Harbor district transformer bearing capacity evaluation method

Abstract

The invention belongs to the technical field of transformer evaluation, and particularly relates to a harbor area transformer bearing capacity evaluation method, which comprises the steps of S1, analyzing voltage and current distribution characteristics and a harmonic component propagation rule of a harbor area microgrid under high-proportion new energy access; analyzing a change rule of a winding leakage magnetic field, an iron core magnetic flux density and winding and iron core loss; s2, calculating the hot spot temperature of the transformer in each historical time period; s3, calculating the predicted residual service life Tb and the predicted actual total service life Tins of the transformer; s4, calculating an aging coefficient B and a primary health index H1 of the transformer; s5, calculating a comprehensive health index Hcom of the transformer; s6, calculating a final health index H of the transformer; s7, if H reaches a preset threshold value, equipment is replaced; and if not, reversely deducing and predicting the residual life Tb1. According to the method, the operation state of the transformer can be accurately evaluated under the working condition of frequent power flow switching.

Description

Technical Field

[0001] The present invention belongs to the technical field of transformer evaluation, and in particular relates to a method for evaluating the carrying capacity of a transformer in a port area. Background Art

[0002] As core equipment in power systems, power transformers undertake critical functions of power transmission and voltage conversion. Their operational reliability is directly related to the safety and economic viability of the power grid. In port power distribution systems, in particular, transformers must cope with high load fluctuations, complex electromagnetic environments, and special operating conditions, posing significant challenges to their insulation performance and lifespan management.

[0003] Traditional research has focused on optimizing equipment materials and controlling losses, but has overlooked the unique energy structure changes inherent in port microgrids. The intermittent output of distributed photovoltaic and wind power, combined with impactful loads such as shore power systems and electric heavy truck charging stations, leads to frequent "source-load" power imbalances in the system, triggering rapid switching of power flow direction and amplitude. This frequent bidirectional power flow exposes traditional unidirectional transformer designs to multiple operational risks: reverse power flow leads to unbalanced core flux distribution, exacerbating magnetic saturation and harmonic distortion; a surge in winding eddy current and iron losses causes localized overheating, accelerating insulation aging; and electromagnetic transients induce mechanical vibration and structural stress concentration.

[0004] Existing research suffers from three core flaws: First, models based on steady-state or unidirectional flow assumptions struggle to accurately characterize the dynamic switching characteristics of bidirectional flow, leading to ineffective harmonic management and inaccurate temperature rise predictions. Second, insufficient research on multi-physics coupling mechanisms prevents quantifying the cumulative damage to insulation from electromagnetic, thermal, and mechanical stresses. Current equivalent aging theories ignore the effects of transient overloads and thermal shock. Third, there is a lack of a dynamic assessment system tailored to port scenarios. Existing operation and maintenance strategies fail to incorporate transformer health into scheduling optimization, and monitoring methods still focus on steady-state parameters while ignoring transient processes. Typical cases show that the misalignment between photovoltaic output and port operations forces step-down transformers to step up, increasing the voltage on the low-voltage side and causing core saturation. Combined with harmonic injection from power electronics, this creates a vicious cycle of strong "electromagnetic-thermal" coupling.

[0005] Therefore, how to accurately evaluate the operating status of the transformer under conditions of frequent power flow switching has become an urgent problem to be solved. Summary of the Invention

[0006] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for evaluating the carrying capacity of a port transformer, which can accurately evaluate the operating status of the transformer under conditions of frequent power flow switching.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0008] A method for evaluating the load-carrying capacity of a transformer in a port area comprises the following steps:

[0009] S1. Construct a grid-connected simulation model for a doubly-fed asynchronous wind turbine generator and distributed photovoltaic power source. Combined with Monte Carlo random sampling technology and harmonic power flow calculation methods, analyze the voltage and current distribution characteristics and propagation patterns of harmonic components when a high proportion of renewable energy is connected to the Xiagang microgrid. Use finite element simulation technology to construct a transformer multi-physics field coupling model to simulate the dynamic distribution of the electromagnetic field and temperature field inside the transformer under conditions of frequent power flow switching, and analyze the changing patterns of winding leakage magnetic field, core magnetic flux density, and winding and core losses.

[0010] S2. Based on the analysis results of S1, taking harmonics and magnetic flux density changes as the key factors affecting transformer losses, an improved model for calculating the hot spot temperature of the transformer when the power flow frequently switches is constructed. The model is used to calculate the hot spot temperature of the transformer at the corresponding time based on the transformer load data and ambient temperature data. The historical load data and ambient temperature data of the transformer are obtained, and the hot spot temperature of the transformer in each historical time period is calculated using the improved hot spot temperature calculation model.

[0011] S3. Analyze the relationship between hotspot temperature and life loss, and build a hotspot temperature rise and life loss model for the transformer based on the analysis results. Calculate the corresponding insulation life loss according to the hotspot temperature of the transformer; Calculate the insulation life loss T according to the hotspot temperature of the transformer in each historical time period obtained in S2 through the hotspot temperature rise and life loss model. N ; Then calculate its estimated remaining service life T b And the expected actual total life T ins , T ins =T suv +T b , T b =T a -T N Where, T suv is the operation time of the transformer; T a The service life of the transformer is preset at the factory;

[0012] S4. Calculate the transformer load level correction factor K L and operating environment correction factor K T , combined with S3 to obtain the expected actual total life of the transformer T ins , calculate the expected service life T of the transformer end ; Based on T end Calculate the transformer's aging factor B and primary health index H1, T b =T a -T0; where H0 is the initial health index, determined by the original information of the transformer; Ta is the year when the health index assessment is conducted, T0 is the year of initial commissioning; T b The remaining life of the equipment;

[0013] S5. Calculate the simplified test health index H of transformer oil 2a , Oil chromatography test health index H 2b , furfural test health index H 2c ; Combined with the primary health index H1 obtained by S4, calculate the comprehensive health index H of the transformer com , H com =f com max{H1,H 2a ,H 2b ,H 2c}; where f com is a preset constant;

[0014] S6. Combine the operating environment information of the transformer to obtain the health correction factor based on the operating status data and the health correction factor based on the fault repair record; calculate the comprehensive correction factor K of the transformer based on the health correction factor and the health correction factor com ; Combined with S5 to obtain the comprehensive health index H com , get the final health index H of the transformer, H = K com H com ;

[0015] S7. If the final health index H reaches the preset threshold H t , it is determined that the equipment needs to be replaced; if it is not reached, the remaining life is predicted Based on the predicted remaining life T b1 and aging factor B to formulate corresponding maintenance plans to optimize the operation and maintenance strategy.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] 1. Dynamic multi-physics coupling modeling. For bidirectional power flow switching conditions, this system integrates Monte Carlo random sampling and finite element simulation techniques to establish a dynamic electromagnetic-temperature field correlation model. This model accurately simulates harmonic propagation, sudden changes in magnetic flux density, and leakage field distortion caused by renewable energy integration, reveals the changing patterns of transient parameters within the transformer, and addresses the problem of magnetic saturation and loss calculation errors caused by ignoring power direction switching in existing technologies.

[0018] 2. Harmonic-flux density synergistic hotspot temperature correction. By introducing harmonic components and magnetic flux density as key correction parameters, a dynamic hotspot temperature calculation model is constructed. Compared with traditional methods that rely solely on load current, this model effectively captures the combined effects of harmonic injection from power electronics and core hysteresis on temperature rise, significantly improving the accuracy of hotspot temperature prediction in frequent power reverse scenarios and providing more reliable input conditions for insulation aging assessment.

[0019] 3. Multi-dimensional life loss quantitative assessment. This approach overcomes the limitations of a single temperature aging model and integrates historical load fluctuations, ambient temperature and humidity, and transient thermal shock parameters to establish a dynamic life loss calculation system. Compared to existing equivalent aging theories, this solution can quantitatively analyze the nonlinear degradation of insulation paper polymerization caused by transient overloads, enabling differentiated predictions of remaining service life and expected operating life, providing data support for differentiated operations and maintenance.

[0020] 4. A health diagnosis system based on multi-source data fusion. An innovative integrated health index model, combining chemical indicators like oil chromatography and furfural testing with electromagnetic parameters, is designed. This model, combined with operating environment correction factors and fault repair records, forms a multi-dimensional health status assessment framework. Compared to traditional single-parameter threshold judgments, this approach, through heterogeneous data feature extraction and combined diagnosis, can identify potential faults such as partial discharge and insulation degradation earlier, reducing the risk of misjudgment.

[0021] 5. Dynamic load-carrying capacity driven operation and maintenance decision-making. Based on the real-time health index and aging coefficient, the reverse model and the predicted remaining life T b1 , then based on the predicted remaining life T b1 and aging factor B to develop a corresponding maintenance plan to optimize the operation and maintenance strategy. Compared with static periodic maintenance strategies, this solution can dynamically adjust the maintenance plan based on the actual load status of the transformer, optimizing the allocation of maintenance resources while ensuring equipment safety, and achieving a paradigm shift from "planned maintenance" to "predictive maintenance."

[0022] In summary, this method solves the problem of insufficient adaptability of traditional assessment systems in bidirectional power flow scenarios through dynamic modeling, multi-physics field coupling analysis, and multi-source data fusion. It can accurately evaluate the operating status of transformers under conditions of frequent power flow switching, and provide the port area power grid with an intelligent operation and maintenance solution that takes into account both reliability and economy.

[0023] Preferably, in S2, the hotspot temperature calculation improved model constructed is:

[0024]

[0025] Where, Indicates the transformer hot spot temperature, Indicates the real-time ambient temperature. Indicates the temperature rise of the top oil temperature over the ambient temperature. It is the temperature rise of the winding hot spot above the top oil temperature;

[0026]

[0027]

[0028] Where, and They represent the initial and final temperature rises of the transformer winding hot spot relative to the top oil within a time interval t; τ h represents the winding time constant; and They represent the initial and final temperature rises of the top oil relative to the environment within a certain time interval t; τ to It represents the top oil time constant, and the calculation formula is:

[0029]

[0030] Where, τ to,R represents the top oil time constant under rated load; n∈[0.8,1] represents the transformer top oil temperature rise calculation index, and when n=1, τ to =τ to,R ; is the rated temperature rise of the top oil relative to the environment;

[0031] in, and satisfy:

[0032]

[0033]

[0034] Where, P R is the resistance loss, P EC is the eddy current loss, P NL is the no-load loss, P LL is the load loss, and the one with R subscript is the rated value of the corresponding variable; It is the rated temperature rise of the winding hot spot relative to the top oil temperature.

[0035] Such a setting, 1. Dynamic staged temperature rise calculation. Innovatively use the exponential function to describe the transition process of initial and final temperature rise, through the time constant τ h and τ to Quantify the dynamic characteristics of temperature changes. Compared to traditional steady-state models, this model accurately captures the evolution of winding and oil temperature gradients during sudden load changes, resolving the prediction lag problem of existing methods under shock loads. It is particularly suitable for frequent power reverse transmission in port areas.

[0036] 2. Dynamic correction of loss parameters. R ) and eddy current loss (P EC ) as an independent variable in the temperature rise calculation, establishing a mapping between loss type and temperature rise. Compared to simplified models that rely solely on load current, this approach can distinguish the contributions of different loss components to hotspot temperatures, significantly improving calculation accuracy for power electronics in harmonic injection scenarios.

[0037] 3. Time constant adaptive adjustment. For the top oil time constant τ to Design a dynamic correction formula containing the exponent n, through the rated value τ to,R The ratio of the oil temperature to the actual loss is used to achieve parameter adaptation. Compared with the traditional method of fixed time constant, it can accurately characterize the nonlinear characteristics of the oil temperature change rate with the load rate, solving the problem of insufficient adaptability of the existing model under low load rate conditions.

[0038] 4. Multi-level temperature rise coupling modeling. A three-level superposition model of ambient temperature, top oil temperature rise, and winding hotspot temperature rise is constructed to isolate the conduction paths of different heat sources. Compared to a single heat path model, this model can clearly quantify the amplifying effect of reduced oil circulation efficiency on hotspot temperature, providing a theoretical basis for targeted heat dissipation optimization and effectively identifying local overheating risks caused by cooling system anomalies.

[0039] In summary, this method breaks through the accuracy bottleneck of traditional models in transient temperature rise calculation through dynamic parameter correction and multi-stage heat conduction modeling, and can accurately evaluate the operating status of the transformer under conditions of frequent power flow switching.

[0040] Preferably, the calculation formula of each loss satisfies:

[0041]

[0042] P EC =F EC k EC P R ;

[0043]

[0044] P LL =P R +P EC +P OSL , P OSL =F B F OSL k OSL P R ;

[0045] Where, F R 、F EC 、F OSLare the harmonic loss calculation factors for resistance loss, eddy current loss, and stray loss respectively; I T is the effective value of the fundamental component phase current of the load current flowing through the transformer; R T is the transformer resistance parameter; k EC ∈

[0046] [0.05,0.15] is the eddy current loss coefficient; δ h , δ e are the hysteresis loss and eddy current loss coefficients in the transformer core respectively; f R The rated frequency of the transformer; B Tm is the maximum value of transformer flux density; P OSL is the stray loss; F B is the magnetic density change compensation coefficient; k OSL ∈[0.05,0.2] is the stray loss coefficient;

[0047] in,

[0048]

[0049]

[0050]

[0051] Where, I1 represents the effective value of the fundamental component of the load current, I k Indicates the effective value of the kth harmonic of the load current, B T is the transformer flux density, B TR is the rated magnetic flux density of the transformer, m B Parameters obtained by electrical test fitting of the transformer are used to quantify the saturation degree and the rate of change of the magnetization characteristics of the transformer core material after exceeding the rated magnetic flux density.

[0052] This setup can: 1. Accurately decouple the loss by type. Innovatively decompose the total loss into resistance loss (P R ), eddy current loss (P EC ), stray loss (P OSL ) and other independent components, through different harmonic factors (F R 、F EC 、F OSL ) enables modeling the correlation between loss type and harmonic order. Compared to traditional methods for calculating total losses, this approach can clearly quantify the amplification effect of high-frequency harmonics on eddy current losses and the contribution of low-frequency harmonics to resistance losses, providing a theoretical basis for targeted loss reduction.

[0053] 2. Dynamic quantification of harmonic loss. Construct a weight coefficient model for each harmonic component, such as the eddy current loss factor FEC Using k 2 Weighted calculation, stray loss factor FOSL uses k 0.8 Weighted, it accurately reflects the nonlinear impact of different harmonic orders on losses. Compared with the traditional equivalent conversion method based on total harmonic distortion (THD), it significantly improves the accuracy of mapping harmonic spectrum distribution characteristics to losses, and is particularly suitable for broadband harmonic scenarios dominated by power electronic equipment.

[0054] 3. Magnetic density compensation adaptive mechanism. Introducing magnetic density change compensation coefficient F B , establish the piecewise function relationship between stray loss and real-time magnetic flux density. T Exceeding rated value B TR When the parameter m is fitted by experiment B Dynamically correct loss calculation effectively avoids the problem of underestimation of stray losses in traditional methods under high magnetic density conditions, and provides a safety warning basis for transformer over-flux operation.

[0055] 4. Multi-dimensional loss superposition modeling. LL =P R +P EC +P OSL The linear superposition structure can realize the physical separation of fundamental wave loss and harmonic loss. Compared with the empirical coefficient method in the existing IEC standard, this model can analyze the influence of different material properties (such as k EC 、k OSL ) on the loss structure, providing a quantitative evaluation tool for transformer core selection, winding optimization and other design links.

[0056] 5. Parameter system scalability. By defining k EC 、k OSL Equal coefficient interval and experimental fitting parameter m B , establishing an open parameter framework that adapts to different transformer models. Compared to general models with fixed coefficients, this solution allows for flexible configuration based on specific equipment parameters such as material properties and cooling methods, significantly improving the model's applicability to diverse transformer scenarios, including oil-immersed and dry-type transformers.

[0057] In summary, this method breaks through the accuracy bottleneck of traditional loss calculation models under complex harmonic conditions through harmonic dynamic weighting, loss type decoupling and magnetic density adaptive compensation, and can accurately evaluate the operating status of the transformer under conditions of frequent power flow switching.

[0058] Preferably, in S3, the hotspot temperature rise and life loss model is:

[0059] The transformer insulation life aging factor at rated load rate and reference temperature is defined as:

[0060]

[0061] Corresponding insulation life loss T N for:

[0062]

[0063] Where, Δt i is the i-th time interval; N is the total number of time intervals experienced; F ins,i is Δt i The corresponding insulation life loss is as follows.

[0064] Such a setting, 1. Dynamic aging factor modeling. Innovative construction based on real-time hot spot temperature Exponential aging factor F ins , a nonlinear mapping between aging rate and temperature rise is achieved by designing a double temperature term in the numerator and denominator. Compared with the traditional Arrhenius model with a single temperature parameter, this formula is The structure strengthens the dynamic comparison between the rated temperature reference point (20°C corresponds to 283K) and the real-time temperature, more accurately reflecting the accelerating effect of temperature fluctuations on the degradation of the polymerization degree of insulation paper.

[0065] 2. Time discretization integration algorithm. This cumulative form converts continuous life loss into a weighted sum of discrete time intervals. Compared to existing integration methods based on average temperature, this solution is more adaptable to unevenly sampled time series data in actual monitoring systems, resolving the problem of cumulative integration errors caused by missing data in traditional methods. This approach is particularly suitable for port scenarios with frequent load fluctuations.

[0066] 3. Quantification of cumulative losses under multiple operating conditions. By calculating and superimposing data by time period, a dynamic correlation between life loss and operating conditions is established. Compared to single-point assessment methods, this method can accurately quantify the cumulative damage of transformers under complex operating conditions such as diurnal load cycles and intermittent photovoltaic output, providing theoretical support for the formulation of differentiated operation and maintenance cycles.

[0067] 4. Exponential function parameter optimization. A symmetrical design with the same coefficient of 15,000 is used in the numerator and denominator of the formula. This not only preserves the temperature sensitivity of the Arrhenius equation but also reduces the risk of overfitting the model to extreme temperature points through parameter balance optimization. Compared to traditional empirical parameter models, this enhances numerical stability while maintaining physical clarity.

[0068] Preferably, S4 includes:

[0069] S41. Obtain historical load data from the substation load monitoring system and calculate the average load rate β;

[0070]

[0071] Where, is the average load size during transformer operation; S N is the rated capacity;

[0072] S42. Calculate the load level correction factor K based on the β obtained in step 1. L ;

[0073]

[0074] S43. Calculate the operating environment correction factor K based on the geographical environment where the transformer equipment is installed. T ;

[0075]

[0076] S44, comparison time T suv With T a and preset advance adjustment time T set The difference between the expected operating life T end ;

[0077]

[0078] S45, based on T end Calculate the aging coefficient B, and calculate the primary health index H1 in combination with the aging coefficient:

[0079]

[0080]

[0081] Such a setting, 1. Dynamic load and environment compensation mechanism. Innovative design load level correction factor K L The piecewise function model is divided into six intervals based on the average load factor β, each with its own corresponding correction coefficient. Compared with traditional static correction models, this model accurately reflects the accelerated aging effect of transformers in different load factor intervals (such as no-load, light-load, and overload). It is particularly suitable for port areas with frequent load fluctuations, and solves the problem that existing methods are not sufficiently adaptable to dynamic loads.

[0082] 2. Multi-factor coupling modeling system. Constructing environmental correction factor K T This multi-dimensional evaluation framework comprehensively considers the installation environment (indoor / outdoor), extreme temperature threshold (39°C), and degree of contamination. Compared to simplified models that only distinguish between indoor and outdoor scenarios, this approach, through coupled analysis of temperature and contamination, more accurately quantifies the synergistic degradation effects of coastal high salt fog and high temperature environments on insulation materials, filling a gap in traditional methods for modeling the interaction of multiple environmental factors.

[0083] 3. Dynamic threshold adjustment for life prediction. end The preset advance adjustment time T is introduced into the calculation set , by comparing T suv The dynamic relationship with Ta enables adaptive switching of the life prediction model. Compared with fixed life calculation formulas, this mechanism can dynamically adjust the assessment strategy based on the remaining life margin of the equipment, effectively avoiding the problem of amplified prediction errors in traditional methods when the equipment is nearing the end of its life.

[0084] 4. Exponential health evolution model. Construct an exponential function based on the aging coefficient B The exponential relationship between the initial health index H0 and the time increment characterizes the evolution of the health state. Compared to the linear degradation assumption, this model better reflects the actual characteristic of insulation material aging, which increases exponentially with temperature and load. This provides a more realistic decay curve for preventive maintenance.

[0085] 5. Multi-source data fusion architecture. This architecture integrates load monitoring data (β), environmental parameters (temperature, contamination), and equipment inventory data (Tins) to build a cross-system data fusion assessment framework. Compared to single-data source assessment methods, this solution significantly improves the comprehensiveness and reliability of health status assessments through multi-dimensional correlation analysis of electrical parameters, environmental parameters, and historical operation and maintenance records.

[0086] Preferably, S5 includes:

[0087] S51. Establish the acid value, breakdown voltage, water content, and dielectric loss parameters based on the simplified oil test results; determine the concentrations of hydrogen, methane, ethane, ethylene, and acetylene dissolved in the transformer oil based on the oil chromatography test; and determine the furfural content (FFA) based on the furfural test.

[0088] S52. Obtain the test parameters acid value, breakdown voltage, micro-water content, and dielectric loss level value V based on the empirical formula 11 、V 12 、V 13 、V 14 , and test parameters of the concentration of hydrogen, methane, ethane, ethylene and acetylene dissolved in transformer oil 21 、V 22 、V 23 、V 24 、V 25 ;

[0089] S53. Based on expert evaluation, obtain the index weights C of the test parameters acid value, breakdown voltage, trace water content, and dielectric loss 11 、C 12 、C 13 、C 14, and the index weight C of the concentration of hydrogen, methane, ethane, ethylene and acetylene dissolved in the transformer oil 21 、C 22 、C 23 、C 24 、C 25 ;

[0090] S54. Calculate the simplified oil test health index H based on the results of S51-S53. 2a , oil chromatography test health index H 2b , furfural test health index H 2c ;

[0091]

[0092]

[0093] H 2c =2.33(FFA) 0.68 ;

[0094] S55, combined with the primary health index H1 obtained in S4, calculate the comprehensive health index H of the transformer com , H com =

[0095] f com max{H1,H 2a ,H 2b ,H 2c}; where f com is a preset constant.

[0096] This setup integrates three heterogeneous data sources: simplified oil testing (acid value, breakdown voltage, etc.), oil chromatography testing (gas concentrations like H2 and CH4), and furfural testing (FFA content), to establish a comprehensive monitoring system covering electrical performance, insulation degradation, and solid insulation cracking. Compared to traditional single-oil chemical indicator evaluation methods, this solution significantly improves the ability to jointly diagnose potential defects such as partial discharge and overheating faults through cross-validation of multiple indicators, avoiding the risk of misjudgment based on a single parameter.

[0097] 2. Dynamic weight allocation mechanism. Use expert evaluation method to assign weights to each test parameter (such as acid value C 11 , acetylene concentration C 25 Differentiated weights are assigned to various indicators (e.g., [e.g., [e.g., [unclear]), [e.g., [unclear]), [unclear]), breaking through the limitations of traditional fixed-weight models. Compared to existing equal-weighted approaches, this mechanism dynamically adapts weights to the characteristic parameters of different failure modes. This allows it to accurately capture early warning signals of key degradation indicators (e.g., a sharp increase in acetylene concentration), improving the sensitivity of early fault identification.

[0098] 3. Nonlinear aging modeling optimization. For furfural test health index H 2c Design 2.33 (FFA) 0.68 A nonlinear calculation formula was developed, using an exponential function to fit the actual relationship between the aging rate of insulating paper and furfural content. Compared to traditional linear conversion methods, this method better reflects the nonlinear law of polymerization decay in solid insulating materials. It can accurately reflect the accelerated aging trend near the critical FFA concentration, avoiding the lifespan prediction bias caused by linear extrapolation.

[0099] 4. Multi-level index fusion strategy. Construct H com =f com max{H1,H 2a ,H 2b ,H 2c The comprehensive health index model uses the maximum value selection principle to enhance the identification of weak link effects. Compared with the weighted average method, this strategy emphasizes the most deteriorating sub-index, ensuring that local serious defects are not masked by the overall health index, significantly improving the timeliness of early warning of sudden failures (such as sudden acetylene surges).

[0100] 5. Modular evaluation architecture design. The health index is decomposed into the primary health index H1 and the three types of test health index independent calculation modules, supporting flexible switching between sub-item evaluation and comprehensive diagnosis. Compared with the integrated calculation model, this architecture not only retains the independent diagnostic value of each test data, but also com Constants achieve system-level state fusion, providing accurate decision-making basis for differentiated operation and maintenance strategies (such as oil treatment and solid insulation replacement).

[0101] Preferably, in S52, the method for determining each level value is:

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111] Preferably, in S53, the weight of each indicator is:

[0112] C 11 =0.2191, C 12 =0.2191, C 13 =0.2191, C 14 =0.3425, C 21 =0.1923, C 22 =

[0113] 0.1154, C 23 =0.1154, C 24 =0.1154, C 25 =0.4615.

[0114] Preferably, in S6, the health status correction factor based on the operating status data includes: the operation time correction coefficient K 11 , Core grounding current correction factor K 12 ; Health status correction factors based on fault repair records include: Transformer appearance grade correction factor K 21 , casing reliability level correction factor K 22 , Cooling method correction coefficient K 23 , Family defect correction coefficient K 24 , Correction coefficient K of the number of failures in the past five years 25 , Near-zone short-circuit correction coefficient K 26 , partial discharge correction factor K 27 ;

[0115] Comprehensive correction factor K com The calculation formula is:

[0116]

[0117] Such a setting, 1. Multi-source data layered correction mechanism. Innovatively divide the correction factor into operating status data (K 1m ) and fault repair records (K 2n ) are two independent dimensions, covering real-time equipment operating parameters (such as core grounding current) and historical operation and maintenance data (such as the number of failures in the past five years). Compared to traditional single-dimensional correction methods, this hierarchical structure systematically distinguishes the impact weights of short-term state fluctuations and long-term aging trends, addressing the accuracy loss caused by existing models that conflate operating data with historical records.

[0118] 2. Full-factor dynamic coupling modeling. Through the product form Construct a comprehensive correction factor to achieve dynamic coupling analysis of parameters in different dimensions. Compared with linear combination methods such as the weighted average method, the product relationship can amplify the short-board effect of abnormal correction factors (such as a sudden drop in the reliable level of the casing), more sensitively capture the chain effect of local defects of the equipment on the overall health status, and avoid key risk indicators being diluted by conventional parameters.

[0119] 3. Data fusion over the entire life cycle. Integrate the operation time correction coefficient K 11 (reflecting the operation years) and the number of faults K 25 (dynamically update data) in the past five years to construct a health evolution tracking system with a time span covering the equipment's commissioning to date. Compared with the static reference value model, it can accurately characterize the non-linear change law of the equipment aging rate with the operation years.

[0120] Preferably, the operation time correction coefficient K 11 satisfies: when 0 ≤ T ≤ 5 years, K 11 = 1; when 5 < T ≤ 10 years, K 11 = 1.01; when 10 < T ≤ 20 years, K 11 = 1.02; when 20 < T ≤ 30 years, K 11 = 1.05; when T > 30 years, K 11 = 1.09; where T is the equipment commissioning time;

[0121] The iron core grounding current correction coefficient K 12 satisfies: when I = 0, K 12 = 1; when 0 < I ≤ 0.1A, K 12 = 1.05; when 0.1A < I ≤ 0.3A, K 12 = 1.1; when I > 0.3A, K 12 = 1.2; where I is the iron core grounding current;

[0122] The calculation formula for the transformer appearance grade correction coefficient K 21 is K 21 = 0.9 + 0.1×L; where L is the highest grade of the ratings for the 4 parts of the main body, cooling system, tap changer, and non-electrical components; the ratings are from 1 to 5;

[0123] In the bushing reliability level correction coefficient K 22 , a single bushing is matched according to the grade R: when R = 1, K R = 0.9, when R = 2, K R = 1, when R = 3, K R = 1.1, when K = 4, K R = 1.2, when R = 5, K R = 1.4; Denote the reliability levels of the three-phase bushings as K Ra 、K Rb 、KRc , if max{K Ra ,K Rb ,K Rc}>1, then K 22 =K Ra +K Rb +K Rc ; otherwise K 22 =

[0124] min{K Ra ,K Rb ,K Rc};

[0125] Cooling method correction factor K 23 In the case of oil-immersed self-cooling or oil-immersed air-cooling, K 23 =1, when the cooling method is forced oil circulation cooling K 23 =0.96, when the cooling method is forced guide oil circulation cooling K 23 =0.95;

[0126] Family defect correction factor K 24 In the same series, if there is no problem with the equipment, K 24 =0.96, when the same series of equipment has a few defects that do not endanger operation 24 =1, when there is a potential risk of repeated failures in the same series of equipment K 24 =1.04;

[0127] Correction coefficient K of the number of failures in the past five years 25 When n=0, K 25 =0.96, when n=1, K 25 =1, when n∈[2,4] 25 =1.04, when n∈[5,10] 25 =1.2, when n>10, K 25 =1.4;

[0128] Nearby short circuit correction factor K 26 If the transformer has a short circuit in the vicinity, K 26 =1.04, otherwise K 26 =1;

[0129] Partial discharge correction factor K 27 In the case of partial discharge in the transformer, K 27 =1.2, when there is no partial discharge in the transformer, K 27 =1.

[0130] This setup, 1. Multi-dimensional parameter coverage system is constructed. Innovatively integrates 9 types of correction factors such as years of operation, core grounding current, and appearance grade, covering four dimensions of equipment operation time, electrical characteristics, mechanical structure, and historical operation and maintenance data. Compared with the traditional simplified model that only considers load rate and oil temperature, this system monitors the abnormality of core grounding current (K 12 ) and partial discharge detection (K 27 ) can simultaneously capture hidden electrical defects and obvious mechanical degradation, solving the evaluation blind spot problem caused by the single parameter dimension of the existing method.

[0131] 2. Dynamic segmentation correction mechanism. 11 ) and the number of failures in the past five years (K 25 ) Designing a nonlinear piecewise function overcomes the limitations of traditional linear aging assumptions. When equipment operates for more than 30 years, the accelerated growth of K11 better reflects the rapid degradation rate of insulation materials in the later stages of aging. Compared to fixed-year attenuation models, this model can provide earlier warning of lifespan inflection points.

[0132] 3. Quantification of latent risk factors. 24 ) and cooling method correction factor (K 23 )'s differentiated design transforms traditional qualitative indicators, such as batch quality risk and cooling system efficiency, for the same model of equipment into calculable, quantitative parameters. Compared to empirical judgment methods, this solution systematically identifies the potential impact of cumulative efficiency degradation in forced oil circulation cooling systems on temperature rise, providing data support for cooling optimization.

[0133] 4. Defect sensitivity enhancement mechanism. In casing reliability level correction (K 22 ) uses a "short board effect" calculation strategy: when high-risk bushings are present, the sum of the three-phase coefficients is taken; otherwise, the minimum value is taken. Compared with the average value calculation method, this strategy can amplify the warning effect of single-phase bushing degradation on the overall health status, prevent critical defects from being diluted by normal phase parameters, and significantly improve the sensitivity of identifying local severe defects.

[0134] 5. Full life cycle tracking capability. Build the operation life (K 11 ) and near-zone short-circuit records (K 26 ) collaborative analysis model, enabling the combined calculation of equipment aging patterns and damage from sudden impact events. Compared to a static baseline model, this model dynamically reflects the differential impact of long-term, gradual aging and transient overload shocks on insulation life, more accurately quantifying the cumulative damage to winding mechanical strength caused by historical short-circuit currents. BRIEF DESCRIPTION OF THE DRAWINGS

[0135] In order to make the purpose, technical solutions and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0136] Figure 1 Flowchart of this method;

[0137] Figure 2 1 is a structural diagram of a grid-connected simulation model of a doubly-fed asynchronous wind turbine generator in an embodiment;

[0138] Figure 3 1 is a structural diagram of a distributed photovoltaic power grid-connected simulation model in an embodiment;

[0139] Figure 4 Schematic diagram of transformer core loss distribution obtained by transformer multi-physics field simulation in the embodiment;

[0140] Figure 5 Schematic diagram of transformer core temperature distribution obtained by transformer multi-physics field simulation in the embodiment. DETAILED DESCRIPTION

[0141] The following is a further detailed description through specific implementation methods:

[0142] Example:

[0143] like Figure 1 As shown, this embodiment discloses a method for evaluating the carrying capacity of a port transformer, comprising the following steps:

[0144] S1. Construct a grid-connected simulation model of a doubly-fed asynchronous wind turbine generator and a distributed photovoltaic power source. Combined with the Monte Carlo random sampling technique and the harmonic power flow calculation method, analyze the voltage and current distribution characteristics and the propagation law of harmonic components when a high proportion of renewable energy is connected to the microgrid in the Xiagang area. Use finite element simulation technology to construct a transformer multi-physics field coupling model to simulate the dynamic distribution of the electromagnetic field and temperature field inside the transformer under the condition of frequent power flow switching, and analyze the changing laws of the winding leakage magnetic field, core magnetic flux density, winding and core losses.

[0145] When implementing it specifically, Figure 2 、 Figure 3 As shown in the figure, this paper studies and analyzes the voltage and current distribution and harmonic generation characteristics of the port area microgrid when the power flow frequently switches under the condition of high proportion of new energy access.

[0146] This study constructs a grid-connected simulation model for a doubly-fed asynchronous wind turbine and distributed photovoltaic power sources. Using Monte Carlo random sampling techniques, the study generates multi-scenario uncertain power samples that incorporate wind and solar output fluctuations and load variations. This model simulates the frequent power flow switching conditions associated with a high proportion of renewable energy connected to the Xiagang microgrid. Harmonic power flow calculation methods are then used to analyze the voltage and current distribution characteristics and harmonic component propagation patterns at each node in the microgrid under different operating scenarios.

[0147] The high proportion of renewable energy access leads to frequent bidirectional switching of current flows in the port area's microgrid, resulting in significant dynamics in voltage and current distribution. During forward power supply, voltage exhibits regional attenuation with load fluctuations, and current distribution is dominated by traditional load characteristics. During reverse power supply, the low-voltage side voltage rises due to the reverse flow, potentially causing an increase in the magnetic flux density or even saturation of the step-down transformer core, leading to increased no-load losses and excitation current distortion. Current amplitude varies significantly with the proportion of renewable energy access and load factor. The higher the reverse load factor, the greater the reverse current amplitude on the low-voltage side. Furthermore, the current waveform exhibits non-sinusoidal characteristics due to the influence of nonlinear loads and power electronics, exacerbating the volatility of the system's electrical parameters.

[0148] The power electronic converters introduced by renewable energy integration are the primary source of harmonics in the port area's microgrid, generating characteristic harmonic components such as the 5th and 7th orders. When a doubly-fed wind turbine is connected to the grid, the 6n±1st harmonics generated by the rotor-side converter account for a significant proportion; photovoltaic inverters, due to their PWM modulation characteristics, generate high-frequency harmonics near the switching frequency. Under reverse power transmission conditions, the reverse flow causes changes in the transformer's excitation characteristics, further inducing core harmonic losses and causing harmonic content to increase with increasing reverse load factors. Research has shown that the harmonic distortion rate increases nonlinearly with the proportion of renewable energy integration, and that at the moment of power flow switching, the short-term harmonic amplitude surges due to transient shocks, resulting in a compound impact on the system's power quality and transformer insulation performance.

[0149] like Figure 4 、 Figure 5 As shown, the present invention studies and analyzes the distribution and change rules of the electric, magnetic and thermal multi-physical fields of the port transformer when the tidal current switches frequently and the impact characteristics when the tidal current switches frequently.

[0150] This study used finite element simulation technology to construct a transformer multi-physics coupling model, simulating the dynamic distribution of the electromagnetic and temperature fields within the transformer under conditions of frequent power flow switching. By setting different reverse load rates and voltage fluctuation scenarios, the changing patterns of winding leakage magnetic field, core flux density, and winding and core losses were analyzed. Combined with electromagnetic-magnetic-thermal coupling simulation, the impact of power flow switching on the transformer multi-physics distribution was explored.

[0151] When the tide switches frequently, the electric, magnetic and thermal multi-physical fields inside the port area transformer show significant dynamic characteristics: the reversal of the tide increases the voltage on the low-voltage side, the magnetic flux density of the iron core increases and is prone to saturation, resulting in excitation current distortion and nonlinear increase in no-load loss. The winding leakage magnetic field is concentrated at the end and increases with the increase of the reverse load rate, exacerbating the eddy current loss; the iron loss density is highest at the corner of the iron core and increases with the increase of magnetic density. The copper loss of the winding increases significantly due to the increase of reverse current and the influence of harmonics, and the loss is concentrated at the end of the low-voltage winding; the hot spot temperature of the winding in the temperature field increases and moves upward with the tide switching, and the coupling of the core and top oil temperature rise is aggravated.

[0152] Frequent switching of currents produces multi-dimensional impact characteristics on transformers: rapid current reversal causes periodic changes in the electromagnetic force of the winding, intensifies mechanical vibration and may cause structural stress concentration and winding displacement; loss mutations accelerate the local temperature rise rate, the insulation system is subjected to periodic thermal stress, and short-term hot spot temperature exceeds the limit, accelerating the aging of oil-paper insulation. Harmonics and magnetic flux fluctuations further enhance the thermal accumulation effect; the high-frequency voltage change rate causes the displacement current in the insulating medium to increase and energy to accumulate rapidly, triggering local discharge and causing the insulation material to deteriorate, significantly shortening the life of the transformer.

[0153] S2. Based on the analysis results of S1, taking harmonics and magnetic flux density changes as the key factors affecting transformer losses, an improved model for calculating the hot spot temperature of the transformer when the power flow frequently switches is constructed. The model is used to calculate the hot spot temperature of the transformer at the corresponding time based on the load data and ambient temperature data of the transformer; the historical load data and ambient temperature data of the transformer are obtained, and the hot spot temperature of the transformer in each historical time period is calculated using the improved hot spot temperature calculation model.

[0154] In specific implementation, the constructed improved hot spot temperature calculation model is:

[0155]

[0156] Where, Indicates the transformer hot spot temperature, Indicates the real-time ambient temperature. Indicates the temperature rise of the top oil temperature over the ambient temperature. It is the temperature rise of the winding hot spot above the top oil temperature;

[0157]

[0158]

[0159] Where, and They represent the initial and final temperature rises of the transformer winding hot spot relative to the top oil within a time interval t; τ h Indicates the winding time constant. This value is only related to the material structure of the transformer winding and can be directly provided by the transformer manufacturer. and They represent the initial and final temperature rises of the top oil relative to the environment within a certain time interval t;

[0160] τ to It represents the top oil time constant, and the calculation formula is:

[0161]

[0162] Where, τ to,RIndicates the top oil time constant under rated load, which can be directly provided by the transformer manufacturer; n∈

[0163] [0.8,1] represents the transformer top oil temperature rise calculation index, and when n=1, τ to =τ to,R ; is the rated temperature rise of the top oil relative to the environment;

[0164] in, and satisfy:

[0165]

[0166]

[0167] Where, P R is the resistance loss, P EC is the eddy current loss, P NL is the no-load loss, P LL is the load loss, and the one with R subscript is the rated value of the corresponding variable; It is the rated temperature rise of the winding hot spot relative to the top oil temperature.

[0168] In this way, the exponential function is innovatively used to describe the transition process between the initial state and the final state temperature rise, and the time constant τ is used to describe the transition process between the initial state and the final state temperature rise. h and τ ti Quantify the dynamic characteristics of temperature changes. Compared with the traditional steady-state model, this model can accurately capture the evolution of the temperature gradient of the winding and oil when the load changes suddenly, solving the problem of prediction lag under impact load in existing methods. It is especially suitable for the frequent power reverse transmission scenario in the port area. In addition, the resistance loss (P R ) and eddy current loss (P EC ) is introduced as an independent variable in the temperature rise calculation to establish a mapping relationship between loss type and temperature rise. Compared with the simplified model that relies only on load current, this solution can distinguish the contribution of different loss components to the hotspot temperature, significantly improving the calculation accuracy of power electronic equipment in the harmonic injection scenario. In addition, for the top oil time constant τ to Design a dynamic correction formula containing the exponent n, through the rated value τ to,R Parameter adaptation is achieved by comparing the ratio of the oil temperature change rate to the actual loss. Compared with the traditional method of fixed time constants, it can accurately characterize the nonlinear characteristics of the oil temperature change rate with the load rate, solving the problem of insufficient adaptability of the existing model under low-load conditions. In addition, a three-level superposition model of ambient temperature, top oil temperature rise, and winding hot spot temperature rise is constructed to separate the conduction paths of different heat sources. Compared with a single heat path model, it can clearly quantify the amplification effect of reduced oil circulation efficiency on hot spot temperature, providing a theoretical basis for targeted heat dissipation optimization and effectively identifying the risk of local overheating caused by cooling system anomalies.

[0169] In specific implementation, the calculation formula of each loss satisfies:

[0170]

[0171] P EC =F EC k EC P R ;

[0172]

[0173] P LL =P R +P EC +P OSL , P OSL =P B F OSL k OSL P r ;

[0174] Where, f R 、F EC 、F OSL are the harmonic loss calculation factors for resistance loss, eddy current loss, and stray loss respectively; I T is the effective value of the fundamental component phase current of the load current flowing through the transformer; R T is the transformer resistance parameter; k EC ∈

[0175] [0.05,0.15] is the eddy current loss coefficient; δ h , δ e are the hysteresis loss and eddy current loss coefficients in the transformer core respectively; f R The rated frequency of the transformer; B Tm is the maximum value of transformer flux density; P OSL is the stray loss; F B is the magnetic density change compensation coefficient; k OSL ∈[0.05,0.2] is the stray loss coefficient;

[0176] in,

[0177]

[0178]

[0179]

[0180] Where, I1 represents the effective value of the fundamental component of the load current, I k Indicates the effective value of the kth harmonic of the load current, B Tis the transformer flux density, B TR is the rated magnetic flux density of the transformer, m B Parameters obtained by electrical test fitting of the transformer are used to quantify the saturation degree and rate of change of magnetization characteristics of the transformer core material after exceeding the rated magnetic flux density.

[0181] In this way, the total loss is decomposed into resistance loss (P R ), eddy current loss (P EC ), stray loss (P OSL ) and other independent components, through different harmonic factors (F R 、F EC 、F OSL ) to achieve the correlation modeling between loss type and harmonic order. Compared with the traditional overall calculation method of total loss, this solution can clearly quantify the amplification effect of high-frequency harmonics on eddy current loss and the contribution of low-frequency harmonics to resistance loss, providing a theoretical basis for targeted loss reduction transformation. In addition, a weight coefficient model is constructed for each harmonic component, such as the eddy current loss factor F EC Using k 2 Weighted calculation, stray loss factor FOSL uses k 0.8 Weighted, accurately reflects the nonlinear effect of different harmonic orders on loss. Compared with the traditional equivalent conversion method based on total harmonic distortion (THD), it significantly improves the mapping accuracy of harmonic spectrum distribution characteristics to loss, especially suitable for broadband harmonic scenarios dominated by power electronic equipment. In addition, the magnetic density change compensation coefficient F is introduced. B , establish the piecewise function relationship between stray loss and real-time magnetic flux density. T Exceeding rated value B TR When the parameter m is fitted by experiment B Dynamically correct loss calculation effectively avoids the problem of underestimation of stray losses in traditional methods under high magnetic density conditions, and provides a safety warning basis for transformer over-flux operation.

[0182] By P LL =P R +P EC +P OsL The linear superposition structure can realize the physical separation of fundamental wave loss and harmonic loss. Compared with the empirical coefficient method in the existing IEC standard, this model can analyze the influence of different material properties (such as k EC 、k OSL ) on the differential impact of loss composition, providing a quantitative evaluation tool for transformer core selection, winding optimization and other design links. In addition, by defining k EC 、k OSL Equal coefficient interval and experimental fitting parameter m B, establishing an open parameter framework that adapts to different transformer models. Compared to general models with fixed coefficients, this solution allows for flexible configuration based on specific equipment parameters such as material properties and cooling methods, significantly improving the model's applicability to diverse transformer scenarios, including oil-immersed and dry-type transformers.

[0183] S3. Analyze the relationship between hotspot temperature and life loss, and build a hotspot temperature rise and life loss model for the transformer based on the analysis results. Calculate the corresponding insulation life loss according to the hotspot temperature of the transformer; Calculate the insulation life loss T according to the hotspot temperature of the transformer in each historical time period obtained in S2 through the hotspot temperature rise and life loss model. N ; Then calculate its estimated remaining service life T b And the expected actual total life T ins , T ins =T suv +T b , T b =T a -T N Where, T suv is the operation time of the transformer; T a The service life of the transformer is preset at the factory.

[0184] In specific implementation, the hotspot temperature rise and life loss model is:

[0185] The transformer insulation life aging factor at rated load rate and reference temperature is defined as:

[0186]

[0187] Corresponding insulation life loss T N for:

[0188]

[0189] Where Δt i is the i-th time interval; N is the total number of time intervals experienced; F ins,i is Δt i The corresponding insulation life loss is as follows.

[0190] In this way, the innovation is based on the real-time hot spot temperature Exponential aging factor F ins , a nonlinear mapping between aging rate and temperature rise is achieved by designing a double temperature term in the numerator and denominator. Compared with the traditional Arrhenius model with a single temperature parameter, this formula is The structure strengthens the dynamic comparison between the rated temperature reference point (20℃ corresponds to 283K) and the real-time temperature, and more accurately reflects the accelerated effect of temperature fluctuations on the degradation of the insulation paper polymerization degree. The cumulative form of the life loss is converted into a weighted sum of discrete time intervals. Compared with existing average temperature-based integration methods, this solution can adapt to non-uniformly sampled time series data in actual monitoring systems, solving the problem of cumulative integration errors caused by missing data in traditional methods. It is particularly suitable for port scenarios with frequent load fluctuations. Furthermore, through time-segment calculation and superposition, a dynamic correlation between life loss and operating conditions is established. Compared with single-point assessment methods, this method can accurately quantify the cumulative damage of transformers under complex operating conditions such as diurnal load cycles and intermittent photovoltaic output, providing theoretical support for the formulation of differentiated operation and maintenance cycles. In addition, the numerator and denominator of the formula adopt a symmetrical structure design with the same coefficient of 15,000, which not only preserves the temperature sensitivity of the Arrhenius equation but also reduces the risk of overfitting the model to extreme temperature points through parameter balance optimization. Compared with traditional empirical parameter models, it enhances numerical stability while maintaining physical clarity.

[0191] S4. Calculate the transformer load level correction factor K L and operating environment correction factor K T , combined with S3 to obtain the expected actual total life of the transformer T ins , calculate the expected service life of the transformer T end ; Based on T end Calculate the transformer's aging factor B and primary health index H1, Among them, H0 is the initial health index, which is determined by the original information of the transformer; T a is the year when the health index assessment is conducted, and T0 is the year of initial commissioning.

[0192] In specific implementation, S4 includes:

[0193] S41. Obtain historical load data from the substation load monitoring system and calculate the average load rate β;

[0194]

[0195] Where, is the average load size during transformer operation; S N is the rated capacity;

[0196] S42. Calculate the load level correction factor K based on the β obtained in step 1. L ;

[0197]

[0198] S43. Calculate the operating environment correction factor K based on the geographical environment where the transformer equipment is installed. T ;

[0199]

[0200] S44, comparison time T suv With T a and preset advance adjustment time T set The difference between the expected operating life T end ;

[0201]

[0202] S45, based on T end Calculate the aging coefficient B, and calculate the primary health index H1 in combination with the aging coefficient:

[0203]

[0204]

[0205] In this way, the innovative design load level correction factor K L The piecewise function model is divided into six intervals according to the average load rate β and corresponds to different correction coefficients. Compared with the traditional static correction model, it can accurately reflect the aging acceleration effect of the transformer in different load rate intervals (no load, light load, overload, etc.), which is especially suitable for the working conditions with frequent load fluctuations in the port area, solving the problem of insufficient adaptability of the existing method to dynamic loads. In addition, the environmental correction factor K is constructed. T The multi-dimensional evaluation framework comprehensively considers the three influencing factors of installation environment (indoor / outdoor), extreme temperature threshold (39°C) and degree of contamination. Compared with the simplified model that only distinguishes indoor and outdoor scenes, this solution more accurately quantifies the synergistic degradation effect of coastal high salt fog and high temperature environment on insulation materials through the coupling analysis of temperature and contamination, filling the gap in traditional methods in modeling the interaction of multiple environmental factors. In addition, within the expected operating life T end The preset advance adjustment time T is introduced into the calculation set , by comparing T suv The dynamic relationship with Ta enables adaptive switching of the life prediction model. Compared with fixed life calculation formulas, this mechanism can dynamically adjust the assessment strategy based on the remaining life margin of the equipment, effectively avoiding the problem of amplified prediction errors in traditional methods when the equipment is nearing the end of its life.

[0206] In addition, an exponential function is constructed based on the aging coefficient B The evolution of health status is characterized by the exponential relationship between the initial health index H0 and the time increment. Compared with the linear degradation assumption, this model is more consistent with the actual characteristics of the aging rate of insulation materials increasing exponentially with temperature and load, providing a decay curve that is more in line with engineering practice for preventive maintenance. In addition, load monitoring data (β), environmental parameters (temperature, contamination) and equipment ledger data (Tins) are integrated to build an evaluation framework for cross-system data fusion. Compared with the evaluation method based on a single data source, this solution significantly improves the comprehensiveness and reliability of health status assessment through multi-dimensional correlation analysis of electrical parameters, environmental parameters, and historical operation and maintenance records.

[0207] S5. Calculate the simplified test health index H of transformer oil 2a , Oil chromatography test health index H 2b , furfural test health index H 2c ; Combined with the primary health index H1 obtained by S4, calculate the comprehensive health index H of the transformer com , H com =f com max{H1,H 2a ,H 2b ,H 2c}; where f com is a preset constant.

[0208] In specific implementation, S5 includes:

[0209] S51. Establish the acid value, breakdown voltage, water content, and dielectric loss parameters based on the simplified oil test results; determine the concentrations of hydrogen, methane, ethane, ethylene, and acetylene dissolved in the transformer oil based on the oil chromatography test; and determine the furfural content (FFA) based on the furfural test.

[0210] S52. Obtain the test parameters acid value, breakdown voltage, micro-water content, and dielectric loss level value V based on the empirical formula 11 、V 12 、V 13 、V 14 , and test parameters of the concentration of hydrogen, methane, ethane, ethylene and acetylene dissolved in transformer oil 21 、V 22 、V 23 、V 24 、V 25 ;

[0211] S53. Based on expert evaluation, obtain the index weights C of the test parameters acid value, breakdown voltage, trace water content, and dielectric loss 11 、C 12 、C 13 、C 14, and the index weight C of the concentration of hydrogen, methane, ethane, ethylene and acetylene dissolved in the transformer oil 21 、C 22 、C 23 、C 24 、C 25 ;

[0212] S54. Calculate the simplified oil test health index H based on the results of S51-S53. 2a , oil chromatography test health index H 2b , furfural test health index H 2c ;

[0213]

[0214]

[0215] H 2c =2.33(FFA) 0.68 ;

[0216] S55, combined with the primary health index H1 obtained in S4, calculate the comprehensive health index H of the transformer com , H com =f com max{H1,H 2a ,H 2b ,H 2c}; where f com is a preset constant.

[0217] In S52, the method for determining each level value is as follows:

[0218]

[0219]

[0220]

[0221]

[0222]

[0223]

[0224]

[0225]

[0226]

[0227] In S53, the weights of each indicator are:

[0228] C 11 =0.2191, C 12 =0.2191, C 13 =0.2191, C 14 =0.3425, C 21 =0.1923, C 22 =

[0229] 0.1154, C 23 =0.1154, C 24 =0.1154, C 25 =0.4615.

[0230] This setup innovatively integrates three heterogeneous data sources: simplified oil test (acid value, breakdown voltage, etc.), oil chromatography test (gas concentrations such as H2, CH4), and furfural test (FFA content), to establish a comprehensive monitoring system covering electrical performance, insulation degradation, and solid insulation cracking. Compared with the traditional single oil index evaluation method, this solution significantly improves the joint diagnosis capability of potential defects such as partial discharge and overheating faults through multi-dimensional index cross-validation, avoiding the risk of misjudgment of a single parameter. In addition, the expert evaluation method is used to evaluate each test parameter (such as acid value C 11 , acetylene concentration C 25 Compared with the existing equal weighting method, this mechanism dynamically adapts the characteristic parameters of different failure modes through weights, accurately captures the early warning signals of key degradation indicators (such as a sharp increase in acetylene concentration), and improves the sensitivity of early fault identification. 2c Design 2.33 (FFA) 0.68 A nonlinear calculation formula was developed, using an exponential function to fit the actual relationship between the aging rate of insulating paper and furfural content. Compared to traditional linear conversion methods, this method better reflects the nonlinear law of polymerization decay in solid insulating materials. It can accurately reflect the accelerated aging trend near the critical FFA concentration, avoiding the lifespan prediction bias caused by linear extrapolation.

[0231] Build H com =f com max{H1,H 2a ,H 2b ,H 2cThe comprehensive health index model of {} adopts the maximum selection principle to strengthen the identification of short - board effects. Compared with the weighted average method, this strategy ensures that local serious defects are not covered by the overall health index by highlighting the most deteriorated sub - index, greatly improving the timeliness of early warning for sudden failures (such as sudden increase in acetylene). Moreover, the health index is decomposed into an independent calculation module of the primary health index H1 and three types of test health indexes, supporting flexible switching between sub - item evaluation and comprehensive diagnosis. Compared with the integrated calculation model, this architecture not only retains the independent diagnostic value of each test data, but also realizes system - level state fusion through the f com constant, providing an accurate decision - making basis for differential operation and maintenance strategies (such as oil treatment, solid insulation replacement).

[0232] S6. Combine the operation environment information of the transformer to obtain the health status correction factor based on the operation condition data and the health status correction factor based on the fault repair records; calculate the comprehensive correction factor K of the transformer based on the health status correction factor and the health status correction factor com ; Then combine with the comprehensive health index H obtained in S5 com to obtain the final health index H of the transformer, H = K com H com .

[0233] Specifically, when implemented, the health status correction factor based on the operation condition data includes: operation time correction coefficient K 11 、core grounding current correction coefficient K 12 ; The health status correction factor based on the fault repair records includes: transformer appearance grade correction coefficient K 21 、bushing reliability grade correction coefficient K 22 、cooling method correction coefficient K 23 、family defect correction coefficient K 24 、number of faults in the past five years correction coefficient K 25 、near - area short - circuit correction coefficient K 26 、partial discharge correction coefficient K 27 ;

[0234] The operation time correction coefficient K 11 satisfies: when 0 ≤ T ≤ 5 years, K 11 = 1; when 5 < T ≤ 10 years, K<​​​​​​​​​​​​​=1; when 0 <I≤0.1A,K 12 =1.05; when 0.1A <I≤0.3A,K 12 =1.1; when I>0.3A, K 12 =1.2; where I is the core grounding current;

[0236] Transformer appearance grade correction factor K 21 The calculation formula is K 21 =0.9+0.1×L; where L is the highest rating for the four parts: the main body, cooling system, tap changer, and non-electrical components; the rating is from 1 to 5;

[0237] Casing reliability level correction factor K 22 In the middle, the single casing is matched according to the level R: when R=1, K R =0.9, R=2 when K R =1, R=3, K R =1.1, R=4 when K R =1.2, R=5 when K R =1.4; the reliability levels of the three-phase bushings are K Ra , K Rb , K Rc , if max{K Ra ,K Rb ,K Rc}>1, then K 22 =K Ra +K Rb +K Rc ; otherwise K 22 =

[0238] min{K Ra ,K Rb ,K Rc};

[0239] Cooling method correction factor K 23 In the case of oil-immersed self-cooling or oil-immersed air-cooling, K 23 =1, when the cooling method is forced oil circulation cooling K 23 =0.96, when the cooling method is forced guide oil circulation cooling K 23 =0.95;

[0240] Family defect correction factor K 24 In the same series, if there is no problem with the equipment, K 24 =0.96, when the same series of equipment has a few defects that do not endanger operation 24 =1, when there is a potential risk of repeated failures in the same series of equipment K 24 =1.04;

[0241] Correction coefficient K of the number of failures in the past five years 25 When n=0, K 25 =0.96, when n=1, K 25 =1, when n∈[2,4] 25 =1.04, when n∈[5,10] 25 =1.2, when n>10, K 25 =1.4;

[0242] Near-zone short-circuit correction coefficient K 26 If the transformer has a short circuit in the vicinity, K 26 =1.04, otherwise K 26 =1;

[0243] Partial discharge correction factor K 27 In the case of partial discharge in the transformer, K 27 =1.2, when there is no partial discharge in the transformer, K 27 =1.

[0244] In this way, the system innovatively integrates 9 types of correction factors, such as operation years, core grounding current, and appearance grade, covering four dimensions of equipment operation time, electrical characteristics, mechanical structure, and historical operation and maintenance data. Compared with the traditional simplified model that only considers load rate and oil temperature, this system monitors the abnormality of core grounding current (K 12 ) and partial discharge detection (K 27 ) linkage analysis can simultaneously capture hidden electrical defects and obvious mechanical degradation, solving the evaluation blind spot problem caused by the single parameter dimension of the existing method. In addition, for the commissioning time (K 11 ) and the number of failures in the past five years (K 25 ) designs a nonlinear piecewise function to break through the limitations of traditional linear aging assumptions. When the equipment has been in operation for more than 30 years, the accelerated growth characteristics of K11 are more in line with the actual law of the rapid degradation rate of insulation materials in the later stage of aging. Compared with the fixed-year attenuation model, it can provide an early warning of the life inflection point. In addition, through the family defect correction factor (K 24 ) and cooling method correction factor (K 23 )'s differentiated design transforms traditional qualitative indicators, such as batch quality risk and cooling system efficiency, for the same model of equipment into calculable, quantitative parameters. Compared to empirical judgment methods, this solution systematically identifies the potential impact of cumulative efficiency degradation in forced oil circulation cooling systems on temperature rise, providing data support for cooling optimization.

[0245] In the casing reliability level correction (K 22) adopts the "short board effect" calculation strategy: when there is a high-risk casing, the sum of the three-phase coefficients is taken, otherwise the minimum value is taken. Compared with the average value calculation method, this strategy can amplify the warning effect of single-phase casing degradation on the overall health status, avoid the dilution of key defects by normal phase parameters, and significantly improve the sensitivity of identifying local serious defects. In addition, the construction of the commissioning period (K 11 ) and near-zone short-circuit records (K 26 ) collaborative analysis model, enabling the combined calculation of equipment aging patterns and damage from sudden impact events. Compared to a static baseline model, this model dynamically reflects the differential impact of long-term, gradual aging and transient overload shocks on insulation life, more accurately quantifying the cumulative damage to winding mechanical strength caused by historical short-circuit currents.

[0246] Comprehensive correction factor K com The calculation formula is:

[0247]

[0248] Through the product form A comprehensive correction factor is constructed to enable dynamic coupled analysis of parameters from different dimensions. Compared to linear combination methods such as weighted averages, the product relationship can amplify the shortcomings of abnormal correction factors (such as a sudden drop in casing reliability), more sensitively capture the chain reaction of local equipment defects on the overall health status, and prevent key risk indicators from being diluted by conventional parameters.

[0249] S7. If the final health index H reaches the preset threshold H t , it is determined that the equipment needs to be replaced; if it is not reached, the remaining life is predicted Based on the predicted remaining life T b1 and aging factor B to formulate corresponding maintenance plans to optimize the operation and maintenance strategy.

[0250] Compared to existing technologies, this method utilizes dynamic multi-physics coupling modeling. For bidirectional power flow switching conditions, this method integrates Monte Carlo random sampling and finite element simulation techniques to establish a dynamic electromagnetic-temperature field correlation model. This model accurately simulates the harmonic propagation, flux density mutations, and leakage field distortion caused by the integration of new energy sources, reveals the transient parameter variations within the transformer, and addresses the problem of magnetic saturation and loss calculation deviations caused by the prior art's neglect of power direction switching. Furthermore, a dynamic hotspot temperature calculation model is constructed by introducing harmonic components and flux density as key correction parameters. Compared to traditional methods that rely solely on load current, this model effectively captures the combined effects of harmonic injection and core hysteresis on temperature rise in power electronics equipment, significantly improving the accuracy of hotspot temperature predictions in frequent power reverse flow scenarios and providing more reliable input conditions for insulation aging assessment. Furthermore, this method overcomes the limitations of a single temperature aging model by integrating historical load fluctuations, ambient temperature and humidity, and transient thermal shock parameters to establish a dynamic life loss calculation system. Compared with the existing equivalent aging theory, this solution can quantitatively analyze the nonlinear attenuation of the polymerization degree of insulation paper caused by transient overload, achieve differentiated prediction of the remaining service life and expected operating years, and provide data support for differentiated operation and maintenance.

[0251] In addition, this method sets up a health diagnosis system that integrates multi-source data. Innovatively design a comprehensive health index model of chemical indicators such as oil chromatography and furfural test and electromagnetic parameters, combined with operating environment correction factors and fault repair records, to form a multi-dimensional health status assessment framework. Compared with the traditional single parameter threshold judgment, this method can identify potential faults such as partial discharge and insulation degradation earlier through heterogeneous data feature extraction and joint diagnosis, thereby reducing the risk of misjudgment. In addition, this method is based on the real-time health index and aging coefficient, and the reverse model and the predicted remaining life T b1 , then based on the predicted remaining life T b1 and aging factor B to develop a corresponding maintenance plan to optimize the operation and maintenance strategy. Compared with static periodic maintenance strategies, this solution can dynamically adjust the maintenance plan based on the actual load status of the transformer, optimizing the allocation of maintenance resources while ensuring equipment safety, and achieving a paradigm shift from "planned maintenance" to "predictive maintenance."

[0252] This method solves the problem of insufficient adaptability of traditional assessment systems in bidirectional power flow scenarios through dynamic modeling, multi-physics field coupling analysis and multi-source data fusion. It can accurately evaluate the operating status of transformers under conditions of frequent power flow switching, and provide the port area power grid with an intelligent operation and maintenance solution that takes into account both reliability and economy.

[0253] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the technical solutions. Those skilled in the art should understand that modifications or equivalent replacements of the technical solutions of the present invention that do not depart from the purpose and scope of the technical solutions of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for evaluating the carrying capacity of a transformer in a port area, characterized in that: The following steps are involved: S1. Construct a grid-connected simulation model for a doubly-fed asynchronous wind turbine generator and distributed photovoltaic power source. Combined with Monte Carlo random sampling technology and harmonic power flow calculation methods, analyze the voltage and current distribution characteristics and propagation patterns of harmonic components when a high proportion of renewable energy is connected to the Xiagang microgrid. Use finite element simulation technology to construct a transformer multi-physics field coupling model to simulate the dynamic distribution of the electromagnetic field and temperature field inside the transformer under conditions of frequent power flow switching, and analyze the changing patterns of winding leakage magnetic field, core magnetic flux density, and winding and core losses. S2. Based on the analysis results of S1, taking harmonics and magnetic flux density changes as the key factors affecting transformer losses, an improved model for calculating the hot spot temperature of the transformer when the power flow frequently switches is constructed. The model is used to calculate the hot spot temperature of the transformer at the corresponding time based on the transformer load data and ambient temperature data. The historical load data and ambient temperature data of the transformer are obtained, and the hot spot temperature of the transformer in each historical time period is calculated using the improved hot spot temperature calculation model. S3. Analyze the relationship between hotspot temperature and life loss, and build a hotspot temperature rise and life loss model for the transformer based on the analysis results. Calculate the corresponding insulation life loss according to the hotspot temperature of the transformer; Calculate the insulation life loss T according to the hotspot temperature of the transformer in each historical time period obtained in S2 through the hotspot temperature rise and life loss model. N ; Then calculate its estimated remaining service life T b And the expected actual total life T ins , T ins =T suv +T b , T b =T a -T N Where, T suv is the operation time of the transformer; T a The service life of the transformer is preset at the factory; S4. Calculate the transformer load level correction factor K L and operating environment correction factor K T , combined with S3 to obtain the expected actual total life of the transformer T ins , calculate the expected service life of the transformer T end ; Based on T end Calculate the transformer's aging factor B and primary health index H1, Among them, H0 is the initial health index, which is determined by the original information of the transformer; T a is the year when the health index assessment is conducted, and T0 is the year of initial commissioning; S5. Calculate the simplified test health index H of transformer oil 2a , Oil chromatography test health index H 2b , furfural test health index H 2c ; Combined with the primary health index H1 obtained by S4, calculate the comprehensive health index H of the transformer com , H com =f com max{H1,H 2a ,H 2b ,H 2c }; where f com is a preset constant; S6. Combine the operating environment information of the transformer to obtain the health correction factor based on the operating status data and the health correction factor based on the fault repair record; calculate the comprehensive correction factor K of the transformer based on the health correction factor and the health correction factor com ; Combined with S5 to obtain the comprehensive health index H com , get the final health index H of the transformer, H = K com H com ; S7. If the final health index H reaches the preset threshold H t , it is determined that the equipment needs to be replaced; if it is not reached, the remaining life is predicted Based on the predicted remaining life T b1 and aging factor B to formulate corresponding maintenance plans to optimize the operation and maintenance strategy.

2. The method for evaluating the load capacity of a transformer in a port area according to claim 1, wherein: In S2, the improved hot spot temperature calculation model is constructed as follows: Where, Indicates the transformer hot spot temperature, Indicates the real-time ambient temperature. Indicates the temperature rise of the top oil temperature over the ambient temperature. It is the temperature rise of the winding hot spot above the top oil temperature; Where, and They represent the initial and final temperature rises of the transformer winding hot spot relative to the top oil within a time interval t; τ h represents the winding time constant; and They represent the initial and final temperature rises of the top oil relative to the environment within a certain time interval t; τ ti It represents the top oil time constant, and the calculation formula is: Where, τ to,R represents the top oil time constant under rated load; n∈[0.8,1] represents the transformer top oil temperature rise calculation index, and when n=1, τ to =τ to,R ; is the rated temperature rise of the top oil relative to the environment; in, and satisfy: Where, P R is the resistance loss, P EC is the eddy current loss, P NL is the no-load loss, P LL is the load loss, and the one with R subscript is the rated value of the corresponding variable; It is the rated temperature rise of the winding hot spot relative to the top oil temperature.

3. The method for evaluating the load capacity of a transformer in a port area according to claim 2, wherein: The calculation formulas for each loss satisfy: P EC =F EC k EC P R ; P LL =P R +P EC +P OCL ,P OSL =F B F OSL k OSL P R ; Where, F R 、F EC 、F OSL are the harmonic loss calculation factors for resistance loss, eddy current loss, and stray loss respectively; I T is the effective value of the fundamental component phase current of the load current flowing through the transformer; R T is the transformer resistance parameter; k EC ∈ [0.05,0.15] is the eddy current loss coefficient; δ h , δ e are the hysteresis loss and eddy current loss coefficients in the transformer core respectively; f R The rated frequency of the transformer; B Tm is the maximum value of transformer flux density; P OSL is the stray loss; F B is the magnetic density change compensation coefficient; k OSL ∈[0.05,0.2] is the stray loss coefficient; in, Where, I1 represents the effective value of the fundamental component of the load current, I k Indicates the effective value of the kth harmonic of the load current, B T is the transformer flux density, B TR is the rated magnetic flux density of the transformer, m B Parameters obtained by electrical test fitting of the transformer are used to quantify the saturation degree and the rate of change of the magnetization characteristics of the transformer core material after exceeding the rated magnetic flux density.

4. The method for evaluating the load capacity of a transformer in a port area according to claim 3, wherein: In S3, the hotspot temperature rise and life loss model is: The transformer insulation life aging factor at rated load rate and reference temperature is defined as: Corresponding insulation life loss T N for: Where, Δt i is the i-th time interval; N is the total number of time intervals experienced; F ins,i is Δt i The corresponding insulation life loss is as follows.

5. The method for evaluating the load capacity of a transformer in a port area according to claim 4, wherein: S4 includes: S41. Obtain historical load data from the substation load monitoring system and calculate the average load rate β; Where, is the average load size during transformer operation; S N is the rated capacity; S42. Calculate the load level correction factor K based on the β obtained in step 1. L ; S43. Calculate the operating environment correction factor K based on the geographical environment where the transformer equipment is installed. t ; S44, comparison time T suv With T a and preset advance adjustment time T set The difference between the expected operating life T end ; S45, based on T end Calculate the aging coefficient B, and calculate the primary health index H1 in combination with the aging coefficient:

6. The method for evaluating the carrying capacity of a transformer in a port area according to claim 5, wherein: include: S51. Establish the acid value, breakdown voltage, water content, and dielectric loss parameters based on the simplified oil test results; determine the concentrations of hydrogen, methane, ethane, ethylene, and acetylene dissolved in the transformer oil based on the oil chromatography test; and determine the furfural content (FFA) based on the furfural test. S52. Obtain the test parameters acid value, breakdown voltage, micro-water content, and dielectric loss level value V based on the empirical formula 11 、V 12 、V 13 、V 14 , and test parameters of the concentration of hydrogen, methane, ethane, ethylene and acetylene dissolved in transformer oil 21 、V 22 、V 23 、V 24 、V 25 ; S53. Based on expert evaluation, obtain the index weights C of the test parameters acid value, breakdown voltage, trace water content, and dielectric loss 11 、C 12 、C 13 、C 14 , and the index weight C of the concentration of hydrogen, methane, ethane, ethylene and acetylene dissolved in the transformer oil 21 、C 22 、C 23 、C 24 、C 25 ; S54. Calculate the simplified oil test health index H based on the results of S51-S53. 2a , oil chromatography test health index H 2b , furfural test health index H 2c ; H 2c =2.33(FFA) 0.68 ; S55, combined with the primary health index H1 obtained in S4, calculate the comprehensive health index H of the transformer com , H com =f com max{H1,H 2a ,H 2b ,H 2c }; where f com is a preset constant.

7. The method for evaluating the load capacity of a transformer in a port area according to claim 6, wherein: In S52, the method for determining each level value is as follows:

8. The method for evaluating the load capacity of a port transformer according to claim 7, wherein: In S53, the weights of each indicator are: C 11 =0.2191、C 12 =0.2191、C 13 =0.2191、C 14 =0.3425、C 21 =0.1923、C 22 =0.1154、C 23 =0.1154、C 24 =0.1154、C 25 =0.4615。 9. The method for evaluating the load-carrying capacity of a port transformer according to claim 8, wherein: In S6, the health status correction factors based on the operating status data include: the operation time correction factor K 11 , Core grounding current correction factor K 12 ; Health status correction factors based on fault repair records include: Transformer appearance grade correction factor K 21 , casing reliability level correction factor K 22 , Cooling method correction coefficient K 23 , Family defect correction coefficient K 24 , Correction coefficient K of the number of failures in the past five years 25 , Near-zone short-circuit correction coefficient K 26 , partial discharge correction factor K 27 ; Comprehensive correction factor K com The calculation formula is:

10. The method for evaluating the load capacity of a transformer in a port area according to claim 9, wherein: Commissioning time correction factor K 11 Satisfy: when 0 ≤ T ≤ 5 years, K 11 = 1; when 5 < T ≤ 10 years, K 11 = 1.01; when 10 < T ≤ 20 years, K 11 = 1.02; when 20 < T ≤ 30 years, K 11 = 1.05; when T > 30 years, K 11 = 1.09; where T is the operation time of the equipment; Core grounding current correction factor K 12 Satisfies: When I=0, K 12 =1; when 0 <I≤0.1A,K 12 =1.05; when 0.1A <I≤0.3A,K 12 =1.1; when I>0.3A, K 12 =1.2; where I is the core grounding current; Transformer appearance grade correction factor K 21 The calculation formula is K 21 =0.9+0.1×L; where L is the highest rating for the four parts: the main body, cooling system, tap changer, and non-electrical components; the rating is from 1 to 5; Casing reliability level correction factor K 22 In the middle, the single casing is matched according to the level R: when R=1, K R =0.9, R=2 when K R =1, R=3, K R =1.1, R=4 when K R =1.2, R=5 when K R =1.4; the reliability levels of the three-phase bushings are K Ra , K Rb , K Rc , if max{K Ra ,K Rb ,K Rc }>1, then K 22 =K Ra +K Rb +K Rc Otherwise K 22 =min{K Ra ,K Rb ,K Rc }; Cooling method correction factor K 23 In the case of oil-immersed self-cooling or oil-immersed air-cooling, K 23 =1, when the cooling method is forced oil circulation cooling K 23 =0.96, when the cooling method is forced guide oil circulation cooling K 23 =0.95; Family defect correction factor K 24 In the same series of equipment, if there is no problem, K 24 =0.96, when the same series of equipment has a few defects that do not endanger operation 24 =1, when there is a potential risk of repeated failures in the same series of equipment K 24 =1.04; Correction coefficient K of the number of failures in the past five years 25 When n=0, K 25 =0.96, when n=1, K 25 =1, n∈[2,4] K 25 =1.04, when n∈[5,10] 25 =1.2, when n>10, K 25 =1.4; Near-zone short-circuit correction coefficient K 26 If the transformer has a short circuit in the vicinity, K 26 =1.04, otherwise K 26 =1; Partial discharge correction factor K 27 In the case of partial discharge in the transformer, K 27 =1.2, when there is no partial discharge in the transformer, K 27 =1.

Images