the sensor corresponding to the ultrasonic pulse signal with the largest peak amplitude.

[0064] In one specific embodiment, the total number N of ultrasound pulse signals contained in an ultrasound pulse signal set is counted, and it is determined whether the total number N of ultrasound pulse signals is greater than a preset number threshold (the value is 3).

[0065] If the number exceeds a preset threshold, then the three ultrasound pulse signals with the highest peak signal-to-noise ratio (SNR) are selected from all ultrasound pulse signals in the ultrasound pulse signal set. The SNR calculation formula is as follows: (Peak amplitude / Noise baseline of the sensor).

[0066] If the SNR requirement cannot be met, select pulses from three different sensors in descending order of peak amplitude.

[0067] Obtain the precise timestamps t1, t2, and t3 (unit: microseconds, accuracy 0.1μs) of these three ultrasonic pulse signals.

[0068] Read the three-dimensional coordinates of the three sensors corresponding to the three ultrasonic pulse signals from the calibration database: (x1, y1, z1), (x2, y2, z2), (x3, y3, z3).

[0069] Obtain the propagation speed v (in m / s) of the ultrasonic wave in the current insulating medium of the device. This speed value depends on the medium type (oil, etc.). Real-time compensation and correction are performed for (epoxy resin, etc.) and temperature.

[0070] Time difference of arrival calculation:

[0071] Using sensor 1 as a reference, calculate the relative time difference: Δt21 = t2 – t1, Δt31 = t3 – t1.

[0072] Establish the system of positioning equations:

[0073] Assuming the coordinates of the discharge point are (x, y, z), a system of equations is established based on the hyperbolic positioning principle:

[0074] ,

[0075] ,

[0076] Iterative solution:

[0077] The Taylor series expansion iterative method is used to solve the problem, and the specific steps are as follows:

[0078] Set initial iteration value Typically, the geometric center of the three sensor coordinates is taken.

[0079] In the k-th iteration, the above nonlinear equations are... Performing a first-order Taylor expansion, we obtain the linearized error equation: , Let be the Jacobian matrix, and let its elements be the partial derivatives of the distance difference with respect to the coordinates. This is the vector of differences between observed and calculated values. This is the coordinate correction amount;

[0080] Solve the linear least squares problem: , This is the transpose of the Jacobian matrix;

[0081] Updated coordinates: ;

[0082] Repeat the iteration until the norm of the coordinate correction is reached. Less than the preset convergence threshold (e.g., 1mm) or the maximum number of iterations is reached;

[0083] The final convergence point P is the coordinate of the initial discharge point.

[0084] If the number is not greater than a preset threshold, select the sensor corresponding to the ultrasonic pulse signal with the largest peak amplitude from all ultrasonic pulse signals in the ultrasonic pulse signal set as the target sensor.

[0085] If multiple ultrasonic pulse signals with similar amplitudes exist (difference <3dB), the sensor corresponding to the pulse with the highest signal-to-noise ratio should be selected first.

[0086] The three-dimensional spatial coordinates (xs, ys, zs) pre-calibrated by the target sensor are directly used as the coordinates of the initial discharge point.

[0087] In summary, the positioning strategy is first automatically determined based on the data abundance of the pulse set (such as the number of ultrasonic pulse signals). For high-quality signals, a nonlinear optimization solution based on the Time Difference of Arrival (TDOA) method is used to achieve high-precision positioning. For weak signals, the positioning is downgraded to rapid positioning with the strongest signal source as a reference, thus ensuring both high precision potential and reliability and robustness under all operating conditions. During the positioning process, the accuracy and stability of the positioning results are improved in multiple dimensions by integrating peak signal-to-noise ratio sensor optimization, sound velocity model with medium temperature compensation, and iterative optimization algorithms based on Taylor expansion and least squares. Each initial discharge point generated carries complete metadata such as coordinates, positioning type, and quality indicators. This not only provides a unified benchmark for accurate spatial correlation with the infrared abnormal temperature rise area, but also the pulse statistical characteristics and positioning reliability contained therein become key dimensions for subsequent multimodal feature fusion. In conclusion, this step efficiently transforms the "temporal accuracy" advantage of ultrasonic detection into a "spatial accuracy" advantage through intelligent algorithms, becoming a core bridge connecting front-end signal processing and back-end intelligent diagnosis, significantly improving the spatial resolution, environmental adaptability, and decision support capabilities of the entire detection system.

[0088] Step S104: Based on a preset image processing strategy, determine the infrared image of the target surface of the power equipment under test within a preset time period, and aggregate each preliminary discharge point based on at least one abnormal temperature rise area in the infrared image of the target surface to obtain at least one set of preliminary discharge points, wherein one set of preliminary discharge points corresponds to one suspected discharge source, and one set of preliminary discharge points is associated with one abnormal temperature rise area.

[0089] In this step, a sequence of surface infrared images of the electrical equipment under test is acquired within a preset time period;

[0090] The target ultrasonic pulse signal sequence with the most ultrasonic pulse signals is selected from various ultrasonic pulse signal sequences, and the candidate surface infrared image corresponding to the target peak occurrence time is selected from the surface infrared image sequence. The target peak occurrence time is the peak occurrence time of each target ultrasonic pulse signal in the target ultrasonic pulse signal sequence.

[0091] The infrared images of each candidate surface are sorted according to the chronological order to obtain the infrared image sequence of the candidate surface, and the image similarity between two adjacent infrared images of the candidate surface in the infrared image sequence is calculated.

[0092] Select the first candidate surface infrared image and the second candidate surface infrared image corresponding to the maximum image similarity among the various image similarities, and identify the first candidate temperature rise region corresponding to the first candidate surface infrared image and the second candidate temperature rise region corresponding to the second candidate surface infrared image based on the temperature field distribution in the first candidate surface infrared image and the second candidate temperature rise region corresponding to the second candidate surface infrared image.

[0093] Spatial registration is performed on the first candidate temperature rise region and the second candidate temperature rise region. After spatial registration, the percentage of overlapping area between the first candidate temperature rise region and the second candidate temperature rise region is calculated, and it is determined whether the percentage of overlapping area is greater than the preset contour similarity threshold.

[0094] If the overlap area ratio is greater than the preset contour similarity threshold, the first candidate temperature rise region and the second candidate temperature rise region are merged to obtain the fused candidate temperature rise region.

[0095] Calculate the first average temperature of the fusion candidate temperature rise region in the infrared image of the first candidate surface and the second average temperature of the fusion candidate temperature rise region in the infrared image of the second candidate surface, and take the maximum value of the first average temperature and the second average temperature as the reference temperature rise value of the fusion candidate temperature rise region.

[0096] Based on the position and outline of the candidate temperature rise region in the infrared images of the first and second candidate surfaces, a preset pseudo-color mapping table is used to map the reference temperature rise value into a unified background infrared image to obtain the target surface infrared image. The background infrared image is the average temperature field image of the surface of the power equipment under test within a preset time period.

[0097] If the overlap area ratio is not greater than the preset contour similarity threshold, the first candidate temperature rise region and the second candidate temperature rise region are determined as two independent temperature rise regions to be analyzed; the first target average temperature in the first candidate temperature rise region and the second target average temperature in the second candidate temperature rise region are calculated; based on the position and contour of the first candidate temperature rise region in the infrared image of the first candidate surface and the position and contour of the second candidate surface infrared image in the infrared image of the first candidate surface, the first target average temperature and the second target average temperature are mapped to a unified background infrared image using a preset pseudo-color mapping table to obtain the target surface infrared image.

[0098] Specifically, based on the position and contour of the candidate temperature rise region in the first and second candidate surface infrared images, a preset pseudo-color mapping table is used to map the reference temperature rise value into a unified background infrared image, resulting in the target surface infrared image including:

[0099] Calculate the pixel-level average temperature of all surface infrared images in the surface infrared image sequence within a preset time period to generate a time-averaged temperature field image, which serves as a unified background infrared image. For fusion candidate temperature rise regions, obtain the set of contour pixel coordinates of the fusion candidate temperature rise regions in the first and second candidate surface infrared images. Establish a pseudo-color mapping table from reference temperature rise values ​​to display colors. Directly map the set of contour pixel coordinates corresponding to the fusion candidate temperature rise regions to the corresponding coordinate positions in the background infrared image. Based on the reference temperature rise values ​​of the fusion candidate temperature rise regions, determine the fill color through the pseudo-color mapping table and draw it in the background infrared image in a semi-transparent overlay manner to obtain an image that fuses the color markers of the fusion candidate temperature rise regions and the background temperature field, i.e., the target surface infrared image.

[0100] In one specific embodiment, an infrared thermal imager, pre-synchronized with the ultrasonic acquisition system, acquires a sequence of infrared images of the surface of the device under test within a preset time period. Each frame of the image carries a microsecond-level timestamp aligned with the ultrasonic data.

[0101] Analyze all ultrasonic pulse signal sequences to identify the target ultrasonic pulse signal sequence with the most ultrasonic pulse signals, which corresponds to the period of most active discharge.

[0102] Extract the peak occurrence time of all pulses in the ultrasonic pulse signal sequence of the target. Using this timestamp as an index, extract the candidate surface infrared image at the corresponding time from the infrared image sequence. If the timestamps do not match perfectly, select the frame with the smallest time difference.

[0103] The initially selected candidate images are sorted by time to form a sequence of candidate surface infrared images. The structural similarity index (SSIM) combined with temperature histogram correlation is used to calculate the image similarity between adjacent frames. The pair of images with the highest similarity is selected as the first candidate surface infrared image and the second candidate surface infrared image.

[0104] For the selected first candidate surface infrared image and the second candidate surface infrared image, an adaptive threshold segmentation algorithm based on local temperature statistics is applied to identify pixel connected regions that are significantly higher than the surrounding background (e.g., more than 3 times the standard deviation of the background temperature), which are then used as the first candidate temperature rise region and the second candidate temperature rise region.

[0105] By utilizing inherent, invariant feature points (such as bolts and corner points of signs) on the surface of the electrical equipment under test, existing feature matching algorithms are used to perform sub-pixel-level spatial registration of the first and second candidate temperature rise regions. Under a unified coordinate system, the intersection-over-union ratio (IoU) of the two candidate regions is calculated as the percentage of overlapping area.

[0106] If the overlap area ratio is greater than the preset contour similarity threshold (e.g., 0.6), it is determined to be the same heat source at two different times, and the two region contours are merged into a fusion candidate temperature rise region.

[0107] For the merged candidate temperature rise regions, the average temperature of the merged candidate temperature rise region in the two original images is calculated separately, and the maximum value is taken as the reference temperature rise value for that region to capture the peak effect of the heat source. For the unmerged independent regions, the first target average temperature and the second target average temperature are calculated separately.

[0108] Calculate the pixel-level time average of all infrared image sequences within the entire preset time period to generate a time-averaged temperature field image reflecting the overall steady-state heating of the device, which serves as the background.

[0109] Map the outlines of all identified regions (merged candidate temperature rise regions or unmerged independent regions) to the same coordinate system of the background reference map.

[0110] Based on the reference temperature rise value of each region, or the first target average temperature and the second target average temperature, it is converted into the corresponding color value through a preset rainbow color or iron red pseudo-color mapping table.

[0111] Using alpha blending (transparency overlay) technology, the colored areas are rendered onto the background reference image in a semi-transparent manner (e.g., 50% transparency) to generate the final infrared image of the target surface. This image simultaneously preserves background temperature field information and prominent areas of abnormal temperature rise.

[0112] In this embodiment, firstly, guided by the activity of ultrasonic signals, key frames highly correlated with discharge events are intelligently screened from massive infrared images, achieving data dimensionality reduction and target focusing, greatly improving processing efficiency. Secondly, through cross-time frame candidate region comparison and fusion, real and continuous discharge heat sources and transient interference are effectively identified, enhancing the robustness of temperature rise region identification. Finally, the generated target surface infrared image is not a simple presentation of the original data, but a "thematic diagnostic map" that integrates temporal dimension information (multi-time verification) and spatial enhancement features (highlighting, pseudo-color quantization). It not only intuitively and clearly marks the location and relative severity of abnormal temperature rise, but also provides high-quality, target-specific input for subsequent cross-modal spatial fusion through the inherent temporal correlation with ultrasonic data, thereby solving as much as possible the technical problems of complex backgrounds, confused heat sources, and difficulty in directly correlating with transient electrical signals in traditional infrared inspection.

[0113] Furthermore, after temperature calibration and noise filtering of the infrared image of the target surface, an adaptive threshold segmentation and region growing algorithm is used to identify at least one abnormal temperature rise region, and the contour polygon of each abnormal temperature rise region is extracted.

[0114] Establish a spatial mapping relationship between the ultrasonic sensor coordinate system and the infrared image pixel coordinate system, and map the three-dimensional coordinates of each initial discharge point to the infrared image pixel coordinate system to obtain the projected coordinates of each initial discharge point.

[0115] Determine whether a certain projected coordinate of a certain initial discharge point is located inside the outline polygon of any abnormal temperature rise region;

[0116] If a certain projected coordinate is located inside the outline polygon of a certain abnormal temperature rise region, then a certain initial discharge point is assigned to the set of a certain initial discharge point corresponding to a certain abnormal temperature rise region.

[0117] If a certain projected coordinate is not located inside the contour polygon of any abnormal temperature rise area, then calculate the shortest distance from the certain projected coordinate to the boundary of each contour polygon, and determine whether the shortest distance is less than the preset spatial tolerance threshold.

[0118] If the shortest distance is less than the preset spatial tolerance threshold, then a certain initial discharge point will be assigned to the set of initial discharge points corresponding to the nearest abnormal temperature rise area.

[0119] If the shortest distance is not less than the preset spatial tolerance threshold, a certain initial discharge point will be recorded as an isolated discharge event and output to the abnormal event log for manual review.

[0120] In another specific embodiment, Otsu's method is used for adaptive threshold segmentation, and combined with existing region growing algorithms, at least one abnormal temperature rise region is accurately extracted, and a list of vertex coordinates of its contour polygon is obtained.

[0121] A mapping model is established from the three-dimensional coordinate system of the ultrasonic sensor to the two-dimensional coordinate system of the infrared image pixels through pre-completed joint calibration. This mapping model is typically a perspective transformation matrix or polynomial model that includes rotation, translation, and scaling.

[0122] Using this mapping model, the three-dimensional coordinates (x, y, z) of each preliminary discharge point obtained in step S103 are converted into projected coordinates (u, v) on the infrared image;

[0123] Use the ray method or polygon filling algorithm to determine whether each projected coordinate (u, v) is located inside the contour polygon of any abnormal temperature rise region;

[0124] For points outside the contour, calculate their Euclidean shortest distance to the boundary of each contour polygon (which consists of a series of line segments).

[0125] If a certain projected coordinate is located inside the contour polygon of a certain abnormal temperature rise region, it is directly assigned to the set of preliminary discharge points corresponding to that contour.

[0126] If a certain projected coordinate is not located inside the contour polygon of any abnormal temperature rise region, and the shortest distance is less than the spatial tolerance threshold (for example, the image pixel distance corresponding to an actual spatial distance of 5 cm), then it is assigned to the set of preliminary discharge points corresponding to the nearest region.

[0127] If a certain projected coordinate is not located inside the outline polygon of any abnormal temperature rise region, and the shortest distance is not less than the spatial tolerance threshold, then the initial discharge point is marked as an isolated discharge event, all its information is recorded and stored in the abnormal event log, awaiting further manual analysis.

[0128] The final output includes several sets of preliminary discharge points, each set is uniquely associated with an abnormal temperature rise region, and they all point to a suspected discharge source.

[0129] In this embodiment, by using coordinate system 1 and mapping, the ultrasonic positioning points calculated based on time difference are accurately projected onto the infrared feature image. Through the dual criteria of "internal inclusion" and "proximity constraint", this strategy intelligently aggregates multiple ultrasonic positioning points that may be generated by the same discharge source and are slightly discrete in space under the same abnormal thermal region. This effectively filters out false points or error points caused by sound wave reflection, refraction or noise in ultrasonic positioning, significantly improving the confidence and clarity of discharge source positioning. Finally, each set of preliminary discharge points output is a highly credible suspected discharge source that has been cross-verified by acoustic and thermal evidence, providing a clean, structured and physically meaningful input for subsequent feature extraction and pattern recognition.

[0130] Step S105: Extract a waveform feature vector corresponding to a certain initial discharge point set and a temperature feature vector of a certain abnormal temperature rise region, and fuse the waveform feature vector and the temperature feature vector to obtain a fused feature vector.

[0131] In this step, a multi-dimensional waveform analysis is performed on all ultrasonic pulse signals corresponding to a certain initial discharge point set to extract a set of time-domain waveform features and frequency-domain waveform features to form a certain waveform feature sub-vector. The time-domain waveform features include pulse amplitude statistical features, pulse time interval statistical features, and pulse energy features. The frequency-domain waveform features include the spectral centroid, spectral width, and energy proportion of a specific frequency band extracted by fast Fourier transform.

[0132] Temperature field analysis is performed on the infrared image region corresponding to a certain abnormal temperature rise area. A set of geometric features and temperature field features are extracted to form a certain temperature feature sub-vector. Among them, the geometric features include the contour area, contour perimeter and contour aspect ratio of the abnormal temperature rise area, and the temperature field features include temperature gradient features and temperature rise statistical features.

[0133] A waveform feature vector and a temperature feature vector are standardized to eliminate the influence of dimensions. The standardized waveform feature vector and the temperature feature vector are then concatenated to obtain a fused feature vector.

[0134] In one specific embodiment, for a certain set of preliminary discharge points (corresponding to a suspected discharge source), this set is composed of multiple sets of ultrasonic pulse signals, and each set of ultrasonic pulse signals contains multiple ultrasonic pulse signals from different sensors, as specifically implemented as follows:

[0135] Step S1051, acquire all associated pulse signals:

[0136] From the initial set of discharge points, obtain the set of all ultrasonic pulse signals contained therein;

[0137] Extract all ultrasonic pulse signals (including timestamps, peak amplitudes, energy, and complete time-domain waveform segments of each ultrasonic pulse signal) from these pulse signal sets.

[0138] Count the total number of ultrasound pulse signals Total number of sensors and the number of pulse signal sets .

[0139] Step S1052, Overall statistical analysis of time-domain waveform characteristics:

[0140] Pulse amplitude statistical characteristics: Calculate all The mean of the peak amplitudes of the ultrasonic pulse signals Standard deviation Maximum value kurtosis and the skewness of the amplitude distribution ;

[0141] Statistical characteristics of pulse time intervals:

[0142] Sort the pulses from all pulse signal sets by global timestamp;

[0143] Calculate the time interval sequence of adjacent pulses, and then calculate the mean of that sequence. Standard deviation ISI std and coefficient of variation ;

[0144] Pulse energy statistical characteristics:

[0145] The energy of each ultrasonic pulse signal is calculated by integral of the waveform squared.

[0146] Calculate the total energy of all ultrasound pulse signals Average energy and energy standard deviation ;

[0147] The average energy of each ultrasonic pulse signal set is calculated, and then the standard deviation of the average energy of these sets is calculated to characterize the energy fluctuation of different discharge events.

[0148] Step S1053, Comprehensive extraction of frequency domain waveform features:

[0149] Representative sample selection: From all ultrasound pulse signals, select the top M pulses ranked by peak amplitude (e.g., ...). () is used as a representative ultrasonic pulse signal.

[0150] Spectrum analysis: Perform a Fast Fourier Transform (FFT) on the time-domain waveform segment of each representative ultrasound pulse signal to obtain its amplitude spectrum.

[0151] Spectral characteristic calculation:

[0152] Spectral centroid SC: Calculate the spectral centroid of each ultrasound pulse signal and then take the average of these centroids. and standard deviation ;

[0153] Spectrum Width (BW): Calculate the -3dB bandwidth of each pulse spectrum and take the average value. ;

[0154] Energy percentage in specific frequency bands: Calculate the ratio of energy in two key frequency bands (e.g., 40-80kHz and 150-300kHz) to the total spectral energy. Calculate the average energy percentage of these pulses in each of the two frequency bands. ;

[0155] Dominant frequency distribution: Identify the frequency component with the largest amplitude (dominant frequency) in the spectrum of each pulse, and count the proportion of pulses whose dominant frequencies fall into the two key frequency bands mentioned above;

[0156] Spectral consistency characteristics: Calculate the average correlation coefficient among all representative pulse spectra to characterize the spectral similarity of different discharge events;

[0157] Arrange the above features in order to form waveform feature sub-vectors.

[0158] Step S1054, for a specific abnormal temperature rise region (from the infrared image of the target surface) associated with the same set of initial discharge points, the following analysis is performed:

[0159] Region and background definition:

[0160] Based on the contour polygon of the region, the region range is precisely defined in the infrared image and raw temperature data of the target surface;

[0161] Define a ring-shaped background region (e.g., a ring extending 10-20 pixels outward from the outer edge of the defined area) around this region, and calculate its average temperature as the ambient background temperature. ;

[0162] Geometric feature calculation:

[0163] Outline area: The total number of pixels inside the outline polygon, converted into the actual physical area;

[0164] Perimeter of the outline: The sum of the lengths of all sides of the outline polygon;

[0165] Aspect ratio: The ratio of the longer side to the shorter side of the smallest bounding rectangle of the region;

[0166] Contour complexity: defined as (twice the contour perimeter) / ( The area of ​​the outline is used to describe the degree of deviation of the outline from the circle.

[0167] Temperature field characteristics calculation:

[0168] Temperature gradient features: The Sobel operator is used to calculate the magnitude and direction of the temperature gradient for each pixel within the region, and the maximum value of the gradient magnitude is taken. and average ;

[0169] Statistical characteristics of temperature rise:

[0170] Regional average temperature rise: , It is the arithmetic mean of the temperature values ​​of all pixels inside the polygonal outline of the abnormal temperature rise region.

[0171] Maximum temperature rise in the region: , The maximum temperature value among all pixels inside the polygonal outline of the abnormal temperature rise region;

[0172] Temperature rise standard deviation: The standard deviation of temperature rise for each pixel within the region;

[0173] Arrange the above features in order to form a temperature feature sub-vector.

[0174] Step S1055: Perform Z-score normalization on the waveform feature vector and the temperature feature vector respectively, and then concatenate the normalized waveform feature vector with the temperature feature vector to form a high-dimensional fused feature vector.

[0175] If the dimensionality of the fused feature vector is too high, principal component analysis (PCA) or an autoencoder can be used to reduce the dimensionality, preserving the main information while reducing computational complexity.

[0176] In summary, the method in this embodiment aggregates and analyzes all pulse signals from multiple ultrasonic pulse signal sets corresponding to the initial discharge point set to extract waveform features that comprehensively reflect the overall discharge mode of the discharge source. Simultaneously, it extracts the geometric and temperature field features of the associated abnormal temperature rise region, providing supplementary information from a thermal effect perspective. This cross-modal feature fusion strategy achieves, for the first time, a joint digital characterization of the acoustic and thermal physical properties of the same discharge source in partial discharge detection. The resulting fused feature vector not only includes the time-frequency characteristics and energy release patterns of the discharge but also integrates the spatial distribution and intensity information of the thermal effect, providing rich and complementary high-dimensional feature inputs for subsequent intelligent recognition models. This significantly improves the model's ability to distinguish different discharge types (such as corona discharge, surface discharge, and internal discharge) and makes the assessment of discharge intensity more accurate and reliable, thus providing crucial decision-making basis for the precise diagnosis and risk assessment of power equipment.

[0177] Step S106: Input the fused feature vector into a preset partial discharge identification model, and the partial discharge identification model outputs the discharge type identification result and discharge intensity evaluation value of a suspected discharge source.

[0178] In this step, a dual-channel deep neural network is used as the preset partial discharge identification model. This network includes:

[0179] Feature preprocessing layer: Receives the fused feature vector output from step S105, and first performs dimension adaptation and preliminary feature interaction through a fully connected layer;

[0180] Dual-branch feature extraction module:

[0181] Ultrasonic Feature Branch: A multilayer perceptron (MLP) specifically processes the portion of the fused feature vector that belongs to the waveform feature subvector, learning the acoustic pattern of the discharge;

[0182] Infrared Feature Branch: Another multilayer perceptron (MLP) specifically processes the part of the fused feature vector that belongs to the temperature feature subvector, and learns the thermal mode of discharge;

[0183] Feature fusion and decision module: This module concatenates the high-level abstract features from the two branches, then passes them through a shared network containing multiple fully connected layers and Dropout layers, and finally connects them to two parallel output heads.

[0184] Type recognition head: A softmax classification layer that outputs the probability distribution of each preset discharge type. K represents the number of discharge types (e.g., K=4, corresponding to "corona discharge", "surface discharge", "internal discharge" and "other discharge").

[0185] Intensity assessment head: A linear regression layer (or ordered classification layer) outputs a continuous discharge intensity assessment value I (e.g., range 0-100) or a discrete intensity level (e.g., "weak", "moderate", "strong", "critical").

[0186] In summary, the method of this application achieves highly sensitive capture and intelligent cross-channel association of discrete discharge events from complex noise backgrounds through multi-channel ultrasonic acquisition and sliding window clustering based on a physical propagation model, providing clean and structured discharge event units for subsequent analysis. Secondly, by employing an infrared image selection and fusion strategy guided by ultrasonic active periods, a "thematic diagnostic map" is generated that highlights abnormal temperature rises related to discharge, solving the problem of dispersed infrared detection targets and difficulty in associating transient events. Furthermore, through spatial aggregation of acoustic and thermal dual-modal evidence chains, ultrasonic positioning points and infrared thermal regions are accurately associated and cross-validated, effectively filtering out spurious signals and positioning errors in single-modal detection and significantly improving the confidence of discharge source positioning. Then, by extracting multi-level and multi-dimensional statistical and physical features from the aggregated multi-event ultrasonic pulse sequence and associated infrared regions and fusing them into a unified feature vector, a fusion feature capable of comprehensively characterizing the essence of discharge is constructed. Through intelligent analysis of the fusion feature using a partial discharge identification model, high-precision automatic identification of discharge types (such as corona, surface, and internal discharge) and quantitative assessment of discharge intensity are achieved. Overall, this invention not only overcomes the limitations of single detection methods, but also achieves earlier, more accurate, and more intuitive perception and diagnosis of the insulation status of power equipment through deep information fusion and intelligent analysis, providing strong technical support for predictive maintenance and safe operation of power systems.

[0187] Please see Figure 2 The diagram shows a structural block diagram of a partial discharge intelligent detection system according to this application.

[0188] like Figure 2 As shown, the partial discharge intelligent detection system 200 includes an acquisition module 210, an analysis module 220, a determination module 230, an aggregation module 240, an extraction module 250, and an output module 260.

[0189] The acquisition module 210 is configured to acquire multiple ultrasonic signals from the power equipment under test within a preset time period, wherein the multiple ultrasonic signals are respectively acquired by multiple sensors; the analysis module 220 is configured to perform pulse analysis on the multiple ultrasonic signals according to preset signal detection rules to obtain ultrasonic pulse signal sequences corresponding to the multiple ultrasonic signals, and to slide a preset sliding window on each ultrasonic pulse signal sequence to cluster the ultrasonic pulse signals that meet the time synchronization condition to obtain at least one ultrasonic pulse signal set, wherein an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal; the determination module 230 is configured to determine the preliminary discharge point corresponding to the at least one ultrasonic pulse signal set; and the aggregation module 240 is configured to determine the power equipment under test based on a preset image processing strategy. The system uses infrared images of the target surface of power equipment within a preset time period. Based on at least one abnormal temperature rise region in the infrared images of the target surface, it aggregates each preliminary discharge point to obtain at least one set of preliminary discharge points. Each set of preliminary discharge points corresponds to a suspected discharge source, and each set of preliminary discharge points is associated with an abnormal temperature rise region. An extraction module 250 is configured to extract a waveform feature sub-vector corresponding to a set of preliminary discharge points and a temperature feature sub-vector of an abnormal temperature rise region. The waveform feature sub-vector and the temperature feature sub-vector are then fused to obtain a fused feature vector. An output module 260 is configured to input the fused feature vector into a preset partial discharge identification model. The partial discharge identification model outputs a discharge type identification result and a discharge intensity evaluation value for a suspected discharge source.

[0190] It should be understood that Figure 2 The modules and references described in the document Figure 1 The steps described in the text correspond to those in the method described above. Therefore, the operations, features, and corresponding technical effects described above also apply to the method described in the text. Figure 2 The various modules in the document will not be described in detail here.

[0191] In other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the partial discharge intelligent detection method in any of the above method embodiments.

[0192] In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows:

[0193] Multiple ultrasonic signals from the electrical equipment under test within a preset time period are acquired, wherein the multiple ultrasonic signals are respectively acquired by multiple sensors;

[0194] Pulse analysis is performed on the multiple ultrasonic signals to obtain ultrasonic pulse signal sequences corresponding to the multiple ultrasonic signals. Based on a preset sliding window, the ultrasonic pulse signal sequences are slid across each ultrasonic pulse signal sequence to cluster the ultrasonic pulse signals that meet the time synchronization condition, thereby obtaining at least one ultrasonic pulse signal set. In this set, an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal.

[0195] Determine the initial discharge point corresponding to the at least one set of ultrasonic pulse signals;

[0196] Based on a preset image processing strategy, the infrared image of the target surface of the power equipment under test is determined within a preset time period. Based on at least one abnormal temperature rise area in the infrared image of the target surface, each preliminary discharge point is aggregated to obtain at least one set of preliminary discharge points. Among them, one set of preliminary discharge points corresponds to one suspected discharge source, and one set of preliminary discharge points is associated with one abnormal temperature rise area.

[0197] Extract a waveform feature vector corresponding to a certain initial discharge point set and a temperature feature vector of a certain abnormal temperature rise region, and fuse the waveform feature vector and the temperature feature vector to obtain a fused feature vector;

[0198] The fused feature vector is input into a preset partial discharge identification model, and the partial discharge identification model outputs the discharge type identification result and discharge intensity evaluation value of a suspected discharge source.

[0199] Computer-readable storage media may include a stored program area and a stored data area, wherein the stored program area may store an operating system and an application program required for at least one function; the stored data area may store data created based on the use of the partial discharge intelligent detection system, etc. Furthermore, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the partial discharge intelligent detection system via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0200] Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 3As shown, the device includes a processor 310 and a memory 320. The electronic device may also include an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 can be connected via a bus or other means. Figure 3 Taking a bus connection as an example, the memory 320 is the computer-readable storage medium described above. The processor 310 executes various server functions and data processing by running non-volatile software programs, instructions, and modules stored in the memory 320, thereby implementing the partial discharge intelligent detection method described in the above embodiment. The input device 330 can receive input digital or character information and generate key signal inputs related to user settings and function control of the partial discharge intelligent detection system. The output device 340 may include a display screen or other display device.

[0201] The aforementioned electronic device can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0202] In one implementation, the above-described electronic device is applied to a partial discharge intelligent detection system as a client, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to:

[0203] Multiple ultrasonic signals from the electrical equipment under test within a preset time period are acquired, wherein the multiple ultrasonic signals are respectively acquired by multiple sensors;

[0204] Pulse analysis is performed on the multiple ultrasonic signals to obtain ultrasonic pulse signal sequences corresponding to the multiple ultrasonic signals. Based on a preset sliding window, the ultrasonic pulse signal sequences are slid across each ultrasonic pulse signal sequence to cluster the ultrasonic pulse signals that meet the time synchronization condition, thereby obtaining at least one ultrasonic pulse signal set. In this set, an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal.

[0205] Determine the initial discharge point corresponding to the at least one set of ultrasonic pulse signals;

[0206] Based on a preset image processing strategy, the infrared image of the target surface of the power equipment under test is determined within a preset time period. Based on at least one abnormal temperature rise area in the infrared image of the target surface, each preliminary discharge point is aggregated to obtain at least one set of preliminary discharge points. Among them, one set of preliminary discharge points corresponds to one suspected discharge source, and one set of preliminary discharge points is associated with one abnormal temperature rise area.

[0207] Extract a waveform feature vector corresponding to a certain initial discharge point set and a temperature feature vector of a certain abnormal temperature rise region, and fuse the waveform feature vector and the temperature feature vector to obtain a fused feature vector;

[0208] The fused feature vector is input into a preset partial discharge identification model, and the partial discharge identification model outputs the discharge type identification result and discharge intensity evaluation value of a suspected discharge source.

[0209] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0210] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent detection of partial discharge, characterized in that, include: Multiple ultrasonic signals from the electrical equipment under test within a preset time period are acquired, wherein the multiple ultrasonic signals are respectively acquired by multiple sensors; Pulse analysis is performed on the multiple ultrasonic signals to obtain ultrasonic pulse signal sequences corresponding to the multiple ultrasonic signals. Based on a preset sliding window, the ultrasonic pulse signal sequences are slid across each ultrasonic pulse signal sequence to cluster the ultrasonic pulse signals that meet the time synchronization condition, thereby obtaining at least one ultrasonic pulse signal set. In this set, an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal. Determine the initial discharge point corresponding to the at least one set of ultrasonic pulse signals; Based on a preset image processing strategy, the infrared image of the target surface of the power equipment under test is determined within a preset time period. Based on at least one abnormal temperature rise area in the infrared image of the target surface, each preliminary discharge point is aggregated to obtain at least one set of preliminary discharge points. Among them, one set of preliminary discharge points corresponds to one suspected discharge source, and one set of preliminary discharge points is associated with one abnormal temperature rise area. Extract a waveform feature vector corresponding to a certain initial discharge point set and a temperature feature vector of a certain abnormal temperature rise region, and fuse the waveform feature vector and the temperature feature vector to obtain a fused feature vector; The fused feature vector is input into a preset partial discharge identification model, and the partial discharge identification model outputs the discharge type identification result and discharge intensity evaluation value of a suspected discharge source.

2. The intelligent partial discharge detection method according to claim 1, characterized in that, Determining the preliminary discharge point corresponding to the at least one set of ultrasonic pulse signals includes: Determine whether the number of ultrasonic pulse signals in a certain set of ultrasonic pulse signals is greater than a preset threshold. If the number exceeds a preset threshold, then select ultrasonic pulse signals from three different spatial location sensors from a certain set of ultrasonic pulse signals; The timestamps and peak amplitudes of ultrasonic pulse signals from three sensors at different spatial positions are obtained, and the arrival time difference between any two ultrasonic pulse signals is calculated based on the pre-calibrated three-dimensional spatial coordinates of the three sensors at different spatial positions. Based on the known propagation speed of ultrasound in the insulating medium of the equipment, the three-dimensional spatial coordinates of three sensors at different spatial positions, and the arrival time difference between any two ultrasonic pulse signals, a set of hyperbolic positioning equations is established. The least squares estimation algorithm is used to solve the hyperbola localization equations to obtain a three-dimensional spatial coordinate corresponding to a certain ultrasonic pulse signal set, which is used as a certain initial discharge point; If the number is not greater than a preset threshold, a target sensor is selected from the sensors corresponding to each ultrasonic pulse signal, and the three-dimensional spatial coordinates of the target sensor are used as reference coordinates to obtain a preliminary discharge point. The target sensor is the sensor corresponding to the ultrasonic pulse signal with the largest peak amplitude.

3. The intelligent partial discharge detection method according to claim 1, characterized in that, The method of determining the target surface infrared image of the power equipment under test within a preset time period based on a preset image processing strategy includes: Acquire the surface infrared image sequence of the power equipment under test within a preset time period; Among the various ultrasonic pulse signal sequences, the target ultrasonic pulse signal sequence with the largest number of ultrasonic pulse signals is selected, and the candidate surface infrared image corresponding to the target peak occurrence time is selected from the surface infrared image sequence, wherein the target peak occurrence time is the peak occurrence time of each target ultrasonic pulse signal in the target ultrasonic pulse signal sequence; The infrared images of each candidate surface are sorted according to the chronological order to obtain a sequence of infrared images of candidate surfaces, and the image similarity between two adjacent infrared images of candidate surfaces in the sequence is calculated. Select the first candidate surface infrared image and the second candidate surface infrared image corresponding to the maximum image similarity among the various image similarities, and identify the first candidate temperature rise region corresponding to the first candidate surface infrared image and the second candidate temperature rise region corresponding to the second candidate surface infrared image based on the temperature field distribution in the first candidate surface infrared image and the second candidate temperature rise region corresponding to the second candidate surface infrared image. Spatial registration is performed on the first candidate temperature rise region and the second candidate temperature rise region. After spatial registration, the percentage of overlapping area between the first candidate temperature rise region and the second candidate temperature rise region is calculated, and it is determined whether the percentage of overlapping area is greater than a preset contour similarity threshold. If the percentage of overlapping area is greater than a preset contour similarity threshold, then the first candidate temperature rise region and the second candidate temperature rise region are merged to obtain a fused candidate temperature rise region. Calculate the first average temperature of the fusion candidate temperature rise region in the infrared image of the first candidate surface, and the second average temperature of the fusion candidate temperature rise region in the infrared image of the second candidate surface, and take the maximum value of the first average temperature and the second average temperature as the reference temperature rise value of the fusion candidate temperature rise region. Based on the position and contour of the fusion candidate temperature rise region in the infrared images of the first and second candidate surfaces, the reference temperature rise value is mapped to a unified background infrared image using a preset pseudo-color mapping table to obtain the infrared image of the target surface. The background infrared image is the average temperature field image of the surface of the power equipment under test within the preset time period.

4. The intelligent partial discharge detection method according to claim 3, characterized in that, After determining whether the percentage of overlapping area is greater than a preset contour similarity threshold, the method further includes: If the percentage of the overlapping area is not greater than the preset contour similarity threshold, then the first candidate temperature rise region and the second candidate temperature rise region are determined as two independent temperature rise regions to be analyzed. Calculate the first target average temperature in the first candidate temperature rise region and the second target average temperature in the second candidate temperature rise region; Based on the position and contour of the first candidate temperature rise region in the infrared image of the first candidate surface and the position and contour of the second candidate surface infrared image in the infrared image of the first candidate surface, a preset pseudo-color mapping table is used to map the average temperature of the first target and the average temperature of the second target into a unified background infrared image to obtain the infrared image of the target surface.

5. The intelligent partial discharge detection method according to claim 3, characterized in that, The step of mapping the reference temperature rise value to a unified background infrared image using a preset pseudo-color mapping table based on the position and contour of the fused candidate temperature rise region in the first and second candidate surface infrared images to obtain the target surface infrared image includes: Calculate the pixel-level average temperature of all surface infrared images in the surface infrared image sequence within the preset time period, and generate a time-averaged temperature field image as a unified background infrared image; For the fusion candidate temperature rise region, obtain the set of contour pixel coordinates of the fusion candidate temperature rise region in the infrared image of the first candidate surface and the infrared image of the second candidate surface; Establish a pseudo-color mapping table from reference temperature rise value to display color; The set of contour pixel coordinates corresponding to the fusion candidate temperature rise region is directly mapped to the corresponding coordinate position of the background infrared image. Based on the reference temperature rise value of the fusion candidate temperature rise region, the fill color is determined by the pseudo-color mapping table and drawn in the background infrared image in a semi-transparent overlay manner to obtain an image that fuses the color marker of the fusion candidate temperature rise region and the background temperature field, i.e., the infrared image of the target surface.

6. The intelligent partial discharge detection method according to claim 1, characterized in that, The aggregation of each preliminary discharge point based on at least one abnormal temperature rise region in the infrared image of the target surface to obtain at least one set of preliminary discharge points includes: After temperature calibration and noise filtering of the infrared image of the target surface, an adaptive threshold segmentation and region growing algorithm is used to identify at least one abnormal temperature rise region, and the contour polygon of each abnormal temperature rise region is extracted. Establish a spatial mapping relationship between the ultrasonic sensor coordinate system and the infrared image pixel coordinate system, and map the three-dimensional coordinates of each initial discharge point to the infrared image pixel coordinate system to obtain the projected coordinates of each initial discharge point. Determine whether a certain projected coordinate of a certain initial discharge point is located inside the outline polygon of any abnormal temperature rise region; If a certain projected coordinate is located inside the outline polygon of a certain abnormal temperature rise region, then the certain initial discharge point is assigned to the set of certain initial discharge points corresponding to the certain abnormal temperature rise region. If a certain projection coordinate is not located inside the contour polygon of any abnormal temperature rise area, then calculate the shortest distance from the projection coordinate to the boundary of each contour polygon, and determine whether the shortest distance is less than the preset spatial tolerance threshold. If the shortest distance is less than the preset spatial tolerance threshold, then the initial discharge point is assigned to the set of initial discharge points corresponding to the nearest abnormal temperature rise area. If the shortest distance is not less than the preset spatial tolerance threshold, then the initial discharge point will be recorded as an isolated discharge event and output to the abnormal event log for manual review.

7. The intelligent partial discharge detection method according to claim 1, characterized in that, The step of extracting a waveform feature vector corresponding to a set of preliminary discharge points and a temperature feature vector of an abnormal temperature rise region, and then fusing the waveform feature vector and the temperature feature vector to obtain a fused feature vector includes: Multi-dimensional waveform analysis is performed on all ultrasonic pulse signals corresponding to a certain initial discharge point set to extract a set of time-domain waveform features and frequency-domain waveform features to form a certain waveform feature sub-vector. The time-domain waveform features include pulse amplitude statistical features, pulse time interval statistical features and pulse energy features. The frequency-domain waveform features include the spectral centroid, spectral width and energy proportion of a specific frequency band extracted by fast Fourier transform. Temperature field analysis is performed on the infrared image region corresponding to the abnormal temperature rise region, and a set of geometric features and temperature field features are extracted to form a temperature feature sub-vector. The geometric features include the contour area, contour perimeter and contour aspect ratio of the abnormal temperature rise region, and the temperature field features include temperature gradient features and temperature rise statistical features. The waveform feature vector and the temperature feature vector are standardized to eliminate the influence of dimensions. The standardized waveform feature vector and the temperature feature vector are then concatenated to obtain a fused feature vector.

8. A partial discharge intelligent detection system, characterized in that, include: The acquisition module is configured to acquire multiple ultrasonic signals from the power equipment under test within a preset time period, wherein the multiple ultrasonic signals are respectively acquired by multiple sensors; The analysis module is configured to perform pulse analysis on the multiple ultrasonic signals according to preset signal detection rules, obtain ultrasonic pulse signal sequences corresponding to the multiple ultrasonic signals, and slide on each ultrasonic pulse signal sequence based on a preset sliding window to cluster the ultrasonic pulse signals that meet the time synchronization condition to obtain at least one ultrasonic pulse signal set, wherein an ultrasonic pulse signal sequence contains at least one ultrasonic pulse signal corresponding to the same ultrasonic signal. The determination module is configured to determine the initial discharge point corresponding to the at least one set of ultrasonic pulse signals; The aggregation module is configured to determine the infrared image of the target surface of the power equipment under test within a preset time period based on a preset image processing strategy, and aggregate each preliminary discharge point based on at least one abnormal temperature rise area in the infrared image of the target surface to obtain at least one set of preliminary discharge points, wherein one set of preliminary discharge points corresponds to one suspected discharge source, and one set of preliminary discharge points is associated with one abnormal temperature rise area. The extraction module is configured to extract a waveform feature corresponding to a set of initial discharge points and a temperature feature of an abnormal temperature rise region, and to fuse the waveform feature and the temperature feature to obtain a fused feature vector. The output module is configured to input a certain fused feature vector into a preset partial discharge identification model, and the partial discharge identification model outputs the discharge type identification result and discharge intensity evaluation value of a certain suspected discharge source.

9. An electronic device, characterized in that, include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method according to any one of claims 1 to 7.

Images