Transformer early fault diagnosis method based on difference between virtual and actual winding leakage magnetic field waveforms

Abstract

The present invention relates to the technical field of transformer fault detection. Disclosed is a transformer early fault diagnosis method based on the difference between virtual and actual winding leakage magnetic field waveforms, comprising: collecting leakage magnetic field information of a non-ideal winding structure transformer by means of fiber-optic magnetic field sensors; decomposing a winding of the non-ideal winding structure transformer into a plurality of regular windings; superposing leakage magnetic fields at the ends of a plurality of regular windings to obtain a non-ideal winding end leakage magnetic field; calculating a leakage flux matrix leakage inductance parameter and a leakage magnetic field output matrix parameter on the basis of the non-ideal winding end leakage magnetic field; establishing a circuit-leakage magnetic field multi-state digital model corresponding to the transformer; and using the difference between an actual transformer leakage magnetic field waveform and a digital transformer model virtual leakage magnetic field waveform as a fault characteristic quantity, completing threshold calculation and early fault sensitivity verification, forming a waveform difference-based early fault diagnosis criterion, and determining a fault position. The present invention can achieve quick and sensitive diagnosis of early faults of transformer windings and improves device operating reliability.

Description

Early Fault Diagnosis Method for Transformers Based on the Difference Between Virtual and Real Waveforms of Leakage Magnetic Field in Windings Technical Field

[0001] This invention relates to the field of transformer fault detection technology, and in particular to a method for early fault diagnosis of transformers that utilizes the difference between the virtual and real waveforms of the leakage magnetic field in the winding. Background Technology

[0002] Early transformer faults include winding deformation, arcing faults, and minor inter-turn faults. Because there is a considerable time gap between insulation failure and a serious transformer fault, and currently there is a lack of effective online monitoring methods for reliable identification of these early faults, early faults are difficult to detect.

[0003] Traditional transformer differential protection systems used for inter-turn short circuits in dry-type transformer windings cannot detect early-stage faults. Existing online transformer monitoring methods have poor reliability, frequently resulting in false alarms or missed detections of early faults. Current transformer modeling is based on ideal geometry, but in practical engineering applications, insulation requirements and other factors lead to uneven winding and gaps between coils, necessitating the establishment of a leakage flux matrix and leakage inductance parameter calculation model for non-ideal windings. Current technologies can only calculate the leakage magnetic field output matrix for transformers with regular windings, and the leakage magnetic field distribution at the upper and lower ends of non-ideal windings, where the leakage magnetic field sensor installation locations differ significantly from those of ideal windings.

[0004] Since air is a linear medium, the distribution of the spatial leakage magnetic field in the winding exhibits linear superposition. Therefore, it is necessary to identify early winding faults by monitoring changes in the spatial leakage magnetic field. Measurement points are fixed on the upper and lower surfaces of the high-voltage winding. Using the structural dimensions of each winding and the distances from the winding to the core and yoke, a linear leakage magnetic field output matrix is ​​calculated to transform the transformer's multi-winding current into a spatial leakage magnetic field. The current of each transformer winding is collected, and using this current as input, a virtual waveform of the radial leakage flux on the upper and lower surfaces of the high-voltage winding is calculated through the leakage magnetic field output matrix. Fiber optic leakage magnetic field sensors at the measurement points on the upper and lower surfaces of the high-voltage winding are used to measure the radial leakage flux in the winding space online. This measurement is then transmitted via fiber optic cable to an early fault diagnosis device based on the virtual waveform of the leakage magnetic field. A criterion for early winding faults based on the difference between the virtual calculated value and the measured value is proposed, and a threshold value for this criterion is calculated to verify the detection sensitivity of early faults.

[0005] Summary of the Invention

[0006] In view of the problems existing in the early fault diagnosis of transformers that utilize the difference between the virtual waveform and the measured waveform of the winding leakage magnetic field, this invention is proposed.

[0007] Therefore, the problem to be solved by this invention is that traditional transformer differential protection cannot reflect early faults, existing online transformer monitoring methods have poor reliability and often falsely report or miss early faults, and how to establish a calculation model for leakage flux matrix leakage inductance parameters of non-ideal windings.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0009] In a first aspect, embodiments of the present invention provide a method for early fault diagnosis of transformers utilizing the difference between the real and virtual waveforms of the leakage magnetic field in the windings. This method includes: acquiring leakage magnetic field information of a non-ideal structure winding transformer using an optical fiber magnetic field sensor; decomposing the windings of the non-ideal structure winding transformer into multiple regular windings according to the superposition theorem and the principle of magnetomotive force balance; obtaining the leakage magnetic field at the ends of the non-ideal windings by superimposing the leakage magnetic fields at the ends of the multiple regular windings; calculating leakage flux matrix leakage inductance parameters and leakage magnetic field output matrix parameters based on the leakage magnetic field at the ends of the non-ideal windings; establishing a circuit-leakage magnetic field multi-state digital model corresponding to the transformer based on the leakage flux matrix leakage inductance parameters and leakage magnetic field output matrix parameters; using the difference between the actual transformer leakage magnetic field waveform and the virtual leakage magnetic field waveform of the digital transformer model as a fault characteristic quantity, completing threshold value calculation and early fault sensitivity verification, forming an early fault diagnosis criterion based on waveform differences, and determining the fault location.

[0010] As a preferred embodiment of the transformer early fault diagnosis method utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field described in this invention, the step of collecting leakage magnetic field information of a non-ideal structure winding transformer by means of installing fiber optic magnetic field sensors on the upper and lower surfaces of the dry-type transformer winding and on the inner side of the yoke, transmitting the optical signal containing the magnetic field to the magnetic differential protection device through optical fiber transmission, converting it into an electrical signal through a photoelectric conversion module, and converting it into a digital signal through an A / D conversion.

[0011] As a preferred embodiment of the transformer early fault diagnosis method utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field described in this invention, when the non-ideal structure winding transformer is operating normally, the high and low voltage magnetomotive forces are balanced, and the leakage magnetic field is distributed linearly in the medium. Based on the superposition theorem, the leakage magnetic field is decomposed into longitudinal leakage magnetic field and transverse leakage magnetic field with ampere-turn balance. The Rockwell height equivalent method is used to convert magnetic field lines with different paths into magnetic field lines with the same calculated height and parallel to the height direction of the core window.

[0012] As a preferred embodiment of the transformer early fault diagnosis method utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field described in this invention, the formula for the circuit-leakage magnetic field multi-state digital model of the non-ideal winding transformer is as follows:

[0013] In the formula, A s Bs The state matrix and input matrix represent different operating states of the transformer, and are composed of the leakage flux inductance matrix and the main flux inductance matrix; C represents the leakage magnetic field output matrix, which is determined by the magnetic field model, and the specific formula is as follows:

[0014] In the formula, R 11 R 22 ∈R 3×3 These are diagonal matrices composed of the resistances of each phase winding on the high-voltage side and the resistances of each phase winding on the low-voltage side of the transformer.

[0015] R load L load ∈R 3×3 Composed of the low-voltage side load impedance, L 11 ,L 22 ,L 12 ,L 21 ∈R 3×3 It is a block matrix of the inductance matrix.

[0016] As a preferred embodiment of the transformer early fault diagnosis method utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field described in this invention, wherein: the leakage magnetic flux induction intensity inside the transformer can be calculated using a single Fourier algorithm, combined with A s B s By solving for the winding current, the relationship between the winding current and the leakage magnetic field can be obtained. x =Ci(t), the overall structure is as follows:

[0017] Among them, C 11 =[c A1 c A2 c A3 c A4 c A5 c A6 c A7 c A8 ] T C 14 =[c a1 c a2 c a3 c a4 c a5 c a6 c a7 c a8 ] T This represents the coefficient related to the high and low voltage winding currents of phase A in the output leakage flux formula for a total of 8 measuring points above and below phase A. The meanings of other matrix parameters are similar, and the output matrix parameters are only related to the transformer structure and the measuring point locations.

[0018] As a preferred embodiment of the transformer early fault diagnosis method utilizing the difference between the virtual and real waveforms of the leakage magnetic field in this invention, the calculation process of the leakage flux inductance matrix includes: selecting an irregular dry-type transformer model with unequal high and low voltage winding heights, preferentially full inner layer of the high voltage winding, non-closely wound outer layer, segmented inner layer with gaps, and gaps between multiple coils in the outer winding for analysis; placing leakage flux measuring points on the upper surface of the high voltage winding, with each measuring point consisting of 4 sensors; defining h w h is the height of the low-voltage winding, h0 is the distance between the winding end and the yoke, and h wp h ws h represents the height of the inner and outer coils of the high-voltage winding after removing the gaps. wp =h1+h2,h ws =h3+7h4, window height h=h w +2h0; Define δ as the distance between the low-voltage winding and the core, c and b as the thicknesses of the high-voltage and low-voltage windings respectively, a as the interval between the high-voltage and low-voltage windings, and Δ as the distance from the high-voltage winding to the neutral line between phase A and phase B. Then the pole pitch τ = a + b + c + δ + Δ.

[0019] As a preferred embodiment of the transformer early fault diagnosis method utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field described in this invention, wherein: state matrix A s and input matrix B s There exists an inductance matrix, which includes the main magnetic flux inductance matrix and the leakage magnetic flux inductance matrix. The inductance matrix is ​​obtained by adding the inductance matrices of the main magnetic flux and the leakage magnetic flux: L = L T +L σ .

[0020] Secondly, embodiments of the present invention provide an early fault diagnosis system for transformers that utilizes the difference between the virtual waveform and the measured waveform of the winding leakage magnetic field. The system includes: a magnetic field acquisition module for acquiring leakage magnetic field information of a transformer with a non-ideal winding structure via an optical fiber magnetic field sensor; a data processing module for decomposing the non-ideal winding into multiple regular windings according to the superposition theorem and the principle of magnetomotive force balance, calculating the leakage flux matrix leakage inductance parameters and leakage magnetic field output matrix parameters, and establishing a multi-state digital model of the transformer circuit and leakage magnetic field; a fault diagnosis module for comparing the actual transformer leakage magnetic field waveform with the virtual waveform of the digital model, using the waveform difference as a fault characteristic quantity, completing threshold value calculation and early fault sensitivity verification, and forming an early fault diagnosis criterion based on waveform differences; and a fault location module for determining the specific location of the early fault in the transformer winding according to the diagnosis criterion.

[0021] Thirdly, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any step of the above-described method for early fault diagnosis of transformers using the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0022] Fourthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the above-described method for early fault diagnosis of transformers using the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0023] The beneficial effects of this invention are as follows: (1) The windings of a multi-winding transformer with a non-ideal geometric structure are split, and the superposition method, single Fourier method, magnetic circuit method, leakage magnetic energy method and other methods are combined to update the solution of the leakage magnetic flux inductance matrix in the state space equation of the non-ideal transformer, and obtain the state space equation of the non-ideal transformer. At the same time, the radial leakage magnetic flux in the internal space of the non-ideal transformer is solved and used as the output quantity to establish a multi-state analytical model. (2) The parameters of the analytical model are continuously optimized by the Pearson correlation coefficient to ensure the consistency between the analytical model and the actual transformer. (3) When the transformer experiences winding deformation or a slight inter-turn fault, there is a difference between the measured value of the sensor and the virtual analytical value output by the digital twin model. As the degree of fault deepens, the measured value shows a regular increase (inter-turn fault) or decrease (winding deformation), but the difference will increase, which can effectively identify the fault and classify the fault type. (4) By utilizing the difference between the virtual calculated value and the measured value of the leakage magnetic field superimposed on the multi-winding space, in the event of a fault in a single or complex working condition, it is possible to accurately identify inter-turn short circuits of less than 1% and axial winding deformation of more than 5%. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 is an overall flowchart of the transformer early fault diagnosis method that utilizes the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0026] Figure 2 shows a non-ideal structure winding calculation model for an early fault diagnosis method of transformers that utilizes the difference between the virtual and real waveforms of the leakage magnetic field in the winding.

[0027] Figure 3 shows the longitudinal magnetic field and magnetomotive force distribution of the transformer early fault diagnosis method that utilizes the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0028] Figure 4 shows the distribution of the transverse magnetic field and magnetomotive force in the inner layer of the high-voltage winding in the early fault diagnosis method of transformer using the difference between the virtual and real waveforms of the leakage magnetic field in the winding.

[0029] Figure 5 shows the distribution of the transverse magnetic field and magnetomotive force of the outer layer of the high-voltage winding in the early fault diagnosis method of transformer using the difference between the virtual and real waveforms of the leakage magnetic field of the winding.

[0030] Figure 6 shows the transformer digital twin modeling and model consistency judgment diagram of the early fault diagnosis method of transformer using the difference between virtual and real waveforms of winding leakage magnetic field.

[0031] Figure 7 is a flowchart of the early fault identification method for transformers, which utilizes the difference between the virtual and real waveforms of the leakage magnetic field in the winding.

[0032] Figure 8 shows a comparison of simulated and calculated waveforms for a transformer early fault diagnosis method that utilizes the difference between virtual and real waveforms of the winding leakage magnetic field during normal transition to a fault outside the three-phase zone.

[0033] Figure 9 shows a comparison of the measured and calculated waveforms of the transformer early fault diagnosis method based on the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0034] Figure 10 is a comparison of the dynamic mode values ​​of different degrees of winding deformation and the normal output values ​​of the analytical model in the early fault diagnosis method of transformers that utilizes the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0035] Figure 11 shows the radial leakage flux amplitude generated at the upper end of the fault component under different degrees of deformation in the early fault diagnosis method of transformer using the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0036] Figure 12 is a comparison of the measured values ​​and the analytical model output values ​​of the transformer early fault diagnosis method based on the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0037] Figure 13 shows the radial leakage flux amplitude generated by the inter-turn fault component in the transformer early fault diagnosis method that utilizes the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0038] Figure 14 shows a comparison between the simulated values ​​of inter-turn short circuits of different degrees and the normal output values ​​of the analytical model of the transformer early fault diagnosis method that utilizes the difference between the virtual and real waveforms of the winding leakage magnetic field.

[0039] Figure 15 shows the variation of radial leakage flux amplitude generated by the fault component in the early fault diagnosis method of transformers that utilizes the difference between the virtual and real waveforms of the winding leakage magnetic field. Detailed Implementation

[0040] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0041] Example 1

[0042] Referring to Figures 1 and 2, the first embodiment of the present invention provides a method for early fault diagnosis of transformers utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field, including:

[0043] S1: Collect leakage magnetic field information of non-ideal structure winding transformers using fiber optic magnetic field sensors.

[0044] Specifically, the information on leakage magnetic field of non-ideal structure winding transformers is collected by fiber optic magnetic field sensors. This includes installing fiber optic magnetic field sensors on the upper and lower surfaces of the dry-type transformer windings and on the inner side of the yoke. The optical signal containing the magnetic field is transmitted through optical fiber to the magnetic differential protection device, converted into an electrical signal by a photoelectric conversion module, and then converted into a digital signal by an A / D converter.

[0045] S2: Based on the superposition theorem and the principle of magnetomotive force balance, the windings of the non-ideal structure winding transformer are decomposed into multiple regular windings.

[0046] Specifically, when a non-ideal structure winding transformer is operating normally, the high and low voltage magnetomotive forces are balanced, and the leakage magnetic field is distributed linearly in the medium. Based on the superposition theorem, the leakage magnetic field is decomposed into longitudinal and transverse leakage magnetic fields with ampere-turn balance. The Rockwell height equivalent method is used to convert magnetic field lines with different paths into magnetic field lines with the same calculated height and parallel to the height direction of the core window.

[0047] S3: Obtain the leakage magnetic field at the end of a non-ideal winding by superimposing the leakage magnetic fields at the ends of multiple regular windings.

[0048] S4: Calculate the leakage flux matrix leakage inductance parameters and leakage magnetic field output matrix parameters based on the leakage magnetic field at the end of the non-ideal winding.

[0049] S5: Based on the leakage inductance parameters of the flux matrix and the leakage magnetic field output matrix parameters, establish a multi-state digital model of the circuit-leakage magnetic field corresponding to the transformer.

[0050] S6: Using the difference between the actual transformer leakage field waveform and the virtual leakage field waveform of the digital transformer model as a fault characteristic quantity, the threshold value is calculated and the early fault sensitivity is verified, forming an early fault diagnosis criterion based on waveform differences to determine the fault location.

[0051] Furthermore, a multi-state analytical model of the circuit and leakage magnetic field of a non-ideal winding transformer is established, and an early fault protection scheme for the non-ideal transformer is constructed based on this model. The analytical formula of the model is shown below:

[0052] In the formula, A s B s The state matrix and input matrix represent different operating states of the transformer, and are composed of the leakage flux inductance matrix and the main flux inductance matrix; C represents the leakage magnetic field output matrix, which is determined by the magnetic field model, and the specific formula is as follows:

[0053] In the formula, R 11 R 22 ∈R 3×3 These are diagonal matrices composed of the resistances of each phase winding on the high-voltage side and the resistances of each phase winding on the low-voltage side of the transformer.

[0054] R load L load ∈R 3×3 Composed of the low-voltage side load impedance, L 11 ,L 22 ,L 12 ,L 21 ∈R 3×3 It is a block matrix of the inductance matrix.

[0055] State matrix A s and input matrix B s There exists an inductance matrix, which includes the main magnetic flux inductance matrix and the leakage magnetic flux inductance matrix. The inductance matrix is ​​obtained by adding the inductance matrices of the main magnetic flux and the leakage magnetic flux: L = L T +L σ .

[0056] The main magnetic flux inductance matrix can be obtained using the equivalent magnetic circuit model of a transformer, and the leakage magnetic flux inductance matrix can be obtained using the leakage magnetic energy method.

[0057] All winding currents of a certain phase are distributed in the winding space through linear transformation to generate leakage magnetic field. The leakage magnetic field distribution of a certain phase winding is obtained by superimposing the leakage magnetic fields of each winding current. In this way, the linear leakage magnetic field output transformation matrix of the winding leakage magnetic field generated by the superposition of each winding current can be established, such as C in formula (1).

[0058] The leakage magnetic flux density inside the transformer can be calculated using the single Fourier algorithm. Combined with the winding current obtained from solving the aforementioned state equations, the relationship between the winding current and the leakage magnetic field (B) can be derived. x =Ci(t), where C is the leakage magnetic field output matrix, and its overall structure is shown below:

[0059] Among them, C 11 =[c A1 c A2 c A3 c A4 c A5 c A6 c A7 c A8 ] T C 14 =[c a1 c a2 c a3 c a4 c a5 c a6 c a7 c a8 ] T This represents the coefficient related to the high and low voltage winding currents of phase A in the output leakage flux formula for a total of 8 measuring points above and below phase A. The meanings of other matrix parameters are similar, and the output matrix parameters are only related to the transformer structure and the measuring point locations.

[0060] Furthermore, the calculation process of the leakage flux inductance matrix is ​​as follows: An irregular dry-type transformer model is selected, characterized by unequal heights of the high and low voltage windings, preferentially full inner winding of the high voltage winding, non-closely wound outer winding, segmented inner winding with gaps, and gaps between multiple coils in the outer winding. Leakage flux measuring points are placed on the upper surface of the high voltage winding, with each measuring point consisting of four sensors, as shown in Figure 2. h is defined as... w h is the height of the low-voltage winding, h0 is the distance between the winding end and the yoke, and h wp h ws h represents the height of the inner and outer coils of the high-voltage winding after removing the gaps. wp =h1+h2,h ws =h3+7h4, window height h=h w +2h0; Define δ as the distance between the low-voltage winding and the core, c and b as the thicknesses of the high-voltage and low-voltage windings respectively, a as the interval between the high-voltage and low-voltage windings, and Δ as the distance from the high-voltage winding to the neutral line between phase A and phase B. Then the pole pitch τ = a + b + c + δ + Δ.

[0061] Furthermore, neglecting the winding excitation current, the high and low voltage magnetomotive forces of the non-ideal structure winding transformer are balanced during normal operation, and the leakage magnetic field distribution is linear. Therefore, the superposition theorem can be used to decompose the non-ideal structure winding into several regular, magnetomotive force-balanced windings, calculate the longitudinal and transverse leakage magnetic fields separately, and then superimpose them. Based on the superposition theorem, the leakage magnetic field is decomposed into longitudinal and transverse leakage magnetic fields with ampere-turn balance. Using the Rockwell height equivalence method, the magnetic field lines with different paths are converted into magnetic field lines with the same calculated height and parallel to the height direction of the core window.

[0062] To determine the leakage inductance of a non-ideal winding structure, the winding magnetomotive force is first decomposed, and then the spatial leakage magnetic energy is integrated to solve for the leakage inductance.

[0063] Using the superposition theorem, the magnetic field is decomposed into a longitudinal magnetic field, a transverse magnetic field in the inner layer of the high-voltage winding, and a transverse magnetic field in the outer layer of the high-voltage winding. The high and low voltage windings are in ampere-turn balance, and N1I1 = N2I2. The actual magnetomotive force in the upper inner layer of the high-voltage winding is... The actual magnetomotive force in the lower inner layer of the high-voltage winding is The actual magnetomotive force of the upper outer layer of the high-voltage winding is The actual magnetomotive force of each segment in the lower outer layer of the high-voltage winding is In Figures 3, 4, and 5, the high-voltage winding is extended to the same height as the low-voltage winding. Due to the significant difference in the distribution of the inner and outer layers of the high-voltage winding, the inner and outer layers are analyzed separately. The leakage magnetomotive force distribution can be obtained by superimposing the leakage magnetic fields of the windings, whether only a longitudinal leakage magnetic field is generated or only a transverse leakage magnetic field is generated.

[0064] The magnetomotive force is in balance, and the magnetomotive force of the inner layer of the high-voltage winding is The outer magnetomotive force is To ensure that the superimposed magnetomotive force matches the actual magnetomotive force of the inner layer of the high-voltage winding, the magnetomotive forces from top to bottom are as follows: h1F1、 (h wp -h1)F1、

[0065] in, The idea behind obtaining the magnetomotive force of the outer winding in Figure 5 is the same as that of the inner winding, and will not be listed again.

[0066] The solution for leakage magnetic energy is divided into inside the window and outside the window. Inside the window is the part of the winding wrapped by the iron yoke, and outside the window is the part in the air.

[0067] The longitudinal leakage magnetic energy of the high and low voltage windings inside the window is as follows:

[0068] In the formula h eq,l,IW The Lopez equivalent height within the window; N1, N2, N 1-in N 1-out The numbers represent the number of turns in the high and low voltage windings, and the number of turns in the inner and outer layers of the high voltage winding, respectively; μ0 is the air permeability.

[0069] Longitudinal leakage magnetic energy W of high and low voltage windings outside the window σA,l,OW W σa,l,OW The formula is the same as above, except that the Rockwell equivalent height is changed to h. eq,IW / ρ l That's it, ρ l This represents the longitudinal Lochte coefficient.

[0070] The transverse leakage magnetic energy inside and outside the inner window of the high-voltage winding is shown in formulas (6) and (7):

[0071] Among them, F i The x-axis is shown in the graph above.

[0072] The formulas for the transverse leakage magnetic energy of the outer high-voltage winding are shown in (8) and (9):

[0073] In summary, the leakage inductance of the high and low voltage windings of phase A can be calculated as shown in formulas (10) and (11):

[0074] Among them l IW l OW The equivalent magnetic path length is inside and outside the window.

[0075] Calculation of leakage flux transformation matrix parameters for non-ideal windings: Taking phase A as an example, the superposition method and single Fourier method are used for analysis, and the solution region is shown in Figure 3. Fourier decomposition of the winding cross-sectional current density is performed, and the current density in the region from x = 0 to x = τ is extended using the method of images to make it a periodic function. Then, the discrete current density J in the solution region is calculated. z (x) can then be expanded into a continuous Fourier series, at which point the computational domain can be divided into three regions, I, II, and III, as shown below:

[0076] At the interfaces between regions I and II and I and III, the tangential components of the vector magnetic potential and magnetic field strength are continuous. By solving the Poisson equation and Laplace equation that satisfy the boundary conditions, the vector magnetic potential of each region can be obtained. Finally, the radial leakage magnetic flux induction intensity of each region of the normal winding can be calculated as follows:

[0077] Similarly, the formulas for radial leakage magnetic flux induction intensity in each region of the inner and outer high-voltage windings can be solved. Taking the measurement point of phase A in Figure 2 as an example, the formulas for region I of the inner high-voltage winding and region III of the outer high-voltage winding can be obtained, as shown in equations (14) and (15):

[0078] In summary, the parameters of the leakage magnetic output matrix representing phase A can be obtained. Taking the measurement point 1 on phase A as an example, the parameter formulas are shown in equations (16) and (17):

[0079] Among them, Q A1 Q A2 Q A3 Q a The coefficients are related to the high and low voltage side currents of phase A; similarly, the parameters of phases B and C can be obtained, only the current density and location parameters are different.

[0080] Transformer early fault protection based on the difference between virtual calculated and measured values ​​of leakage magnetic field:

[0081] To verify the accuracy of digital twin modeling, a consistency evaluation criterion needs to be established to accurately reflect the differences between the multi-state analytical model and the physical entity. If the consistency requirements are met, the solved leakage flux is output; if not, the parameters are modified until the model parameters are optimal. The Pearson correlation coefficient is used to verify the correctness of the modeling theory mentioned above. Generally, a Pearson correlation coefficient between 0.8 and 1.0 is considered to indicate a strong correlation between the two sets of waveforms.

[0082] If the following formula is satisfied, the multi-state analytical model and the physical model are considered to be consistent. If not, the parameters are optimized. The overall modeling approach is shown in Figure 6.

[0083] 0.8≤r(φ 上1 ,φ 上2 )≤1 0.8≤r(φ 下1 ,φ 下2 )≤1 (18)

[0084] (1) Criteria for early faults in transformer windings

[0085] The difference between the measured value of the sensor and the virtual analytical value output by the digital twin model is defined as the fault signal Δφ. x The specific formula is as follows:

[0086] Δφ x =abs(φ xf -φ xn (19)

[0087] In the formula φ xf φ xn These are the measured values ​​of the dynamic model when the transformer has an internal fault and the radial leakage flux output by the analytical model when no internal fault has occurred, respectively, where abs is an absolute value function.

[0088] The amplitude of the fault signal is extracted using the Fast Fourier Transform (FFT) method, and the maximum fault signal amplitude Δφ is taken. xmax When the amplitude of the maximum fault signal is greater than the start-up threshold Δφ set When the protection mechanism is activated, the specific fault identification process is shown in Figure 7.

[0089] (2) Early fault protection setting

[0090] Solve the fault model of the transformer when winding deformation and inter-turn short circuit occur. When winding deformation occurs, the radial leakage magnetic flux induction intensity generated by the fault component is:

[0091] The radial leakage flux generated by the axial winding deformation fault component is ΔB x =ΔB x1 +ΔB x2 .

[0092] When an inter-turn short circuit occurs, the ampere-turns are no longer balanced, and a DC component exists compared to the winding deformation. The radial leakage flux induction intensity generated by this fault component is:

[0093] The radial leakage flux generated by the inter-turn short-circuit fault component is ΔB′. x =ΔB′ x1 +ΔB′ x2 .

[0094] The expressions for D1 and D2 are as follows:

[0095] In the above formula: J″ m内 、J″ m外 The harmonic component amplitude when the discrete current density is expanded into a continuous Fourier series within the solution region for winding deformation fault components; J″′ 0内 、J″′ 0外 、J″′ m内 、J″′ m外 The amplitudes of the DC and harmonic components when expanding the discrete current density into a continuous Fourier series within the solution region for inter-turn short-circuit fault components.

[0096] In minor early winding faults, minor winding deformation has a smaller fault signal than inter-turn short circuit. Therefore, the winding deformation condition is used to perform setting analysis on early internal faults of the transformer.

[0097] Information on the position parameters of each measuring point at the upper end and the harmonic component amplitude J″ when a 5% winding deformation occurs. m Substituting (5%) into formula (20), the radial leakage flux at the upper end is calculated and the maximum fault signal amplitude is extracted to obtain the basic tuning formula. Considering the influence of environmental factors, sensor installation position errors, etc., the reliability coefficient K is taken. rel =0.85, the formula is as follows:

[0098] Δφ set =K relmax{abs[FFT(Δφ x (J″ m (5%)]} (22)

[0099] In summary, the present invention has the following beneficial effects: (1) The windings of a multi-winding transformer with a non-ideal geometric structure are split, and the superposition method, single Fourier method, magnetic circuit method, leakage magnetic energy method and other methods are combined to update the solution of the leakage magnetic flux inductance matrix in the state space equation of the non-ideal transformer, and obtain the state space equation of the non-ideal transformer. At the same time, the radial leakage magnetic flux in the internal space of the non-ideal transformer is solved and used as the output quantity to establish a multi-state analytical model. (2) The parameters of the analytical model are continuously optimized by the Pearson correlation coefficient to ensure the consistency between the analytical model and the actual transformer. (3) When the transformer experiences winding deformation or a slight inter-turn fault, there is a difference between the measured value of the sensor and the virtual analytical value output by the digital twin model. As the degree of fault deepens, the measured value shows a regular increase (inter-turn fault) or decrease (winding deformation), but the difference will increase, which can effectively identify the fault and classify the fault type. (4) By utilizing the difference between the virtual calculated value and the measured value of the leakage magnetic field superimposed on the multi-winding space, in the event of a fault in a single or complex working condition, it is possible to accurately identify inter-turn short circuits of less than 1% and axial winding deformation of more than 5%.

[0100] Example 2

[0101] Referring to the figure, based on the first embodiment, this embodiment further provides a transformer early fault diagnosis system utilizing the difference between the virtual waveform and the measured waveform of the winding leakage magnetic field, including: a magnetic field acquisition module, used to acquire leakage magnetic field information of a non-ideal structure winding transformer through an optical fiber magnetic field sensor; a data processing module, used to decompose the non-ideal winding into multiple regular windings according to the superposition theorem and the magnetomotive force balance principle, calculate the leakage flux matrix leakage inductance parameters and leakage magnetic field output matrix parameters, and establish a transformer circuit-leakage magnetic field multi-state digital model; a fault diagnosis module, used to compare the actual transformer leakage magnetic field waveform with the virtual waveform of the digital model, use the waveform difference as a fault feature quantity, complete the threshold value calculation and early fault sensitivity verification, and form an early fault diagnosis criterion based on the waveform difference; and a fault location module, used to determine the specific location of the early fault in the transformer winding according to the diagnosis criterion.

[0102] This embodiment also provides a computer device applicable to the early fault diagnosis method of transformers utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the early fault diagnosis method of transformers utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field as proposed in the above embodiment.

[0103] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0104] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the transformer early fault diagnosis method based on the difference between the virtual and real waveforms of the winding leakage magnetic field, as proposed in the above embodiment.

[0105] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0106] Example 3

[0107] Referring to Figures 8 to 15, this is the third embodiment of the present invention. Based on the first two embodiments, this embodiment provides a method for early fault diagnosis of transformers by utilizing the difference between the virtual and real waveforms of the winding leakage magnetic field. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation and dynamic model verification.

[0108] Specifically, the wiring parameters for the dynamic model test are as follows: Infinite power supply voltage E = 1kV, ignoring system internal resistance. Line parameters: Line length l = 200km, total line resistance 2.54Ω, total inductive reactance 34.94Ω, total capacitive reactance 6.72μF. Transformer parameters: Rated capacity S = 50kVA, rated voltage ratio U1 / U2 = 1 / 0.4kV, where U1 and U2 are the high and low voltage sides of the dynamic model test transformer, respectively. Rated frequency is 50Hz, and the winding parameters referred to the high voltage side are R... A =0.5899Ω, R B = 0.5896Ω, R C =0.5903Ω, X A =X B =X C= 1.52Ω. The impedance of the load parameter is Z. L =3.199Ω.

[0109] The dynamic model transformer has certain limitations in terms of operating condition settings. Therefore, a simulation model was established in Ansys software at a 1:1 scale with the actual dry-type transformer to supplement the experimental data.

[0110] First, the consistency of the digital twin model is verified.

[0111] 1. Consistency Verification of ANSYS Finite Element Simulation Model. Figure 8 shows a comparison between the simulated values ​​of the simulation model and the solution values ​​of the digital twin model under the three-phase short-circuit fault condition outside the normal operating zone of the transformer. As can be seen from Figure 8, the digital twin model still maintains a high degree of consistency with the simulation model when the transformer's operating state changes.

[0112] To further verify the consistency between the simulation model and the digital twin model, the results of both were verified under normal operation and no-load closing conditions. The verification results are shown in Table 1.

[0113] Table 1 Consistency Verification between Analytical Model and Simulation Model

[0114] As shown in the table above, the correlation coefficients between the analytical model and the simulation model under various working conditions are all within the allowable range, indicating that the two are consistent.

[0115] 2. Verification of consistency between dynamic model and experimental model. Figure 9 shows the comparison between the magnetic flux values ​​obtained from the measurement data of the upper and lower measuring points of phase A of the dynamic model test transformer under no-load closing and the output data of the digital twin circuit-leakage magnetic field multi-state analytical model. Due to the limited conditions of the dynamic model experiment, in order to protect the test transformer, the correlation coefficient analysis between the dynamic model data and the predicted values ​​of the analytical model is only performed on the normal operation and no-load closing conditions, as shown in Table 2.

[0116] Table 2 Consistency Verification between Analytical Model and Dynamic Model Physical Entity

[0117] As shown in Table 2, under normal operation and no-load closing conditions, the correlation coefficients between the predicted values ​​obtained by the digital twin analytical model and the measured values ​​of the dynamic model entity are all within the range of [0.8, 1], indicating that the digital twin analytical model and the physical transformer are also consistent.

[0118] Secondly, the early fault protection criteria for transformers were verified.

[0119] The protection setting is adjusted using formula (22). Substituting the relevant parameters, the basic setting value for phase A is calculated to be 328.2mT*mm^2. Considering a reliability coefficient of 0.85, the fault protection threshold is calculated as follows: Δφ set =278.97mT·mm 2 .

[0120] Figures 10 and 11 show the changes in leakage flux calculated from the leakage flux induction intensity measured at the upper end measuring point during normal operation, as calculated by the digital twin multi-state analytical model, and the changes in leakage flux obtained by the fault signal curve verification diagram when different degrees of axial compression winding deformation faults occur at the upper end of the high-voltage side winding of phase A of the moving model transformer. As can be seen from Figures 10 and 11, when axial compression deformation of 5% or more occurs at the end of the winding, the fault signal amplitude is greater than the set threshold, and the protection will activate.

[0121] Figures 12 and 13 show a comparison of the radial leakage flux of the transformer during the dynamic model test when an inter-turn short circuit occurs simultaneously with the solution value of the normal no-load closing operation output by the digital twin multi-state analytical model, as well as the change in the fault signal amplitude. Figures 12 and 13 show that the transformer experiences an inter-turn short circuit fault during the first 74ms of no-load closing operation. After 74ms, the fault is cleared, and the transformer's operating state transitions to normal no-load closing operation.

[0122] Figures 14 and 15 show the changes in leakage flux calculated from the leakage flux induction intensity measured at each measuring point on the upper end during normal operation, as calculated by the digital twin multi-state analytical model, and the changes in leakage flux obtained by the simulation model when different degrees of inter-turn short circuits occur in the high-voltage side winding of phase A. The figures also include fault signal curve verification diagrams. As shown in Figures 14 and 15, when an inter-turn short circuit of 0.5% or more occurs in the winding, the fault signal amplitude is greater than the set threshold, triggering the protection mechanism.

[0123] Combining Figures 10 and 11, and Figures 14 and 15, it can be seen that, based on the solution value of the multi-state analytical model, the measured value of the dynamic model is smaller than the solution value when the winding is deformed, but larger than the solution value when there is an inter-turn short circuit. This pattern can effectively identify the fault type. Simulation and dynamic model tests verify the accuracy of the model established in this paper and demonstrate the feasibility and universality of the protection criteria.

[0124] Then, simulation was used to verify the protection scheme. Due to the limited conditions of the dynamic model test, the simulation model established above, which is consistent with the dynamic model test transformer, was used to obtain early fault data of different windings for fault identification. Table 3 lists the Δφ at the upper and lower ends of phase A when an early winding fault occurs under different operating conditions. xmaxConditions: Under winding deformation conditions, A3% represents a 3% axial compression deformation at the upper end of the A-phase high-voltage winding; under inter-turn short circuit conditions, A+A0.5% outside the zone represents a single-phase ground fault in the A-phase line outside the zone, accompanied by a one-turn inter-turn short circuit in the A-phase high-voltage winding of the transformer. No-load closing +A0.5% represents a no-load closing condition, accompanied by a one-turn inter-turn short circuit fault in the A-phase high-voltage winding.

[0125] Table 3 shows the maximum fault signal amplitude for different internal faults.

[0126] Analyzing the data in Table 3, we can draw the following conclusions:

[0127] 1) In this invention, under different fault conditions, Δφ x max Both increase with the severity of the fault, therefore Δφ x max It has a good fault indication function.

[0128] 2) Although it can still operate when the winding is deformed by 4%, considering the interference of various noises in actual operation, and to improve sensitivity, it is believed that the present invention can effectively identify winding deformation of 5% and above.

[0129] 3) This invention can effectively identify inter-turn short circuits in extreme scenarios such as inter-turn short circuits of 0.5% or higher, and inter-turn short circuits occurring simultaneously with grounding faults on external lines. Compared to winding deformation, it has higher sensitivity for inter-turn short circuits under various operating conditions.

[0130] 4) This invention is not affected by inrush current.

[0131] Finally, dynamic model experiments were conducted to verify the protection scheme. To highlight the superiority of the method proposed in this invention, five operating states were selected: no-load closing, 5% and 10% winding deformation, and 1.5% and 2.5% inter-turn short circuit under no-load closing. The present invention and traditional longitudinal differential protection for transformers were compared and analyzed under these operating states. As the main protection for the transformer, the longitudinal differential protection needs to avoid the unbalanced current flowing through the differential circuit and must consider the influence of factors such as reliability coefficient, type coefficient, non-periodic component coefficient, and error. Therefore, the operating threshold of the differential protection was set to I. set =8.26A, and the comparison results of the two protection methods are shown in Table 4.

[0132] Table 4. Maximum fault signal amplitude under different operating conditions in dynamic model testing.

[0133] As shown in Table 4, the proposed solution does not operate when the transformer does not experience an internal fault; the differential protection does not operate under any operating conditions; during a fault, Δφ x maxAll exceed the threshold, and the leakage flux characteristics change significantly. It can accurately identify winding deformation of more than 5% and inter-turn short circuit of 1.5% occurring under no-load closing conditions. Its sensitivity is higher than that of differential protection.

[0134] In summary, this invention decomposes the windings of a multi-winding transformer with a non-ideal geometry and combines methods such as the superposition method, single Fourier method, magnetic circuit method, and leakage magnetic energy method to update the solution of the leakage flux inductance matrix in the state-space equation of the non-ideal transformer, thus obtaining the state-space equation of the non-ideal transformer. Simultaneously, it solves for the radial leakage flux within the internal space of the non-ideal transformer and uses it as the output quantity to establish a multi-state analytical model. This invention continuously optimizes the parameters of the analytical model through the Pearson correlation coefficient, ensuring consistency between the analytical model and the actual transformer. When the transformer experiences winding deformation or a minor inter-turn fault, there is a difference between the measured values ​​of the sensors and the virtual analytical values ​​output by the digital twin model. As the severity of the fault increases, the measured values ​​show a regular increase (inter-turn fault) or decrease (winding deformation), but the difference always increases, enabling effective fault identification and classification. Furthermore, by utilizing the difference between the virtual calculated value and the measured value of the leakage magnetic field superimposed on the multi-winding space, this invention can accurately identify inter-turn short circuits of less than 1% and axial winding deformations of more than 5% in the event of a fault under single or complex operating conditions.

[0135] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for early fault diagnosis of transformers based on the difference between the virtual and real waveforms of the leakage magnetic field in the winding, characterized in that: Includes the following steps: Leakage magnetic field information of non-ideal structure winding transformers is collected using fiber optic magnetic field sensors; Based on the superposition theorem and the principle of magnetomotive force balance, the windings of the non-ideal structure winding transformer are decomposed into multiple regular windings. The leakage magnetic field at the end of a non-ideal winding is obtained by superimposing the leakage magnetic fields at the ends of multiple regular windings; Calculate the leakage flux matrix leakage inductance parameters and leakage magnetic field output matrix parameters based on the leakage magnetic field at the end of the non-ideal winding. Based on the leakage inductance parameters of the flux matrix and the leakage magnetic field output matrix parameters, a multi-state digital model of the circuit-leakage magnetic field corresponding to the transformer is established. Using the difference between the actual transformer leakage field waveform and the virtual leakage field waveform of the digital transformer model as a fault characteristic quantity, threshold value calculation and early fault sensitivity verification are completed, forming an early fault diagnosis criterion based on waveform differences to determine the fault location.

2. The transformer early fault diagnosis method based on the difference between virtual and real waveforms of winding leakage magnetic field as described in claim 1, characterized in that: The method of collecting leakage magnetic field information of non-ideal structure winding transformers by means of fiber optic magnetic field sensors includes installing fiber optic magnetic field sensors on the upper and lower surfaces of the dry transformer windings and on the inner side of the yoke, transmitting the optical signal containing the magnetic field through optical fiber to the magnetic differential protection device, converting it into an electrical signal through a photoelectric conversion module, and then converting it into a digital signal through an A / D converter.

3. The method for early fault diagnosis of transformers based on the difference between virtual and real waveforms of winding leakage magnetic field as described in claim 2, characterized in that: When a non-ideal structure winding transformer is operating normally, the high and low voltage magnetomotive forces are balanced, and the leakage magnetic field is distributed linearly in the medium. Based on the superposition theorem, the leakage magnetic field is decomposed into longitudinal and transverse leakage magnetic fields with ampere-turn balance. Using the Rockwell height equivalent method, magnetic field lines with different paths are converted into magnetic field lines with the same calculated height and parallel to the height direction of the core window.

4. The transformer early fault diagnosis method based on the difference between the virtual and real waveforms of the winding leakage magnetic field as described in claim 3, characterized in that: The formula for the circuit-leakage magnetic field multi-state digital model of the non-ideal winding transformer is as follows: In the formula, A s B s The state matrix and input matrix represent different operating states of the transformer, and are composed of the leakage flux inductance matrix and the main flux inductance matrix; C represents the leakage magnetic field output matrix, which is determined by the magnetic field model, and the specific formula is as follows: In the formula, R 11 R 22 ∈R 3×3 These are diagonal matrices composed of the resistances of each phase winding on the high-voltage side and the resistances of each phase winding on the low-voltage side of the transformer; R load L load ∈R 3×3 Composed of the low-voltage side load impedance, L 11 ,L 22 ,L 12 ,L 21 ∈R 3×3 It is a block matrix of the inductance matrix.

5. The transformer early fault diagnosis method based on the difference between the virtual and real waveforms of the winding leakage magnetic field as described in claim 4, characterized in that: The leakage flux induction intensity inside the transformer is calculated using the single Fourier algorithm, combined with A s B s Solving for the obtained winding current, we can obtain the relationship B between the winding current and the leakage magnetic field. x =Ci(t), the overall structure is as follows: Among them, C 11 =[c A1 c A2 c A3 c A4 c A5 c A6 c A7 c A8 ] T C 14 =[c a1 c a2 c a3 c a4 c a5 c a6 c a7 c a8 ] T This represents the coefficient related to the high and low voltage winding currents of phase A in the output leakage flux formula for a total of 8 measuring points above and below phase A.

6. The method for early fault diagnosis of transformers using the difference between virtual and real waveforms of winding leakage magnetic field as described in claim 5, characterized in that: The calculation process of the leakage flux inductance matrix includes, An irregular dry-type transformer model is selected for analysis, characterized by unequal heights of high and low voltage windings, preferential full winding of the inner layer of the high voltage winding and non-close winding of the outer layer, segmented inner layer with gaps, and gaps between multiple coils of the outer winding. The leakage flux measurement point is placed on the upper surface of the high voltage winding, and each measurement point consists of 4 sensors; Define h w h is the height of the low-voltage winding, h0 is the distance between the winding end and the yoke, and h wp h ws h represents the height of the inner and outer coils of the high-voltage winding after removing the gaps. wp =h1+h2,h ws =h3+7h4, window height h=h w +2h0; Define δ as the distance between the low-voltage winding and the core, c and b as the thicknesses of the high-voltage and low-voltage windings respectively, a as the interval between the high-voltage and low-voltage windings, and Δ as the distance from the high-voltage winding to the neutral line between phase A and phase B. Then the pole pitch τ = a + b + c + δ + Δ.

7. The method for early fault diagnosis of transformers based on the difference between virtual and real waveforms of winding leakage magnetic field as described in claim 6, characterized in that: State matrix A s and input matrix B s There exists an inductance matrix, which includes a main flux inductance matrix and a leakage flux inductance matrix. The inductance matrix is ​​obtained by adding the inductance matrices of the main flux and the leakage flux: L = L T +L σ .

8. A transformer early fault diagnosis system utilizing the difference between the virtual waveform and the measured waveform of the winding leakage magnetic field, based on the transformer early fault diagnosis method utilizing the difference between the virtual and measured waveforms of the winding leakage magnetic field as described in any one of claims 1 to 7, characterized in that: include, The magnetic field acquisition module is used to acquire leakage magnetic field information of non-ideal structure winding transformers through fiber optic magnetic field sensors; The data processing module is used to decompose the non-ideal winding into multiple regular windings according to the superposition theorem and the principle of magnetomotive force balance, calculate the leakage flux matrix leakage inductance parameters and leakage magnetic field output matrix parameters, and establish a multi-state digital model of transformer circuit-leakage magnetic field. The fault diagnosis module is used to compare the actual transformer leakage magnetic field waveform with the virtual waveform of the digital model, use the waveform difference as a fault feature quantity, complete the threshold value calculation and early fault sensitivity verification, and form an early fault diagnosis criterion based on waveform difference. The fault location module is used to determine the specific location of early-stage faults in the transformer windings based on diagnostic criteria.

Images