GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integration method

Abstract

The invention relates to the technical field of power equipment state monitoring and three-dimensional visualization, and discloses a GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integration method, which comprises the following steps: S1, constructing a GIS digital twinborn model; s2, performing multi-source acquisition on partial discharge signals by adopting an ultrahigh frequency and ultrasonic sensor, and performing real-time transmission through WebSocket after wavelet denoising and feature extraction; s3, mapping the discharge amplitude and frequency to the surface texture of the model based on a Three.js shader, and calculating the coordinates of a discharge point by using a fusion positioning method; s4, constructing a multi-dimensional interaction analysis function; and S5, intelligent retrieval positioning based on BERT semantic analysis and R * tree index is provided. Through multi-source data fusion and a three-dimensional visualization technology, positioning and dynamic display of partial discharge are realized, space query, time backtracking and working condition simulation are supported, and the operation and maintenance efficiency is improved in combination with intelligent retrieval.

Description

Technical Field

[0001] This invention relates to the field of power equipment condition monitoring and three-dimensional visualization technology, and in particular to an integrated method for real-time three-dimensional visualization diagnosis and positioning of partial discharge in GIS. Background Technology

[0002] As a core component of the power system, the insulation status of GIS equipment directly affects the safe operation of the power grid, and partial discharge is an important indicator of insulation defects. Existing partial discharge monitoring technologies have significant limitations: First, the positioning accuracy is insufficient; traditional time-of-flight methods using ultra-high frequency technology are affected by on-site interference, making it difficult to accurately locate defects. Second, visualization methods are limited, relying heavily on two-dimensional maps that require professional interpretation and cannot intuitively present the spatial distribution of discharges. Third, interactivity is poor, lacking multi-dimensional analysis capabilities and making it difficult to trace the development trend of discharges.

[0003] While existing technologies employ 3D visualization, they largely remain static model displays, failing to achieve dynamic correlation with real-time monitoring data. Furthermore, the integration of multi-source data fusion and lightweight web deployment is insufficient, limiting system usability. Therefore, there is an urgent need for a GIS-based real-time visualization method for partial discharge that enables real-time data mapping, multi-dimensional interaction, and cross-platform application, thereby improving diagnostic accuracy and operational efficiency. Summary of the Invention

[0004] The purpose of this invention is to overcome the problems of low positioning accuracy, poor visualization effect and weak interactivity of GIS partial discharge monitoring in the prior art, and to provide an integrated method for real-time three-dimensional visualization diagnosis and positioning of GIS partial discharge, which is applicable to partial discharge monitoring and insulation status assessment of GIS at all voltage levels.

[0005] The objective of this invention is achieved through the following technical solutions: Figure 1 As shown, a method for real-time three-dimensional visualization diagnosis and localization of partial discharge in GIS includes the following steps: S1: Construct a GIS digital twin model in GLTF format and load it using Draco compression and InstancedMesh optimization; S2: Partial discharge signals are acquired from multiple sources using ultra-high frequency and ultrasonic sensors, and then transmitted in real time via WebSocket after wavelet denoising and feature extraction; S3: Based on the Three.js shader, the discharge amplitude and frequency are mapped to the surface texture of the model, and the coordinates of the discharge point are calculated using the fusion positioning method; S4: Build multi-dimensional interactive analysis capabilities; S5: Provides intelligent retrieval and positioning based on BERT semantic parsing and R-tree indexing.

[0006] Further, the step S1 comprises the following sub-steps: S11: using Blender to model the key components of GIS equipment in detail, constructing a fine three-dimensional digital twin model, and ensuring that the model geometric features are consistent with the actual equipment; S12: exporting the model to GLTF format and using Draco algorithm for binary compression to optimize the model file size and reduce loading time; S13: loading the model through Three.js GLTFLoader and using instance rendering technology for repeated components to reduce rendering resource occupation and improve loading efficiency.

[0007] Further, the key components in the step S11 include circuit breakers and disconnectors.

[0008] Further, the step S2 comprises the following sub-steps: S21: deploying UHF sensors and ultrasonic sensors to synchronously collect partial discharge signals and cover key monitoring areas of GIS equipment; S22: pre-processing the original signals through edge computing nodes, including wavelet denoising and feature parameter extraction; S23: transmitting the pre-processed data to the front end in JSON format through WebSocket protocol, and the data frame contains timestamp, device ID, and discharge location coordinate key information to ensure control transmission delay.

[0009] Further, the frequency range of the UHF sensor in the step S21 is 300-1500MHz, and the frequency range of the ultrasonic sensor is 40-80kHz.

[0010] Further, the step S3 comprises the following sub-steps: S31: based on Three.js shader technology, establishing the mapping relationship between discharge parameters and model surface visual features, displaying high-risk areas in red, and simulating discharge arcs combined with particle special effects; S32: combining UHF time difference positioning method and ultrasonic amplitude positioning method to solve the three-dimensional coordinates of the discharge point through least squares method The positioning formula is: (7); Wherein, the first is the UHF sensor, the second is the ultrasonic sensor, is the number of sensors and at least 3 UHF sensors + 1 ultrasonic sensor, is the coordinate of the first sensor, is the signal arrival time difference, The speed of electromagnetic wave propagation. The speed at which ultrasound propagates; S33: The extracted multidimensional feature parameters are reduced in dimensionality by PCA and displayed by overlaying them using the Points particle system of Three.js. The particle size is proportional to the discharge energy, and the color is displayed differently depending on the discharge type.

[0011] Step S4 includes the following sub-steps: S41: Spatial Dimension Interaction: Using Raycaster ray detection technology, the model can be clicked to interact. After clicking, a pop-up window will display the PRPD / PRPS map and insulation aging index at that location. (10); in For high-amplitude discharge times, This represents the total number of discharges. The average discharge amplitude, For the maximum amplitude, The proportion of high-frequency discharge; S42: Time Dimension Interaction: Develop a timeline control to support historical discharge data backtracking. Combine a K-means clustering and Adam-optimized LSTM model with 3 hidden layers, the number of neurons being 64, 128 and 64 respectively, a dropout rate of 0.2, 50 training rounds, predict the discharge trend for the next 24 hours and display it with a heatmap overlay. S43: Operating Condition Dimension Interaction: Design a slider control to adjust the operating voltage and load parameters, and dynamically update the internal electric field distribution of the model. (12); in Spatial coordinates The electric field distribution is dynamically updated. As a reference electric field distribution, For the rated value, This is the actual operating voltage. This represents the actual load current.

[0012] Furthermore, the insulation aging index mentioned in step S41 It was determined to be high risk at that time.

[0013] Furthermore, step S5 includes the following sub-steps: S51: Construct a semantic tag library for GIS equipment and establish associations between GIS equipment components and 3D model nodes; S52: Employs a BERT fine-tuning model to parse user input and combines it with an R-tree index to quickly locate the target component; S53: After positioning is completed, the viewing angle is automatically adjusted, and the historical discharge area is highlighted.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. Improve positioning accuracy by using multi-source data fusion and optimization algorithms to trace the source of defects; 2. Three-dimensional visualization intuitively presents the discharge distribution, and dynamic effects enhance data readability; 3. Multi-dimensional interaction supports comprehensive analysis, reducing manual interpretation costs and improving operation and maintenance efficiency; 4. The lightweight web architecture supports cross-platform access, reduces deployment costs, and is suitable for mobile operation and maintenance scenarios. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0016] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a flowchart of the three-dimensional model construction and loading optimization process of the present invention; Figure 3 This is a diagram of the multi-source data acquisition and transmission architecture of the present invention; Figure 4 This is a schematic diagram illustrating the principle of visual mapping of discharge parameters in this invention. Figure 5 This is a schematic diagram of the multi-dimensional interactive interface of the present invention; Figure 6 This is a flowchart of the intelligent retrieval and positioning process of the present invention. Detailed Implementation

[0017] 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, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0018] In the description of this invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0019] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. Furthermore, the technical features involved in the different embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0020] Figure 1 This is a flowchart of the method of the present invention. First, a three-dimensional digital twin model of the GIS equipment is constructed. Partial discharge signals are collected and preprocessed through multi-source sensors. Three.js is used to realize the dynamic mapping between discharge parameters and the model. Combined with multi-dimensional interactive controls and intelligent retrieval modules, real-time visualization diagnosis and analysis of partial discharge in GIS are finally realized.

[0021] A method for real-time three-dimensional visualization diagnosis and localization of partial discharge in GIS includes the following steps: S1: Construct a GIS digital twin model in GLTF format and load it using Draco compression and InstancedMesh optimization.

[0022] Traditional GIS equipment models are mostly static displays, resulting in large file sizes, slow loading times, and excessive rendering resources consumed by repetitive components, leading to lag and stuttering. This step addresses the issues of insufficient consistency between the model and the actual equipment, as well as low loading efficiency, through refined modeling and lightweight optimization.

[0023] like Figure 2 As shown, step S1 includes the following sub-steps: S11: Use Blender to build a 1:1 scale 3D model based on GIS equipment design drawings, and perform detailed modeling of key components of circuit breakers (including arc extinguishing chambers and operating mechanisms) and disconnect switches to ensure geometric errors and that the model structure is consistent with the actual equipment.

[0024] S12: Export the model to GLTF format, enable binary compression, use the Draco algorithm to set the compression rate to 60%-80%, reduce the model file size and reduce loading time.

[0025] S13: Loads models using Three.js's GLTFLoader, employs instancedMesh rendering technology for repetitive parts, reduces rendering batches, lowers GPU resource consumption, and maintains model loading and interaction frame rates.

[0026] S2: Partial discharge signals are acquired from multiple sources using ultra-high frequency and ultrasonic sensors, and then transmitted in real time via WebSocket after wavelet denoising and feature extraction.

[0027] Single sensors are susceptible to field interference, resulting in raw signals with high noise and unclear characteristics. Direct transmission of these signals leads to data redundancy and high latency. This step improves signal quality and reduces transmission load through multi-source fusion and edge preprocessing.

[0028] Figure 3 The multi-source data acquisition and transmission architecture diagram for step S2 is as follows: Step S2 includes the following sub-steps: S21: Deploy UHF and ultrasonic sensors in key areas of the busbar room and circuit breaker compartment of the GIS equipment. The sensor sampling rate is set to 100MS / s. The frequency range of the UHF sensor is 300-1500MHz, and the frequency range of the ultrasonic sensor is 40-80kHz. Synchronization error is used to ensure signal time alignment.

[0029] S22: Edge computing nodes preprocess the raw signal: The wavelet denoising process employs a db4 wavelet basis, with a decomposition level of 5, and uses a soft thresholding function to process the noise figure. (1); in, These are the original wavelet coefficients. For the threshold, The standard deviation of noise. represents the high-frequency coefficients of the first-level wavelet decomposition, and 0.6745 is the bias correction coefficient. This is the signal length.

[0030] After wavelet denoising, the signal-to-noise ratio was improved by 25±3dB.

[0031] Feature extraction: Calculating feature vectors: (2); in, , The peak value is the amplitude. The pulse width. Discharge frequency; S23: Encapsulate the preprocessed data into JSON format, with the data frame structure as follows: (3); in, For timestamps, For device ID, Here are the coordinates of the discharge location. For feature vectors, Data frames are pushed to the front end via the WebSocket protocol to control transmission latency and reduce data transmission volume.

[0032] S3: Based on the Three.js shader, the discharge amplitude and frequency are mapped to the surface texture of the model, and the discharge point is calculated using the fusion localization method; Traditional two-dimensional maps struggle to visually represent the spatial distribution of discharges, and the parameters lack dynamic correlation with the model. This step utilizes shader technology and multi-algorithm fusion to establish a real-time mapping between discharge features and 3D visual effects. The principle diagram of the discharge parameter visualization mapping of this invention is shown below. Figure 4 As shown.

[0033] Step S3 includes the following sub-steps: S31: Write a program based on a custom shader (ShaderMaterial) in Three.js to establish the mapping relationship between discharge parameters and visual features of the model surface: (4); in A is the maximum discharge amplitude, A is the actual discharge value, and H is the hue component in the HSV color model.

[0034] When the hue component is in the low amplitude range ( When the hue component is in the mid-range, it displays a blue hue. When the hue component is in the high amplitude range, it displays a yellow hue. When displayed, it is in red.

[0035] Particle effects: Dynamic particles are generated using the Points particle system, with particle size proportional to the discharge energy. (5); Where 0.02 is the baseline size, This refers to the discharge energy.

[0036] Particle lifecycle varies with discharge frequency Dynamic adjustment: (6); S32: Integrating UHF time difference positioning and ultrasonic amplitude positioning, the three-dimensional coordinates (x, y, z) of the discharge point are calculated using the least squares method. The positioning formula is as follows: (7); Among them, the The first is a UHF sensor, the second... One is an ultrasonic sensor. The number of sensors must be at least 3 UHF sensors + 1 ultrasonic sensor. For the first Each sensor has known three-dimensional coordinates. Let be the three-dimensional coordinates of the discharge source to be determined. To discharge the power to the first The straight-line distance between the sensors For the signal arrival time difference, The speed of electromagnetic wave propagation. This represents the speed at which ultrasound propagates.

[0037] S33: Extracted multidimensional feature parameters After PCA dimensionality reduction, the covariance matrix is ​​calculated: (8); Where n is the number of samples. For the first The feature vector of each sample This is the mean vector of all samples.

[0038] Let the covariance matrix be... eigenvalues Where m is the feature dimension, and the cumulative contribution rate of the first k principal components is: (9); in, For the front The cumulative contribution rate of each principal component Covariance matrix The Each feature value.

[0039] Take the satisfaction The minimum value is usually taken from the first two principal components, which are then superimposed on the corresponding positions in the model using the Points particle system in Three.js, as shown in Table 1. The particle colors are differentiated according to the discharge type, using different hues for distinction. The discharge types are corona discharge, surface discharge, and internal discharge.

[0040] Table 1. Correspondence between discharge types and visualization features: S4: Build multi-dimensional interactive analysis capabilities; Existing systems lack multi-dimensional analysis tools, making it difficult to achieve the linkage between spatial querying, temporal tracing, and operational condition simulation. This step improves data interpretation efficiency through the design of cross-dimensional interactive controls.

[0041] Step S4 includes the following sub-steps: S41: Spatial Dimension Interaction: Based on Three.js's Raycaster ray detection technology, implement model click interaction. After the click event is triggered: Calculate the coordinates of the intersection point ; Query historical discharge data within a 0.5m radius of the location to generate PRPD (phase-amplitude) and PRPS (phase-amplitude-frequency) maps. Calculate the insulation aging index: (10); in For high-amplitude discharge times, This represents the total number of discharges. The average discharge amplitude, For the maximum amplitude, The proportion of high-frequency discharge; Insulation aging index If the risk is deemed high, a pop-up information window will display the above information to control the response time.

[0042] S42: Interaction in the Time Dimension Develop a sliding timeline control that supports adjustment of the time dimension. After selecting a historical period, the system loads the corresponding data and updates the model display. The LSTM model, improved by K-means clustering and Adam optimization, contains three hidden layers with 64, 128, and 64 neurons respectively. The activation function is ReLU, the output layer uses Linear, and the dropout rate is 0.2. The training run is 50 epochs, using firing data from the past 72 hours as input, to predict the firing trend for the next 24 hours. (11); in, L represents the predicted discharge trend at future times, and L is the sequence length. For model parameters, The historical discharge data sequence is used as input to the model. The prediction results are displayed on the model surface in the form of a heat map, and the color depth represents the risk probability. S43: Operating Condition Interaction: Design a slider control to adjust the operating voltage U and load current I parameters. The voltage U is 0.8-1.2 times the rated voltage, and the load current I is 0-1.1 times the rated current.

[0043] The system is based on a preset electric field distribution model: (12); in Spatial coordinates The electric field distribution is dynamically updated. As a reference electric field distribution, For the rated value, This is the actual operating voltage. This is the actual load current. This refers to the electric field distribution quantity related to the load current.

[0044] A multi-dimensional interactive analysis function is constructed to dynamically update the visualization of the electric field distribution within the model, such as... Figure 5 As shown, the discharge evolution process under different operating conditions can be simulated.

[0045] S5: Provides intelligent retrieval and location based on BERT semantic parsing and R-tree indexing. Traditional device location relies on manual searching, which is inefficient and prone to errors. This step achieves rapid device location and historical data association through semantic parsing and spatial indexing, such as... Figure 6 Flowchart for intelligent search and location.

[0046] Step S5 includes the following sub-steps: S51: Construct a GIS equipment semantic tag library, using a hierarchical structure to describe equipment components: (13); S52: Employs a BERT fine-tuning model to parse the device name entered by the user. Through keyword matching and synonym expansion, combined with R-tree indexing, it quickly locates the 3D coordinates of the target component. ; S53: After positioning is completed, the camera angle is automatically adjusted to center the target area and the historical discharge concentration area at that location is marked with a bright material. The emissive intensity of the bright material is set to 0.8.

[0047] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method, characterized in that, The method comprises the following steps: S1: constructing a GIS digital twin model in GLTF format and loading it through Draco compression and InstancedMesh optimization; S2: collecting partial discharge signals using multi-source ultrahigh frequency and ultrasonic sensors, and transmitting them in real time through WebSocket after wavelet denoising and feature extraction; S3: mapping the discharge amplitude and frequency to the model surface texture based on the Three.js shader, and calculating the discharge point coordinates using the fusion positioning method; S4: constructing a multi-dimensional interactive analysis function; S5: providing intelligent search positioning based on BERT semantic analysis and R-tree indexing.

2. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 1, characterized in that, The step S1 comprises the following sub-steps: S11: using Blender to model the key components of GIS equipment in detail, constructing a fine three-dimensional digital twin model, and ensuring that the model geometry is consistent with the actual equipment; S12: exporting the model to GLTF format, using Draco algorithm for binary compression, optimizing the model file size, and reducing the loading time; S13: loading the model through Three.js's GLTFLoader, using instance rendering technology for repeated components to reduce resource occupation and improve loading efficiency.

3. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 2, characterized in that, The key components in step S11 include circuit breakers and disconnectors.

4. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 1, characterized in that, The step S2 comprises the following sub-steps: S21: deploying ultrahigh frequency sensors and ultrasonic sensors to synchronously collect partial discharge signals, covering key monitoring areas of GIS equipment; S22: preprocessing the original signals through edge computing nodes, including wavelet denoising and feature parameter extraction; S23: transmitting the preprocessed data to the front end in JSON format through WebSocket protocol, with data frames containing timestamp, device ID, and discharge location coordinate key information to ensure control transmission delay.

5. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 4, characterized in that, The frequency range of the ultrahigh frequency sensor in step S21 is 300-1500MHz; the frequency range of the ultrasonic sensor is 40-80kHz.

6. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 1, characterized in that, The step S3 comprises the following sub-steps: S31: based on Three.js shader technology, establishing the mapping relationship between discharge parameters and model surface visual features, displaying high-risk areas in red, and simulating discharge arcs with particle effects; S32: fuse UHF time difference positioning method and ultrasonic amplitude positioning method, and solve three-dimensional coordinates of discharge point by least square method The positioning formula is: （7）； Wherein, the first is a UHF sensor, the second is an ultrasonic sensor, is the number of sensors and at least 3 UHF sensors + 1 ultrasonic sensor, is the known three-dimensional coordinates of the first sensor, is the three-dimensional coordinates of the discharge source to be solved, is the straight-line distance from the discharge source to the first sensor, is the signal arrival time difference, is the electromagnetic wave propagation speed, is the ultrasonic wave propagation speed; S33: reducing the multi-dimensional feature parameters through PCA, and displaying them through Three.js's Points particle system, with particle size proportional to discharge energy and color differentiating according to discharge type.

7. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 1, characterized in that, The step S4 comprises the following sub-steps: S41: spatial dimension interaction: realizing model click interaction through Raycaster ray detection technology, and displaying the PRPD / PRPS spectrum and insulation aging index at the clicked position in a pop-up window; （10）； wherein is the number of high amplitude discharges, is the total number of discharges, is the average amplitude of discharges, is the maximum amplitude, is the high frequency discharge ratio, ; S42: time dimension interaction: developing a time axis control, supporting historical discharge data backtracking, and using a LSTM model with 3 hidden layers, 64, 128, and 64 neurons respectively, a dropout rate of 0.2, and 50 training rounds to predict the discharge trend in the next 24 hours and display it as a heat map. S43: Working condition dimension interaction: design a sliding bar control to adjust the operating voltage and load parameters, and dynamically update the internal electric field distribution of the model: （12）； wherein for the spatial coordinates the dynamically updated electric field distribution at the location, for the reference electric field distribution, for the rated value, for the actual operating voltage, for the actual load current.

8. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 7, characterized in that, The insulation aging index in the step S41 is determined as high risk.

9. The GIS partial discharge real-time three-dimensional visualization diagnosis and positioning integrated method according to claim 1, characterized in that, The step S5 includes the following sub-steps: S51: Build a GIS device semantic label library to associate GIS device components with three-dimensional model nodes; S52: Analyze user input using a BERT fine-tuning model and quickly locate the target component using R-tree indexing; S53: After positioning, automatically adjust the viewing angle and highlight the historical discharge area.

Images