Method and device for predicting residual life of transformer bushing based on digital twinning

Abstract

The invention provides a transformer bushing residual life prediction method and device based on digital twinning. The method comprises the step of performing prediction according to temperature field distribution corresponding to eddy current loss power. The step of calculating the eddy current loss power comprises the following steps of: obtaining guide rod current frequency, material resistivity and geometric structure parameters; calculating the cross sectional area and the direct-current resistance of the guide rod; determining alternating-current resistance by combining the frequency, the resistivity, the outer diameter and the inner diameter; dividing the AC resistance by the DC resistance to obtain a skin effect coefficient; acquiring the amplitude of current flowing through the guide rod; and calculating the eddy current loss power according to the skin effect coefficient, the square of the current amplitude and the direct current resistance. Through detailed calculation of the eddy current loss power and the temperature field distribution, the residual life of the sleeve is accurately predicted, and the perspectiveness and the accuracy of operation and maintenance decisions of a power system are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of transformer monitoring, and particularly relates to a transformer bushing residual life prediction method and device based on digital twinning. BACKGROUND

[0002] Oil-immersed transformers play a crucial role in power systems, enabling the conversion of different voltage levels to meet industrial and residential electricity demand. As a key component for support and insulation, the operating state of the high-voltage bushing directly affects the safety and stability of the power system.

[0003] However, due to the particularity of the bushing structure and the lack of internal monitoring means, it is currently difficult to effectively monitor the key parameters inside the bushing, resulting in an inability to accurately assess the insulation performance and predict the residual life of the bushing. Traditional high-voltage equipment maintenance methods are costly and have low prediction accuracy, failing to meet the requirements of modern power systems for forward-looking and accurate operation and maintenance decisions. In particular, material aging, internal oil loss, and moisture in the bushing during long-term operation can cause abnormal heating, thereby affecting the residual life of the bushing. SUMMARY

[0004] The purpose of the present application is to overcome the defects in the prior art, and to provide a transformer bushing residual life prediction method and device based on digital twinning.

[0005] The present application provides a transformer bushing residual life prediction method based on digital twinning, which predicts the residual life of the bushing according to the temperature field distribution corresponding to the eddy current loss power. The calculation of the eddy current loss power includes:

[0006] Obtaining the current frequency, material resistivity, and geometric structure parameters of the transformer bushing guide rod, the geometric structure parameters including the guide rod outer diameter, guide rod inner diameter, and guide rod length;

[0007] According to the guide rod outer diameter and guide rod inner diameter, the guide rod cross-sectional area is calculated;

[0008] According to the material resistivity, the guide rod length, and the cross-sectional area, the guide rod DC resistance is calculated; based on the current frequency, material resistivity, guide rod outer diameter, and guide rod inner diameter, the guide rod AC resistance is calculated; the AC resistance is divided by the DC resistance to obtain the skin effect coefficient;

[0009] Obtaining the current amplitude flowing through the transformer bushing guide rod;

[0010] According to the skin effect coefficient, the square of the current amplitude, and the DC resistance, the eddy current loss power of the transformer bushing guide rod is calculated.

[0011] Optionally, the current frequency, material resistivity and geometric parameters of the transformer bushing lead are obtained, the geometric parameters include an outer diameter of the lead, an inner diameter of the lead and a length of the lead, and the current frequency of the transformer bushing lead is obtained by:

[0012] The AC current waveform of the lead is measured in real time through a current transformer installed at an outgoing line of the transformer bushing, and a power frequency is extracted from the AC current waveform as the current frequency.

[0013] Optionally, the current frequency, material resistivity and geometric parameters of the transformer bushing lead are obtained, the geometric parameters include an outer diameter of the lead, an inner diameter of the lead and a length of the lead, and the geometric parameters of the transformer bushing lead are obtained by:

[0014] The design drawings and test reports of the transformer bushing are retrieved, the outer diameter of the lead and the inner diameter of the lead are obtained from the design drawings, and the length of the lead is obtained from the test reports.

[0015] Optionally, the AC resistance of the lead is calculated based on the current frequency, material resistivity, outer diameter of the lead and inner diameter of the lead, and the AC resistance of the lead is calculated by:

[0016] An electromagnetic field analytical model of the thin-walled conductor is established, and the AC resistance of the lead is calculated by solving the boundary value problem of Maxwell's equations in the hollow cylindrical coordinate system in the electromagnetic field analytical model.

[0017] Optionally, the method further comprises:

[0018] The eddy current loss power is input into a temperature field model, and the temperature field distribution of the transformer bushing is calculated based on the Joule heat loss of the lead and the dielectric loss of the insulating medium.

[0019] Optionally, the eddy current loss power is input into a temperature field model, and the temperature field distribution of the transformer bushing is calculated based on the Joule heat loss of the lead and the dielectric loss of the insulating medium, and the temperature field distribution of the transformer bushing is calculated by:

[0020] The flow state of the transformer oil is calculated by the Navier-Stokes equation, and the heat conduction equation and the convection heat transfer equation are coupled according to the flow state of the transformer oil to solve the steady-state temperature distribution function.

[0021] Optionally, the method further comprises:

[0022] The temperature field distribution corresponding to the eddy current loss power is input into a trained machine learning model, and the time for the transformer bushing to reach a preset failure threshold is predicted as the remaining life.

[0023] Optionally, the machine learning model comprises:

[0024] The training is performed by taking the simulation characteristic parameter matrix generated by the digital twin as input and taking the time for the bushing to reach the failure threshold as label to train the support vector machine model.

[0025] The application also provides a transformer bushing residual life prediction device based on digital twin, comprising:

[0026] An eddy current loss module is configured to predict the residual life of the bushing according to a temperature field distribution corresponding to the eddy current loss power, and to calculate the eddy current loss power, the eddy current loss module comprising:

[0027] An acquisition unit is configured to acquire the current frequency, material resistivity and geometric structure parameters of the transformer bushing guide rod, the geometric structure parameters including the guide rod outer diameter, guide rod inner diameter and guide rod length;

[0028] An area unit is configured to calculate the guide rod cross-sectional area according to the guide rod outer diameter and guide rod inner diameter;

[0029] A direct current unit is configured to calculate the guide rod direct current resistance according to the material resistivity, the guide rod length and the cross-sectional area;

[0030] An alternating current unit is configured to calculate the guide rod alternating current resistance based on the current frequency, material resistivity, guide rod outer diameter and guide rod inner diameter;

[0031] A coefficient unit is configured to divide the alternating current resistance by the direct current resistance to obtain the skin effect coefficient;

[0032] An amplitude unit is configured to acquire the current amplitude flowing through the transformer bushing guide rod;

[0033] A loss unit is configured to calculate the eddy current loss power of the transformer bushing guide rod according to the skin effect coefficient, the square of the current amplitude and the direct current resistance.

[0034] Optionally, the acquisition unit acquires the current frequency, material resistivity and geometric structure parameters of the transformer bushing guide rod, and the geometric structure parameters include the guide rod outer diameter, guide rod inner diameter and guide rod length, and the acquisition of the current frequency of the transformer bushing guide rod comprises:

[0035] The current frequency is extracted from an alternating current waveform of the guide rod measured in real time by a current transformer installed at an outgoing line of the transformer bushing.

[0036] Optionally, the acquisition unit acquires the current frequency, material resistivity and geometric structure parameters of the transformer bushing guide rod, and the geometric structure parameters include the guide rod outer diameter, guide rod inner diameter and guide rod length, and the acquisition of the geometric structure parameters of the transformer bushing guide rod comprises:

[0037] Obtaining a design drawing and a test report of the transformer bushing, obtaining an outer diameter of the guide rod and an inner diameter of the guide rod from the design drawing, and obtaining a length of the guide rod from the test report.

[0038] Optionally, the AC unit calculates the AC resistance of the guide rod based on the current frequency, material resistivity, the outer diameter of the guide rod, and the inner diameter of the guide rod, including:

[0039] An electromagnetic field analytical model of the thin-wall conductor is established, and the AC resistance of the guide rod is calculated by solving a boundary value problem of Maxwell equations in a hollow cylindrical coordinate system in the electromagnetic field analytical model.

[0040] Optionally, the method further includes:

[0041] The temperature unit inputs the eddy current loss power into a temperature field model, and the temperature field model calculates a temperature field distribution of the transformer bushing based on Joule heat loss and dielectric loss of the guide rod.

[0042] Optionally, the temperature unit inputs the eddy current loss power into a temperature field model, and the temperature field model calculates a temperature field distribution of the transformer bushing based on Joule heat loss and dielectric loss of the guide rod, including:

[0043] The transformer oil flow state is calculated by Navier-Stokes equation, and a steady-state temperature distribution function is solved by coupling a heat conduction equation and a convection heat transfer equation according to the transformer oil flow state.

[0044] Optionally, the method further includes:

[0045] The prediction unit inputs the temperature field distribution corresponding to the eddy current loss power into a trained machine learning model, and predicts a time when the transformer bushing reaches a preset failure threshold as a remaining life.

[0046] Optionally, the machine learning model includes:

[0047] The training module takes a simulation feature parameter matrix generated by the digital twin as input, takes a time when the bushing reaches a failure threshold as a label, and trains a support vector machine model.

[0048] The application has the following beneficial effects:

[0049] The application provides a transformer bushing residual life prediction method based on digital twinning, residual life of the bushing is predicted according to a temperature field distribution corresponding to eddy current loss power, and the eddy current loss power is calculated by the following steps: acquiring a current frequency, a material resistivity and a geometric structure parameter of a bushing guide rod of a transformer, the geometric structure parameter including an outer diameter of the guide rod, an inner diameter of the guide rod and a length of the guide rod; calculating a cross-sectional area of the guide rod according to the outer diameter of the guide rod and the inner diameter of the guide rod; calculating a direct current resistance of the guide rod according to the material resistivity, the length of the guide rod and the cross-sectional area; calculating an alternating current resistance of the guide rod based on the current frequency, the material resistivity, the outer diameter of the guide rod and the inner diameter of the guide rod; dividing the alternating current resistance by the direct current resistance to obtain a skin effect coefficient; and acquiring a current amplitude flowing through the bushing guide rod of the transformer; and calculating the eddy current loss power of the bushing guide rod of the transformer according to the skin effect coefficient, the square of the current amplitude and the direct current resistance. The application realizes accurate prediction of the residual life of the transformer bushing by detailed calculation of the eddy current loss power and the temperature field distribution, and improves the foresight and accuracy of the operation and maintenance decision of the power system. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is a process flow diagram for calculating the eddy current loss power in the application;

[0051] Figure 2 is a schematic diagram of residual life prediction of the transformer bushing in the application;

[0052] Figure 3 is a schematic diagram of a sensing module in the application;

[0053] Figure 4 is a correction schematic diagram of the digital twinning module in the application. DETAILED DESCRIPTION

[0054] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it is understood that various forms implement the present disclosure and should not be limited by the embodiments set forth herein. On the contrary, the embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0055] The application constructs a transformer bushing residual life prediction system based on digital twinning technology, including an entity sensing module, a digital twinning body and a life prediction module.

[0056] The entity sensing module collects temperature, current and voltage data in real time through temperature sensors, current transformers and voltage transformers installed at key parts of the bushing and at the outlet; the memory integrates data one month ago into a service period characteristic parameter matrix, and recent data into a sensing characteristic parameter matrix.

[0057] The digital twin is composed of a structure model, a parameter model, and a multi-physical field model: the structure model is constructed by CAD software according to the physical parameters of the bushing (terminal, oil pillow, capacitor core, flange, etc.) to build a two-dimensional model; the parameter model is based on test reports and maintenance reports to fit the curves of the electrical field parameters (dielectric constant ε, conductivity σ) of mineral oil, oil paper and other materials, the temperature field parameters (heat transfer coefficient h) and the flow field parameters (dynamic viscosity ξ, fluid density ρ) with temperature variation; the multi-physical field model realizes electromagnetic field-temperature field-fluid field coupling simulation through Maxwell equations, Navier-Stokes equations and heat transfer equations, and outputs a simulation characteristic parameter matrix.

[0058] A direct correlation rule between the number of bushing capacitor core layers and the furfural influence coefficient is established: the number of layers is changed by changing the oil channel structure, which forcibly corrects the curvature radius of the conductivity with temperature variation, so that the same temperature sensing data will generate differentiated conductivity fitting curves in different numbers of layers of the sleeve, thereby binding the electrical response characteristics of the material to the physical degradation structure depth of the bushing.

[0059] The digital twin correction process compares the difference between the simulation matrix and the service cycle matrix: if the relative error is ≤ a set threshold, the final simulation quantity characteristic matrix is output; if the error is > the threshold, the material parameters are corrected and the simulation is re-performed. The remaining life prediction module uses a support vector machine model to input the simulation characteristic parameter matrix (including temperature field and electric field distribution) to predict the time when the bushing reaches the failure threshold.

[0060] Please refer to Figure 2 Based on the above system, the application provides a transformer bushing residual life prediction method based on digital twin, which predicts the residual life of the bushing according to the temperature field distribution corresponding to the eddy current loss power.

[0061] Please refer to Figures 2 to 4 As shown, the calculation of the eddy current loss power includes:

[0062] S101, obtain the current frequency, material resistivity and geometric structure parameters of the transformer bushing guide rod, the geometric structure parameters including guide rod outer diameter, guide rod inner diameter and guide rod length;

[0063] The current frequency is obtained by real-time measurement of the guide rod alternating current waveform through the current transformer installed at the bushing outlet, and the power frequency f (unit: Hz) is extracted.

[0064] The geometric structure parameters (guide rod outer diameter r1, guide rod inner diameter r2, guide rod length l) are directly obtained by calling design drawings and test reports.

[0065] Wherein, r1 and r2 are derived from the size (unit: meter) marked on the design drawing, and l is derived from the axial length (unit: meter) of the guide rod recorded in the test report. The material resistivity p (unit: Ω·m) is obtained according to the resistivity data of the conductor in the material parameter library of the casing.

[0066] S102, calculating the cross-sectional area of the guide rod according to the outer diameter of the guide rod and the inner diameter of the guide rod;

[0067] The cross-sectional area s (unit: square meter) of the guide rod is calculated through the geometric relationship of the hollow cylinder:

[0068]

[0069] Wherein r1 is the outer diameter of the guide rod, and r2 is the inner diameter of the guide rod. The formula defines the effective current-carrying area of the conductor, which provides a geometric basis for the calculation of the direct current resistance.

[0070] S103, calculating the direct current resistance of the guide rod according to the material resistivity, the length of the guide rod and the cross-sectional area;

[0071] The direct current resistance R (unit: ohm) of the guide rod is calculated based on Ohm's law:

[0072]

[0073] Wherein p is the material resistivity, l is the length of the guide rod, and s is the cross-sectional area. This calculation represents the inherent resistance characteristics of the conductor under direct current conditions.

[0074] S104, calculating the alternating current resistance of the guide rod based on the current frequency, material resistivity, outer diameter of the guide rod and inner diameter of the guide rod;

[0075] An electromagnetic field analytical model of the thin-walled conductor is established to solve the boundary value problem of Maxwell's equations in the hollow cylindrical coordinate system.

[0076] The differential form of Maxwell's equations is:

[0077]

[0078] ▽·B=0

[0079] ▽·D=ρ

[0080] The constitutive equation is:

[0081] D=εE

[0082] B=μH

[0083] J=σE

[0084] Wherein, μ is the magnetic permeability (unit: H / m), E is the electric field intensity (unit: V / m), and p is the charge density (unit: C / m3 ), B is magnetic induction (unit: T), H is magnetic field intensity (unit: A / m), and ε is dielectric constant (unit: F / m).

[0085] In the calculation of the eddy current field, there is a formula:

[0086] J e = σE

[0087]

[0088] where ω is the angular frequency of the current.

[0089] In the eddy current field model, because σ » ωε (ω = 2πf is the angular frequency) under the power frequency condition, the displacement current can be ignored.

[0090] The magnetic vector potential A is introduced The eddy current field equation is obtained:

[0091]

[0092] The Coulomb gauge is applied which is simplified as:

[0093]

[0094] The boundary conditions are that the air domain outside the bushing B = 0 and A = 0; and the medium interface satisfies B 2n = B 1n and H 2t = H 1t . By solving the model, the alternating current resistance R ac (unit: ohm) is obtained.

[0095] S105, the alternating current resistance is divided by the direct current resistance to obtain a skin effect coefficient;

[0096] The skin effect coefficient k s is defined as:

[0097]

[0098] where R ac is the alternating current resistance, and R is the direct current resistance. The coefficient quantifies the degree of additional loss caused by the skin effect of high-frequency current.

[0099] S106, obtaining the current amplitude flowing through the transformer bushing guide rod;

[0100] The current waveform of the guide rod is measured in real time by a current transformer, and the peak value thereof is extracted as the current amplitude I (unit: ampere).

[0101] S107, calculate the eddy current loss power of the transformer bushing lead according to the skin effect coefficient, the square of the current amplitude and the direct current resistance.

[0102] The formula for calculating the eddy current loss power P (unit: watt) is:

[0103] P=k s I 2 R

[0104] wherein k s is the skin effect coefficient, I is the current amplitude, and R is the direct current resistance. This formula quantifies the heat generation power of the lead under high-frequency current, which is the core heat source input for temperature field simulation.

[0105] Further, the temperature field distribution is calculated:

[0106] The eddy current loss power P is input into the temperature field model, combined with the Joule heat loss of the lead and the dielectric loss (unit volume heat generation power Q), and the transformer oil flow state is solved by the Navier-Stokes equation:

[0107]

[0108] wherein v is the flow velocity (unit: m / s), F is the mass force (unit: N), p is the fluid pressure (unit: Pa), and ξ is the dynamic viscosity (unit: kg / (m·s)).

[0109] Coupling the heat conduction equation:

[0110]

[0111] The convection heat transfer equation is:

[0112] Q=hA(T w -T f )

[0113] wherein ρ is the density (unit: kg / m 3 ), C is the specific heat capacity (unit: J / (kg·K)), λ is the thermal conductivity (unit: W / (m·K)), h is the convective heat transfer coefficient (unit: W / (m 2 ·℃)), T w and T f are the wall and fluid temperatures (unit: ℃), respectively, and A is the surface area (unit: m 2 ).

[0114] The thermal radiation power is calculated as E=εC0(T / 100) 4 (C0=5.67W\cdotpm -2 \cdotpm -4). The final steady-state temperature distribution function is T = f(x, y, z).

[0115] It should be noted that in this application, when the temperature error is out of limit, the oil paper heat exchange coefficient correction is triggered, and the correction range is determined by the sleeve oil channel inclination characteristic; when the radial electric field error is out of limit, the dielectric constant is adjusted, and the correction step is strictly proportional to the aluminum foil spacing between the capacitor core. In particular, when the temperature and electric field error are out of limit at the same time, the fluid viscosity negative feedback adjustment mechanism controlled by the number of flange bolts forms a cooperative correction path across the physical field. This three-way binding decision rule of error type, correction parameter and structure characteristic fundamentally reconstructs the topological relationship of the traditional correction process.

[0116] Remaining life prediction:

[0117] The temperature field distribution is input into the trained machine learning model (support vector machine) to predict the time when the sleeve reaches the preset failure threshold. The training process includes: taking the simulation characteristic parameter matrix generated by the digital twin as input (including temperature field, electric field distribution), and taking the sleeve failure time as label; dividing the training set and test set and cross validating; evaluating the model by accuracy, recall rate and F1 value; after parameter tuning and optimization, the remaining life prediction result is output.

[0118] Further, the curvature extreme point of the axial electric field distribution is extracted, and its amplitude is multiplied by the flange length to convert into a dimensionless distortion factor. The dominant position of this distortion factor in the machine learning model: its weight proportion is dynamically calculated from the sleeve operating voltage value, and when the operating voltage exceeds the reference value, the distortion factor will squeeze the weight space of other features, so that the model decision logic completely focuses on the physical characterization of the layered defects of the capacitor core. This feature weight dynamic allocation mechanism based on structure parameters completely changes the input structure of the conventional machine learning model.

[0119] The application also provides a transformer sleeve remaining life prediction device based on digital twinning, comprising:

[0120] An eddy current loss module predicts the remaining life of the sleeve according to the temperature field distribution corresponding to the eddy current loss power, and calculates the eddy current loss power, the eddy current loss module comprising:

[0121] An acquisition unit acquires the current frequency, material resistivity and geometric structure parameters of the transformer sleeve guide rod, the geometric structure parameters including the guide rod outer diameter, guide rod inner diameter and guide rod length;

[0122] An area unit calculates the cross-sectional area of the guide rod according to the guide rod outer diameter and guide rod inner diameter;

[0123] A direct current unit calculates the direct current resistance of the guide rod according to the material resistivity, the guide rod length and the cross-sectional area;

[0124] The alternating current unit calculates the AC resistance of the lead based on the current frequency, material resistivity, lead outer diameter and lead inner diameter;

[0125] The coefficient unit obtains the skin effect coefficient by dividing the AC resistance by the DC resistance;

[0126] The amplitude unit obtains the current amplitude flowing through the transformer bushing lead;

[0127] The loss unit calculates the eddy current loss power of the transformer bushing lead according to the skin effect coefficient, the square of the current amplitude and the DC resistance.

[0128] Further, the obtaining unit obtains the current frequency, material resistivity and geometric structure parameters of the transformer bushing lead, the geometric structure parameters including the lead outer diameter, the lead inner diameter and the lead length, and the obtaining of the current frequency of the transformer bushing lead comprises:

[0129] The AC current waveform of the lead is measured in real time through the current transformer installed at the outgoing line of the transformer bushing, and the power frequency is extracted from the AC current waveform as the current frequency.

[0130] Further, the obtaining unit obtains the current frequency, material resistivity and geometric structure parameters of the transformer bushing lead, the geometric structure parameters including the lead outer diameter, the lead inner diameter and the lead length, and the obtaining of the geometric structure parameters of the transformer bushing lead comprises:

[0131] The design drawings and test reports of the transformer bushing are called, the lead outer diameter and the lead inner diameter are obtained from the design drawings, and the lead length is obtained from the test reports.

[0132] Further, the alternating current unit calculates the AC resistance of the lead based on the current frequency, material resistivity, lead outer diameter and lead inner diameter, comprising:

[0133] An electromagnetic field analytical model of the thin-walled conductor is established, and the AC resistance of the lead is calculated by solving the boundary value problem of Maxwell's equation in the hollow cylindrical coordinate system in the electromagnetic field analytical model.

[0134] Further, it further comprises:

[0135] The temperature unit inputs the eddy current loss power into a temperature field model, and the temperature field model calculates the temperature field distribution of the transformer bushing based on the Joule heat loss of the lead and the dielectric loss.

[0136] Further, the temperature unit inputs the eddy current loss power into a temperature field model, and the temperature field model calculates the temperature field distribution of the transformer bushing based on the Joule heat loss of the lead and the dielectric loss, comprising:

[0137] The transformer oil flow state is calculated by the Navier-Stokes equation, the heat conduction equation and the convection heat transfer equation are coupled according to the transformer oil flow state, and a steady-state temperature distribution function is solved.

[0138] Further, it also includes:

[0139] The prediction unit inputs the temperature field distribution corresponding to the eddy current loss power into the trained machine learning model, predicts the time when the transformer bushing reaches the preset failure threshold, and takes the residual life.

[0140] Further, the machine learning model includes:

[0141] The training module takes the simulation characteristic parameter matrix generated by the digital twin as input, takes the time when the bushing reaches the failure threshold as a label, and trains the support vector machine model.

[0142] The above description of the embodiments is to facilitate the understanding and application of the application by those skilled in the art. Those skilled in the art will obviously make various modifications to the above embodiments, and apply the general principles described herein to other embodiments without having to go through creative labor. Therefore, the present application is not limited to the above embodiments, and the improvements and modifications made by those skilled in the art to the present application according to the disclosure of the present application should be within the scope of protection of the present application.

Claims

1. A method for predicting the remaining life of transformer bushings based on digital twins, characterized in that, The remaining life of the bushing is predicted based on the temperature field distribution corresponding to the eddy current loss power. The calculation of the eddy current loss power includes: The current frequency, material resistivity, and geometric parameters of the transformer bushing guide rod are obtained, including the outer diameter, inner diameter, and length of the guide rod. Calculate the cross-sectional area of ​​the guide rod based on its outer diameter and inner diameter. Calculate the DC resistance of the guide rod based on the resistivity of the material, the length of the guide rod, and the cross-sectional area; Calculate the AC resistance of the guide rod based on the current frequency, material resistivity, outer diameter of the guide rod, and inner diameter of the guide rod. Divide the AC resistance by the DC resistance to obtain the skin effect coefficient; Obtain the current amplitude flowing through the bushing guide rod of the transformer; The eddy current loss power of the transformer bushing conductor is calculated based on the skin effect coefficient, the square of the current amplitude, and the DC resistance.

2. The method for predicting the remaining life of transformer bushings based on digital twins according to claim 1, characterized in that, Obtain the current frequency, material resistivity, and geometric parameters of the transformer bushing conductor, including: The AC current waveform of the transformer bushing conductor is measured in real time by a current transformer installed at the outlet of the transformer bushing, and the power frequency is extracted from the AC current waveform as the current frequency.

3. The method for predicting the remaining life of transformer bushings based on digital twins according to claim 1, characterized in that, Obtain the current frequency, material resistivity, and geometric parameters of the transformer bushing conductor, including: Retrieve the design drawings and test reports of the transformer bushing, obtain the outer diameter and inner diameter of the guide rod from the design drawings, and obtain the length of the guide rod from the test reports.

4. The method for predicting the remaining life of transformer bushings based on digital twins according to claim 1, characterized in that, Based on the current frequency, material resistivity, outer diameter of the guide rod, and inner diameter of the guide rod, calculate the AC resistance of the guide rod, including: An analytical electromagnetic field model of a thin-walled conductor is established. By solving the boundary value problem of Maxwell's equations in the hollow cylindrical coordinate system within the analytical electromagnetic field model, the AC resistance of the conductor rod is calculated.

5. The method for predicting the remaining life of transformer bushings based on digital twins according to claim 1, characterized in that, After calculating the eddy current loss power of the transformer bushing guide rod based on the skin effect coefficient, the square of the current amplitude, and the DC resistance, the method further includes: The eddy current loss power is input into the temperature field model, which calculates the temperature field distribution of the transformer bushing based on the Joule heat loss of the conductor and the insulation dielectric loss.

6. The method for predicting the remaining life of transformer bushings based on digital twins according to claim 5, characterized in that, The eddy current loss power is input into the temperature field model, which is based on the Joule heat loss of the conductor and the insulation dielectric loss, to calculate the temperature field distribution of the transformer bushing, including: The flow state of transformer oil is calculated using the Navier-Stokes equations. Based on the coupled heat conduction and convection heat transfer equations of the transformer oil flow state, the steady-state temperature distribution function is solved.

7. The method for predicting the remaining life of transformer bushings based on digital twins according to claim 5, characterized in that, The eddy current loss power is input into the temperature field model, which is based on the Joule heat loss of the conductor and the insulation dielectric loss. After calculating the temperature field distribution of the transformer bushing, the model further includes: The temperature field distribution corresponding to the eddy current loss power is input into a trained machine learning model to predict the time when the transformer bushing reaches a preset failure threshold, which is then used as the remaining lifespan.

8. The method for predicting the remaining life of transformer bushings based on digital twins according to claim 7, characterized in that, The machine learning model includes the following training steps: The simulation feature parameter matrix generated by the digital twin is used as input, and the time when the bushing reaches the failure threshold is used as the label to train the support vector machine model.

9. A transformer bushing remaining life prediction device based on digital twin, characterized in that, include: An eddy current loss module predicts the remaining life of the bushing based on the temperature field distribution corresponding to the eddy current loss power, and calculates the eddy current loss power. The eddy current loss module includes: The acquisition unit acquires the current frequency, material resistivity, and geometric parameters of the transformer bushing guide rod, including the guide rod outer diameter, guide rod inner diameter, and guide rod length. The area unit calculates the cross-sectional area of ​​the guide rod based on its outer diameter and inner diameter. The DC unit calculates the DC resistance of the guide rod based on the resistivity of the material, the length of the guide rod, and the cross-sectional area. The AC unit calculates the AC resistance of the guide rod based on the current frequency, material resistivity, guide rod outer diameter, and guide rod inner diameter; The coefficient unit is obtained by dividing the AC resistance by the DC resistance to obtain the skin effect coefficient. Amplitude unit, to obtain the current amplitude flowing through the transformer bushing guide rod; The loss unit calculates the eddy current loss power of the transformer bushing guide rod based on the skin effect coefficient, the square of the current amplitude, and the DC resistance.

10. A transformer bushing remaining life prediction device based on digital twin according to claim 9, characterized in that, The acquisition unit acquires the current frequency, material resistivity, and geometric parameters of the transformer bushing conductor, including: The AC current waveform of the transformer bushing conductor is measured in real time by a current transformer installed at the outlet of the transformer bushing, and the power frequency is extracted from the AC current waveform as the current frequency.

11. A transformer bushing remaining life prediction device based on digital twin according to claim 9, characterized in that, The acquisition unit acquires the current frequency, material resistivity, and geometric parameters of the transformer bushing conductor, including: Retrieve the design drawings and test reports of the transformer bushing, obtain the outer diameter and inner diameter of the guide rod from the design drawings, and obtain the length of the guide rod from the test reports.

12. The transformer bushing remaining life prediction device based on digital twin according to claim 9, characterized in that, The AC unit calculates the AC resistance of the guide rod based on the current frequency, material resistivity, outer diameter of the guide rod, and inner diameter of the guide rod, including: An analytical electromagnetic field model of a thin-walled conductor is established. By solving the boundary value problem of Maxwell's equations in the hollow cylindrical coordinate system within the analytical electromagnetic field model, the AC resistance of the conductor rod is calculated.

13. The transformer bushing remaining life prediction device based on digital twin according to claim 9, characterized in that, After calculating the eddy current loss power of the transformer bushing guide rod based on the skin effect coefficient, the square of the current amplitude, and the DC resistance, the loss unit further includes: The temperature unit inputs the eddy current loss power into the temperature field model, which calculates the temperature field distribution of the transformer bushing based on the Joule heat loss of the conductor and the insulation dielectric loss.

14. A transformer bushing remaining life prediction device based on digital twin according to claim 13, characterized in that, The temperature unit inputs the eddy current loss power into the temperature field model, which calculates the temperature field distribution of the transformer bushing based on the Joule heat loss of the conductor and the insulation dielectric loss, including: The flow state of transformer oil is calculated using the Navier-Stokes equations. Based on the coupled heat conduction and convection heat transfer equations of the transformer oil flow state, the steady-state temperature distribution function is solved.

15. A transformer bushing remaining life prediction device based on digital twin according to claim 13, characterized in that, The temperature unit inputs the eddy current loss power into the temperature field model. This temperature field model, based on the Joule heat loss of the conductor and the insulation dielectric loss, calculates the temperature field distribution of the transformer bushing and then further includes: The prediction unit inputs the temperature field distribution corresponding to the eddy current loss power into a trained machine learning model to predict the time when the transformer bushing reaches a preset failure threshold, which is taken as the remaining life.

16. A transformer bushing remaining life prediction device based on digital twin according to claim 15, characterized in that, The machine learning model includes the following training steps: The training module takes the simulation feature parameter matrix generated by the digital twin as input and the time when the bushing reaches the failure threshold as a label to train the support vector machine model.

Images