A transformer comprehensive analysis method and device, a terminal and a storage medium

Abstract

The application discloses a transformer comprehensive analysis method, and discloses a device, a terminal and a storage medium with the transformer comprehensive analysis method, wherein the transformer comprehensive analysis method determines the current working state and aging degree of the transformer by inputting the data of the insulation oil, the insulation oil paper and the reliability and short-circuit resistance of the transformer into a trained multi-scale attention convolution GRU neural network.

Description

Technical Field

[0001] This invention relates to the field of electrical equipment monitoring technology, and in particular to a method, apparatus, terminal and storage medium for analyzing the performance of transformer oil paper. Background Technology

[0002] Transformers, as crucial equipment in power systems, undertake key tasks such as voltage transformation, power distribution, and transmission, playing a vital role in providing high-quality power services and ensuring the safe, reliable, high-quality, and economical operation of the power system. The stable operation of power equipment under long-term, high-load, and all-weather conditions plays a fundamental role in national production and the economy. Therefore, ensuring the stable operation of power equipment is essential, and one means of ensuring this is through various methods of fault diagnosis. Power equipment fault diagnosis technology is a practical discipline that combines engineering practice and plays a vital role in power production applications.

[0003] Currently, dynamic information monitoring technology mainly uses single-dimensional signals such as vibration, ultrasound, temperature, light signals, radiation, and electromagnetic signals for analysis, but lacks a comprehensive monitoring and detection technology for transformer status. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art. To this end, this invention proposes a transformer condition monitoring and detection method, which can continuously monitor the operating status of transformers at all times and under various load conditions.

[0005] The present invention also proposes an apparatus having the above-mentioned transformer condition monitoring and detection method.

[0006] The transformer comprehensive analysis method according to a first aspect of the present invention is characterized by comprising the following steps:

[0007] Several transformer parameters were obtained using sensors.

[0008] Based on the transformer parameters, determine the multi-dimensional transformer operating status corresponding to each transformer parameter;

[0009] A neural network using multi-scale attention convolutional GRU is used to determine the current true state of the transformer based on the multi-dimensional transformer operating state.

[0010] The transformer comprehensive analysis method according to the present invention has at least the following beneficial effects: The method proposed in this application performs comprehensive analysis on data from multiple dimensions of the transformer, trains a neural network model using the organized data, and then establishes the correspondence between the transformer's operating status and aging condition under multi-dimensional data conditions, providing a data foundation for determining the transformer's operating status under complex conditions in the future.

[0011] According to some embodiments of this application, the transformer parameters include at least:

[0012] At least one of the following: insulating oil performance parameters, insulating paper performance parameters, transformer reliability parameters, and short-circuit withstand capability parameters.

[0013] According to some embodiments of this application, the process of determining the transformer's operating state based on the insulating oil performance parameters specifically includes:

[0014] The aging status assessment parameters are obtained by the aging status assessment scoring method based on matter-element theory;

[0015] The basic score of the transformer is obtained through factory testing, acceptance testing, and safe operation records.

[0016] By using the group component status, defect management, and defect content in the transformer family record, the family defect correction parameters of the transformer can be obtained.

[0017] By analyzing transformer outlet short circuit, overload, and overexcitation, correction parameters for adverse operating conditions are obtained.

[0018] The evaluation results of the insulating oil paper are determined based on the aging condition assessment parameters, the transformer's basic score, the family defect correction parameters, and the adverse operating condition correction parameters.

[0019] According to some embodiments of this application, the formula used in the step of determining the evaluation result of the insulating oil paper based on the aging condition assessment parameters, the transformer's basic score, the family defect correction parameters, and the adverse operating condition correction parameters is as follows:

[0020] G = (w1B + w2T) × E × F

[0021] Wherein, B is the basic score of the transformer, F is the family defect correction parameter of the transformer, E is the adverse operating condition correction parameter, T is the aging condition assessment parameter, and w1 and w2 are the weights of the basic score B and the aging condition assessment parameter T of the transformer, respectively.

[0022] According to some embodiments of this application, w1 = 0.5; w2 = 0.5.

[0023] According to some embodiments of this application, the process of determining the operating state of a transformer based on the transformer reliability parameters specifically includes:

[0024] Obtain technical compliance assessment data from transformer manufacturers;

[0025] Obtain a summary of historical transformer issues;

[0026] The technical compliance assessment data and historical problem summary data are processed.

[0027] According to some embodiments of this application, the step of processing the technology compliance assessment data and historical problem summary data specifically includes:

[0028] The data is then classified and standardized.

[0029] Perform a rationality analysis on the data and then repair it;

[0030] Encrypt the processed data.

[0031] According to some embodiments of this application, the steps for obtaining the transformer operating state based on transformer reliability parameters include:

[0032] Collect and process transformer data to obtain processed transformer data;

[0033] The processed transformer data is analyzed to obtain the transformer reliability status.

[0034] According to some embodiments of this application, the step of analyzing the processed transformer data to obtain the transformer reliability status includes:

[0035] Based on data from the power grid resource business platform, information on different operating states of transformers is collected.

[0036] Based on the technical compliance assessment, a comprehensive reliability score is determined for each piece of equipment;

[0037] Based on the equipment's entire life cycle data, the importance classification of the equipment and its components within the station is determined.

[0038] The reliability index of a group of equipment of different types, manufacturers and models is calculated by using equipment defect rate, failure rate, maintenance cycle and remaining life.

[0039] The processed transformer data includes information on technical compliance assessment, equipment lifecycle data, equipment defect rate, failure rate, maintenance cycle, and remaining lifespan.

[0040] According to some embodiments of this application, the process of determining the transformer operating state based on the short-circuit withstand capability parameters specifically includes:

[0041] Based on the short-circuit parameters, determine the magnetic flux density;

[0042] Based on the magnetic flux density, calculate the short-circuit impedance and the magnitude of the Ampere force;

[0043] Calculate the average annular tensile and compressive stress of the coil conductor, the maximum axial bending stress of the coil conductor, and the maximum axial bending stress of the coil conductor.

[0044] Calculate the maximum short-circuit current and verify it based on the current stress conditions.

[0045] A second aspect of this application provides a transformer comprehensive analysis device, comprising:

[0046] The insulating oil performance analysis module can analyze the performance of insulating oil and obtain its performance data.

[0047] The insulating oil paper performance analysis module can analyze the performance of insulating oil paper and obtain its performance data.

[0048] The transformer reliability analysis module can analyze the processed transformer data to obtain transformer reliability data;

[0049] The short-circuit withstand capability parameter analysis module can calculate and determine the transformer's short-circuit current withstand performance data based on the short-circuit withstand parameters;

[0050] The data processing module can preprocess the performance data of the insulating oil, the performance data of the insulating paper, and the reliability of the transformer to obtain preprocessed transformer data.

[0051] The neural network training module can use a multi-scale attentional convolutional GRU network to learn the preprocessed data (vibration, acoustic signature, oil chromatography, and other state data) to build a transformer fault database.

[0052] A third aspect of this application provides a terminal comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aforementioned transformer comprehensive analysis method.

[0053] According to a fourth aspect of this application, a computer-readable storage medium is provided that stores computer-executable instructions for performing the above-described transformer comprehensive analysis method.

[0054] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0055] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0056] Figure 1This is a schematic diagram illustrating the steps of the transformer comprehensive analysis method according to an embodiment of the present invention;

[0057] Figure 2 for Figure 1 The diagram shows the breakdown steps of step S200 in the schematic diagram of the transformer comprehensive analysis method.

[0058] Figure 3 for Figure 1 The diagram shows the breakdown steps of step S301 in the schematic diagram of the transformer comprehensive analysis method.

[0059] Figure 4 for Figure 1 The diagram shows the breakdown steps of step S302 in the schematic diagram of the transformer comprehensive analysis method.

[0060] Figure 5 This is a breakdown diagram of step S400 in the schematic diagram of the transformer comprehensive analysis method shown in this application.

[0061] Figure 6 This is a schematic diagram illustrating the radial force and circumferential stress experienced by the coil as described in the embodiments of this application;

[0062] Figure 7 This is a schematic diagram illustrating the bending of a coil under axial force as described in the embodiments of this application;

[0063] Figure 8 This is a schematic diagram illustrating the bending of a coil under radial force as described in the embodiments of this application;

[0064] Figure 9 A schematic diagram illustrating the steps of a method for determining the performance and reliability status of various components of a transformer, as provided in an embodiment of this application.

[0065] Figure 10 This is a schematic diagram of the structure of the transformer comprehensive analysis device provided in the embodiments of this application. Detailed Implementation

[0066] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0067] In the description of this invention, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0068] As a crucial component of the power system, the stable operation of transformers significantly impacts the overall stability of the power system. However, current technologies often analyze transformer operating conditions using only one dimension, such as vibration, ultrasound, temperature, optical signals, radiation, or electromagnetic fields. This approach can introduce substantial errors under complex operating conditions and extreme environments.

[0069] To address the limitations of existing technologies, this application proposes a comprehensive transformer analysis method. This method utilizes sensors to measure several transformer parameters, then determines the transformer state corresponding to each parameter, and finally uses a trained neural network to determine the current transformer state.

[0070] Example 1: This example proposes a method based on four parameters: insulating oil performance parameters, insulating paper performance parameters, transformer reliability parameters, and short-circuit parameters.

[0071] Reference Figure 1 This includes the following steps:

[0072] Step S100: Analyze the performance of the insulating oil to obtain its performance data.

[0073] The performance of insulating oil itself is analyzed using an oil cup. The evaluation items include lightning impulse breakdown characteristics, water absorption and dehydration characteristics, partial discharge characteristics, dielectric properties, compatibility with typical materials, and oil performance after mixing.

[0074] The oil cups used for testing different properties of insulating oil are different, and corresponding patents have been applied for. They belong to existing technology, and their principles will not be elaborated here.

[0075] Step S200: Analyze the performance of the insulating oil paper to obtain its performance data.

[0076] By mining data from the entire transformer process, a scientific evaluation method was established. This included developing a transformer aging health index, a transformer insulation aging status assessment technology based on dielectric response, an absolute insulation life prediction model, a reliability life prediction model, and an economic life prediction model. Furthermore, a transformer maintenance and decommissioning decision-making model based on aging status assessment and life prediction was established.

[0077] Step S300: Analyze the processed transformer data to obtain transformer reliability data.

[0078] By implementing reliability-centered condition assessment technology, we will conduct equipment reliability analysis, establish a data system, and collect reliability-centered condition maintenance-related information based on the digital archives of equipment health management. This includes information on the entire process of design, manufacturing, installation, acceptance, and operation. We will build a reliability data system with a wide range of data sources, comprehensive business coverage, and complete condition parameters to support the development of equipment reliability analysis, failure mode analysis, and failure risk assessment.

[0079] Step S400: Calculate based on the short-circuit parameters to determine the transformer short-circuit current performance data.

[0080] The magnetic field of the coil region is calculated using the finite element method. The coil disc and oil channel regions are merged into an equivalent calculation region and the average current density is considered. The model of the calculation region is relatively simplified, and the radial dimension and axial height of the coil can be directly input to maintain consistency with the design.

[0081] Step S500: Preprocess the performance data of the insulating oil, the performance data of the insulating paper, the transformer reliability data, and the short-circuit data to obtain preprocessed transformer data.

[0082] For various data types, data preprocessing is required to obtain a robust transformer operating condition monitoring and fault diagnosis model. The main processes include filtering, resampling, and data normalization. Filtering removes some electrical noise, and resampling ensures that signal variables at different sampling rates remain consistent across the time dimension; this typically involves both downsampling and upsampling.

[0083] Step S600: Use the Shenxing Network of Multi-Scale Attention Convolutional GRU to learn the preprocessed data and construct a transformer fault database.

[0084] To address the complex relationships between different sensor signals across different time spans and scales, this method extracts features using convolutional neural networks corresponding to different scales at multiple time points, and learns the correlation between key time / feature scales through an attention-based convolutional gate recurrent unit (Att-ConvG RU).

[0085] The steps S100 to S600 above illustrate the general steps of the method described in this application. For easier understanding, the specific process of each step will be explained in more detail below.

[0086] Step S100: Analyze the performance of the insulating oil to obtain its performance data.

[0087] The properties of insulating oil are determined by using different test oil cups. The structure and principle of the oil cups have been described in other applications and will not be repeated here.

[0088] This step establishes the correspondence between various indicators and properties of insulating oil.

[0089] Step S200: Analyze the performance of the insulating oil paper to obtain its performance data.

[0090] Insulating oil paper is an important component of transformers. Its aging condition often reflects the aging condition of the transformer itself. (Refer to...) Figure 2 The assessment of aging status is mainly based on four parameters, namely:

[0091] Step S201: Obtain aging status assessment parameters using the aging status assessment scoring method based on matter-element theory.

[0092] The aging status assessment parameter T is obtained by the aging status assessment scoring method based on matter-element theory;

[0093] Step S202: Obtain the basic score of the transformer by using information such as transformer factory tests, handover tests, and safe operation records.

[0094] The transformer's basic score of B is obtained by using information such as factory tests, acceptance tests, and safe operation records.

[0095] Step S203: Obtain the family defect correction parameters of the transformer by using information such as the status of components, defect management, and defect content in the transformer family record.

[0096] The transformer family defect correction parameter F is obtained by using information such as component status, defect management, and defect content in the transformer family record;

[0097] Step S204: Obtain the correction parameters for adverse operating conditions by detecting short circuit, overload, and overexcitation at the transformer outlet.

[0098] By analyzing conditions such as transformer outlet short circuit, overload, and overexcitation, the correction parameter E for adverse operating conditions is obtained.

[0099] Step S205: Determine the evaluation result of the insulating oil paper based on the aging condition assessment parameters, the transformer's basic score, the family defect correction parameters, and the adverse operating condition correction parameters.

[0100] The calculation is performed based on the above parameters, and the formula used is:

[0101] G=(w1B+w2T)×E×F (1)

[0102] Among them, w1 and w2 are the weights of the transformer's basic score B and aging condition assessment parameter T, respectively.

[0103] Flowchart of the overall aging status assessment index system for transformers (refer to the flowchart) Figure 2 The live-line evaluation index system includes: 1) Oil chromatography analysis, including H2 content, C2H2 content, absolute / relative total hydrocarbon production rate, total hydrocarbon content, and CO content; 2) Oil chemicalization test, including oil micro-water, breakdown voltage, and furfural; 3) Operating conditions, including overload, temperature rise level, near-zone short circuit, and overvoltage; 4) Maintenance history, including the maintenance history of this transformer and similar transformers; 5) Visual inspection, including whether there is oil leakage, abnormal sound, and traces of discharge and flashover.

[0104] The power outage assessment index system includes: 1) Electrical test items, including absorption ratio or polarization index, winding leakage current, winding DC resistance, dielectric loss and core grounding current / resistance; 2) Bushing test items, including oil chromatography analysis (selecting H2 and C2H2), initial capacitance difference, dielectric loss and end screen grounding resistance; 3) Tap changer inspection test items, including number of operations, voltage division ratio and mechanical performance; 4) Other accessory status, including cooling system, testing equipment and protection devices.

[0105] Step S300: Analyze the processed transformer data to obtain transformer reliability data.

[0106] The preferred approach, based on extensive practical experience, involves two parts when collecting transformer data: document review and on-site review. Before each review, it is necessary to collect, organize, and examine 4 to 6 transformer documents. Each transformer document is divided into 43 categories, totaling nearly 100 documents. Furthermore, it is necessary to extract and organize key data from each document. Therefore, the document evaluation work alone is quite tedious.

[0107] To better address the data integration issue, step S300 can be divided into the following two steps:

[0108] Step S301: Collect and process transformer data to obtain processed transformer data.

[0109] When reviewing transformer data, the complexity of the data and the need to carry a large amount of paper documents during on-site assessments due to environmental limitations reduce efficiency. Therefore, a simple and efficient data collection and processing method is proposed in practice. (Refer to...) Figure 3 ,include:

[0110] Step S3011: Obtain technical compliance assessment data from the transformer manufacturer.

[0111] Collect technical compliance assessment data submitted by provincial companies and transformer manufacturers.

[0112] Step S3012: Obtain summary data of historical transformer problems.

[0113] Before the joint meeting, provincial companies, together with the supervising units and municipal companies, will collect historical operational issues of supplier products during the operation and maintenance process and compile a "Summary Table of Historical Issues" to provide feedback to suppliers. The "Summary Table of Historical Issues" should include a brief description of the fault, the location of the fault, an analysis of the cause of the fault, and suggestions for follow-up work.

[0114] Step S3013: Process the technology compliance assessment data and the historical problem summary data.

[0115] The collected data is often inconsistent in format and contains errors and omissions. Further improvements are needed. Specifically, this includes:

[0116] Step S3013-A: Classify and standardize the data.

[0117] A large amount of technical compliance assessment data has been collected, and it is necessary to classify and standardize the collected technical compliance assessment review data and other data.

[0118] Step S3013-B: Perform a rationality analysis on the data and repair it.

[0119] For the acquired data, techniques such as statistics, clustering, association analysis, and time series analysis are used to verify data quality and impute missing values ​​to ensure the validity, consistency, and completeness of the data.

[0120] Step S3013-C: Encrypt the processed data.

[0121] To prevent data from being misused during transmission, the national standard SM4 encryption is used. The key consists of the dedicated machine's CU P serial number, disk serial number, and authorization code. The encrypted data packets (CDs) generated by the dedicated machine can only be operated on the local machine; other machines have no access to them, thus preventing unauthorized modification or leakage of data during transmission.

[0122] Preferably, the acquired data is typically filtered to include key parameters, such as power plant analysis, magnetic field analysis, temperature field analysis, wave process calculation, overexcitation capacity report, service life analysis report, seismic calculation report, oil tank mechanical strength calculation report, DC bias magnetic tolerance, overload capacity, noise calculation, etc. Other non-important data is discarded directly. This simplifies the database information and reduces storage pressure. Furthermore, data templates are established based on the key parameters to simplify the process when collecting and entering data again later.

[0123] Furthermore, according to some embodiments of this application, the data is stored in the form of Table 1:

[0124] Table 1:

[0125]

[0126]

[0127]

[0128]

[0129] Step S302: Analyze the processed transformer data to obtain transformer reliability data. (Refer to...) Figure 4 This step specifically includes:

[0130] Step S3021: Based on the data from the power grid resource business platform, collect information on different operating states of transformers.

[0131] Based on data from the power grid resource business platform, we collect equipment status-related information, including information from the entire process of design, manufacturing, installation, acceptance, and operation. We have established a reliability data system with broad data sources, comprehensive business coverage, and complete status parameters to support equipment reliability analysis, failure mode analysis, and failure risk assessment.

[0132] Step S3022: Based on the technical compliance assessment, determine the comprehensive reliability score for each device.

[0133] Based on the technical compliance assessment, supplier information is reviewed from aspects such as standard implementation, product design, test reports of the same model, and key raw materials and components. Each piece of equipment is scored according to the results to obtain a basic score for the inherent reliability of the equipment. The results of the transformer short-circuit withstand capability verification are integrated, and the transformer short-circuit withstand capability is used as an influencing factor in the inherent reliability assessment to improve the accuracy of the inherent reliability assessment.

[0134] Based on the digital archives of equipment health management, we collect information related to condition-based maintenance centered on reliability, research and formulate equipment reliability indicators, identify key characteristic quantities such as transformer design parameters, supervision records, raw material parameters, factory test data, and factory report data, and analyze the inherent reliability of individual equipment.

[0135] Step S3023: Based on the equipment's full life cycle data, determine the importance classification of the equipment and its components within the station.

[0136] Based on equipment lifecycle data, and considering dynamically changing online monitoring, defect, potential hazard, testing, and operational data, the importance level of equipment and its components within the station is determined using the analytic hierarchy process (AHP) and by comprehensively considering factors such as failure risk and economic efficiency. After quantifying the status analysis of the assets and equipment, equipment reliability is estimated by combining equipment failure probability calculations, providing a basis for equipment lifecycle cost management.

[0137] Step S3024: Calculate the reliability index of group equipment of different equipment types, manufacturers and models using equipment defect rate, failure rate, maintenance cycle and remaining life.

[0138] This study calculates the reliability index of group equipment for different equipment types, manufacturers, and models using indicators such as equipment defect rate, failure rate, maintenance cycle, and remaining life. Principal component analysis (PCA) is used for data dimensionality reduction to generate principal component vectors for multiple indicators. This results in different clusters for each transformer, with each cluster labeled with a reliability tag for a single transformer, categorized as normal, warning, abnormal, and critical. The number of transformers, area, and density within each cluster are calculated for different classifications. The statistical results of the clusters are standardized, categorized as either "normal" or "abnormal," with larger values ​​considered better and smaller values ​​considered better. Finally, weights for different states are determined using the analytic hierarchy process (AHP) and expert experience to calculate the overall equipment reliability index.

[0139] Step S400: Calculate and determine the transformer short-circuit current performance data based on the short-circuit parameters. This includes:

[0140] Step S401: Determine the magnetic induction intensity based on the short-circuit parameters.

[0141] Based on the actual number of turns and rated current value of each section of the transformer coil, the equivalent current density value of each section is obtained:

[0142]

[0143] Among them, J seg N represents the piecewise equivalent current density of the coil used for magnetic field calculations. seg I represents the actual number of turns in the coil segment. rcoil The rated current of the coil is corresponding to the segment of the coil, S seg The area of ​​the coil segment (= coil segment width × coil segment height).

[0144] The equivalent solution equation for the magnetic vector potential A, derived from Maxwell's magnetic field equations, is as follows:

[0145]

[0146] Where μ is the permeability and J is the current density. In the segmented region of the coil, J is...seg The rest of the area is 0.

[0147] For the transformer coil region in a two-dimensional axisymmetric rz coordinate system, equation (3) is expressed as:

[0148]

[0149] Based on equation (4) and the boundary conditions of the solution region for the equivalent current density distribution of each segment of the coil, the boundary condition for the core window is that the magnetic lines of force are perpendicular; if copper / aluminum shielding of the box wall is considered, the box wall is set to have parallel magnetic lines of force. The solution region is meshed using the finite element method. The Delaunay triangular meshing method is adopted, and the meshing boundary lines are set at the inner and outer radii of each coil and the upper and lower limits of each segment to form a set of finite element solution equations. The distribution of magnetic vector potential A of each node is obtained by solving, and then the amplitude and axial components B of the magnetic induction intensity of each element can be obtained. re B ze .

[0150] Step S402: Calculate the short-circuit impedance and the magnitude of the Ampere force based on the magnetic induction intensity.

[0151] Step S4021: Calculate the short-circuit impedance.

[0152] Based on the magnetic flux density obtained in step S401, calculate the transformer short-circuit impedance:

[0153] The magnetic field energy of each subdivided unit is:

[0154]

[0155] By summing the magnetic field energies of each unit, the magnetic field energy of the entire region can be obtained.

[0156]

[0157] Where, N lim Given the number of core columns in the transformer, the transformer reactance is:

[0158]

[0159] Where ω is the angular frequency, I rcoil Where R is the effective value of the coil's rated current, L is the transformer inductance, and the transformer's short-circuit impedance percentage is:

[0160]

[0161] Where S r This refers to the rated capacity of the transformer.

[0162] Step S4022: Calculate the short-circuit impedance.

[0163] Based on the results of the radial and axial components Bre and Bze of the magnetic induction intensity within each unit, the magnitude of the Ampere force on each unit can be calculated (the coil is an axisymmetric structure, and each unit can be considered as a circular wire rotating once):

[0164]

[0165]

[0166] Among them, F re and F ze These represent the radial and axial components of the Ampere force acting on a specific element within a segment of the coil being solved, respectively. e The distance from the centroid of this unit to the iron core is 2πr. e This is the equivalent circumference of the circular wire loop when the unit rotates one revolution, S e J is the unit area. seg S e This is the magnitude of the current in the equivalent circular wire of the unit. It is the multiple of the peak short-circuit current relative to the rated current.

[0167] Step S403: Calculate the average circumferential tensile and compressive stress of the coil conductor, the maximum axial bending stress of the coil conductor, and the maximum axial bending stress of the coil conductor. Specifically, this includes:

[0168] Under the influence of a transformer short-circuit current, the inner coil will experience an inward radial pressure, and the outer coil will experience an outward radial tension. Since the magnetic field distribution along the circumference is basically the same (ignoring the local influence of the core window), this radial force can be considered to be uniformly distributed along the circumference of the coil. In this case, the coil can be represented by a ring uniformly stressed in the circumferential direction, such as... Figure 6 As shown. Based on the knowledge of mechanical principles, the calculation expression for the annular stress can be obtained as follows:

[0169]

[0170] Where σ cirseg f represents the average annular compressive / tensile stress in the coil segments. rseg D is the average radial force per unit length of a single conductor within a coil segment. mseg The average diameter of the coil segments, S con This represents the cross-sectional area of ​​a single conductor in the coil (conductor width × conductor height).

[0171] For a single conductor, the amplitude force f per unit length r From the previous expression for calculating electromagnetic force (2), we can obtain:

[0172] f r =B ai (12)

[0173] Among them, B a The average axial magnetic flux density component of the coil region is given by i, where i is the current in the coil.

[0174] When the software calculates the average annular compressive / tensile stress in the conductor within the coil segment, it uses formula (11) for calculation. Combining this with the Ampere force calculation formula in Section 3, and considering the maximum annular tensile / tensile stress within the coil segment, we have:

[0175]

[0176] Where F raxsegMAX ×h seg For the segment of the coil, with the maximum F raxsegMAX The total amplitude force, N, is considered when taking into account the total amplitude force. conseg F represents the number of conductors within this segment. raxsegMAX ×h seg Divide by N conseg This is the average radial force of each conductor, divided by 2πr. mseg This represents the radial force per unit length of the conductor within the coil segment. The circumferential stress is calculated as follows:

[0177]

[0178] The above calculations represent the tensile and compressive stresses of a single-layer annular structure. If the coil is not layered, the tensile and compressive stresses of a single-layer annular structure are the stresses experienced by the coil. If the coil is layered due to the presence of axial oil channels, U-shaped arrangement, or other reasons, then the tensile and compressive stresses of the entire coil need to be calculated separately for each layer based on the structure.

[0179] For general layered coils, since the stress is well transmitted between layers, they can be considered as a single coil, and the annular tensile and compressive stress within the coil segment should be the average value of each layer. For layered coils with too small axial oil channels, there is a voltage difference between the layers at the same axial height. When the coil deforms, the small interlayer distance can easily lead to interlayer breakdown. Therefore, in this case, the risk of interlayer insulation failure should be considered, and the verification should be appropriately strict. The maximum value of each layer should be taken when calculating the annular tensile and compressive stress. For U-shaped coils with large axial oil channels, considering that the layers of the coil are relatively independent, the maximum value of each layer should be taken when calculating the annular tensile and compressive stress within the coil segment.

[0180] When the coil is subjected to axial electromagnetic force, it will bend and deform between the two pads, generating axial bending stress in the coil. Figure 7 As shown.

[0181] The calculation of axial bending of the coil is based on a straight beam with a uniform load fixed at both ends. Using mechanical theory, the bending moment of the coil can be obtained as follows:

[0182]

[0183] Where f y The radial force per unit length of the conductor refers to the force exerted on the conductor, and l refers to the length of the conductor between the two pads.

[0184]

[0185] Where n s Number of padding blocks between the lines, w s The width of the pad.

[0186] The axial bending stress in the conductor can also be calculated:

[0187]

[0188] Among them W y The axial section modulus of a single conductor:

[0189]

[0190] Where h and b refer to the axial and radial dimensions of a single conductor, respectively, substituting into equation (17) yields:

[0191]

[0192] When the software calculates axial bending stress, it takes the axial force per unit length of the conductor within the segment, and combines it with the Ampere force calculation results, considering the maximum axial force on the conductor within the coil segment, then:

[0193]

[0194] in, It is considered to base the area S of the entire segmented region seg The calculated maximum axial force per unit volume is converted to N based on the actual area of ​​the conductor within the segment. conseg S con The maximum axial force per unit volume, multiplied by the area of ​​a single conductor, gives f. zseg Substituting, we get:

[0195]

[0196] For a coil, the maximum axial bending stress within different segments is the axial bending stress of the coil.

[0197] When the inner winding has rigid support, if the conductor's bending resistance is insufficient under circumferential compressive stress, the coil will deform, and the damage will be caused by forced warping. The purpose of checking the radial bending stress is to assess the winding's bending resistance and prevent forced warping.

[0198] Support bars are installed on the inner and outer sides of the coil as a supporting structure. Under the action of radial electromagnetic force, the coil disc between the support bars may bend and deform, thereby generating bending stress within the coil disc. Figure 8 As shown.

[0199] The radial bending stress mainly concerns the bending deformation of the inner coil under radial pressure. This is because the outer coil, under tension, tends to expand outwards. Since the outer support bars are generally not well fixed, the outer coil primarily generates circumferential tensile stress under radial tension, without considering bending deformation and bending stress. For the inner coil, its inner side is generally supported by support bars against the core column. If we consider the inner support bars to be well-fixed and rigid (i.e., the support bars have small compressive deformation and can be considered approximately incompressible; another extreme case is: if we consider the support bars to be absolutely soft and the inner coil unsupported, then the inner coil coil only experiences circumferential compressive stress without considering bending deformation and bending stress), then the coil between adjacent support bars can be considered as a straight beam with uniform load fixed at both ends. Using mechanical theory, the bending moment of the coil in this case can be obtained as:

[0200]

[0201] Where f r The radial force per unit length of the conductor refers to the force exerted on the conductor, and l refers to the length of the conductor between the two supports.

[0202]

[0203] Where D refers to the inner diameter of the coil, and n st The number of inner support bars of the coil, w st The width of the support bar.

[0204] Existing technologies provide calculation methods for bending stress in coil conductors:

[0205]

[0206] Where Wr refers to the radial section modulus of a single conductor:

[0207]

[0208] Where h and b refer to the axial and radial dimensions of a single conductor, respectively;

[0209] Therefore, the formula for calculating radial bending stress is:

[0210]

[0211] When calculating the radial bending stress, the software takes the radial force f per unit length of the conductor within the segment as the value. rseg Then we have:

[0212]

[0213] Where σ brseg f is the radial bending stress of the conductor within the coil segment. rseg l represents the radial force per unit length of a single conductor within a coil segment. strip The spacing between coil pads ( N strip To determine the number of supports, w strip (Width of the support bar).

[0214] Based on the Ampere force calculation results, considering the maximum amplitude force on the conductor within the coil segment, we have:

[0215] Step S500: Preprocess the performance data of the insulating oil, the performance data of the insulating paper, and the reliability of the transformer to obtain preprocessed transformer data.

[0216] For various transformer monitoring sensor data, preprocessing is required to obtain robust transformer operating condition monitoring and fault diagnosis models. The main processes include filtering, resampling, and data normalization. Filtering removes some electrical noise, and resampling ensures consistency of signal variables at different sampling rates across the time dimension, typically involving downsampling and upsampling. Data normalization primarily employs three methods: normalization, centering, and standardization.

[0217] Step S600: Use the Shenxing network of multi-scale attention convolutional GRU to learn the preprocessed data (vibration, acoustic signature, oil chromatography and other state data) to construct a transformer fault database.

[0218] To address the complex relationships between different sensor signals across varying time spans and scales, this method employs a convolutional neural network (CNN) with multiple time points corresponding to different scales for feature extraction. It then utilizes an attention-based convolutional gate recurrent unit (Att-ConvG RU) to learn the correlations between key time / feature scales. This CNN consists of five time points, each with 512 sampling points as input data (i.e., input data is Y). 512×NN represents the number of sensors. Each time step corresponds to a different convolutional neural network for feature extraction, used to extract the relationships between different sensors and the features of the time-domain signal within a single sensor. Since the dependence of the transformer's operating state on the data decays with increasing observation history, features corresponding to large-scale time steps undergo large-scale pooling to reduce feature detail and optimize subsequent feature fusion. A 2x2 Att-ConvGRU neural network is used to extract the correlation between the transformer's operating state time steps. The attention mechanism here effectively suppresses incoherent feature variables across different time spans. Finally, features from each time step are fused using 1×1 convolutions, and the transformer's operating state probability is obtained through a fully connected network.

[0219] The attention mechanism incorporates convolutional GRU units. The GRU network is used to extract the temporal dependencies of the transformer's operating states. The ConvGRU unit uses convolution to replace multiplication, thus enhancing the feature representation of the GRU. This method further emphasizes temporal dependencies by adding an attention mechanism.

[0220] The attention-based convolutional GRU unit can be described by the following formula:

[0221]

[0222]

[0223]

[0224]

[0225] X' t =A t ⊙X t (32)

[0226]

[0227] H t =λ(1-Z) t )⊙H t-1 +Z t ⊙H′ t (34)

[0228] The attention mechanism is as follows:

[0229]

[0230] Where Q is the query matrix, K is the key matrix, V is the value matrix, and d is the column vector dimension of the key matrix. By introducing an attention mechanism, the ConvGRU neural network can focus on time span and scale features, reducing the impact of insignificant features on transformer state estimation.

[0231] Based on statistical analysis of past UHV transformer equipment quality problems and in conjunction with typical cases, it can be seen that transformer failures mainly occur in raw materials and manufacturing processes, while design and testing problems also occur frequently.

[0232] In order to reduce quality problems during the production of transformer equipment, the following technical measures for supervision are proposed:

[0233] (1) Manufacturing process

[0234] 1) Strictly implement cleanliness control requirements: There is a widespread problem of inadequate implementation of process requirements across various plants. For example, air showers are only open during leadership visits, rendering environmental monitoring ineffective; some operational processes are not adequately controlled, such as strict control during main operations but not during storage and temporary processing. The supervision team should strengthen its supervision and inspection efforts in the future.

[0235] 2) Protection during product transfer: Damage to semi-finished products remains relatively common during the transfer of products between different processes, such as protecting coil leads, handling semi-finished products during transfer, and storing semi-finished products. Personnel's quality awareness should be further strengthened, work schedules should be rationally arranged, and protective measures should be enhanced.

[0236] 3) Tooling and process improvement: The transformer industry has many production and manufacturing links and a low level of automation, which is the main reason for the difficulty in quality control. We should study in all aspects the measures to improve the quality of operation and strengthen the monitoring / inspection of operation results, such as monitoring the transformer body under the tank, checking the transformer body positioning and installation, automatic core binding, and monitoring short circuits between coil winding strands.

[0237] 4) Further improve oil treatment process standards: UHV products involve multiple layers of insulation and a large amount of insulating oil. Oil treatment involves costs and efficiency, and has a crucial impact on product quality. Various manufacturers commonly employ multiple measures such as hot oil circulation, hot oil flushing, hot operation flushing, and oil treatment at increased temperatures. However, none of the manufacturers have strictly controlled the oil treatment effect according to optimal quality control standards. There have been instances where partial discharge tests failed despite low levels of partial discharge, but after further strengthening the oil treatment, the tests passed smoothly. Subsequent manufacturers should optimize their oil treatment plans to avoid situations where substandard oil treatment leads to excessive partial discharge.

[0238] 5) The drying process should be strictly followed when drying the equipment body, and the vacuum data should be based on the reading of the vacuum gauge on the drying tank body.

[0239] (2) Raw material components

[0240] Key raw materials and components required for transformer manufacturing include: silicon steel sheets, electromagnetic wire, insulating molded parts, outgoing line devices, tap changers, bushings, current transformers, gas relays, pressure relief valves, temperature controllers, coolers, bladders, and oil tanks. Currently, the quality management system for materials and components heavily relies on the production processes and inspections of suppliers, lacking effective quality control measures. The large number of material and component suppliers, with varying quality levels, and the numerous and challenging environments involved in the production, inspection, transportation, secondary processing, assembly, and final use processes are the main reasons for the prominent problems with raw material components. Going forward, control should be strengthened in the following aspects to improve quality:

[0241] 1) Strengthen extended supervision and quality spot checks: In subsequent UHV projects, manufacturers should conduct extended supervision of key raw materials / components such as insulating molded parts, outgoing line devices, high-voltage bushings, conductors, and silicon steel sheets. Supervision units should conduct extended supervision or spot checks on insulating molded parts, outgoing line devices, high-voltage bushings, conductors, etc.

[0242] 2) Environmental Control of Insulation Material Packaging, Transportation, and Storage: Foreign object discharge in insulation materials is a major factor contributing to product insulation failures. This is due to the large quantity of insulation materials and the numerous processing and manufacturing stages; differences in quality control among different suppliers and transformer manufacturers for different types of insulation materials are also a primary cause of foreign object discharge problems. In subsequent UHV projects, each manufacturer should conduct a special campaign for comprehensive quality control of insulation materials throughout the entire process, taking into account the specific process characteristics and limitations of each plant, and adopting targeted quality improvement measures.

[0243] 3) Coil winding inter-strand insulation monitoring: Electromagnetic wire problems during manufacturing mainly manifest as inter-strand short circuits, primarily caused by issues with environmental control and equipment processing quality at the conductor manufacturer. On the one hand, transformer manufacturers should strengthen extended supervision to urge improvements in manufacturing quality. On the other hand, online monitoring devices should be installed during the coil winding process to provide early warnings of inter-strand short circuits. Simultaneously, relevant process requirements and countermeasures should be strictly implemented, and inter-strand short circuit tests should be conducted before conductor welding and after hot pressing.

[0244] 4) Moisture prevention measures for silicon steel sheets: Silicon steel sheets are prone to moisture damage in factory areas with high air humidity. At the same time, manufacturers in northern regions have also encountered similar problems in the production process due to a lack of attention to the problem of moisture damage to silicon steel sheets and a lack of protective measures. Management and protection should be further strengthened.

[0245] 5) Mixed use of insulation materials: Although the current situation of mixed or incorrect use of insulation materials has decreased, it still exists. Supervisors should combine full-process witnessing with key attention, assign responsibilities to individuals, increase witnessing methods and content, and establish quality tracking files for key raw materials and components.

[0246] 6) Bushings: Chromatographic data must be provided for each bushing leaving the factory, and the chromatographic test accuracy should be higher than 0.01 ppm. In principle, tandem testing is not allowed. If tandem testing is necessary, it must be reported to the UHV Department for approval in advance. Bushings that have not been tested after tandem testing must be tested individually, and extended supervision of the bushing manufacturing process must be carried out.

[0247] 7) Standardize the quality control of outsourced parts, key components, and especially domestically produced components, strengthen incoming inspection, strictly review factory test reports and other relevant materials, and strengthen the storage inspection of domestically produced components.

[0248] (3) Factory testing phase

[0249] 1) All testing equipment should be inspected regularly, and the frequency of equipment maintenance should be increased to prevent problems with testing equipment from affecting transformer product testing. In addition to normal inspection, testing instruments and meters should be checked before ultra-high voltage product testing, especially the test transformers should be tested regularly with oil chromatography. In addition, highly reliable compensation reactors should be used, and compensation capacitors should have redundant circuits.

[0250] 2) Standardization of test preparation. The manufacturer should standardize and solidify the relative positions of the test specimens and test equipment, the wire gauge and connection method of the high-voltage leads, the selection and arrangement of the grounding circuit, and the selection of the shielding method on the top of the oil tank during the UHV transformer test, so as to eliminate potential risks that may affect the first-pass yield of the test.

[0251] 3) Testing phase: The partial discharge test is strictly recorded throughout the entire process, without overlooking any suspicious discharges, and strictly adheres to the requirement of no partial discharge. Any interference during the test process must be jointly confirmed by the supervision team and the manufacturer before it can continue.

[0252] (4) Design phase

[0253] With the successful commissioning and safe and stable operation of various ultra-high voltage projects, the design of ultra-high voltage transformers has been tested in practice. The overall design structure is reasonable, safe and reliable, and some design problems found in the early stages of the project have been gradually and thoroughly resolved. However, with a spirit of continuous improvement, further optimization and improvement are still needed in the following aspects:

[0254] 1) Low potential and high field strength problem: Strengthen the insulation verification of clamping parts, bolts, support insulation and other materials in low potential areas, and strictly limit the use area and scope of laminated wood, electrical wood and other materials.

[0255] 2) Design improvement and refinement: By further carrying out lean design verification work, the design is refined and improved, and the structural dimensions are further refined, so as to fundamentally avoid operational errors such as loose horn pads and overlapping shielding caps.

[0256] 3) Incorporate the review and verification of relevant documents before the design is frozen into the standardized review system for supervision, and urge the manufacturer to actively communicate with relevant units and expert groups to ensure the accuracy and effectiveness of design drawings and related documents.

[0257] 4) Each manufacturer should maintain its existing strengths in the structure. When making major structural changes such as changing from three columns to two columns or adding a lung lobe magnetic shunt, special attention should be paid to design verification and expert review.

[0258] (5) Storage and transportation stage

[0259] 1) A final joint inspection should be conducted before product shipment, focusing on the standardization of the 3D impact recorder's installation position and verifying the instrument's accuracy. For inflatable transport, the dew point value should be checked for normality, the condition of accessories disassembly and packaging, and the post-test process assurance measures.

[0260] 2) For component packaging and transportation, reinforcement measures should be taken for key components such as outgoing line devices, bushings, and oil tanks.

[0261] 3) Regarding the management of factory storage and transit storage, the manufacturer and the supervision team should determine the possible installation time in advance, reverse the schedule, and ensure that the equipment is completed on time and does not need to be stored in the factory for a long time.

[0262] 4) When using raw materials and components, warning signs should be placed in the temporary storage area of ​​the workshop, and effective measures should be taken to avoid collisions. For example, corner rings should be placed on packaging boxes or material racks, end rings, cardboard, etc. should be stacked neatly, and components should be placed separately with protection around them.

[0263] 5) The product must be securely fastened during transportation, and a safety height warning device should be set up to prevent scratches and collisions during transportation.

[0264] Furthermore, to determine the performance and reliability status of each component of the transformer, a method is proposed. This method can analyze and evaluate the changing characteristics of key performance indicators of equipment and components by conducting standard-specified tests, tests under special operating conditions (DC bias, overvoltage, harmonics, etc.), and state-based analysis-based tests. It can also statistically analyze the patterns of performance degradation or deterioration, and provide performance status evaluation indicators for equipment and components to support performance improvement and reliability enhancement. Specifically:

[0265] A100. Performance evaluation of oil filling equipment and its components.

[0266] Conduct standard-specified tests on raw materials / components used in power transformers and reactors, including insulating molded parts, outgoing line devices, high-voltage bushings, tap changers, electromagnetic wires, silicon steel sheets, pressure relief valves, thermometers, voltage / current transformers, and oil tanks. Compare the factory test results, analyze their quality deterioration indicators, and evaluate their performance stability.

[0267] A200. Performance Equivalent Simulation.

[0268] Based on equivalent simulation of DC voltage and thermal equivalent current, long-term reliability analysis of bushing under the combined action of high voltage and high current is carried out, and the insulation performance, thermal performance and mechanical performance of bushing are comprehensively evaluated.

[0269] A300. Reliability analysis under special operating conditions.

[0270] Conduct performance reliability analysis of tap changers under special operating conditions, including DC bias, harmonics, overvoltage, and low temperature switching test analysis, analyze their performance degradation characteristics under extreme operating conditions, and evaluate their operational reliability.

[0271] A400. Comprehensive Assessment.

[0272] Conduct comprehensive condition assessment performance testing and analysis, statistically analyze the changes in the mechanical / electrical / physicochemical properties of equipment and components, and evaluate their performance stability over a long time scale.

[0273] Another embodiment of this application provides a transformer comprehensive analysis device, such as... Figure 9 As shown, the device 20 includes: an insulating oil analysis module 201, an insulating oil paper analysis module 202, a transformer data collection module 203, a transformer data analysis module 204, a data comprehensive processing module 205, and a neural network training module 206.

[0274] The insulating oil performance analysis module 201 can analyze the performance of insulating oil and obtain the performance data of insulating oil.

[0275] The insulating oil paper performance analysis module 202 can analyze the performance of the insulating oil paper and obtain the performance data of the insulating oil paper.

[0276] The new transformer reliability analysis module 203 can collect and process transformer data to obtain processed transformer data.

[0277] The short-circuit withstand capability parameter analysis module 204 can calculate and determine the transformer's short-circuit current withstand performance data based on the short-circuit withstand parameters.

[0278] The data processing module 205 can preprocess the performance data of the insulating oil, the performance data of the insulating paper, the transformer reliability, and the short-circuit current withstand performance data to obtain preprocessed transformer data.

[0279] The neural network training module 206 is able to learn the preprocessed data (vibration, acoustic signature, oil chromatography and other state data) using the Shenxing network of multi-scale attention convolutional GRU to construct a transformer fault database.

[0280] This application embodiment analyzes the transformer's comprehensive operating status by analyzing data from multiple dimensions, including the state of the transformer oil, transformer oil paper, and the transformer itself. It acquires transformer operating data based on sensors and inputs the data corresponding to typical operating states into a neural network model based on a multi-scale attention convolutional GRU. This establishes a multi-scale comprehensive transformer analysis device.

[0281] Another embodiment of this application provides a terminal, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described transformer comprehensive analysis method.

[0282] Specifically, the processor can be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0283] Specifically, the processor connects to the memory via a bus, which may include a path for transmitting information. The bus can be a PCI bus or an EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc.

[0284] The memory may be ROM or other types of static storage devices that can store static information and instructions, RAM or other types of dynamic storage devices that can store information and instructions, or EEPROM, CD-ROM or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto.

[0285] Optionally, the memory stores the code of the computer program that executes the scheme of this application, and the execution is controlled by the processor. The processor executes the application code stored in the memory to implement... Figure 5 The operation of the transformer comprehensive analysis device provided in the illustrated embodiment.

[0286] Another embodiment of this application provides a computer-readable storage medium storing computer-executable instructions, which are used to perform the above-described... Figure 1 The transformer comprehensive analysis method shown is illustrated.

[0287] This application analyzes data on transformer insulating oil, insulating paper, and the transformer itself, collecting a complete set of correspondences between transformer operating states and transformer data. Then, it trains the transformer data corresponding to different operating states and different aging degrees using a neural network to obtain the operating states and aging degrees corresponding to different transformer data.

[0288] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0289] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0290] The above is a detailed description of the preferred embodiments of this application. However, this application is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A comprehensive analysis method for transformers, characterized in that, Includes the following steps: Several transformer parameters are acquired using sensors; the transformer parameters include at least one of the following: insulating oil performance parameters, insulating paper performance parameters, transformer reliability parameters, and short-circuit withstand capability parameters. Based on the transformer parameters, a multi-dimensional transformer operating state corresponding to each transformer parameter is determined; the process of determining the transformer operating state based on the insulating oil paper performance parameters specifically includes: The aging status assessment parameters are obtained by the aging status assessment scoring method based on matter-element theory; The basic score of the transformer is obtained through factory testing, acceptance testing, and safe operation records. By using the group component status, defect management and defect content in the transformer family record, the family defect correction parameters of the transformer can be obtained. By analyzing transformer outlet short circuit, overload, and overexcitation, correction parameters for adverse operating conditions are obtained. The evaluation results of the insulating oil paper are determined based on the aging condition assessment parameters, the transformer's basic score, the family defect correction parameters, and the adverse operating condition correction parameters. The formula used in the step of determining the evaluation result of the insulating oil paper based on the aging condition assessment parameters, the transformer's basic score, the family defect correction parameters, and the adverse operating condition correction parameters is: Where B is the transformer's basic score, F is the transformer's family defect correction parameter, E is the adverse operating condition correction parameter, and T is the aging condition assessment parameter. and These are the weights of the transformer's basic score B and the aging condition assessment parameter T, respectively. A neural network using multi-scale attention convolutional GRU is used to determine the current true state of the transformer based on the multi-dimensional transformer operating state.

2. The method according to claim 1, characterized in that, =0.5； =0.5。 3. The method according to claim 1, characterized in that, The processed transformer data is analyzed to obtain transformer reliability parameters, including: Collect and process transformer data to obtain processed transformer data; The processed transformer data is analyzed to obtain transformer reliability parameters.

4. The method according to claim 3, characterized in that, The step of collecting and processing transformer data to obtain processed transformer data specifically includes: Obtain technical compliance assessment data from transformer manufacturers; Obtain a summary of historical transformer issues; The technical compliance assessment data and historical problem summary data are processed.

5. The method according to claim 4, characterized in that, The steps for processing the technology compliance assessment data and historical problem summary data specifically include: The data is then classified and standardized. Perform a rationality analysis on the data and then repair it; Encrypt the processed data.

6. The method according to claim 3, characterized in that, The step of analyzing the processed transformer data to obtain transformer reliability parameters includes: Based on data from the power grid resource business platform, information on different operating states of transformers is collected. Based on the technical compliance assessment, a comprehensive reliability score is determined for each piece of equipment; Based on the equipment's entire life cycle data, the importance classification of the equipment and its components within the station is determined. The reliability index of a group of equipment of different types, manufacturers and models is calculated by using equipment defect rate, failure rate, maintenance cycle and remaining life. The processed transformer data includes information on technical compliance assessment, equipment lifecycle data, equipment defect rate, failure rate, maintenance cycle, and remaining lifespan.

7. The method according to claim 2, characterized in that, The process of determining the transformer's operating state based on the short-circuit withstand capability parameters specifically includes: Based on the short-circuit withstand capability parameters, determine the magnetic flux density; Based on the magnetic flux density, calculate the short-circuit impedance and the magnitude of the Ampere force; Calculate the average annular tensile and compressive stress of the coil conductor, the maximum axial bending stress of the coil conductor, and the maximum axial bending stress of the coil conductor. Calculate the maximum short-circuit current and verify it based on the current stress conditions.

8. A transformer comprehensive analysis device, characterized in that, include: The insulating oil performance analysis module can analyze the performance of insulating oil and obtain its performance data. The insulating oil paper performance analysis module can analyze the performance of insulating oil paper and obtain its performance data; specifically, it includes: The aging status assessment parameters are obtained by the aging status assessment scoring method based on matter-element theory; The basic score of the transformer is obtained through factory testing, acceptance testing, and safe operation records. By using the group component status, defect management and defect content in the transformer family record, the family defect correction parameters of the transformer can be obtained. By analyzing transformer outlet short circuit, overload, and overexcitation, correction parameters for adverse operating conditions are obtained. The evaluation results of the insulating oil paper are determined based on the aging condition assessment parameters, the transformer's basic score, the family defect correction parameters, and the adverse operating condition correction parameters. The formula used in the step of determining the evaluation result of the insulating oil paper based on the aging condition assessment parameters, the transformer's basic score, the family defect correction parameters, and the adverse operating condition correction parameters is: Where B is the transformer's basic score, F is the transformer's family defect correction parameter, E is the adverse operating condition correction parameter, and T is the aging condition assessment parameter. and These are the weights of the transformer's basic score B and the aging condition assessment parameter T, respectively. The transformer reliability analysis module can analyze processed transformer data to obtain transformer reliability data; The short-circuit withstand capability parameter analysis module can calculate and determine the transformer's short-circuit current withstand performance data based on the short-circuit withstand capability parameters. The data processing module can preprocess the performance data of the insulating oil, the performance data of the insulating paper, the transformer reliability data, and the short-circuit current performance data to obtain preprocessed transformer data. The neural network training module is able to learn the preprocessed data using a multi-scale attention convolutional GRU neural network to construct a transformer fault database.

9. A terminal, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing computer-executable instructions for performing the method of any one of claims 1 to 7.

Images