Transformer operation state evaluation method and system based on multi-source data

By collecting and analyzing multi-source data, combined with insulation damage index calculation and fault coding, the problem of single evaluation indicators and lack of quantitative evaluation in transformer condition assessment is solved. It realizes quantitative assessment of transformer insulation condition and fault root cause location, and is applicable to transformer condition assessment under frequent switching conditions.

CN123116072APending Publication Date: 2026-11-17JIANGYIN XINGCHENG SPECIAL STEEL WORKS CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611302685.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing transformer condition assessment methods rely on single assessment indicators and lack quantitative assessment tools for the cumulative effects of operational shocks. The assessment and diagnosis processes are isolated, making it difficult to effectively guide the location of root causes of faults.

Method used

A transformer operation status assessment system based on multi-source data is adopted, including multi-source data acquisition, multi-source feature extraction, multi-source fusion assessment and early warning decision modules. By collecting and analyzing dissolved gas components in oil, harmonic current and voltage, vacuum circuit breaker operation records and power supply system data in real time, the system calculates cumulative insulation damage index, fault code and resonant amplification, and realizes multi-dimensional health score weighted fusion.

Benefits of technology

It enables quantitative assessment of transformer insulation status, accurately locates the root cause of faults, is applicable to transformer status assessment under frequent switching conditions, and provides clear maintenance recommendations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116072A_ABST
    Figure CN123116072A_ABST
Patent Text Reader

Abstract

The application discloses a transformer operation state evaluation method and system based on multi-source data and belongs to the technical field of fault diagnosis. The method comprises a multi-source data acquisition module, a multi-source feature extraction module, a multi-source fusion evaluation module and a pre-warning decision module. The multi-source feature extraction module is used for obtaining a current intercept value in the vacuum circuit breaker operation record, equivalent excitation inductance of the transformer, equivalent ground capacitance and lightning impulse withstand voltage, and taking the product of the current intercept value and the square root of the equivalent excitation inductance divided by the equivalent ground capacitance as the amplitude of the current intercept overvoltage, taking the 10th power of the ratio of the amplitude of the current intercept overvoltage to the lightning impulse withstand voltage as the insulation damage equivalent of single operation, and adding up the insulation damage equivalents of each operation in a statistical period to obtain a cumulative insulation damage index. The application solves the technical problem that the prior art cannot quantitatively evaluate the cumulative damage degree of the transformer insulation caused by frequent operation of the vacuum circuit breaker.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, specifically to a method and system for evaluating the operating status of transformers based on multi-source data. Background Technology

[0002] As a core hub in the power system, the operating status of transformers directly affects the safe and stable operation of the power grid. Traditional transformer condition assessment mainly relies on dissolved gas analysis (DGA) in transformer oil, which determines the type of internal fault by detecting the content of characteristic gases such as hydrogen, methane, ethane, ethylene, and acetylene in the transformer oil. However, traditional DGA methods are often limited to threshold-based judgment frameworks, resulting in overly absolute classification boundaries and a tendency to misclassify or miss cases with missing codes. Existing assessment methods based on frequency domain dielectric spectroscopy do not consider the non-uniform aging effect of oil-paper insulation, and their assessment results are average values ​​of the insulation aging state, while the insulation at transformer hotspots is the weakest and most prone to equipment failure. Traditional single-physics field analysis methods are insufficient to comprehensively reflect the true state of transformers under multi-stress coupling.

[0003] Existing transformer condition assessment methods generally suffer from the following shortcomings: The assessment indicators are singular; most methods rely solely on oil gas content or single electrical quantities for judgment, failing to incorporate multi-source data such as circuit breaker operation frequency, harmonic distortion rate, and system impedance characteristics into a unified assessment framework. Existing methods lack quantitative assessment tools for the cumulative effects of circuit breaker operation impacts, failing to transform the well-known qualitative understanding of "operational overvoltage damaging insulation" into calculable quantitative indicators, leading to the long-term neglect of cumulative insulation damage under frequent switching conditions. Condition assessment and fault diagnosis are typically performed as two independent steps, with assessment results failing to effectively guide fault root cause localization.

[0004] Therefore, the present invention aims to provide a method and system for evaluating the operating status of transformers based on multi-source data, so as to solve the defects existing in the prior art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for evaluating the operating status of transformers based on multi-source data, so as to solve the technical problems in the prior art, such as the single evaluation index, lack of quantitative evaluation means for the cumulative effect of operational shocks, and the isolation of evaluation and diagnosis links.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution, specifically including: The transformer operation status assessment system based on multi-source data includes: a multi-source data acquisition module, a multi-source feature extraction module, a multi-source fusion assessment module, and an early warning decision module.

[0007] The multi-source data acquisition module is used to collect multi-dimensional operating data of the transformer in real time, including dissolved gas composition data in the oil, namely the content of hydrogen, methane, ethane, ethylene, acetylene, total hydrocarbons, carbon monoxide, and carbon dioxide, with the content unit being ppm. Among the above gas components, acetylene is a key characteristic gas for characterizing discharge faults, and its content changes have important indicative significance for fault diagnosis.

[0008] High-voltage side harmonic current and voltage data, namely the current amplitude and voltage amplitude of integer harmonics from the 2nd to the 25th order and interharmonics. Refining furnaces generate a large amount of harmonic current during the smelting process, with a frequency spectrum covering the 2nd to 25th orders or even higher. Each harmonic generates a corresponding harmonic voltage drop on the transformer windings. The instantaneous peak value of the harmonic current can reach tens of times the maximum value with a 95% probability, posing a significant threat to the transformer insulation.

[0009] Vacuum circuit breaker operation records include the number of opening and closing operations, operation times, and load status. The refining furnace transformer is operated an average of about 7 times per hour, far exceeding the switching frequency of several times per month for ordinary power transformers. Over a long period, the cumulative number of operations can reach hundreds of thousands to millions of times.

[0010] The data includes short-circuit capacity, filter branch switching status, and bus voltage on the power supply system side. This data reflects the grid-side operating mode and system impedance characteristics, directly affecting the parallel resonance characteristics between the filter branch and the system.

[0011] The multi-source feature extraction module includes four sub-processing units: an operational impact accumulation unit, an oil gas analysis unit, a harmonic feature extraction unit, and a resonant amplification and tracking unit.

[0012] The operational impact accumulation unit is used to acquire the current-cutting value, the equivalent magnetizing inductance, equivalent ground capacitance, and lightning impulse withstand voltage of the transformer from the operation record of the vacuum circuit breaker, and calculates the cumulative insulation damage index according to the following method: the square root of the product of the current-cutting value and the equivalent magnetizing inductance divided by the equivalent ground capacitance is taken as the current-cutting overvoltage amplitude; the 10th power of the ratio of the current-cutting overvoltage amplitude to the lightning impulse withstand voltage is taken as the insulation damage equivalent of a single operation; and the insulation damage equivalents of each operation within the statistical period are summed to obtain the cumulative insulation damage index. The 10th power is derived from the insulation life index in the inverse power model of the insulating material.

[0013] Specifically, the current-cutting overvoltage amplitude U tr Calculate using the following formula:

[0014] Among them, I cutThe current-cutting value is expressed in amperes; the current-cutting value of a vacuum circuit breaker depends on the characteristics of the contact material. L is the equivalent magnetizing inductance of the transformer, expressed in Henrys. C is the equivalent capacitance to ground of the transformer, expressed in farads.

[0015] Insulation damage equivalent D in a single operation op Calculate using the following formula:

[0016] Among them, U BIL This is the transformer's lightning impulse withstand voltage, measured in volts. This parameter represents the transformer's inherent withstand voltage level at the time of manufacture and reflects the transformer's insulation's ability to resist steep-front impulses. The index 10 is the insulation life index, derived from the inverse power model of solid insulation materials.

[0017] Cumulative Insulation Damage Index D cum Calculate using the following formula:

[0018] Where, N op D represents the total number of operations within the statistical period. op,j This represents the insulation damage equivalent of the j-th operation. The operation impact accumulation unit also triggers an early warning when the average number of operations per hour within the statistical period exceeds a preset threshold.

[0019] The oil gas analysis unit is used to perform fault coding on dissolved gas data in the oil using the three-ratio method and to calculate the absolute gas production rate of acetylene. An early warning is triggered when the absolute gas production rate of acetylene exceeds a preset threshold. Absolute gas production rate γ i Calculate using the following formula:

[0020] Among them, C i (t1) and C i (t2) represents the content of the i-th characteristic gas at times t1 and t2, respectively. In one specific implementation, the preset threshold for the absolute acetylene production rate is 0.2 ppm / day.

[0021] The three-ratio coding rules are as follows: If the ratio of methane to hydrogen is less than 0.1, the code is 0; if the ratio is between 0.1 and 1, the code is 1; if the ratio is between 1 and 3, the code is 1; if the ratio is greater than or equal to 3, the code is 2. Similarly, if the ratio of acetylene to ethylene is less than 0.1, the code is 0; if the ratio is between 0.1 and 1, the code is 0; if the ratio is between 1 and 3, the code is 1; if the ratio is greater than or equal to 3, the code is 2. Likewise, if the ratio of ethylene to ethane is less than 0.1, the code is 0; if the ratio is between 0.1 and 1, the code is 0; if the ratio is between 1 and 3, the code is 1; if the ratio is greater than or equal to 3, the code is 2. The three code combinations form a three-digit fault code. Code 102 corresponds to arc discharge, and code 122 corresponds to arc discharge combined with overheating.

[0022] The harmonic feature extraction unit is used to calculate the 95% probability superimposed peak value and the superimposed peak value of the maximum values ​​of each harmonic voltage: , among which, U h,95% U is the maximum value with a 95% probability of being the h-th harmonic voltage. h,max This represents the maximum value of the h-th harmonic voltage. This unit also calculates the peak-to-peak ratio of the harmonic currents for each harmonic: , among which, I h,max I is the maximum value of the h-th harmonic current. h,95% This represents the 95% probability maximum value of the h-th harmonic current. The 95% probability maximum value means that this value is not exceeded 95% of the sampling times within the statistical period, reflecting the typical level of the harmonic current; the maximum value reflects the instantaneous impact level. The ratio R between the two is... h The larger the value, the stronger the impact of that harmonic.

[0023] The resonant amplification tracking unit is used to obtain the parallel resonant frequency and the corresponding amplification factor by solving the parallel resonant equation. For a system containing n filter branches, the parallel resonant frequency f res This is obtained by solving the following equation:

[0024] Among them, X C,k X is the capacitive reactance of the k-th filter branch. L,k Let X be the inductive reactance of the k-th filter branch. S (f res ) is the system at the resonant frequency f res The equivalent reactance is given by the equation. This equation describes the physical condition under parallel resonance conditions where the parallel combination of the system reactance and the filter branch reactance tends to infinity. The amplification factor at the resonance point is calculated by the following formula:

[0025] Among them, Z eq(f res Z is the equivalent impedance of the system at the resonant frequency. S (f res The amplification factor is the source impedance of the system at the resonant frequency. This unit triggers an early warning when the amplification factor exceeds a preset threshold. In one specific implementation, the preset threshold for the amplification factor is 6.

[0026] The multi-source fusion evaluation module is used to weight and fuse the output of the multi-source feature extraction module to obtain a comprehensive health index. The comprehensive health index is a weighted sum of normalized health scores from four dimensions: oil gas analysis, harmonic characteristics, operational impact accumulation, and resonant amplification tracking. The sum of the weight coefficients of each normalized health score is 1. A higher comprehensive health index value indicates a healthier transformer operating condition. Root cause identification is achieved by comparing the normalized health scores of the four dimensions; the fault type corresponding to the lowest score is determined as the root cause.

[0027] The early warning decision module is used to issue graded early warnings and output maintenance suggestions based on the comprehensive health index. The graded early warning levels include at least normal state, alert state, abnormal state, and danger state, with each level corresponding to different preset value ranges of the comprehensive health index. As a specific implementation, a normal state is output when the comprehensive health index is greater than or equal to 80; an alert state is output when the comprehensive health index is greater than or equal to 60 and less than 80; an abnormal state is output when the comprehensive health index is greater than or equal to 40 and less than 60; and a danger state is output when the comprehensive health index is less than 40.

[0028] The transformer operating status assessment method based on multi-source data includes the following steps: S1. Multi-source data acquisition steps: Real-time acquisition of dissolved gas composition data in transformer oil, high-voltage side harmonic current and harmonic voltage data, vacuum circuit breaker operation records, and power supply system data; S2. Multi-source feature extraction step: Obtain the current-cutting value, equivalent magnetizing inductance, equivalent ground capacitance, and lightning impulse withstand voltage from the vacuum circuit breaker operation record. Calculate the cumulative insulation damage index as follows: The square root of the product of the current-cutting value and the equivalent magnetizing inductance divided by the equivalent ground capacitance is used as the current-cutting overvoltage amplitude. The 10th power of the ratio of the current-cutting overvoltage amplitude to the lightning impulse withstand voltage is used as the insulation damage equivalent for a single operation. The insulation damage equivalents for each operation within the statistical period are summed to obtain the cumulative insulation damage index. Simultaneously, the dissolved gas data in the oil is used for fault coding using the three-ratio method, and the absolute acetylene production rate is calculated. The peak value of each harmonic voltage superposition and the peak ratio of the harmonic current are calculated for the harmonic current and harmonic voltage data. The parallel resonant frequency and corresponding amplification factor are calculated for the power supply system data. S3, Multi-source fusion assessment step: The four output results of step S2 are weighted and fused to obtain a comprehensive health index and output the root cause localization results; S4. Early warning decision-making step: Based on the comprehensive health index output in step S3, classify and issue early warnings and output maintenance suggestions.

[0029] Compared with the prior art, the beneficial effects achieved by the present invention are: This invention incorporates the cumulative effect of vacuum circuit breaker operation impact into the transformer insulation condition assessment system. It establishes for the first time a quantitative mapping model between circuit breaker current-cutting impact and cumulative transformer insulation damage, filling a technical gap in assessing transformer insulation life from the perspective of operation impact. This model is particularly suitable for transformer condition assessment under frequent switching conditions.

[0030] A nonlinear cumulative damage quantitative calculation method based on the physical mechanism of current throttling is proposed. Based on a physical model of the current throttling phenomenon in vacuum circuit breakers, the amplitude of the current-throttling overvoltage is calculated using the equivalent magnetizing inductance and equivalent ground capacitance of the transformer. Furthermore, a lifetime exponent is introduced based on an inverse power model of the insulation material to perform nonlinear damage equivalent conversion for a single impact. The cumulative insulation damage index is obtained through accumulation, achieving a leap from qualitative understanding to quantitative assessment. This method does not require a large amount of historical fault data for training and can be directly applied to transformers of different voltage levels and capacities.

[0031] It achieves accurate location of the root cause of the fault. By weighted and fused normalized health scores from four dimensions—oil gas analysis, harmonic characteristics, operational shock accumulation, and resonant amplification tracking—and comparing the scores of each dimension, the root cause is determined. The lowest score item is the dominant factor causing the anomaly, providing maintenance personnel with a clear direction for troubleshooting. Attached Figure Description

[0032] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating the overall architecture of the transformer operation status assessment system based on multi-source data of this invention. Figure 2 This is a flowchart of the steps in the transformer operation status assessment method based on multi-source data of the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Please see Figures 1-2 The present invention provides the following technical solution: like Figure 1 As shown, a multi-source data acquisition module is deployed at the steel plant refining furnace transformer site, including an online oil gas monitoring unit, a power quality monitoring unit, a circuit breaker operation record unit, and a power supply system data interface unit.

[0035] The online gas monitoring unit in the oil continuously collects the content of characteristic gases such as dissolved hydrogen, methane, ethane, ethylene, acetylene, total hydrocarbons, carbon monoxide, and carbon dioxide in the transformer oil through an oil circulation system installed on the transformer body, at a sampling frequency of once a day. The core components of the online gas monitoring unit are a gas chromatography column and a thermal conductivity detector. After the mixed gases in the oil sample are separated by the chromatographic column, they enter the detector sequentially, and qualitative and quantitative analysis is performed based on the retention time and response peak area of ​​each component.

[0036] The power quality monitoring unit is installed at the 35kV busbar on the high-voltage side of the transformer. It continuously collects three-phase voltage and current waveform data at a sampling rate of 256 points per cycle, with an equivalent sampling frequency of 12800Hz, which meets the Nyquist sampling requirements for harmonic analysis from the 2nd to the 25th order. The amplitude and phase information of each harmonic are extracted from the sampled waveform using Fast Fourier Transform, with a frequency resolution of 5Hz, effectively distinguishing between integer harmonics and interharmonics.

[0037] The circuit breaker operation recording unit records the timestamp and operation type of each opening and closing operation by connecting to the auxiliary contacts of the vacuum circuit breaker. The auxiliary contacts operate synchronously with the main contacts; when the main contacts close or open, the auxiliary contacts correspondingly conduct or open. The operation time can be accurately recorded by detecting the level change in the auxiliary contact circuit. This unit also calculates the average number of operations per hour and the total number of cumulative operations based on the operation records.

[0038] The power supply system data interface unit obtains data such as the current 220kV side short-circuit capacity, 35kV bus voltage, and the switching status of each filter branch from the substation monitoring system via a communication network. The substation monitoring system is based on the IEC 61850 communication protocol, with a data refresh cycle of 1 second. The interface unit interacts with the monitoring system through the manufacturing message specification protocol.

[0039] A multi-source feature extraction module, a multi-source fusion evaluation module, and an early warning decision-making module are deployed on a cloud server. These modules communicate via a data bus, using a publish-subscribe model to facilitate data flow.

[0040] like Figure 2 As shown, the implementation process is broken down as follows: Step S1: Multi-source data acquisition. Taking the refining furnace transformer in a steel plant as an example, the transformer has a capacity of 32MVA, a rated voltage of 35kV / 0.415kV, a connection group of Yd11, and a forced oil circulation air cooling method.

[0041] Oil gas monitoring data showed that the acetylene content continuously increased from an initial value of 0.11 ppm, reaching 4.72 ppm, 13.44 ppm, and finally 24.67 ppm. Acetylene is a characteristic gas produced by the cracking of transformer oil-paper insulation under the action of electric arc discharge. Its generation requires the arc temperature to reach thousands of degrees Celsius, causing the carbon-carbon bonds in the oil molecules to break and recombine. Acetylene should not be detected or should be present in extremely low amounts in normally operating transformer oil; its continuous increase indicates the presence of developing arc discharge defects inside the transformer.

[0042] The harmonic current data collected by the power quality monitoring unit are as follows: the 95% probability maximum value of the 2nd harmonic current is 6.7A, and the maximum value is 239.6A; the 95% probability maximum value of the 3rd harmonic current is 18.0A, and the maximum value is 100.9A; the 95% probability maximum value of the 4th harmonic current is 3.2A, and the maximum value is 54.5A. The ratios of the maximum value to the 95% probability maximum value are 35.8, 5.6, and 17.0, respectively. The peak values ​​of the 2nd and 4th harmonics are particularly prominent, indicating the presence of significant impulsive harmonics during the smelting process. The maximum value of the fundamental line voltage of the 35kV bus is approximately 37.24kV, which is about 6.4% higher than the rated voltage of 35kV.

[0043] Circuit breaker operation records show that the transformer is operated an average of approximately 7 times per hour. The operation of a refining furnace transformer differs fundamentally from that of a conventional power transformer: conventional power transformers operate continuously for extended periods, with monthly operations typically limited to a few times; however, a refining furnace transformer requires frequent switching in accordance with the refining furnace's smelting cycle. Each smelting cycle lasts approximately 15 minutes, with one closing and one opening operation completed within each cycle, translating to approximately 7 operations per hour. Since its commissioning, this transformer has accumulated approximately 600,000 operations.

[0044] Power supply system data shows that the 220kV side is currently operating at minimum capacity. After the main transformer step-down, the equivalent short-circuit capacity of the 35kV bus side is 760MVA. The complete set of equipment includes a second-order C-type filter branch, a third-order single-tuned filter branch, a fourth-order single-tuned filter branch, and a fifth-order single-tuned filter branch. All four filter branches are in operation.

[0045] Step S2: Multi-source feature extraction. Gas analysis in oil: The absolute acetylene production rate was calculated using the following formula. The acetylene content increased from 4.72 ppm to 25.15 ppm over a period of 92 days.

[0046] The preset threshold for the absolute acetylene production rate is 0.2 ppm / day. The measured value of 0.222 ppm / day has exceeded this threshold, triggering a warning signal.

[0047] A three-ratio method is used for fault coding. Taking a certain sampling data as an example: methane content is 25.91 ppm, hydrogen content is 36.21 ppm, the ratio of the two is 25.91 / 36.21≈0.716, which is between 0.1 and 1, so according to the coding rules, the first code is 1; acetylene content is 13.44 ppm, ethylene content is 60.63 ppm, the ratio of the two is 13.44 / 60.63≈0.222, which is between 0.1 and 1, so according to the coding rules, the second code is 0; ethylene content is 60.63 ppm, ethane content is 8.84 ppm, the ratio of the two is 60.63 / 8.84≈6.86, which is greater than 3, so according to the coding rules, the third code is 2. The combination of the three codes results in fault code 102, corresponding to an arc discharge fault.

[0048] Harmonic feature extraction: The fundamental impedance of this transformer is approximately 6.87Ω. The harmonic voltages are extracted according to U... h =I h ∙hZ1 is calculated, where h is the harmonic order.

[0049] Second harmonic: U 2,95% =6.7×2×6.87≈92V, U 2,max =239.6×2×6.87≈3292V.

[0050] 3rd harmonic: U 3,95% =18.0×3×6.87≈371V, U 3,max =100.9×3×6.87≈2079V.

[0051] The voltages of other harmonics are calculated using the same method. The 95% probability maximum values ​​of the voltages from the 2nd to the 25th harmonics are superimposed to obtain U. harm,95% ≈2083V; By summing the maximum values, we get U. harm,max ≈18216V, which is about 76% of the transformer's rated phase voltage of 24000V.

[0052] Peak ratios of harmonic currents: R2 = 239.6 / 6.7 ≈ 35.8, R3 = 100.9 / 18.0 ≈ 5.6, R4 = 54.5 / 3.2 ≈ 17.0. The peak ratios of the 2nd and 4th harmonics far exceed the threshold of 5, indicating that the transformer has been subjected to significant 2nd and 4th harmonic impacts.

[0053] Accumulated operational shocks. The equivalent magnetizing inductance L of the transformer is calculated using the following formula: Substituting the parameters U=35000V, f=50Hz, S=32×10 6Given VA and Z% = 0.0696, we get L ≈ 0.0848H. The equivalent capacitance to ground is C = 2 × 10⁻⁶. -9 F, interception value I cut Take 5A as the single-current-cutting overvoltage amplitude:

[0054] Lightning impulse withstand voltage U BIL =200000V, Single-instance damage equivalent: D op =(32557 / 200000) 10 ≈1.31×10 -8 A total of 600,000 operations were performed. If all were calculated based on a cutoff value of 5A, D cum =600000×1.31×10 -8 ≈0.00786.

[0055] In actual operation, the current-cutting value of vacuum circuit breakers is not constant and exhibits a certain degree of statistical dispersion. Considering that approximately 10% of operations have a current-cutting value of 10A, the damage equivalent for these operations needs to be recalculated: when I... cut When =10A, U tr =10×6511.5=65115V, D op,10A =(65115 / 200000) 10 ≈1.34×10 -5 10% of the operations total 60,000, and the cumulative damage for this portion is 60,000 × 1.34 × 10⁻⁶. -5 ≈0.804. 90% of the cumulative damage from the operation is 540,000 × 1.31 × 10⁻⁶. -8 ≈0.00707. Total cumulative damage index D cum ≈0.804+0.00707=0.811.

[0056] The preset threshold for the cumulative insulation damage index is 0.1, while the measured cumulative damage index of 0.811 exceeds this threshold. The operational impact accumulation unit triggers an alarm when the average number of operations per hour exceeds a preset threshold. In this case, the preset threshold is 5 times / hour, while the actual operation frequency is 7 times / hour, which exceeds the threshold.

[0057] Resonant amplification tracking. By solving the parallel resonance equation, the amplification factors at each parallel resonance point are obtained: 1.25 at 87Hz, 7.75 at 131Hz, 7.4 at 169Hz, and 3.46 at 227Hz. The amplification factors at 131Hz and 169Hz both exceed the preset threshold of 6. 131Hz is close to the 3rd harmonic (150Hz) frequency band, and 169Hz falls into the interharmonic frequency band. The harmonic currents at these frequencies will be significantly amplified, leading to a further increase in the transformer terminal voltage and exacerbating the electrical stress on the winding insulation. Based on this, the unit triggers a resonant amplification over-limit warning.

[0058] Step S3: Multi-source fusion assessment. The health scores for the four dimensions are determined as follows: Oil Gas Analysis Unit: Fault code 102 corresponds to arc discharge, and the absolute acetylene production rate exceeds the preset threshold. DGA =60.

[0059] Harmonic feature extraction unit: U harm,max =18216V, U base =24000V, after accumulating the peak values ​​of each harmonic, f harm =62.

[0060] Operational shock accumulation unit: cumulative damage index D cum =0.811, f op =100×exp(-0.811 / 0.1)=100×exp(-8.11)≈0.03.

[0061] Resonant amplification tracking unit: maximum amplification factor is 7.75, f res =100-min(100,(7.75-6)×20)=65.

[0062] Take the weighting coefficient ω DGA =0.35, ω harm =0.25, ω op =0.25, ω res =0.15, Comprehensive Health Index: HI=0.35×60+0.25×62+0.25×0.03+0.15×65=46.26, Operational Impact Cumulative Score f op =0.03 is the lowest among the four dimensions, forming a precipitous gap with the other three dimensions, and the root cause is determined to be the cumulative damage from operational shocks.

[0063] Step S4, Early Warning Decision. The comprehensive health index of 46.26 falls within the 40-60 range, and the early warning decision module outputs an abnormal status. The output maintenance recommendations include: scheduling transformer shutdown for maintenance, focusing on checking for deformation of the high-voltage coil and loose insulation pads between coils; conducting a current-cutting value test on the vacuum circuit breaker to assess its breaking performance; adjusting the closing position from a high to a low position to reduce electrodynamic impact; and checking the independence and reliability of all grounding systems.

[0064] The above steps S1 to S4 are executed cyclically, completing a full data acquisition, feature extraction, fusion evaluation, and early warning decision-making process in each sampling cycle. After adopting the evaluation method and system provided by this invention, the steel plant conducted targeted repairs and maintenance on the transformer, focusing on adjusting the closing tap position, reinforcing the coils, and inspecting the grounding system. After completing the above rectification measures, the transformer operated normally for more than two years without experiencing similar faults, verifying the effectiveness and accuracy of the evaluation method of this invention.

[0065] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0066] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A transformer operating status assessment system based on multi-source data, characterized in that, include: Multi-source data acquisition module, multi-source feature extraction module, multi-source fusion evaluation module, and early warning decision-making module; The multi-source data acquisition module is used to collect data on dissolved gas composition in transformer oil, high-voltage side harmonic current and harmonic voltage, vacuum circuit breaker operation records, and power supply system data in real time. The multi-source feature extraction module is used to obtain the current-cutting value, the equivalent magnetizing inductance, the equivalent ground capacitance, and the lightning impulse withstand voltage from the operation record of the vacuum circuit breaker. The current-cutting overvoltage amplitude is obtained by multiplying the current-cutting value by the square root of the equivalent magnetizing inductance divided by the equivalent ground capacitance. The ratio of the current-cutting overvoltage amplitude to the lightning impulse withstand voltage is exponentially calculated according to a preset insulation life index to obtain the insulation damage equivalent of a single operation. The insulation damage equivalents of each operation within the statistical period are accumulated to obtain the cumulative insulation damage index. The multi-source fusion evaluation module is used to weight and fuse the output of the multi-source feature extraction module to obtain a comprehensive health index. The early warning decision module is used to issue graded early warnings based on the comprehensive health index and output maintenance suggestions.

2. The transformer operation status assessment system based on multi-source data according to claim 1, characterized in that, In the operational impact accumulation unit, the overvoltage amplitude U is cut off. tr Calculate using the following formula: , Among them I cut The current-cutting value is given by L, where L is the equivalent magnetizing inductance of the transformer, and C is the equivalent capacitance to ground of the transformer; the insulation damage equivalent of a single operation is given by D. op Calculate using the following formula: U BIL This refers to the lightning impulse withstand voltage of the transformer. Cumulative Insulation Damage Index D cum Calculate using the following formula: , where N op D represents the total number of operations within the statistical period. op,j Let be the insulation damage equivalent of the j-th operation.

3. The transformer operation status assessment system based on multi-source data according to claim 1, characterized in that, The operation impact accumulation unit triggers an early warning when the average number of operations per hour within the statistical period exceeds a preset threshold.

4. The transformer operation status assessment system based on multi-source data according to claim 1, characterized in that, The multi-source feature extraction module also includes an oil gas analysis unit, which is used to perform fault coding on dissolved gas data in oil using the three-ratio method and calculate the absolute gas production rate of acetylene. When the absolute gas production rate of acetylene exceeds a preset threshold, an early warning is triggered.

5. The transformer operation status assessment system based on multi-source data according to claim 1, characterized in that, The multi-source feature extraction module also includes a harmonic feature extraction unit, which is used to calculate the superimposed peak value of each harmonic voltage under typical operating conditions and the superimposed peak value under extreme operating conditions, and to calculate the ratio between the instantaneous maximum value of each harmonic current and its statistical typical value.

6. The transformer operation status assessment system based on multi-source data according to claim 1, characterized in that, The multi-source feature extraction module also includes a resonant amplification tracking unit, which is used to obtain the parallel resonant frequency and the corresponding amplification factor by solving the parallel resonant equation, and trigger an early warning when the amplification factor exceeds a preset threshold.

7. The transformer operation status assessment system based on multi-source data according to claim 1, characterized in that, In the multi-source fusion assessment module, the comprehensive health index is a weighted sum of normalized health scores from four dimensions: oil gas analysis, harmonic characteristics, operational shock accumulation, and resonant amplification tracking. Root cause localization is achieved by comparing the normalized health scores of the four dimensions, and the fault type corresponding to the lowest score is determined as the root cause.

8. The transformer operation status assessment system based on multi-source data according to claim 1, characterized in that, In the early warning decision module, the levels of the graded early warning include at least normal state, attention state, abnormal state and dangerous state, and each level corresponds to a different preset value range of the comprehensive health index; the larger the value of the comprehensive health index, the healthier the transformer's operating state.

9. A method for evaluating the operating status of transformers based on multi-source data, characterized in that, Includes the following steps: S1. Multi-source data acquisition steps: Real-time acquisition of dissolved gas composition data in transformer oil, high-voltage side harmonic current and harmonic voltage data, vacuum circuit breaker operation records, and power supply system data; S2. Multi-source feature extraction step: Obtain the current-cutting value, the equivalent magnetizing inductance, the equivalent ground capacitance, and the lightning impulse withstand voltage from the operation record of the vacuum circuit breaker. The product of the current-cutting value and the square root of the equivalent magnetizing inductance divided by the equivalent ground capacitance is used as the current-cutting overvoltage amplitude. The ratio of the current-cutting overvoltage amplitude to the lightning impulse withstand voltage is exponentially calculated according to the preset insulation life index to obtain the insulation damage equivalent of a single operation. The insulation damage equivalents of each operation within the statistical period are accumulated to obtain the cumulative insulation damage index. S3, Multi-source fusion evaluation step: The output of step S2 is weighted and fused to obtain the comprehensive health index; S4. Early warning decision-making step: Based on the comprehensive health index output in step S3, classify and issue early warnings and output maintenance suggestions.

10. The transformer operating status assessment method based on multi-source data according to claim 9, characterized in that, Step S2 also includes: using the three-ratio method to code faults in the dissolved gas data in the oil and calculating the absolute gas production rate of acetylene; calculating the peak value of each harmonic voltage superposition and the peak value ratio of harmonic current for the harmonic current and harmonic voltage data; and calculating the parallel resonant frequency and the corresponding amplification factor for the power supply system data.