Converter transformer bushing discharge analysis method, system and device based on multi-physics field data fusion

Abstract

The invention relates to the technical field of power testing, and discloses a converter transformer bushing discharge analysis method, system and device based on multi-physics field data fusion. The invention aims to solve the technical problems that the traditional observation means cannot go deep into the discharge process and the internal mechanism is difficult to reveal. According to the method, a multi-physical field combined observation and analysis framework is constructed, and an instantaneous and irreversible complex physical discharge process is converted into a four-dimensional digital image which can be played back, analyzed and quantized. The core of the method is that a plurality of discrete and mature inversion reconstruction algorithms are integrated under a unified space-time reference for collaborative data processing for the first time. An omnibearing and high-fidelity digital analysis tool is provided for researching the high-voltage discharge physical process, and the depth and efficiency of discharge experiment research are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of power testing technology, and more specifically, to a method, system, and device for analyzing bushing discharge of converter transformers based on multi-physics data fusion. Background Technology

[0002] Converter transformer bushings are core components in ultra-high voltage direct current (UHVDC) transmission projects. Their function is to pass the high-voltage leads inside the transformer through the grounded tank casing, while simultaneously undertaking the dual critical roles of electrical insulation and mechanical fixation. Due to long-term operation under complex electrical stresses from superimposed DC, AC, and harmonics, as well as the harsh environment of multi-field coupling of electricity, heat, and force, the inner and outer insulation interfaces of the bushing, especially the oil-paper composite insulation structure, are the weakest link in the overall insulation performance of the equipment. Partial discharge occurring in this area is a major cause of bushing insulation degradation and even catastrophic flashover or breakdown accidents. Therefore, in-depth research into the physical mechanism of bushing discharge processes is of paramount importance for improving equipment reliability and ensuring the safe and stable operation of the power grid.

[0003] Currently, research on bushing discharge processes mainly relies on direct observation through the construction of physical experimental platforms. Traditional experimental methods typically reproduce specific discharge scenarios in a simulation device and use one or a few sensors for data acquisition. While these methods can reproduce the discharge phenomenon to a certain extent and obtain some electrical characteristic parameters, they also provide some basis for insulation condition assessment.

[0004] However, existing traditional observation methods have fundamental limitations in revealing the deep mechanisms of discharge. Experiments often focus only on certain specific physical quantities, and data collected by different sensors even require manual alignment. The data from different sensors are like isolated data islands, requiring researchers to perform in-depth data processing to correlate different physical data for subsequent analysis, which greatly slows down the research progress. Secondly, measurement methods are mostly limited to the macroscopic level. Existing technologies can usually only obtain data on the external physical changes of discharge, while the core internal physical states that determine the initiation and development of discharge, such as the electric field distribution, spatial electric accumulation, and current density in key areas, need to be inferred from the test results. Furthermore, data analysis is disconnected from the physical process. There is a huge cognitive gap between the raw waveform and image data and the physical process of discharge. Data interpretation relies heavily on human experience, making it difficult to conduct quantitative and systematic mechanistic exploration.

[0005] In summary, existing technologies generally face a common bottleneck: traditional observation methods cannot penetrate deep into the discharge process and reveal its intrinsic mechanisms. Therefore, a novel technical solution is urgently needed to address these issues. Summary of the Invention

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] Therefore, this invention provides a method, system, and device for analyzing bushing discharge of converter transformers based on multi-physics data fusion, solving the technical problem that existing traditional observation methods cannot penetrate into the discharge process and cannot reveal its intrinsic mechanism.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for analyzing bushing discharge in converter transformers based on multi-physics data fusion, which includes the following steps: S1: Establish a unified spatiotemporal benchmark: Based on the discharge test environment, build a static physical model in the virtual computing environment, establish a test scenario based on the spatial location information of the sensors in the test environment, and establish a unified test time benchmark for the sensors. S2: Coordinated Acquisition and Event Triggering: Coordinate the operation of the sensor and the discharge signal generator to acquire the data stream continuously generated by the sensor according to the coordinated acquisition frequency, and set the time reference for the data stream with the discharge trigger signal issued by the generator; S3: Reconstruct a four-dimensional data field: Process the data stream to reconstruct a four-dimensional physical field set that can characterize the dynamic distribution of the discharge process in the test scenario; S4: Fusion Analysis and Source Tracing: The four-dimensional physical field set is integrated into the static physical scene to form a dynamic physical scene of the discharge process, and then the dynamic physical scene is comprehensively analyzed.

[0009] As a preferred embodiment of the converter transformer bushing discharge analysis method based on multi-physics field data fusion described in this invention, the process of building a static physical model in step S1 specifically includes: calling a preset test device model, importing the three-dimensional model of the converter transformer bushing to be tested, configuring the corresponding electrode model according to the selected discharge mode to form an uncorrected digital model, using a visual sensor to obtain the spatial installation information of the physical entity, correcting the digital model through the spatial installation information, and assigning physical attribute parameters to the corrected digital model.

[0010] As a preferred embodiment of the converter transformer bushing discharge analysis method based on multi-physics data fusion described in this invention, step S2, which coordinates the operation of the sensor and the discharge signal generating device, further includes: by analyzing the optimal acquisition frequency of each sensor, setting a reference time period in the test time reference to ensure that all sensors acquire data synchronously at each node of the reference time period; presetting the discharge trigger signal at a preset node of the reference time period; determining three consecutive data acquisition windows around the preset node for acquiring background, dynamic change, and recovery phases; and setting an appropriate working mode for the sensor according to the target characteristics of each data acquisition window.

[0011] As a preferred embodiment of the converter transformer bushing discharge analysis method based on multi-physics field data fusion described in this invention, the process of reconstructing the four-dimensional physical field set in step S3 further includes constructing multiple four-dimensional physical fields to characterize different physical dimensions. The four-dimensional physical fields include: a morphological field that intuitively reproduces the three-dimensional geometric shape of the discharge channel and its evolution over time; a temperature field that depicts the dynamic changes in the temperature distribution of the discharge region and its surroundings; a plasma parameter field that quantitatively characterizes the distribution of microscopic physical properties inside the discharge channel; an electric field that characterizes the local electric stress distribution and its dynamic evolution during the discharge process; an electromagnetic field that directly reflects the spatial propagation path of electromagnetic waves during the test; a sound pressure field used to characterize the propagation of acoustic pressure generated by the discharge in space; and a space charge field that reveals the state of charge accumulation on the surface and inside the insulating medium.

[0012] As a preferred embodiment of the converter transformer bushing discharge analysis method based on multi-physics field data fusion described in this invention, the specific reconstruction process of the multiple four-dimensional physical fields is as follows: A morphological field is established based on multi-view image data acquired by a high-speed camera using a tomographic imaging algorithm; a temperature field is constructed using thermal imaging data acquired by an infrared thermal imager; the plasma parameter field is obtained by combining the spectral data acquired by the spectrometer with the morphological field through Abelian inversion; the electric field, electromagnetic field, sound pressure field, and space charge field are reconstructed by inverting the light polarization change data acquired by the optical electric field sensor, the electromagnetic waveform data acquired by the ultra-high frequency electromagnetic sensor, the acoustic signal acquired by the acoustic sensor, and the charge induction signal acquired by the sensor, respectively; and the current density field is calculated based on the multi-point magnetic field data acquired by the optical magnetic field sensor array.

[0013] As a preferred embodiment of the converter transformer bushing discharge analysis method based on multi-physics field data fusion described in this invention, the comprehensive analysis in step S4 further includes a causal tracing process for the discharge initiation; the causal tracing process for the discharge initiation analyzes the electromagnetic field and the sound pressure field through time difference positioning, and performs cross-verification with the initial luminous point in the morphological field to finally determine the precise three-dimensional spatial coordinates of the discharge initiation point; and retrieves the data of the electric field and the space charge field before the discharge initiation point to determine the dominant physical cause that triggers the discharge.

[0014] As a preferred embodiment of the converter transformer bushing discharge analysis method based on multi-physics field data fusion described in this invention, the comprehensive analysis in step S4 further includes a dynamic analysis process for discharge development. This dynamic analysis process first tracks the motion of the discharge channel head in three-dimensional space within the morphological field, analyzes the propagation path of the discharge channel head, and quantifies the development rate. Based on the position of the channel head, it combines other four-dimensional physical fields for analysis. This combined analysis includes: extracting the field distribution in front of the head through the electric field and the space charge field to analyze the magnitude and direction of the electric driving force; extracting the current vector flowing through the channel from the current density field to evaluate the channel's conductivity; combining the temperature field and the sound pressure field to analyze the energy distribution and quantify the conversion efficiency of discharge energy into thermal and mechanical energy; extracting electron density and temperature from the plasma parameter field to characterize the channel's microscopic physical state; and revealing and verifying the core physical mechanism driving discharge development by establishing a dynamic correlation model between the above physical quantities and the propagation path and development rate.

[0015] As a preferred embodiment of the converter transformer bushing discharge analysis method based on multi-physics data fusion described in this invention, wherein: the dynamic digital scene formed in step S4 is an interactive four-dimensional simulation scene, presented as a three-dimensional view that can be observed and scaled from multiple perspectives; the interactive four-dimensional simulation scene is provided with a time axis that can be dragged arbitrarily, used to review the discharge process and to conduct a detailed examination of the spatiotemporal profile of the discharge process; the interactive four-dimensional simulation scene is also provided with a physics field control panel, used to perform visualization rendering control of different physics field distributions.

[0016] A converter transformer bushing discharge analysis system based on multi-physics data fusion, characterized in that the system is used to execute the method of any one of claims 1 to 8, specifically including: This invention also provides a converter transformer bushing discharge analysis system based on multi-physics data fusion, used to execute the above method, specifically including the following modules: Calibration module: Analyzes data from the vision sensor, obtains spatial installation information of the test device, and sends model calibration information to the subsequent scene definition and construction module; Scene building module: Receives the model calibration information sent by the calibration module, and according to the real test scenario, retrieves the required geometric model from the experimental model library for virtual assembly, and calls the simulation physics library to assign physical properties to the model in order to build a calibrated static physical model; Main control clock module: provides a synchronized clock signal for the discharge controller and the data stream acquisition unit; Acquisition planning module: Analyzes the optimal acquisition frequency of each sensor, sets the reference time period, defines the acquisition window around the trigger node, and plans and sets the working mode of each sensor for the window; Physics field reconstruction module: Analyzes the data stream collected by the data stream acquisition device and generates a four-dimensional physics field set; Four-dimensional rendering module: merges and renders the static physical model with the four-dimensional physical field set to generate a dynamic physical scene; Comprehensive Analysis Module: It has two built-in algorithms for causal tracing and dynamic analysis. It is responsible for combining the four-dimensional physical field set to analyze the location of the discharge initiation point and the determination of the cause, as well as the discharge development process, and generating analysis results. Interactive analysis platform: Receives the dynamic physical scene generated by the four-dimensional rendering module and the analysis results generated by the comprehensive analysis module, and provides a graphical user interface. The interface embeds the dynamic physical scene and has a controllable timeline and a physics field control panel. The analysis results are integrated into the interface in the form of graphics and text for display.

[0017] The present invention also provides a converter transformer bushing discharge analysis system based on multi-physics field data fusion, used to execute the above method and as an experimental carrier for the above system, specifically including: a housing unit, the housing unit including a shell, a manhole opened on the top of the shell, an oil inlet pipe and a sensor wiring port respectively provided on both sides of the manhole on the top of the shell, a fixed flange provided on the side of the shell, and an oil outlet pipe connected to the geometric center of the bottom of the shell. The observation unit includes a main observation window that is movably mounted on the top of the manhole via a hinge, an air outlet pipe connected to the main observation window, four secondary observation windows that are provided through the side of the housing, a positioning grid that is provided on the inner side wall of the housing, and sensor mounting posts that are provided on the grid points of the positioning grid. The discharge unit includes a grounding post fixedly installed on the internal axis of the housing, a discharge platform provided on the top of the grounding post, an insulating plate fixedly installed inside the housing above the discharge platform, and a through hole opened in the insulating plate at the geometric center of the discharge platform; The discharge unit further includes: A connecting rod, wherein a slot is provided at one end of the top of the through hole, and a fixing hole is provided at the end of the connecting rod that connects to the sleeve; The discharge electrode has a threaded rod at its top, which passes through the through hole and the slot in sequence, and engages with a nut for locking the connecting rod and the insulating plate.

[0018] The beneficial effects of this invention are as follows: By constructing a multi-physics joint observation and analysis framework, a transient and irreversible complex physical discharge process is transformed into a replayable, analyzable, and quantifiable four-dimensional digital image. Its core lies in the first-time integration of multiple discrete and mature inversion and reconstruction algorithms under a unified spatiotemporal reference for collaborative data processing. This provides a comprehensive and high-fidelity digital analysis tool for studying high-voltage discharge physical processes, significantly improving the depth and efficiency of discharge experimental research. Attached Figure Description

[0019] 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a converter transformer bushing discharge analysis method based on multi-physics data fusion.

[0021] Figure 2 A schematic diagram of the specific technical process for reconstructing a four-dimensional data field.

[0022] Figure 3 This is a schematic diagram illustrating the specific technical process for integrated analysis and tracing.

[0023] Figure 4 This is a schematic diagram of the functional module architecture of a converter transformer bushing discharge analysis system based on multi-physics field data fusion.

[0024] Figure 5 This is a schematic diagram of the data flow during system operation.

[0025] Figure 6 This is a schematic diagram of the overall structure of the converter transformer bushing discharge analysis device based on multi-physics field data fusion in Example 3.

[0026] Figure 7 This is a cross-sectional view of a converter transformer bushing discharge analysis device based on multi-physics data fusion.

[0027] Figure 8 This is a top view of a converter transformer bushing discharge analysis device based on multi-physics data fusion.

[0028] Figure 9 Detailed diagram of the positioning grid and sensor mounting posts.

[0029] Figure 10 This is an engineering drawing of a partial discharge electrode.

[0030] Figure 11 This is a scene diagram of a partial discharge model.

[0031] Figure 12 This is a schematic diagram of the overall structure of the initial version of the discharge analysis device in Example 4.

[0032] Figure 13 This is a cross-sectional view of the initial version of the discharge analysis device.

[0033] Figure 14 This is a schematic diagram of the overall structure of the improved discharge analysis device in Example 5.

[0034] Figure 15 This is a cross-sectional view of the improved discharge analysis device. Detailed Implementation

[0035] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0036] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0037] Secondly, the term "one embodiment" or "example" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the invention. The appearance of an embodiment in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that mutually excludes other embodiments.

[0038] Example 1 Reference Figures 1-3 as well as Figure 11This is one embodiment of the present invention, which provides a method for analyzing the bushing discharge of a converter transformer based on multi-physics data fusion, including the following steps: S1. Establish a unified spatiotemporal benchmark: Based on the discharge test environment, a static physical model is built in the virtual computing environment. A test scenario is established based on the spatial location information of the sensors in the test environment, and a unified test time benchmark is established for the sensors. A static physical model is built, a preset test device model is called, the three-dimensional model of the bushing of the converter transformer under test is imported, and the corresponding electrode model is configured according to the selected discharge mode to form an uncorrected digital model. The spatial installation information of the physical entity is obtained by using a vision sensor, and the digital model is corrected by using the spatial installation information. Physical attribute parameters are assigned to the corrected digital model.

[0039] After setting up the realistic test environment for the converter transformer bushing, the process of building the physical model of the test site was initiated. First, based on the test equipment used in this experiment, the pre-stored physical models of the corresponding test equipment were retrieved from the test equipment library to serve as the experimental environment for the physical simulation scenario.

[0040] Then, locate the digital model of the converter transformer bushing used in this test, which was prepared during the design phase, import the bushing model into the physical simulation scene and install it. Based on the discharge mode selected for the test, select the physical model of the corresponding discharge scene in the database and install it in the discharge area of ​​the test device.

[0041] After the physical scene is built, the vision sensor is activated to perform a high-precision scan of the flange and discharge platform to obtain the actual spatial information of the bushing, electrode tip, and insulating cardboard. Based on this obtained spatial information, the digital model in the physical scene is then corrected.

[0042] Next, based on the material types of each structure defined during the bushing modeling, the corresponding physical property parameters of the materials are retrieved from the physical simulation library, and these physical property parameters are assigned to the corresponding structures in the bushing model (e.g., dielectric constant, volume conductivity, breakdown field strength, etc. of oil-impregnated insulating paper; relative dielectric constant, conductivity, etc. of insulating oil). Then, the specific parameters of the specific models of insulating oil and insulating paperboard used in the test are selected and imported from the library, and a preliminary physical fitting is performed between the two.

[0043] At this point, a static physical model that is accurate in space and complete in physics has been constructed.

[0044] As required by the experiment, multiple different physical field simulation software were launched, and a blank scene was built in the field simulation software. The digital model of the test device was imported into the blank scene of each field simulation software. According to the different physical parameters recorded by the sensors, corresponding sensors were set in different simulation scenes. The actual sensing points of the sensors were used as data anchor points at the corresponding positions in the simulation scene to set the boundary conditions of the simulation calculation or as a calibration benchmark to verify the accuracy of the simulation results.

[0045] Finally, connect all the data acquisition cards of the sensors used in the test to a unified global master clock (e.g., a rubidium atomic clock synchronized with GPS) to ensure that every data point collected by all subsequent sensors has a unified, high-resolution absolute timestamp.

[0046] S2. Coordinated Acquisition and Event Triggering: The operation of the coordinated sensor and the discharge signal generator is used to acquire the data stream continuously generated by the sensor according to the acquisition frequency after coordination, and the discharge trigger signal issued by the generator is used as the time reference for the data stream. The operation of the coordinated sensor and discharge signal generator is achieved by analyzing the optimal acquisition frequency of each sensor and setting a reference time period in the test time reference to ensure that all sensors acquire data synchronously at each node of the reference time period. The discharge trigger signal is preset at a preset node of the reference time period, and three consecutive data acquisition windows are determined around the preset node for acquiring background, dynamic change and recovery phases. According to the target characteristics of each data acquisition window, an appropriate working mode is set for the sensor.

[0047] The optimal sampling frequency of the sensors is directly stated in the instruction manual accompanying the product. It's necessary to import the optimal sampling frequencies of all sensors into the system for unified planning. For example, ultra-high frequency electromagnetic sensors require sampling rates of several GHz to distinguish pulse waveforms, high-speed cameras require frame rates of hundreds of thousands of fps to capture streamer development, while infrared thermal imagers only require sampling rates of tens to several Hz. Based on these sampling frequencies, a high-frequency reference time period (e.g., 0.1 nanoseconds) is determined, ensuring that the actual sampling frequency of all sensors is an integer multiple or fraction of this period.

[0048] Then, the working state of the sensors is planned according to the actual determined acquisition frequency. Even if the acquisition rates of each sensor are different, they will inevitably achieve synchronous acquisition at the node of each reference time period.

[0049] Subsequently, the discharge trigger signal of the high-voltage generator is precisely preset at a predetermined node within the reference time period, and this node is defined as the time origin (T=0) for all data streams. Around this node, three consecutive data acquisition windows are defined, and based on the target characteristics of each window, a corresponding operating mode is set for the sensor. Background data acquisition window ( T = -10μs to T=-1μs Within this window, the sensors operate in quasi-static monitoring mode. In this mode, the optical electric field sensor and space charge sensor perform scanning or multi-point measurements to obtain the initial electric field distribution and charge accumulation state of the system before discharge, providing a background benchmark for subsequent analysis.

[0050] Dynamic data acquisition window ( T=-1μs to T = +10μs This window is the core acquisition area, where the sensors are switched to high-frequency sampling mode. All high-speed sensors (ultra-high frequency electromagnetic sensors, high-speed cameras, optical magnetic field / electric field sensors, etc.) operate at their highest speed, densely recording the complete dynamic process of discharge occurrence, development, and extinction with the highest time resolution.

[0051] Restore data acquisition window ( T = +10μs to T=+1s ): In this window, the sensor switches to low-frequency long-term monitoring mode. In this mode, the infrared thermal imager, acoustic emission sensor, and dissolved gas analyzer in oil will continuously record at a lower frequency to analyze the cumulative effects and relaxation processes such as heat diffusion after discharge, mechanical stress wave propagation, and chemical product formation.

[0052] S3. Reconstructing the four-dimensional data field: Process the data stream and reconstruct a four-dimensional physical field set that can characterize the dynamic distribution of the discharge process in the test scenario. The process of reconstructing the four-dimensional physical field set further includes constructing multiple four-dimensional physical fields to characterize different physical dimensions. The four-dimensional physical fields include: a morphological field that intuitively reproduces the three-dimensional geometric shape of the discharge channel and its evolution over time; a temperature field that depicts the dynamic changes in the temperature distribution of the discharge region and its surroundings; a plasma parameter field that quantitatively characterizes the distribution of microscopic physical properties inside the discharge channel; an electric field that characterizes the distribution of local electric stress and its dynamic evolution during the discharge process; an electromagnetic field that directly reflects the spatial propagation path of electromagnetic waves during the test; a sound pressure field used to characterize the propagation of acoustic pressure generated by the discharge in space; and a space charge field that reveals the state of charge accumulation on the surface and inside the insulating medium. The specific reconstruction process of multiple four-dimensional physical fields involves establishing a morphological field based on multi-view image data acquired by a high-speed camera using a tomographic imaging algorithm, and constructing a temperature field using thermal imaging data acquired by an infrared thermal imager. Combining spectral data acquired by a spectrometer with the morphological field, a plasma parameter field is obtained through Abelian inversion. Inversion algorithms are then used to invert light polarization change data acquired by an optical electric field sensor, electromagnetic waveform data acquired by a UHF electromagnetic sensor, acoustic signals acquired by an acoustic sensor, and charge induction signals acquired by a sensor, reconstructing the electric field, electromagnetic field, sound pressure field, and space charge field. Finally, the current density field is calculated based on multi-point magnetic field data acquired by an optical magnetic field sensor array.

[0053] After the discharge test is completed, the collected data stream is processed.

[0054] Image data is analyzed using tomographic imaging algorithms. A morphological field is reconstructed from multi-view images acquired by a high-speed camera from different perspectives to visually reproduce the geometry, bifurcation structure, and growth and extinction processes of the discharge channel over time. A temperature field is constructed by processing thermal imaging data sequences acquired by an infrared thermal imager to obtain the dynamic changes in temperature distribution within and around the discharge region.

[0055] Then, the maximum optical radius at the discharge channel cross-section is obtained from the morphological field. As a geometric constraint of the integration path, and combined with the spectral data collected by the spectrometer (such as the Stark broadening information of the Balmer series spectral lines of hydrogen atoms), the electron density and electron temperature distributed radially inside the channel are calculated by performing Abelian inversion, thereby obtaining the plasma parameter field and realizing the quantification of the distribution of microscopic physical properties inside the discharge channel.

[0056] in, Indicates the radius r The actual physical quantity distribution at that location This represents the measured integral intensity; Combined with Stark broadening analysis of spectral lines to determine local electron density hour, Indicates the spectral width after chord integral; The electron temperature distribution was calculated by comparing the intensities of two specific spectral lines. hour, This represents the integral value of the intensity ratio.

[0057] By combining the inversion algorithm and the corresponding inversion formula, a space physics field is constructed.

[0058] By analyzing the light polarization change data acquired by the optical electric field sensor using the electro-optic effect inversion formula, the electric field can be reconstructed. This allows for direct observation of localized electric stress concentration caused by space charge accumulation or geometric effects before discharge, and precise quantification of the dynamic evolution of the electric field during discharge, providing the most direct basis for tracing the initiation cause of discharge.

[0059] in, Indicates electric field strength; It represents the change in the phase (or polarization angle) of light actually measured by the photoelectric detection system of the optical electric field sensor; It is an electro-optic constant, the value of which is determined by a combination of factors such as the electro-optic coefficient of the sensing crystal, the laser wavelength, and the optical path structure, and is obtained through pre-calibration.

[0060] By combining time-domain inversion formulas and analyzing electromagnetic waveform data collected by ultra-high frequency electromagnetic sensors, the electromagnetic field is reconstructed. The propagation, reflection, and superposition processes of the electromagnetic pulses generated by the discharge within the device are clearly reproduced. Based on the wavefront focusing principle, the initial position of the discharge can also be located with high precision.

[0061] in, This is the set focus point, usually 0; It is the wave speed of electromagnetic waves; This indicates the actual recorded waveform data. The actual intensity at that moment; For each assumed source point, calculate that point. To the location of each sensor The distance between them was calculated, and the time it took for the wave emitted from that point to reach each sensor was also calculated. t Then, extract the waveform data from the actual recorded waveform data. The actual intensity at any moment The process involves coherent superposition (summation of waves), followed by calculation of the energy intensity (squared modulus) of the superimposed signal, and then numerical comparison. Determine the grid point with the highest energy intensity value. By combining several adjacent grid points with similarly high energy intensities, a region that is most likely the location of the discharge source is determined.

[0062] Based on the acoustic inversion formula, the acoustic signals collected by the acoustic emission sensor are analyzed to reconstruct the sound pressure field. The propagation process of the mechanical shock wave generated by the discharge in the insulating oil is visualized. This is used to assess the mechanical stress effect of the discharge on the surrounding solid medium, and can also be combined with electromagnetic fields for cooperative positioning to improve positioning accuracy.

[0063] Specifically, it is necessary to obtain the time when the sound wave signal arrives at the sensor from the data recorded by the sensor. (That is, the points where the values ​​change drastically in the acoustic waveform), and then assume a series of source points. Calculate the distance from each source point to each pair of sensors ( and The difference in distance between the two sensors is divided by the difference in their receiving times. Obtain an apparent speed of sound value If the assumed source point deviates significantly from the actual source point, the obtained apparent speed of sound value will be... The degree of dispersion will be very high. If the assumed source point is close to the real source point, then all apparent sound velocity values ​​of the assumed source point will be highly discrete. The points will be relatively close and all tend to a stable and reasonable physical value (i.e., the actual speed of sound in insulating oil, about 1480 m / s). By selecting the points that are closest to each other, we can obtain another area that is most likely the location of the power source.

[0064] Based on the charge inversion formula, the charge-induced signals collected by electrostatic probes or pulsed electroacoustic sensors are analyzed to reconstruct the space charge field. This reveals the charge accumulation state on and inside the insulating medium, and is a key link in understanding the fundamental mechanism of insulation aging under DC or AC / DC combined voltage.

[0065] in, This represents the space charge density along the depth within the insulating medium. A one-dimensional distribution; It is the speed at which sound waves propagate in the insulating medium being tested; This represents the pressure wave (acoustic) signal actually recorded by the piezoelectric sensor; It is a comprehensive proportionality constant obtained through prior experimental calibration, and is related to a series of system parameters such as the applied nanosecond-level pulse electric field intensity, the sensitivity of the piezoelectric sensor, and the gain of the signal amplifier; Through calculation The time it takes for the sound waves generated at the location to reach the sensor is obtained based on the specific time. The pressure wave (acoustic) signal actually recorded by the piezoelectric sensor from the location can then be used to deduce... The actual space charge density at that location.

[0066] By collecting multi-point magnetic field data from an optical magnetic field sensor array, and solving an inverse problem based on Ampere's law, the current density field is calculated and constructed. This elevates the one-dimensional total current measurement to the level of a three-dimensional vector field, quantitatively characterizing the true transport path and intensity distribution of the current inside the discharge channel, facilitating in-depth research on the channel's conductivity characteristics.

[0067] in, It is the vacuum permeability. Representing a spatial point ,time The three-dimensional current density vector, It is a three-dimensional magnetic induction intensity vector field obtained by interpolation and reconstruction of data collected by an optical magnetic field sensor array; By performing curl operations on the magnetic flux density vector field, its local circulation density at various points in space can be quantified. Then, based on the differential form of Ampere's law (right-hand screw law), the current distribution that is its source can be calculated in reverse.

[0068] S4. Fusion Analysis and Source Tracing: The four-dimensional physical field set is integrated into the static physical scene to form a dynamic physical scene of the discharge process, and then the dynamic physical scene is comprehensively analyzed. The comprehensive analysis further includes a causal tracing process for the initiation of discharge. This process involves analyzing the electromagnetic and acoustic pressure fields using time-difference positioning, cross-validating the initial luminous point in the morphological field, and ultimately determining the precise three-dimensional spatial coordinates of the discharge initiation point. Data on the electric field and spatial charge field before the discharge initiation point are retrieved to identify the dominant physical cause that triggered the discharge. It also includes a dynamic analysis process for discharge development. This process first tracks the motion of the discharge channel head in three-dimensional space within a morphological field, analyzes the propagation path of the discharge channel head, and quantifies the development rate. Based on the position of the channel head, it combines other four-dimensional physical fields for analysis. This analysis includes: extracting the field distribution in front of the head through the electric field and space charge field to analyze the magnitude and direction of the electric driving force; extracting the current vector flowing through the channel from the current density field to assess the channel's conductivity; analyzing the energy distribution by combining the temperature field and sound pressure field to quantify the conversion efficiency of discharge energy into thermal and mechanical energy; extracting electron density and temperature from the plasma parameter field to characterize the channel's microscopic physical state; and revealing and verifying the core physical mechanism driving discharge development by establishing a dynamic correlation model between the above physical quantities, propagation path, and development rate. The dynamic digital scene is an interactive four-dimensional simulation scene, presented as a three-dimensional view that can be observed and zoomed from multiple perspectives. The interactive four-dimensional simulation scene has a timeline that can be dragged at will, which is used to review the discharge process and to conduct a detailed examination of the spatiotemporal profile of the discharge process. The interactive four-dimensional simulation scene also has a physics field control panel, which is used to visualize and render different physics field distributions.

[0069] After the four-dimensional physical field set is built, a unified spatiotemporal framework is established based on the established unified test time benchmark. Under the unified spatiotemporal framework, it is combined with the static physical scene to form a dynamic physical scene of the final discharge process.

[0070] The physics control panel allows for the independent display, hiding, or overlay of different physical field distributions such as electric fields, temperature fields, and morphological fields with varying transparency, enabling visualized rendering control for independent analysis of different fields.

[0071] By combining the dynamic physical scenario of the final discharge process, we comprehensively analyze the dominant causes of discharge and the core mechanisms driving its development.

[0072] Time-difference localization analysis was performed on the electromagnetic field and the sound pressure field. For example, the arrival time of the first over-threshold pulse detected in the electromagnetic field was taken as time zero. The time delay between this signal and the sound wave signals received by each sensor in the sound pressure field was calculated. Combined with the known speed of sound in the insulating oil (approximately 1480 m / s), the three-dimensional coordinates of the discharge region were initially calculated. This electroacoustic localization result was spatially compared and fused with the initial luminous point position captured by the morphological field at time zero, ultimately determining a discharge initiation point accurate to the sub-millimeter level after triple cross-validation.

[0073] After determining the starting point, the electric field and space charge field data of that coordinate point at the moment before discharge (e.g., T=-1μs) are retrieved. This mainly involves extracting the background electric field intensity at that coordinate point and the surface charge accumulation density of the space charge field near that point. If the discharge occurs when the electric field intensity is close to or exceeds the theoretical breakdown threshold of the insulating material, it is a discharge caused by the geometric field enhancement effect; if the space charge density in the discharge region is significantly higher than that in the surrounding region, it is a discharge caused by the space charge distortion effect.

[0074] Dynamic analysis of discharge development reveals the complete physical picture of discharge channel propagation. By comparing volume data across consecutive time frames in a morphological field, the three-dimensional motion of the discharge channel head's centroid is identified and tracked, allowing analysis of the three-dimensional propagation path of the discharge channel head and calculation of its instantaneous velocity at each position. Based on the channel head's position, the system simultaneously extracts data from other four-dimensional physical fields for multi-dimensional correlation analysis.

[0075] The electric field vector distribution and charge polarity in a small region in front of the head are extracted from the electric field and space charge field, and the electrostatic force acting on the space charge of the head is calculated to facilitate quantitative analysis of the magnitude and direction of the electric driving force and its guiding effect on the propagation path. The current density vector flowing through the head of the channel is obtained in the current density field. The instantaneous current through the head is obtained by integrating over a small cross section, and then the local conductivity of the channel is evaluated. By combining the temperature field and the sound pressure field, based on the temperature rise rate along the channel path and the peak sound pressure in the surrounding medium, the energy dissipation power generated by the Joule heating effect and the mechanical impact energy generated by the rapid thermal expansion of the channel are quantified. In the plasma parameter field, the radial distribution profile of electron density and electron temperature inside the channel is extracted and correlated with the macroscopic conductivity calculated by the current density field to verify the accuracy of the plasma physics model, thereby characterizing the microscopic physical state of the channel.

[0076] Finally, by plotting and modeling the correlation between all the synchronously extracted, time-evolving multidimensional physical quantities and the development rate, the core physical mechanism driving the development of discharge is revealed and verified, generating a detailed dynamic analysis report containing multidimensional correlation curves and quantitative indicators.

[0077] Example 2 Reference Figure 4 , Figure 5 and Figure 11 These are two embodiments of the present invention. This embodiment provides a converter transformer bushing discharge analysis system based on multi-physics field data fusion. This system is typically deployed on a high-performance computing server to execute the steps described in Embodiment 1.

[0078] The system's software architecture consists of a set of highly collaborative functional modules, specifically including: Calibration module: Analyzes data from the vision sensor, obtains spatial installation information of the test device, and sends model calibration information to the subsequent scene definition and construction module; Scene building module: Receives model calibration information sent by the calibration module, and according to the real test scenario, retrieves the required geometric model from the experimental model library for virtual assembly, and calls the simulation physics library to assign physical properties to the model in order to build a calibrated static physical model; Main control clock module: provides a synchronized clock signal for the discharge controller and the data stream acquisition unit; Acquisition planning module: Analyzes the optimal acquisition frequency of each sensor, sets the baseline time period, defines the acquisition window around the trigger node, and plans and sets the working mode of each sensor for the window; Physics Field Reconstruction Module: Analyzes the data stream collected by the data stream acquisition device and generates a four-dimensional physics field set; 4D rendering module: merges and renders static physical models with 4D physical field sets to generate dynamic physical scenes; Comprehensive Analysis Module: It has two built-in algorithms for causal tracing and dynamic analysis. It is responsible for combining the four-dimensional physical field set to analyze the location of the discharge initiation point and the determination of the cause, as well as the discharge development process, and generating analysis results. Interactive Analysis Platform: Receives dynamic physical scenes generated by the 4D rendering module and analysis results generated by the comprehensive analysis module. It provides a graphical user interface with embedded dynamic physical scenes, a controllable timeline, and a physics field control panel. The analysis results are integrated into the interface for display in the form of graphics and text.

[0079] After setting up and completing the realistic test scenario, a new project is launched through the interactive analysis platform. The system will then invoke the scenario building module.

[0080] First, select the basic model of the discharge observation device and a specific type of converter transformer bushing model from the experimental model library. Then, select a discharge mode according to the experimental purpose. At this time, the scene building module will automatically retrieve the corresponding "needle-plate-electrode" model from the model library and perform preliminary assembly of the three components in virtual space. Afterward, the scene building module will query the simulation physics library and assign physical parameters under standard operating conditions to each component based on the materials defined during the modeling process, thus constructing an initial physical model.

[0081] Then, the calibration module is activated, and the vision sensor connected to the system (such as a structured light 3D scanner) is activated to scan the inside of the device and obtain 3D point data of the actual installation location. The 3D point data is analyzed by the built-in spatial registration algorithm to generate model calibration information, which is then sent to the scene building module. The scene building module then precisely corrects the position and orientation of each structure in the initial physical model, ultimately generating a static physical model that is completely consistent with physical reality.

[0082] After the static physical model is built, acquisition commands are issued through the interactive analysis platform. At this time, the main control clock module will continuously distribute high-precision synchronization clock signals to the discharge controller and data stream acquisition unit.

[0083] The data acquisition planning module is activated to analyze the characteristics of all currently connected sensors, automatically calculate and set an optimal reference time period. Then, based on the experimental procedure, three data acquisition windows are planned: background, dynamic changes, and recovery. The operating mode that each sensor should switch to is set for each window.

[0084] Before the discharge begins, all sensors are activated, and the data stream acquisition unit processes data streams from all sensors at different rates in parallel. Each received data point is appended with a precise timestamp provided by the main control clock module, and the separated, timestamped data streams from each sensor are independently and in real time written to the high-speed storage array.

[0085] Once the discharge begins, the discharge controller executes the pre-input instructions from the acquisition and planning module, and sends a trigger pulse to the discharge signal generator at a preset reference time period node to initiate the discharge.

[0086] After the discharge recovery process ends, the data received by the sensor tends to stabilize. After receiving stable sensor data, the data stream acquisition unit will shut down the sensor and send a discharge end message to the system. At this time, the system will activate the physical field reconstruction module.

[0087] First, the various data streams generated by the data stream acquisition unit are read from the storage array, and a series of parallel and specialized algorithm engines are activated to process the data and finally generate a four-dimensional physical field set.

[0088] After all four-dimensional field reconstructions are completed, the system enters the final analysis and presentation stage, activating the four-dimensional rendering module and the comprehensive analysis module.

[0089] The four-dimensional rendering module combines a static physical model with a four-dimensional physical field set, and uses a built-in high-performance rendering engine to perform fusion rendering in virtual space to generate an interactive dynamic physical scene, which is then output to an interactive analysis platform.

[0090] At the same time, the comprehensive analysis module will activate the built-in algorithms for causal attribution and dynamic analysis. It will automatically perform complex analyses such as multi-field co-location of the discharge initiation point and multi-dimensional physical quantity correlation calculations during the development process, and output the analysis results (such as attribution conclusions, dynamic correlation curves, and lists of key parameters) in a structured manner to the interactive analysis platform.

[0091] Finally, a complete, integrated view of all information can be seen on the unified graphical user interface of the interactive analysis platform. At the center of the interface is a freely manipulated 3D view of the 4D scene provided by the 4D rendering module. The sides of the interface clearly display various analysis results generated by the comprehensive analysis module in the form of charts, data panels, and text summaries. The timeline on the interface can be dragged to replay the discharge process, and different physical fields can be selectively overlaid on the physics field control panel for further research.

[0092] Example 3 Reference Figures 6-10These are three embodiments of the present invention. One embodiment provides a converter transformer bushing discharge analysis device based on multi-physics data fusion. This device is used to execute the method described in Embodiment 1 and serves as an experimental platform for the system described in Embodiment 2. Specifically, it includes: The housing unit 100 includes a housing 101. A manhole 102 is provided on the top of the housing 101. An oil inlet pipe 103 and a sensor wiring port 104 are respectively provided on both sides of the manhole 102 on the top of the housing 101. A fixing flange 105 is provided on the side of the housing 101. An oil outlet pipe 106 is connected to the geometric center of the bottom of the housing 101. The observation unit 200 includes a main observation window 202 that is movably installed on the top of the manhole 102 via a hinge 201. An air outlet pipe 203 is connected to the main observation window 202. Four secondary observation windows 204 are provided through the side of the housing 101. A positioning grid 205 is provided on the inner wall of the housing 101. Sensor mounting posts 206 are provided on the grid points of the positioning grid 205. The discharge unit 300 includes a grounding post 301 fixedly installed on the internal axis of the housing 101. A discharge platform 302 is provided on the top of the grounding post 301. An insulating plate 303 is fixedly installed inside the housing 101 above the discharge platform. A through hole 304 is opened in the insulating plate 303 at the geometric center of the discharge platform 302. The discharge unit 300 also includes: A connecting rod 305 has a slot 306 at one end of the top of the through hole 304 and a fixing hole 307 at the end of the connecting rod 305 that connects to the sleeve. The discharge electrode 308 has a threaded rod 309 on its top. The threaded rod 309 passes through the through hole 304 and the slot 306 in sequence, and is engaged with a nut for locking the connecting rod 305 and the insulating plate 303.

[0093] In the experimental preparation phase, the types and quantities of sensors required for the test were first determined, and the installation positions of each sensor were planned. All sensor probes with transmission lines were inserted into the internal reaction chamber of the housing 101 through sensor connection ports 104. Each sensor probe was then fixed to its corresponding sensor mounting post 206 through the manhole 102, and the wiring was secured using the sensor mounting posts 206 without sensors installed. After the internal sensors were installed, the sensor connection ports 104 were sealed with sealing material. The sensors were connected to the acquisition planning module, and in the virtual test scenario, the sensor probes were installed onto their corresponding virtual sensor mounting posts 206. A high-speed camera was then fixed to the observation window.

[0094] After installing the discharge bushing via the fixing flange 105, the connecting rod 305 is fixedly connected to the metal discharge tube of the bushing by bolts and nuts passing through the fixing hole 307. The required discharge electrode 308 is selected, and insulating paper is placed between the discharge electrode 308 and the discharge platform 302. The threaded rod 309 at the top of the discharge electrode 308 passes through the through hole 304 and the slot 306 in sequence, and is fixedly installed on the connecting rod 305 and the insulating plate 303 by nuts 310. The grounding post 301 is connected to the grounding wire for grounding.

[0095] After completing the internal installation, close and seal the main observation window 202 at the top of the manhole 102, shut off the oil outlet pipe 106, open the oil inlet pipe 103 and the vent pipe 203, inject insulating oil into the housing 101, and quickly expel the internal gas through the vent pipe 203. During this process, the system will complete the calibration of the physical model, coordinate the sensor acquisition frequencies, build multiple simulation scenarios, and import the corresponding sensor coordinates.

[0096] After the oil filling is completed, the oil inlet pipe 103 is closed, and the gas outlet pipe 203 is connected to the gas analysis device. The discharge signal generator is energized for the bushing at the preset discharge time. The sensor and high-speed camera record the discharge process. The gas analysis device analyzes the gas generated by the ionization of the insulating oil during the discharge process. After the discharge is completed, the power is cut off, and the insulating oil is recovered through the oil outlet pipe 106 for subsequent operations.

[0097] Preferably, the secondary observation window 204 has a top-down viewing angle of 90° to improve the accuracy of building stereo vision. The angles between the four sensors and the horizontal plane are 0°, 45°, 30°, and 45° in clockwise order.

[0098] Preferably, the secondary observation window 204 is located inside the housing 101 and sealed at one end by the observation glass. If it is located on the outside, air will be trapped inside the observation window after oil is injected, which will affect the observation quality.

[0099] Example 4 Reference Figure 12 and Figure 13 This is one of the four embodiments of the present invention. This embodiment is the initial version of the converter transformer bushing discharge analysis device based on multi-physics data fusion in Embodiment 3, and specifically includes: The housing 101 has a main observation window 201, an oil inlet pipe 103, and an air outlet pipe 203 on its top. A fixed flange 105 is located at the beveled surface and geometric center of the front end of the housing 101. A secondary observation window 204 is provided on the side of the housing 101. An oil outlet pipe 106 is connected to the side of the housing 101 near the bottom. A ground electrode 401 is provided on the back of the housing. A conductive rod 402 connects the ground electrode 401 and the grounding post 301. The remaining structure is the same as in Embodiment 3.

[0100] The usage process is largely the same as that of Example 3. To observe the discharge process within the housing 101, a main observation window 201 and a secondary observation window 204 were established, and the discharge process was captured by a high-speed camera. However, the captured image data could not be used for spatial modeling due to insufficient three-dimensional perspective and lack of specific modeling references. Therefore, a comprehensive analysis combining images from each observation window was necessary during experimental analysis. Furthermore, there were no standardized installation procedures for the sensors. These problems were discovered during the actual use of the initial version of the device, leading to improvements and the final device of Example 3.

[0101] Example 5 Reference Figure 9 , Figure 14 and Figure 15 These are five embodiments of the present invention. Compared to embodiment 3, the main structure of the housing 101 is changed to a standard sphere, the internal positioning grid 205 is a standard spherical coordinate system grid, and the sensor mounting posts 206 are all arranged radially along the sphere. Compared to embodiment 3, the positioning effect of this embodiment is better, which can greatly improve the efficiency of physical field construction. However, the production process requirements are more stringent. It is necessary to ensure that the inner wall of the housing 101 is a standard spherical surface without any deviation, the drawing of the positioning grid 205 must also be accurate, each sensor mounting post 206 is arranged radially along the sphere, and the sensor mounting posts 206 must be calibrated before each experiment.

[0102] In summary, by constructing a multiphysics joint observation and analysis framework, a transient and irreversible complex physical discharge process is transformed into a replayable, analyzable, and quantifiable four-dimensional digital image. Its core lies in the first-ever integration of multiple discrete and mature inversion and reconstruction algorithms under a unified spatiotemporal reference for collaborative data processing. This provides a comprehensive and high-fidelity digital analysis tool for studying high-voltage discharge physical processes, significantly improving the depth and efficiency of discharge experimental research.

[0103] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for analyzing bushing discharge in converter transformers based on multi-physics data fusion, characterized in that, Performed by a computer device, including the following steps: S1: Establish a unified spatiotemporal benchmark: Based on the discharge test environment, build a static physical model in the virtual computing environment, establish a test scenario based on the spatial location information of the sensors in the test environment, and establish a unified test time benchmark for the sensors; S2: Coordinated Acquisition and Event Triggering: Coordinate the operation of the sensor and the discharge signal generator to acquire the data stream continuously generated by the sensor according to the coordinated acquisition frequency, and set the time reference for the data stream with the discharge trigger signal issued by the generator. S3: Reconstruct a four-dimensional data field: Process the data stream to reconstruct a four-dimensional physical field set that can characterize the dynamic distribution of the discharge process in the test scenario; S4: Fusion Analysis and Source Tracing: The four-dimensional physical field set is integrated into the static physical scene to form a dynamic physical scene of the discharge process, and then the dynamic physical scene is comprehensively analyzed.

2. The method for analyzing converter transformer bushing discharge based on multi-physics data fusion according to claim 1, characterized in that, The process of building a static physical model in step S1 specifically includes: calling a preset test device model, importing the three-dimensional model of the bushing of the converter transformer to be tested, configuring the corresponding electrode model according to the selected discharge mode to form an uncorrected digital model, using a visual sensor to obtain the spatial installation information of the physical entity, correcting the digital model through the spatial installation information, and assigning physical attribute parameters to the corrected digital model.

3. The method for analyzing bushing discharge of converter transformers based on multi-physics data fusion according to claim 1, characterized in that, The coordinated operation of the sensor and the discharge signal generating device in step S2 further includes: by analyzing the optimal acquisition frequency of each sensor, setting a reference time period in the test time reference to ensure that all sensors acquire data synchronously at each node of the reference time period; preset the discharge trigger signal at a preset node of the reference time period, determine three consecutive data acquisition windows around the preset node for acquiring background, dynamic change and recovery phases, and set an appropriate working mode for the sensor according to the target characteristics of each data acquisition window.

4. The method for analyzing bushing discharge of converter transformers based on multi-physics data fusion according to claim 1, characterized in that, The process of reconstructing the four-dimensional physical field set in step S3 further includes constructing multiple four-dimensional physical fields to characterize different physical dimensions. The four-dimensional physical fields include: a morphological field that intuitively reproduces the three-dimensional geometric shape of the discharge channel and its evolution over time; a temperature field that depicts the dynamic changes in the temperature distribution of the discharge region and its surroundings; a plasma parameter field that quantitatively characterizes the distribution of microscopic physical properties inside the discharge channel; an electric field that characterizes the distribution of local electric stress and its dynamic evolution during the discharge process; an electromagnetic field that directly reflects the spatial propagation path of electromagnetic waves during the test; a sound pressure field used to characterize the propagation of acoustic pressure generated by the discharge in space; and a space charge field that reveals the state of charge accumulation on the surface and inside the insulating medium.

5. The method for analyzing bushing discharge of converter transformers based on multi-physics data fusion according to claim 4, characterized in that, The specific reconstruction process of the multiple four-dimensional physical fields is as follows: a morphological field is established based on multi-view image data collected by a high-speed camera using a tomographic imaging algorithm, and a temperature field is constructed using thermal imaging data collected by an infrared thermal imager; the plasma parameter field is obtained by combining the spectral data collected by the spectrometer with the morphological field through Abelian inversion; the electric field, electromagnetic field, sound pressure field, and space charge field are reconstructed by inversion algorithms on the light polarization change data collected by the optical electric field sensor, the electromagnetic waveform data collected by the ultra-high frequency electromagnetic sensor, the acoustic signal collected by the acoustic sensor, and the charge induction signal collected by the sensor; and the current density field is calculated based on the multi-point magnetic field data collected by the optical magnetic field sensor array.

6. The method for analyzing bushing discharge of converter transformers based on multi-physics data fusion according to claim 4, characterized in that, The comprehensive analysis in step S4 further includes a causal tracing process for the initiation of discharge. The causal tracing process for the initiation of discharge analyzes the electromagnetic field and the sound pressure field through time difference positioning, and performs cross-verification with the initial luminous point in the morphological field to finally determine the precise three-dimensional spatial coordinates of the discharge initiation point. The data of the electric field and the spatial charge field before the discharge initiation point are retrieved to determine the dominant physical cause that triggers the discharge.

7. The method for analyzing bushing discharge of converter transformers based on multi-physics data fusion according to claim 4, characterized in that, The comprehensive analysis in step S4 also includes a dynamic analysis process for discharge development; the dynamic analysis process for discharge development first tracks the motion process of the discharge channel head in three-dimensional space in the morphological field, analyzes the propagation path of the discharge channel head, and quantifies the development rate. Based on the position of the channel head, analysis is performed in conjunction with other four-dimensional physical fields. This combined analysis includes: extracting the field distribution in front of the head through the electric field and the space charge field to analyze the magnitude and direction of the electric driving force; extracting the current vector flowing through the channel from the current density field to evaluate the channel's conductivity; analyzing the energy distribution by combining the temperature field and the sound pressure field to quantify the conversion efficiency of discharge energy into thermal and mechanical energy; extracting electron density and temperature from the plasma parameter field to characterize the channel's microscopic physical state; and revealing and verifying the core physical mechanism driving discharge development by establishing a dynamic correlation model between the above physical quantities and the propagation path and the development rate.

8. The method for analyzing bushing discharge of converter transformers based on multi-physics data fusion according to claim 1, characterized in that, The dynamic digital scene formed in step S4 is an interactive four-dimensional simulation scene, presented as a three-dimensional view that can be observed and zoomed from multiple perspectives. The interactive four-dimensional simulation scene has a timeline that can be dragged at will, which is used to review the discharge process and to conduct a detailed inspection of the spatiotemporal profile of the discharge process. The interactive four-dimensional simulation scene also has a physics field control panel, which is used to visualize and render different physics field distributions.

9. A converter transformer bushing discharge analysis system based on multi-physics data fusion, characterized in that, The system is used to perform the method according to any one of claims 1 to 8, specifically including: Calibration module: Analyzes data from the vision sensor, obtains spatial installation information of the test device, and sends model calibration information to the subsequent scene definition and construction module; Scene building module: Receives the model calibration information sent by the calibration module, and according to the real test scenario, retrieves the required geometric model from the experimental model library for virtual assembly, and calls the simulation physics library to assign physical properties to the model in order to build a calibrated static physical model; Main control clock module: provides a synchronized clock signal for the discharge controller and the data stream acquisition unit; Acquisition planning module: Analyzes the optimal acquisition frequency of each sensor, sets the reference time period, defines the acquisition window around the trigger node, and plans and sets the working mode of each sensor for the window; Physics field reconstruction module: Analyzes the data stream collected by the data stream acquisition device and generates a four-dimensional physics field set; Four-dimensional rendering module: merges and renders the static physical model with the four-dimensional physical field set to generate a dynamic physical scene; Comprehensive Analysis Module: It has two built-in algorithms for causal tracing and dynamic analysis. It is responsible for combining the four-dimensional physical field set to analyze the location of the discharge initiation point and the determination of the cause, as well as the discharge development process, and generating analysis results. Interactive analysis platform: Receives the dynamic physical scene generated by the four-dimensional rendering module and the analysis results generated by the comprehensive analysis module, and provides a graphical user interface. The interface embeds the dynamic physical scene and has a controllable timeline and a physics field control panel. The analysis results are integrated into the interface in the form of graphics and text for display.

10. A converter transformer bushing discharge analysis device based on multi-physics data fusion, characterized in that, The apparatus is used to perform the method of any one of claims 1 to 8 and serves as an experimental carrier for the system of claim 9, specifically comprising: The housing unit (100) includes a housing (101), a manhole (102) is provided on the top of the housing (101), an oil inlet pipe (103) and a sensor wiring port (104) are respectively provided on both sides of the manhole (102) on the top of the housing (101), a fixing flange (105) is provided on the side of the housing (101), and an oil outlet pipe (106) is connected to the geometric center of the bottom of the housing (101). The observation unit (200) includes a main observation window (202) that is movably installed on the top of the manhole (102) via a hinge (201). An air outlet pipe (203) is connected to the main observation window (202). Four secondary observation windows (204) are provided through the side of the housing (101). A positioning grid (205) is provided on the inner wall of the housing (101). Sensor mounting posts (206) are provided on the grid points of the positioning grid (205). The discharge unit (300) includes a grounding post (301) fixedly installed on the internal axis of the housing (101), a discharge platform (302) is provided on the top of the grounding post (301), and an insulating plate (303) is fixedly installed inside the housing (101) above the discharge platform. The insulating plate (303) has a through hole (304) at the geometric center of the discharge platform (302). The discharge unit (300) further includes: A connecting rod (305) is provided with a slot (306) at one end of the top of the through hole (304), and a fixing hole (307) is provided at the end of the connecting rod (305) that is connected to the sleeve. The discharge electrode (308) has a threaded rod (309) on its top. The threaded rod (309) passes through the through hole (304) and the slot (306) in sequence, and is engaged with a nut for locking the connecting rod (305) and the insulating plate (303).

Images