Method and system for judging moisture diffusion degree of transformer bushing

Abstract

The invention provides a method and system for discriminating the moisture diffusion degree of a transformer bushing, and the method comprises the steps: carrying out the frequency domain dielectric spectrum test of the transformer bushing, and obtaining the corresponding field data information; wherein the field data information at least comprises an FDS curve, field test temperature and oil pressure; inputting the field data information into a preset transformer oil-paper bushing moisture diffusion degree discrimination model to discriminate the moisture diffusion degree, and obtaining a final discrimination result; according to the method, the diffusion degree in the casing pipe is obtained through the finite element simulation technology, the damping mode is typical, space distribution is obvious, a large amount of collected data under different damping conditions can be generated, and judgment of the field collected data on the space distribution and the damping degree of moisture in the casing pipe is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of water diffusion degree determination, and in particular to a method and system for determining the water diffusion degree of a transformer bushing. Background Art

[0002] In the power system, the transformer is an important equipment for voltage conversion, power transmission and distribution, and its operating status is directly related to the safety and stability of the power grid. Oil-impregnated paper bushing is widely used as a transformer outlet device and is one of the key equipment of the ultra-high voltage transmission system. It is widely used in transformers of 110kV and above. However, due to unreasonable sealing structure and failure of sealing materials, oil-impregnated paper bushings often suffer from water ingress and moisture. Insulation moisture has become one of the direct causes of unplanned bushing shutdowns. It is crucial to study the moisture problem of oil-paper bushings. Regarding the diagnosis method of oil-paper bushing moisture, traditional detection methods have shortcomings, such as being affected by leakage current and temperature changes, and it is difficult to accurately reflect the insulation condition. Studies have shown that when the moisture content increases, the real part of the complex capacitance increases significantly in the low-frequency band, and the ratio of the real part of the complex capacitance is used to diagnose the degree of moisture. However, a single characteristic quantity is difficult to fully evaluate the comprehensive degradation state of the bushing, and multiple characteristic quantities are required for comprehensive diagnosis. Current research focuses on the overall moisture situation and ignores the uneven distribution of internal moisture, which may lead to accelerated degradation in local areas. Therefore, it is very important to determine the spatial distribution of moisture and the degree of moisture inside the oil-paper casing. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for determining the moisture diffusion degree of a transformer bushing, so as to solve the technical problem of how to accurately determine the moisture diffusion degree of the transformer bushing.

[0004] On the one hand, a method for determining the moisture diffusion degree of a transformer bushing is provided, comprising:

[0005] Performing a frequency domain dielectric spectrum test on the transformer bushing to obtain corresponding field data information; wherein the field data information at least includes an FDS curve, field test temperature, and oil pressure;

[0006] The field data information is input into a preset transformer oil-paper bushing moisture diffusion degree discrimination model to discriminate the moisture diffusion degree and obtain a final discrimination result.

[0007] Preferably, the transformer oil-paper bushing moisture diffusion degree discrimination model is preset by the following steps:

[0008] Through the preset multi-physics field simulation model of the transformer, the water diffusion simulation data in the bushing under different conditions and at different times are obtained; wherein the water diffusion simulation data at least includes the frequency domain dielectric spectrum curve under different bushing water distribution and the simulated temperature and oil pressure information;

[0009] Preprocess the moisture diffusion simulation data, establish the label relationship corresponding to the simulated moisture diffusion degree, and obtain the data set;

[0010] Based on the convolutional neural network, by mining the features of the data set and establishing a functional relationship between the data set and the moisture diffusion degree of the bushing, obtain the final discriminant model for the moisture diffusion degree of the transformer oil-paper bushing.

[0011] Preferably, the multi-physical field simulation model of the transformer is preset through the following steps

[0012] Establish the corresponding geometric model of the transformer oil-paper bushing according to the size, material properties and physical parameters of the oil-paper bushing; wherein, the geometric model includes at least an electric field, a thermal field, a flow field and a dilute matter transfer field;

[0013] Solve the relevant physical field equations through a preset simulation tool to calculate the moisture diffusion degree under the influence of the multi-physical field.

[0014] Preferably, it further includes solving the relevant physical field equations according to the following formula

[0015] Solve the electric field distribution inside the bushing according to the following formula

[0016]

[0017] where ε is the dielectric constant is the electric potential, and ρ is the charge density;

[0018] Calculate the temperature distribution inside the bushing according to the following formula

[0019]

[0020] where κ is the thermal conductivity, T is the temperature, and Q is the heat source;

[0021] Calculate the behavior of the insulating oil flow according to the following formula

[0022]

[0023] where ρ is the fluid density, v is the flow velocity, p is the pressure, μ is the viscosity, and F is the liquid pressure.

[0024] Preferably, it further includes determining the behavior of moisture in the solid insulation medium according to the following formula in the oil-impregnated paper

[0025]

[0026] where c1 is the moisture concentration in the oil-impregnated paper and D1 is the diffusion coefficient in the oil-impregnated paper is the Laplace operator, and t is the moment of behavior;

[0027] In insulating oil, the behavior of moisture in the liquid insulating medium is determined according to the following formula:

[0028]

[0029] where c2 is the moisture concentration in the insulating oil, v is the liquid flow rate, D2 is the diffusion coefficient in the insulating oil, is the Laplace operator, and t is the moment of behavior.

[0030] Preferably, the obtaining of the moisture diffusion simulation data in the bushing under different conditions and at different times includes,

[0031] Taking the complex capacitance as the analysis object, the real part characterizes the strength of dielectric polarization, and the imaginary part characterizes the losses caused by medium conductance and relaxation polarization;

[0032] Applying an alternating voltage with a preset frequency between the conductive rod and the end screen in the multi-physical field simulation model, recording the response current and the phase difference between the voltage and the current, and obtaining the real and imaginary parts of the complex capacitance;

[0033] By changing the frequency of the alternating voltage within a preset range, the frequency-domain dielectric spectrum under different bushing moisture distributions is obtained. The simulated temperature and oil pressure are used as multi-dimensional data, and the distribution of the moisture concentration obtained by simulating the response moisture diffusion is used as the dataset label.

[0034] Preferably, the preprocessing of the moisture diffusion simulation data and the establishment of the label relationship corresponding to the degree of simulated moisture diffusion to obtain the dataset include,

[0035] Normalizing the collected data and transforming the multi-dimensional data to a unified scale;

[0036] Calculating the moisture concentration at different positions in the bushing, using the moisture concentration as the label; and dividing the moisture concentration into multiple levels from low to high according to the set threshold;

[0037] According to the geometric structure of the bushing itself and the empirical moisture-affected distribution, the inside of the bushing is divided into multiple moisture-affected areas according to characteristic positions such as in the insulating oil, at the bottom of the core, the inner layer of the middle part of the core, the outer layer of the middle part of the core, and the top of the core;

[0038] Combining the processed multi-dimensional data of the oil-paper bushing and combining it with the moisture-affected discrimination label to form a dataset.

[0039] On the other hand, a discrimination system for the moisture diffusion degree of a transformer bushing is also provided to implement the discrimination method for the moisture diffusion degree of a transformer bushing, including:

[0040] A data acquisition module is used to perform frequency-domain dielectric spectroscopy tests on transformer bushings to obtain corresponding on-site data information. Among them, the on-site data information at least includes FDS curves, on-site test temperature, and oil pressure.

[0041] A discrimination module is used to input the on-site data information into a preset discrimination model for the moisture diffusion degree of transformer oil-paper bushings to discriminate the moisture diffusion degree and obtain a final discrimination result.

[0042] Preferably, it further includes a model generation module, which is used to obtain simulated moisture diffusion data in the bushing under different conditions and at different times through a preset multi-physics field simulation model of the transformer. Among them, the simulated moisture diffusion data at least includes frequency-domain dielectric spectroscopy curves under different bushing moisture distributions and simulated temperature and oil pressure information.

[0043] Preprocess the simulated moisture diffusion data, establish a label relationship corresponding to the simulated moisture diffusion degree, and obtain a data set.

[0044] Based on a convolutional neural network, by mining the features of the data set and establishing a functional relationship between the data set and the moisture diffusion degree of the bushing, a final discrimination model for the moisture diffusion degree of transformer oil-paper bushings is obtained.

[0045] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0046] The discrimination method and system for the moisture diffusion degree of transformer bushings provided by the present invention obtain the diffusion degree in the bushing through finite element simulation technology. The moisture absorption mode is typical and the spatial distribution is obvious, and a large number of acquisition data under different moisture absorption conditions can be generated, realizing the discrimination of the spatial distribution and moisture absorption degree of the internal moisture of the bushing based on on-site acquisition data. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, obtaining other drawings based on these drawings still belongs to the scope of the present invention.

[0048] Figure 1 It is a main flow schematic diagram of a discrimination method for the moisture diffusion degree of a transformer bushing in an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0049] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings.

[0050] AsFigure 1 As shown in the figure, it is a schematic diagram of an embodiment of a method for discriminating the moisture diffusion degree of a transformer bushing provided by the present invention. In this embodiment, the method includes the following steps:

[0051] Step S1: Perform frequency-domain dielectric spectroscopy testing on the transformer bushing to obtain the corresponding on-site data information. Among them, the on-site data information at least includes the FDS curve (Frenzel curve), on-site test temperature, and oil pressure. That is, perform frequency-domain dielectric spectroscopy testing on the on-site transformer oil-paper bushing to obtain multi-dimensional data information such as its FDS curve, on-site test temperature, and oil pressure, and discriminate the moisture diffusion degree.

[0052] Step S2: Input the on-site data information into a preset discrimination model for the moisture diffusion degree of the transformer oil-paper bushing to discriminate the moisture diffusion degree and obtain the final discrimination result. It can be understood that a convolutional neural network is constructed, and a functional relationship is established between the multi-dimensional collected data and the moisture diffusion degree of the bushing through feature mining, and a discrimination model for the moisture diffusion degree of the transformer oil-paper bushing is established; and the on-site data information is processed.

[0053] In one embodiment, the discrimination model for the moisture diffusion degree of the transformer oil-paper bushing is preset through the following steps. Through a preset multi-physical field simulation model of the transformer, obtain the moisture diffusion simulation data in the bushing under different conditions and at different times. Among them, the moisture diffusion simulation data at least includes the frequency-domain dielectric spectroscopy curve under different bushing moisture distributions and the simulated temperature and oil pressure information. Preprocess the moisture diffusion simulation data, establish a label relationship corresponding to the simulated moisture diffusion degree, and obtain a data set. Based on the convolutional neural network, through feature mining of the data set and establishing a functional relationship between the data set and the moisture diffusion degree of the bushing, obtain the final discrimination model for the moisture diffusion degree of the transformer oil-paper bushing. It can be understood that a multi-physical field simulation model of electricity, heat, force, and flow of the transformer oil-paper bushing is constructed through finite element analysis to obtain a large amount of moisture diffusion simulations in the bushing under different conditions and at different times; reuse the moisture diffusion simulation results to obtain the frequency-domain dielectric spectroscopy curve under different bushing moisture distributions and the simulated temperature and oil pressure information; preprocess the data obtained by the simulation, establish a label relationship with the corresponding simulated moisture diffusion degree, and construct a machine learning data set; construct a convolutional neural network, and establish a functional relationship between the multi-dimensional collected data and the moisture diffusion degree of the bushing through feature mining, and establish a discrimination model for the moisture diffusion degree of the transformer oil-paper bushing.

[0054] Specific embodiments, the multi-physical field simulation model of the transformer is preset through the following steps. A geometric model of the transformer oil-paper bushing is established according to the size, material properties and physical parameters of the oil-paper bushing. Among them, the geometric model includes at least an electric field, a thermal field, a flow field and a dilute matter transfer field. The relevant physical field equations are solved by a preset simulation tool to calculate the degree of moisture diffusion under the influence of the multi-physical field. Among them, the relevant physical field equations are solved according to the following formula,

[0055] The electric field distribution inside the bushing is solved according to the following formula:

[0056]

[0057] Among them, ε is the dielectric constant, is the electric potential, ρ is the charge density; that is, according to the Poisson equation, the electric field distribution inside the bushing is solved.

[0058] The temperature distribution inside the bushing is calculated according to the following formula:

[0059]

[0060] Among them, κ is the thermal conductivity, T is the temperature, Q is the heat source; that is, the heat conduction equation is used to calculate the temperature distribution inside the bushing.

[0061] The behavior of the insulating oil flow is calculated according to the following formula:

[0062]

[0063] Among them, ρ is the fluid density, v is the flow velocity, p is the pressure, μ is the viscosity, F is the liquid pressure; that is, the Navier-Stokes equation is used to calculate the behavior of the insulating oil flow.

[0064] In the oil-impregnated paper, the behavior of moisture in the solid insulating medium is determined according to the following formula:

[0065]

[0066] Among them, c1 is the moisture concentration in the oil-impregnated paper, D1 is the diffusion coefficient in the oil-impregnated paper, is the Laplace operator, t is the behavior time; that is, Fick's diffusion law is used to describe the behavior of moisture in the solid insulating medium.

[0067] In the insulating oil, the behavior of moisture in the liquid insulating medium is determined according to the following formula:

[0068]

[0069] Among them, c2 is the moisture concentration in the insulating oil, v is the liquid flow velocity, D2 is the diffusion coefficient in the insulating oil, is the Laplace operator, and t is the time of behavior. The convection-diffusion equation is used to describe the behavior of moisture in liquid insulating media.

[0070] In a specific embodiment, obtaining the simulation data of moisture diffusion in the bushing under different conditions and at different times includes taking the complex capacitance as the analysis object. The real part represents the strength of dielectric polarization, and the imaginary part represents the loss caused by dielectric conductance and relaxation polarization. An alternating voltage with a preset frequency is applied between the conductive rod and the end screen in the multi-physics field simulation model, and the response current and the phase difference between the voltage and the current are recorded to obtain the real and imaginary parts of the complex capacitance. By changing the frequency of the alternating voltage within a preset range, the frequency-domain dielectric spectrum under different moisture distributions in the bushing is obtained. The simulated temperature and oil pressure are used as multi-dimensional data, and the distribution of moisture concentration obtained by simulating moisture diffusion is used as the dataset label. It can be understood that in actual measurement, the complex capacitance is directly taken as the analysis object. The real part represents the strength of dielectric polarization, and the imaginary part represents the loss caused by dielectric conductance and relaxation polarization. An alternating voltage U with a frequency of ω is applied between the conductive rod and the end screen of the bushing simulation model, and the response current I and the phase difference between the voltage and the current are recorded. The real and imaginary parts of the complex capacitance are obtained.

[0071] Among them, the expression is as follows:

[0072]

[0073]

[0074] The expression of the tangent value of the dielectric loss angle is

[0075]

[0076] By changing the frequency between 10 -3 and 10 3 Hz, the frequency-domain dielectric spectrum under different moisture distributions in the bushing is plotted, and at the same time, the simulated temperature and oil pressure are recorded as multi-dimensional data. The distribution of moisture concentration obtained by simulating moisture diffusion is recorded as the training set label.

[0077] In a specific embodiment, preprocessing the moisture diffusion simulation data and establishing the label relationship corresponding to the degree of simulated moisture diffusion to obtain the dataset includes standardizing the collected data and transforming the multi-dimensional data to a unified scale; calculating the moisture concentration at different positions in the bushing and using the moisture concentration as the label; and dividing the moisture concentration into multiple levels from low to high according to a set threshold; according to the geometric structure of the bushing itself and the empirical moisture distribution, dividing the inside of the bushing into multiple moisture-affected areas according to characteristic positions such as in the insulating oil, at the bottom of the core, the inner layer of the middle part of the core, the outer layer of the middle part of the core, and the top of the core; combining the processed multi-dimensional data of the oil-paper bushing and combining it with the moisture-affected discrimination label to form a dataset.

[0078] Among them, the collected data is standardized by Z-Score to transform multi-dimensional data to a unified scale, avoiding the influence of scale differences between features on model training. The method is as follows:

[0079]

[0080] Among them, μ is the mean of the data in this dimension, and σ is the standard deviation of this dimension.

[0081] Calculate the moisture concentration at different positions in the casing, and use the moisture concentration as the label. According to the set threshold, the moisture concentration is divided into multiple levels from low to high. According to the self-geometric structure of the casing and the empirical moisture distribution, the inside of the casing is divided into multiple moisture-affected areas according to characteristic positions such as in insulating oil, at the bottom of the core, the inner layer of the middle of the core, the outer layer of the middle of the core, and the top of the core. The processed multi-dimensional data of the oil-paper bushing is combined in 1D and combined with the moisture-affected discrimination label to form a data set.

[0082] In one embodiment, the convolutional neural network is a 1D convolutional neural network to realize feature extraction of 1D data. The specific structure is as follows:

[0083] 1. Input layer: The input dimension is (n_features, 1), where n_features represents the number of multi-dimensional data features.

[0084] 2. Convolutional layer 1: The convolutional kernel size is 3, the number of output channels is 32, and the activation function is ReLU.

[0085] 3. Convolutional layer 2: The convolutional kernel size is 3, the number of output channels is 32, and the activation function is ReLU.

[0086] 4. Pooling layer 1: Max pooling, and the pooling window size is 2.

[0087] 5. Convolutional layer 3: The convolutional kernel size is 3, the number of output channels is 64, and the activation function is ReLU.

[0088] 6. Pooling layer 2: Max pooling, and the pooling window size is 2.

[0089] 7. Flatten layer: Flatten the multi-dimensional input into one dimension.

[0090] 8. Fully connected layer 1: The number of nodes is 128, and the activation function is ReLU.

[0091] 9. Output layer: The number of output categories is the number of moisture diffusion degree discrimination types, and the activation function is softmax.

[0092] Divide the multi-dimensional data data set into a training set and a test set, input it into the convolutional neural network for training, and obtain a discrimination model for the moisture diffusion degree of the transformer oil-paper bushing.

[0093] An embodiment of the present invention further provides a discrimination system for the moisture diffusion degree of a transformer bushing to implement the discrimination method for the moisture diffusion degree of the transformer bushing, including:

[0094] A data acquisition module, configured to perform a frequency-domain dielectric spectroscopy test on the transformer bushing to obtain corresponding on-site data information; wherein, the on-site data information at least includes an FDS curve, on-site test temperature, and oil pressure;

[0095] A discrimination module, configured to input the on-site data information into a preset discrimination model for the moisture diffusion degree of the transformer oil-paper bushing to discriminate the moisture diffusion degree and obtain a final discrimination result.

[0096] A specific embodiment further includes a model generation module, configured to obtain simulation data on moisture diffusion inside the bushing under different conditions and at different times through a preset multi-physical field simulation model of the transformer; wherein, the moisture diffusion simulation data at least includes frequency-domain dielectric spectroscopy curves under different bushing moisture distributions and simulated temperature and oil pressure information; preprocess the moisture diffusion simulation data, establish a label relationship corresponding to the simulated moisture diffusion degree, and obtain a data set; based on a convolutional neural network, mine the features of the data set and establish a functional relationship between the data set and the moisture diffusion degree of the bushing to obtain a final discrimination model for the moisture diffusion degree of the transformer oil-paper bushing.

[0097] It should be noted that the system described in the above embodiment corresponds to the method described in the above embodiment. Therefore, the parts not detailed in the system described in the above embodiment can be obtained by referring to the content of the method described in the above embodiment, and will not be elaborated here.

[0098] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0099] The discrimination method and system for the moisture diffusion degree of the transformer bushing provided by the present invention obtain the diffusion degree in the bushing through finite element simulation technology. The moisture absorption mode is typical and the spatial distribution is obvious. A large number of acquisition data under different moisture absorption conditions can be generated, realizing the discrimination of the spatial distribution and moisture absorption degree of the moisture inside the bushing based on on-site acquisition data.

[0100] The above-disclosed are only the preferred embodiments of the present invention. Of course, the scope of the rights of the present invention cannot be limited thereby. Therefore, equivalent changes made according to the claims of the present invention still fall within the scope covered by the present invention.

Claims

1. A method for determining the degree of moisture diffusion of a transformer bushing, characterized in that: include: Performing a frequency domain dielectric spectrum test on the transformer bushing to obtain corresponding field data information; wherein the field data information at least includes an FDS curve, field test temperature, and oil pressure; The field data information is input into a preset transformer oil-paper bushing moisture diffusion degree discrimination model to discriminate the moisture diffusion degree and obtain a final discrimination result.

2. The method according to claim 1, characterized in that The transformer oil-paper bushing moisture diffusion degree discrimination model is preset by the following steps: Through the preset multi-physics field simulation model of the transformer, the water diffusion simulation data in the bushing under different conditions and at different times are obtained; wherein the water diffusion simulation data at least includes the frequency domain dielectric spectrum curve under different bushing water distribution and the simulated temperature and oil pressure information; Preprocessing the water diffusion simulation data, establishing a label relationship corresponding to the simulated water diffusion degree, and obtaining a data set; Based on the convolutional neural network, the final transformer oil-paper bushing moisture diffusion degree discrimination model is obtained by mining the features of the data set and establishing a functional relationship between the data set and the bushing moisture diffusion degree.

3. The method according to claim 2, characterized in that The multi-physics simulation model of the transformer is preset by the following steps: Establishing a corresponding geometric model of the transformer oil-paper bushing according to the size, material properties and physical parameters of the oil-paper bushing; wherein the geometric model at least includes an electric field, a thermal field, a flow field and a dilute material transfer field; The relevant physical field equations are solved through the preset simulation tools to calculate the degree of moisture diffusion under the influence of multiple physical fields.

4. The method according to claim 3, characterized in that It also includes solving the relevant physical field equations according to the following formula, The electric field distribution inside the casing is solved according to the following formula: Where ε is the dielectric constant, is the electric potential, ρ is the charge density; The temperature distribution inside the casing is calculated according to the following formula: Among them, κ is thermal conductivity, T is temperature, and Q is heat source; The insulating oil flow behavior is calculated according to the following formula: Among them, ρ is the fluid density, v is the flow velocity, p is the pressure, μ is the viscosity, and F is the liquid pressure.

5. The method according to claim 4, characterized in that It also includes the determination of the behavior of water in solid insulating media in oil-impregnated paper according to the following formula: Where c1 is the water concentration in the oil-impregnated paper, D1 is the diffusion coefficient in the oil-impregnated paper, is the Laplace operator, t is the action time; In insulating oils, the behavior of water in the liquid insulating medium is determined by the following formula: Where c2 is the water concentration in the insulating oil, v is the liquid flow rate, D2 is the diffusion coefficient in the insulating oil, is the Laplace operator, and t is the action time.

6. The method according to claim 5, characterized in that The obtaining of the simulation data of water diffusion in the casing under different conditions and at different times includes: Taking complex capacitance as the analysis object, the real part represents the strength of dielectric polarization, and the imaginary part represents the loss caused by dielectric conductivity and relaxation polarization; An AC voltage of a preset frequency is applied between the conductive rod and the end screen in the multi-physics field simulation model, and the response current and the phase difference between the voltage and the current are recorded to obtain the real part and the imaginary part of the complex capacitance; By changing the frequency of the AC voltage within a preset range, the frequency domain dielectric spectra under different casing moisture distributions are obtained. The simulated temperature and oil pressure are used as multidimensional data, and the distribution of moisture concentration obtained by the response moisture diffusion simulation is used as the data set label.

7. The method according to claim 6, characterized in that The water diffusion simulation data is preprocessed to establish a label relationship corresponding to the simulated water diffusion degree, and the obtained data set includes: Standardizing the collected data and transforming the multidimensional data to a uniform scale; Calculate the moisture concentration at different locations in the casing and use the moisture concentration as a label; And according to the set threshold, the moisture concentration is divided into multiple levels from low to high; According to the geometric structure of the bushing itself and the empirical moisture distribution, the inside of the bushing is divided into multiple moisture-affected areas according to characteristic positions such as the insulating oil, the bottom of the core, the inner layer of the middle of the core, the outer layer of the middle of the core, and the top of the core; The processed multi-dimensional data of the oil-paper casing are combined and combined with the moisture discrimination label to form a data set.

8. A system for determining the degree of moisture diffusion of a transformer bushing, used to implement the method according to any one of claims 1 to 7, characterized in that: include: A data acquisition module is used to perform a frequency domain dielectric spectrum test on the transformer bushing to obtain the corresponding field data information; wherein the field data information at least includes an FDS curve, field test temperature, and oil pressure; The discrimination module is used to input the field data information into a preset transformer oil-paper bushing moisture diffusion degree discrimination model to discriminate the moisture diffusion degree and obtain a final discrimination result.

9. The system according to claim 8, characterized in that It also includes a model generation module, which is used to obtain the water diffusion simulation data in the bushing under different conditions and at different times through a preset multi-physical field simulation model of the transformer; wherein the water diffusion simulation data at least includes the frequency domain dielectric spectrum curve under different bushing water distribution and the simulated temperature and oil pressure information; Preprocessing the water diffusion simulation data, establishing a label relationship corresponding to the simulated water diffusion degree, and obtaining a data set; Based on the convolutional neural network, the final transformer oil-paper bushing moisture diffusion degree discrimination model is obtained by mining the features of the data set and establishing a functional relationship between the data set and the bushing moisture diffusion degree.

Images