Transformer winding fault diagnosis method, terminal and storage medium

Abstract

The invention belongs to the technical field of power equipment maintenance, and particularly relates to a transformer winding fault diagnosis method, a terminal and a storage medium, and the method comprises the steps: obtaining frequency response curves of a transformer winding in a normal state and an abnormal state through a frequency sweep test, the head end of the primary side winding is connected with a sweep frequency signal and the tail end is suspended; determining the number of units of an equivalent network according to the number of the frequency response curve characteristic frequencies of the windings in the normal state, and reconstructing a normal winding equivalent network composed of a plurality of RLC units; according to the method, the high-precision equivalent network is constructed, the frequency response residual error analysis is performed, and the Pearson's correlation coefficient is combined to screen the high-correlation element, so that the specific position of the winding deformation or fault can be accurately positioned, the fault type is distinguished, and the limitation that the traditional frequency response method can only judge the overall deformation degree is thoroughly solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power equipment maintenance, and particularly relates to a transformer winding fault diagnosis method, a terminal, and a storage medium. Background Art

[0002] In the power system, as a key device, the safe and stable operation of the transformer is crucial. However, transformer winding faults are one of the main factors affecting the reliability of the transformer, which may lead to a decline in the performance of the transformer or even damage, thus threatening the safe and stable operation of the power system. Therefore, accurately and efficiently diagnosing transformer winding faults is of great significance.

[0003] Currently, the frequency response analysis method (frequency response method) is widely used in the power system to diagnose winding deformation. This method applies a swept-frequency sine signal within a certain frequency range to the winding and measures the sine steady-state response at the end of the winding to obtain the frequency response curve of the winding. By comparing the current frequency response curve with the frequency response curve at the time of factory, it is possible to preliminarily determine whether the winding is deformed. However, the traditional frequency response method can only diagnose the severity of the overall deformation of the winding, and cannot accurately obtain the exact location of the winding deformation, the type of fault, and the severity of different parts. Summary of the Invention

[0004] Aiming at the defect in the prior art that the traditional frequency response analysis method can only preliminarily determine whether the winding is deformed and cannot accurately obtain the exact location of the winding deformation, the type of fault, and the severity of different parts, the present invention provides a transformer winding fault diagnosis method, a terminal, and a storage medium to solve the above technical problems.

[0005] In a first aspect, the present invention provides a transformer winding fault diagnosis method, including: Step S1: Obtain the frequency response curves of the transformer winding in the normal state and the abnormal state through swept-frequency testing, where during the testing, the secondary winding is short-circuited and grounded, the first end of the primary winding is connected to the swept-frequency signal and the second end is suspended, and the test frequency band is 1 kHz to 2 MHz; Step S2: Determine the number of units of the equivalent network according to the number of characteristic frequencies of the frequency response curve of the normal state winding, and reconstruct the normal winding equivalent network composed of multiple RLC units; Step S3: Calculate the voltage and current parameters of all branches based on the reconstructed equivalent network, and construct a Jacobian matrix for characterizing the sensitivity of the frequency response function to network parameter changes; Step S4: Calculate the frequency response residuals between the normal frequency response curve and the abnormal frequency response curve, decompose and reconstruct the Jacobian matrix and the frequency response residuals respectively, obtain the corresponding reconstructed real frequency response residuals and real Jacobian matrix, and construct a real relationship equation between the reconstructed real Jacobian matrix and the real frequency response residuals; Step S5: Analyze the Pearson correlation coefficients between the column vectors of the reconstructed real Jacobian matrix and the real frequency response residuals, sort the column vectors in descending order of correlation, and record the corresponding network element numbers; Step S6: Select the network elements corresponding to several column vectors with the highest correlation, construct a least squares equation to solve the corrected values of the fault element parameters, update the equivalent network parameters, and recalculate the frequency response curve; Step S7: Iteratively execute Steps S3 to S6 until the residual between the updated frequency response curve and the measured abnormal curve is less than a preset first residual threshold, and determine the fault location, type, and severity based on the finally corrected network element numbers and parameter changes.

[0006] A further improvement of this technical solution is that Step S1 specifically includes: Step S11: The test frequency band is from 1 kHz to 2 MHz. The swept-frequency signal is generated by a signal generator and input through the head end of the primary winding, and the tail end is kept floating; Step S12: The secondary winding of the transformer under test is grounded through a short-circuit wire to eliminate the interference of the secondary-side distributed parameters on the frequency response curve; Step S13: Calculate the frequency response function by measuring the spectral data of the current at the head end and the voltage at the tail end of the primary side , and the calculation formula is: ; where is the frequency response function, which characterizes the impedance characteristics of the transformer winding under high-frequency signals; is the voltage amplitude at the tail end of the primary winding of the transformer at frequency f, with the unit of volt; is the current amplitude at the head end of the primary winding of the transformer at frequency f, with the unit of ampere; is the phase factor, which reflects the phase difference between the voltage and the current ; is the imaginary unit, that is, it satisfies , which characterizes the phase shift characteristics of inductance and capacitance.

[0007] A further improvement of this technical solution is that Step S2 includes: Step S21: Jointly analyze the amplitude-frequency characteristics and phase-frequency characteristics of the normal-state frequency response curve, and statistically count the total number N of characteristic frequencies through extreme point detection and inflection point identification; Extreme points include peaks and valleys; Step S22: Compensate the mutual inductance effect in the high-frequency band according to a preset ratio based on the total number N of characteristic frequencies, and dynamically determine the number n of equivalent network units of the winding; The calculation formula is: , where is the floor function; Step S23: Construct a winding equivalent network with a chain - type series structure using the RLC units with the determined number of units; Step S24: Based on the pre - stored least - squares optimization algorithm, with the measured frequency response curve as the benchmark, iteratively adjust the resistance, inductance, and capacitance parameters of each unit so that the frequency response residual between the frequency response curve of the winding equivalent network and the measured curve converges to a preset second residual threshold; the objective function of the pre - stored least - squares optimization algorithm is: ; where is the frequency response function measured by sweep - frequency testing, which characterizes the frequency response characteristics of the winding under normal conditions; is the voltage amplitude at the end of the primary - side winding of the transformer in the equivalent network at frequency , with the unit of volt; is the current amplitude at the beginning of the primary - side winding of the transformer in the equivalent network at frequency , with the unit of ampere; is the phase factor, which reflects the phase difference between the voltage and current in the equivalent network .

[0008] A further improvement of this technical solution is that step S3 includes: Step S31: Based on the reconstructed equivalent network, solve the voltage values and current values of the RLC elements in all branches through circuit analysis; Step S32: For each RLC element, calculate the sensitivity of the frequency response function to the parameters according to its voltage or current value. Among them, the sensitivity of the inductance parameter is characterized by the product of the square of the branch current value and the angular frequency, the sensitivity of the capacitance parameter is characterized by the product of the square of the branch voltage value and the angular frequency, the sensitivity of the resistance parameter is directly characterized by the square of the branch current value, and the sensitivity of the mutual - inductance parameter is characterized by the product of the product of the currents in the associated inductance branches and the angular frequency; Step S33: Fill all the sensitivity values into a matrix according to the pre - stored test frequency points and parameter indices to form the Jacobian matrix A; Step S34: Compare the measured sensitivity with the elements of the Jacobian matrix, and when the deviation between the measured sensitivity and the matrix elements exceeds the preset deviation threshold, re - execute steps S31 to S33 until the deviation converges.

[0009] A further improvement of this technical solution is that step S4 includes: Step S41: Calculate the frequency response residual between the normal frequency response curve and the abnormal frequency response curve, decompose the real part and the imaginary part of the frequency response residual, and then vertically splice the real part and the imaginary part of the frequency response residual to obtain the real - number residual vector ; Step S42: Decompose the Jacobian matrix A into a real part and an imaginary part, and vertically splice the real part and the imaginary part of the Jacobian matrix A to obtain the reconstructed real - number Jacobian matrix ; Step S43: Correlate the reconstructed real - valued Jacobian matrix with the real - valued residual vector and construct its real - valued relationship equation: , where is the network parameter correction value vector, which is used to characterize the parameter change caused by the fault.

[0010] A further improvement of this technical solution is that step S5 includes: Step S51: Calculate the correlation between each column vector constituting the real - valued Jacobian matrix and the real - frequency response residual according to the Pearson correlation coefficient PCC formula: ; The Pearson correlation coefficient formula is: , where is the number of test frequency points; is the value of the i - th element in the column vector ; is the value of the i - th element in the real - frequency response residual ; Step S52: Perform a traversal operation on all column vectors of the real - valued Jacobian matrix , and substitute each column vector into the Pearson correlation coefficient formula to calculate its correlation coefficient value with the real - frequency response residual ; During the calculation process, record the correlation coefficient value corresponding to each column vector and the column index g of this column vector in the real - valued Jacobian matrix , and at the same time establish a mapping relationship between the column index g and the equivalent network parameter number of the transformer winding, that is, each column index g corresponds to one or a group of component parameters in the winding equivalent network; Step S53: Sort all column vectors of the real - valued Jacobian matrix in descending order according to the calculated correlation coefficient values; During the sorting process, according to the recorded column index and the mapping relationship between the column index and the equivalent network parameter number of the transformer winding, output the equivalent network parameter number information corresponding to each column vector in order to form a parameter number sequence.

[0011] A further improvement of this technical solution is that step S6 includes: Step S61: Based on the Pearson correlation coefficient sorting result, select the column vectors of the real - valued Jacobian matrix whose correlation with the real - frequency response residual ranks in the top 5%, record their corresponding network component numbers, and form a set S of fault component numbers; Step S62: According to the real - valued Jacobian matrix Extract the column vectors to construct a subset of the Jacobian matrix ; Step S63: Extract the elements corresponding to the set S of fault component numbers from the network parameter correction value vector to form a sub-vector of network parameter correction values ; Step S64: Based on the subset of the Jacobian matrix and the sub-vector of network parameter correction values construct a least squares equation: ; Step S65: Solve the constructed least squares equation to obtain the correction values of the network component parameters , and its calculation formula is: , where is the transpose matrix of the subset of the Jacobian matrix ; Step S66: Based on the correction values of the network component parameters obtained by solving update the parameters of the winding equivalent network; and based on the updated network parameters, recalculate the frequency response curve.

[0012] A further improvement of this technical solution is that step S7 includes: After the iteration termination, the finally corrected network component numbers and parameter changes, combined with the pre-stored fault feature library, determine the location, type and severity of the fault; the fault feature library contains the corresponding relationships between the fault types, fault locations corresponding to different component numbers, and the parameter changes and the fault severity. According to the parameter changes of the network components, judge the fault type and severity; the fault types include short circuit, open circuit and parameter drift, and the severity includes mild fault, moderate fault and severe fault.

[0013] In a second aspect, the present invention provides a terminal, including: a processor and a memory, wherein the memory is used to store a computer program, the processor is used to call and run the computer program from the memory, so that the terminal executes the method of the above-mentioned terminal.

[0014] In a third aspect, the present invention provides a computer storage medium, in which instructions are stored, and when it runs on a computer, it enables the computer to execute the methods described in the above aspects.

[0015] The beneficial effects of the present invention are as follows: Precisely locate the fault position and type, breaking through the limitations of traditional methods: By constructing a high-precision equivalent network and frequency response residual analysis, and combining the Pearson correlation coefficient to screen highly correlated components, this method can accurately locate the specific position of winding deformation or faults and distinguish the fault types, completely solving the limitation that the traditional frequency response method can only judge the overall deformation degree.

[0016] Dynamically optimize the equivalent network and improve the matching accuracy: Dynamically determine the number of units based on the characteristic frequency, compensate for the high-frequency mutual inductance effect, avoid overfitting or underfitting, and ensure that the electromagnetic characteristics of the winding are covered by the equivalent network in the full frequency band. Iteratively optimize the parameters through the least squares algorithm to make the residual between the network frequency response and the measured curve converge to the threshold, improving the accuracy matching accuracy of the equivalent network and laying a reliable foundation for subsequent fault analysis.

[0017] The present invention calculates the parameter sensitivity based on the branch voltage and current, accurately quantifies the sensitivity of the frequency response to the parameters, and avoids the errors of manual experience. By statistically ranking the correlations (such as the top 5% components), eliminating the interference of irrelevant parameters, the fault location range is narrowed down to key components, improving the calculation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] 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, for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative efforts.

[0019] Figure 1 It is a schematic flowchart of the method according to an embodiment of the present invention.

[0020] Figure 2 It is a schematic diagram of the winding equivalent network according to an embodiment of the present invention.

[0021] Figure 3 It is a schematic structural diagram of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] In order to make the objectives, features, and advantages of the present invention more obvious and understandable, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the drawings in the specific embodiments of the present invention. Obviously, the following described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of this application.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs. The terms used in the description of this invention herein are for the purpose of describing specific embodiments only and are not intended to limit the invention.

[0024] Figure 1 It is a schematic flowchart of a transformer winding fault diagnosis method provided by this invention. According to different requirements, the order of steps in this flowchart can be changed and some can be omitted.

[0025] As Figure 1 shown, this method includes: Step S1: Obtain the frequency response curves of the transformer winding in the normal state and the abnormal state through sweep frequency testing. During the testing, the secondary winding is short-circuited and grounded, the leading end of the primary winding is connected to the sweep frequency signal and the trailing end is left floating, and the test frequency band is from 1 kHz to 2 MHz; Step S2: Determine the number of units of the equivalent network according to the number of characteristic frequencies of the frequency response curve of the normal state winding, and reconstruct the equivalent network of the normal winding composed of multiple RLC units; Step S3: Calculate the voltage and current parameters of all branches based on the reconstructed equivalent network, and construct a Jacobian matrix for characterizing the sensitivity of the frequency response function to the change of network parameters; Step S4: Calculate the frequency response residuals between the normal frequency response curve and the abnormal frequency response curve, decompose and reconstruct the Jacobian matrix and the frequency response residuals respectively, obtain the corresponding reconstructed real frequency response residuals and real Jacobian matrix, and construct a real relationship equation between the reconstructed real Jacobian matrix and the real frequency response residuals; Step S5: Analyze the Pearson correlation coefficients between the column vectors of the reconstructed real Jacobian matrix and the real frequency response residuals, sort the column vectors from high to low according to the correlation and record the corresponding network element numbers; Step S6: Select the network elements corresponding to several column vectors with the highest correlation, construct a least squares equation to solve the correction values of the fault element parameters, update the equivalent network parameters and recalculate the frequency response curve; Step S7: Iteratively execute steps S3 to S6 until the residual between the updated frequency response curve and the measured abnormal curve is less than a preset first residual threshold, and determine the fault location, type and severity according to the finally corrected network element numbers and parameter changes.

[0026] For the convenience of understanding this invention, the principle of the transformer winding fault diagnosis method of this invention is described below, and in combination with the process of diagnosing the transformer winding fault based on iterative calculation in the embodiment, the transformer winding fault diagnosis method provided by this invention is further described.

[0027] Among them, step S1 specifically includes: Step S11: The test frequency band is from 1 kHz to 2 MHz. The swept-frequency signal is generated by a signal generator and input through the first-end of the primary winding, while the end is kept floating. Step S12: The secondary winding of the transformer under test is grounded through a short-circuit wire to eliminate the interference of secondary-side distributed parameters on the frequency response curve. Step S13: Calculate the frequency response function by measuring the spectral data of the current at the first-end and the voltage at the end of the primary side. The calculation formula is: ; Where, is the frequency response function, which characterizes the impedance characteristics of the transformer winding under high-frequency signals; is the voltage amplitude at the end of the primary winding of the transformer at frequency f, with the unit of volt; is the current amplitude at the first-end of the primary winding of the transformer at frequency f, with the unit of ampere; is the phase factor, which reflects the phase difference between the voltage and the current ; is the imaginary unit, that is, it satisfies , which characterizes the phase shift characteristics of inductance and capacitance.

[0028] In the winding test of a power transformer with the model of SFZ-10000 / 110, the frequency response characteristic test operation is carried out. The specific steps are as follows: Step S11: Setting and input of the swept-frequency signal.

[0029] Select the Keysight 33500B signal generator, set its output swept-frequency signal, with the frequency range from 1 kHz to 2 MHz, the step frequency of 10 kHz, and the output level of 10 Vpp (peak-to-peak). Connect the swept-frequency signal to the first-end (H1 terminal) of the primary side (high-voltage side) winding of the transformer through a coaxial cable, and the end (H2 terminal) of the primary side is floating. For example, when testing a 110 kV / 10 kV distribution transformer, the primary winding adopts a star connection, the H1 terminal is connected to the signal, and the H2 terminal is floating to ensure the unidirectional transmission of the signal.

[0030] Step S12: Grounding treatment of the secondary winding.

[0031] Use a copper-core wire with a cross-sectional area of 4 mm² to short-circuit all terminals (X1, X2, X3) of the secondary side (low-voltage side) and then connect them to the grounding pile, with the grounding resistance ≤ 0.1 Ω. Compare the frequency response curves before and after grounding. When not grounded, an abnormal resonance peak (amplitude deviation +5 dB) appears at 1.5 MHz, and the resonance peak disappears after grounding, and the smoothness of the curve is improved.

[0032] Step S13: Calculation of the frequency response function.

[0033] Select a high-precision current probe with the model number Hioki 9016-01 to measure the current at the first-end of the primary side. , with an accuracy of ±0.5%. Select a high-voltage differential probe with the model number Tektronix P5200A to measure the voltage at the end of the primary side. , with an accuracy of ±0.2%. Use an oscilloscope of the Keysight MXR series to measure the voltage-current phase difference. , for example, the voltage-current phase difference is 80 degrees at 1 MHz, and calculate the frequency response function corresponding to this frequency point.

[0034] The advantages of selecting the test frequency band from 1 kHz to 2 MHz in the present invention are as follows: The low-frequency band (1 kHz to 100 kHz) is used to capture the overall inductance characteristics of the winding (such as inter-turn looseness); the high-frequency band (greater than 500 kHz) is used to excite the inter-turn capacitance and distributed parameters to locate partial discharge or insulation deterioration. The present invention tests by inputting at the first-end and leaving the end open, which can avoid signal reflection interference, reduce the measured signal attenuation rate, and improve the signal-to-noise ratio of the first-end input compared with the intermediate tap input method. The secondary side winding is grounded, which can eliminate the interference of the secondary side distributed parameters on the frequency response curve and improve the curve stability.

[0035] In addition, step S2 includes: Step S21: Jointly analyze the amplitude-frequency characteristic and phase-frequency characteristic of the frequency response curve in the normal state, and statistically count the total number N of characteristic frequencies through extreme point detection and inflection point identification; the extreme points include peaks and valleys; Step S22: Compensate the mutual inductance effect in the high-frequency band according to a preset ratio based on the total number N of characteristic frequencies, and dynamically determine the number n of winding equivalent network units; the calculation formula is: , where is the floor function; Step S23: Construct a winding equivalent network with a chain series structure from the RLC units with the determined number of units; Step S24: Based on the pre-stored least squares optimization algorithm, with the measured frequency response curve as the reference, iteratively adjust the resistance, inductance, and capacitance parameters of each unit to make the frequency response residual between the frequency response curve of the winding equivalent network and the measured curve converge to a preset second residual threshold; the objective function of the pre-stored least squares optimization algorithm is: ; where is the frequency response function measured through frequency sweep testing, which characterizes the frequency response characteristics of the winding in the normal state; is the voltage amplitude at the end of the primary side winding of the transformer in the equivalent network at the frequency , with the unit of volt; is the current amplitude at the first-end of the primary side winding of the transformer in the equivalent network at the frequency , with the unit of ampere; is the phase factor, reflecting the phase difference between voltage and current in the equivalent network .

[0036] In the winding test of a power transformer with the model number SFZ - 10000 / 110 (rated voltage 110 kV / 10 kV, and the high - voltage winding adopts a chain - type structure), 12 significant extreme points (6 peaks and 6 valleys) were detected in the amplitude - frequency curve of the normal winding; 8 inflection points (phase change rate ≥ 5° / Hz) were identified in the phase - frequency curve; the total number of characteristic frequencies N = 12 + 8 = 20. For example, the mutual - inductance element M of the high - voltage winding 1(n+1) triggers an implicit resonance peak at 1.8 MHz, which requires additional compensation units to cover. The compensation rule is as follows: , that is, 5 additional winding equivalent network units are added to specifically compensate for the mutual - inductance effect in the high - frequency band (1 MHz to 2 MHz) (the resonance point associated with the mutual - inductance element M of the high - voltage winding 1(n+1) ). After compensation, the winding equivalent network diagram as shown in Figure 2 is obtained. The compensated winding equivalent network consists of 25 winding equivalent network units (i.e., RLC units), and each unit contains a resistor R w , an inductor L w , a ground capacitance C gw and an inter - turn capacitance C sw . The component layout is as shown in Figure 2. A mutual - inductance coefficient M wk is set between adjacent inductors (as marked in the red dashed - box area in Figure 2 ) to reflect the magnetic - field interaction between winding turns; the chain - type series structure (refer to the high - voltage and low - voltage winding layout in Figure 2 ), with the sweep - frequency signal connected to the first end and the last end left floating.

[0037] Resistor R w : Calculate the DC resistance based on the cross - sectional area of the copper wire (120 mm²) and add skin - effect correction; Inductor L w : Initialize according to the number of winding turns (for example, 500 turns on the high - voltage side), geometric dimensions, and mutual - inductance distribution in the image; Ground capacitance C gw , inter - turn capacitance C sw : Calculate based on the dielectric constant of the inter - turn insulation material (such as epoxy resin).

[0038] Through the dynamic compensation mechanism, the present invention additionally compensates 5 winding equivalent network units to accurately characterize the high - frequency resonance points (such as the amplitude depression at 1.8 MHz) caused by the mutual - inductance M wk in the image, avoiding the problem of frequency - band omission in traditional tests. Through dynamic unit - number compensation, chain - type topology construction, and closed - loop parameter optimization, step S2 of the present invention solves the defect of insufficient modeling of high - frequency resonance points in traditional test methods and achieves high - precision matching in the full frequency band (1 kHz to 2 MHz).

[0039] In addition, step S3 includes: Step S31: Based on the reconstructed equivalent network, solve for the voltage values and current values of the RLC components in all branches through circuit analysis; Step S32: For each RLC component, calculate the sensitivity of the frequency response function to the parameters according to its voltage or current value. Among them, the sensitivity of the inductance parameter is characterized by the product of the square of the branch current value and the angular frequency, the sensitivity of the capacitance parameter is characterized by the product of the square of the branch voltage value and the angular frequency, the sensitivity of the resistance parameter is directly characterized by the square of the branch current value, and the sensitivity of the mutual inductance parameter is characterized by the product of the product of the currents in the associated inductance branches and the angular frequency; Step S33: Fill all the sensitivity values into a matrix according to the pre-stored test frequency points and parameter indices to form the Jacobian matrix A; Step S34: Compare the measured sensitivity with the elements of the Jacobian matrix, and when the deviation between the measured sensitivity and the matrix elements exceeds the preset deviation threshold, re-execute steps S31 to S33 until the deviation converges.

[0040] Based on the winding equivalent network constructed in step S2, calculate the voltage of each branch through Kirchhoff's law and current value ; According to the first-order derivative rule of the frequency response function with respect to the network parameters defined above, calculate respectively: Inductance parameter sensitivity: , where is the effective value of the current in the w-th inductance branch; Resistance parameter sensitivity: , where is the effective value of the current in the w-th resistance branch; Capacitance parameter sensitivity: , where is the effective value of the voltage in the w-th capacitance branch; Mutual inductance parameter sensitivity: , where , are the effective values of the currents in the mutually inductive associated branches.

[0041] Fill the calculated sensitivity values of each parameter into the Jacobian matrix A according to the correspondence between the frequency point i and the parameter w: , where the matrix element is defined as: , represents the w-th network parameter (including ).

[0042] Verify the matrix consistency through frequency response sensitivity simulation: If the deviation between the calculated value and the simulated value of the parameter sensitivity at any frequency point i exceeds 10%, then recalibrate the calculation process of the branch voltage and current; output the complete Jacobian matrix A for subsequent frequency response residual correlation analysis and fault location.

[0043] The present invention constructs a highly reliable Jacobian matrix through precise branch parameter calculation, sensitivity formula mapping, and matrix calibration, solving the problem of insufficient modeling of high-frequency mutual inductance effects in traditional methods. Combining the chain topology in the image, it comprehensively covers the electromagnetic coupling characteristics of all components, providing a reliable data basis for fault diagnosis.

[0044] Further, step S4 includes: Step S41, calculate the frequency response residuals of the normal frequency response curve and the abnormal frequency response curve, and decompose the real part and the imaginary part of the frequency response residuals. Then, vertically splice the real part and the imaginary part of the frequency response residuals to obtain a real residual vector ; Step S42, decompose the Jacobian matrix A into a real part and an imaginary part, and vertically splice the real part and the imaginary part of the Jacobian matrix A to obtain a reconstructed real Jacobian matrix ; Step S43, correlate the reconstructed real Jacobian matrix with the real residual vector and construct its real relationship equation: , where is the network parameter correction value vector, used to characterize the parameter change amount caused by the fault.

[0045] The real part of the decomposed frequency response residual characterizes the amplitude difference, reflecting the change in resistance or mutual inductance; the imaginary part of the decomposed frequency response residual characterizes the phase difference, reflecting the inductor or capacitor fault. By separating the real part and the imaginary part, the fault contributions of resistance / mutual inductance and inductor / capacitor are independently quantified; the real part and the imaginary part residuals are vertically spliced into a vector of dimension 2n×1.

[0046] The real part of the Jacobian matrix A corresponds to the sensitivity of the frequency response amplitude to the parameter; the imaginary part of the Jacobian matrix A corresponds to the sensitivity of the frequency response phase to the parameter; the real part and the imaginary part matrices are vertically spliced to form a matrix of dimension 2n×m.

[0047] For example, for the mutual inductance parameter M 12 in the image, its sensitivity element is: ; By solving the least squares, is obtained, indicating that the mutual inductance value has decreased by 15%.

[0048] Through complex residual decomposition and vertical splicing, the present invention resolves the high-frequency resonance characteristics into independent amplitude / phase analysis in the real number domain, solving the technical bottleneck that the traditional frequency response method cannot distinguish multi-dimensional fault sources. Combining the mutual inductance symbols and chain topologies in the image, it accurately quantifies the magnetic field coupling effect, supporting the rapid positioning of faulty components and design optimization.

[0049] After that, step S5 includes: Step S51, calculate the correlation between each column vector constituting the real Jacobian matrix and the real frequency response residual ; the Pearson correlation coefficient formula is: where, is the number of test frequency points; is the value of the i-th element in the column vector ; is the value of the i-th element in the real frequency response residual ; is the value of the i-th element in the real frequency response residual Step S52, perform a traversal operation on all column vectors of the real Jacobian matrix , and substitute each column vector into the Pearson correlation coefficient formula in turn to calculate its correlation coefficient value with the real frequency response residual ; during the calculation process, record the correlation coefficient value corresponding to each column vector and the column index g of this column vector in the real Jacobian matrix , and at the same time establish a mapping relationship between the column index g and the transformer winding equivalent network parameter number, that is, each column index g corresponds to one or a group of component parameters in the winding equivalent network; Step S53, according to the calculated correlation coefficient values, sort all column vectors of the real Jacobian matrix from high to low; during the sorting, according to the recorded column index and the mapping relationship between the column index and the transformer winding equivalent network parameter number, output the transformer winding equivalent network parameter number information corresponding to each column vector in order to form a parameter number sequence.

[0050] For example, for a power transformer of model SFZ-10000 / 110, the high-voltage winding includes inductance, resistance, and capacitance to ground, and the low-voltage winding includes inductance, mutual inductance, and capacitance between windings; extract the column vectors of each network parameter from the reconstructed real Jacobian matrix , and calculate the correlation between each column vector constituting the real Jacobian matrix and the real frequency response residual according to the Pearson correlation coefficient PCC formula, and for all network parameters ( The Pearson coefficients of () are sorted by absolute value. For example, the generated sequence is: , mutual inductance If the correlation coefficient of (corresponding to the mutual inductance coupling region between the high-voltage winding L1 and the low-voltage winding L11) is 0.92, it indicates that its value has decreased (such as the coupling weakening caused by winding displacement); If the correlation coefficient of (corresponding to the capacitance coupling region between the high-voltage winding L1 and the low-voltage winding L11) is 0.85, it indicates that its value has increased (such as the capacitance value rising due to insulation deterioration).

[0051] Through the calculation and sorting of the Pearson correlation coefficient, the present invention accurately locates the mutual inductance and capacitance coupling fault regions marked in the image, solving the problem of diagnostic ambiguity of traditional methods in complex coupling scenarios.

[0052] In addition, step S6 includes: Step S61, based on the sorting result of the Pearson correlation coefficient, select the real Jacobi matrix whose correlation with the real frequency response residual ranks in the top 5%, record the corresponding network element numbers, and form a set S of fault element numbers; Step S62, construct a subset of the Jacobi matrix according to the column vectors extracted from the real Jacobi matrix ; Step S63, extract the elements corresponding to the set S of fault element numbers from the network parameter correction value vector to form a sub-vector of the network parameter correction value; Step S64, construct a least-squares equation based on the subset of the Jacobi matrix and the sub-vector of the network parameter correction value: ; Step S65, solve the constructed least-squares equation to obtain the correction value of the network element parameters, and its calculation formula is: , where is the transposed matrix of the subset of the Jacobi matrix ; Step S66, update the parameters of the winding equivalent network based on the obtained correction value of the network element parameters; and recalculate the frequency response curve based on the updated network parameters.

[0053] For example, for a power transformer of model SFZ-10000 / 110, the equivalent network as shown in Figure 2 is obtained. If the least-squares equation is solved, the correction value of the network element parameters is obtained: , etc., and the corresponding parameters are updated as follows: correct the inductance value of L5 to L5 + 10 μH to optimize the power frequency impedance characteristics; adjust M 1(n+1) from 0.5 mH to 0.3 mH to suppress high-frequency coupling crosstalk; the resonance peak deviation of the corrected frequency response curve at 1.5 MHz is reduced from ±5 dB to ±0.5 dB.

[0054] Through high-correlation component screening, least squares solution, and parameter update, the present invention directly and precisely corrects the mutual inductance coupling and capacitor fault areas in the image, improving the fault repair efficiency and system stability.

[0055] Finally, step S7 includes: After the iteration termination, the finally corrected network element numbers and parameter changes, combined with the pre-stored fault feature library, determine the location, type, and severity of the fault; the fault feature library contains the corresponding relationship between the fault type, fault location, and parameter changes and fault severity for different element numbers, and the fault type and severity are judged according to the parameter changes of the network elements; the fault types include short circuit, open circuit, and parameter drift, and the severity includes mild fault, moderate fault, and severe fault.

[0056] For example, when testing a power transformer of model SFZ-10000 / 110, the equivalent network shown in Figure 2 is obtained. If the finally corrected parameters are: , and ; the corresponding fault location, type, and severity are queried through the pre-stored fault feature library; among them, the discrimination criteria for severity are: Mild fault: , ; Moderate fault: , ; Severe fault: , .

[0057] : The matching fault type is winding displacement / deformation, corresponding to the mechanical looseness between the high-voltage winding L1 and the low-voltage winding L 26 in the image; it belongs to a moderate fault.

[0058] : The matching fault type is insulation oil deterioration, corresponding to the area of the capacitance C g5 between the high-voltage winding L5 and the ground in the image; it belongs to a severe fault.

[0059] : The matching fault type is turn - to - turn looseness, corresponding to the axial displacement of the high - voltage section L in the image 15 ; it belongs to a minor fault.

[0060] Through parameter matching, grading rules, and a dynamic feature library, the present invention accurately associates the component numbers in the image with physical faults, achieving a closed - loop diagnosis from "parameter correction" to "operation and maintenance decision - making". Combining the mutual inductance coupling region and capacitor layout significantly improves the fault location efficiency and repair accuracy. Based on severity grading, high - risk faults (such as severe insulation deterioration) are preferentially processed to reduce downtime.

[0061] Figure 3 FIG. 9 is a schematic structural diagram of a terminal 300 provided by an embodiment of the present invention. The terminal 300 can be used to execute the transformer winding fault diagnosis method provided by the embodiment of the present invention.

[0062] Among them, the terminal 300 may include: a processor 310, a memory 320, and a communication module 330. These components communicate through one or more buses. Those skilled in the art can understand that the structure of the server shown in the figure does not constitute a limitation to the present invention. It can be a bus - shaped structure, a star - shaped structure, and may also include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0063] Among them, the memory 320 can be used to store the execution instructions of the processor 310. The memory 320 can be implemented by any type of volatile or non - volatile storage terminal or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read - only memory (EEPROM), erasable programmable read - only memory (EPROM), programmable read - only memory (PROM), read - only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disk. When the execution instructions in the memory 320 are executed by the processor 310, the terminal 300 can execute some or all of the steps in the above - mentioned method embodiments.

[0064] The processor 310 serves as the control center of the storage terminal, connecting various parts of the entire electronic terminal through various interfaces and circuits. By running or executing software programs and / or modules stored in the memory 320, and by invoking data stored in the memory, it performs various functions of the electronic terminal and / or processes data. The processor may be composed of an integrated circuit (IC), for example, it may be composed of a single packaged IC, or it may be composed of multiple packaged ICs with the same or different functions connected together. For example, the processor 310 may include only a central processing unit (CPU). In the embodiments of the present invention, the CPU may be a single-core processor or may include multiple cores.

[0065] The communication module 330 is used to establish a communication channel so that the storage terminal can communicate with other terminals. It receives user data sent by other terminals or sends user data to other terminals.

[0066] The present invention also provides a computer storage medium. The computer storage medium can store a program, and when the program is executed, it may include some or all of the steps in the embodiments provided by the present invention. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0067] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions in the embodiments of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc., which can store program codes, and includes several instructions to enable a computer terminal (which may be a personal computer, a server, or a second terminal, a network terminal, etc.) to execute all or part of the steps of the methods described in the embodiments of the present invention.

[0068] For the same or similar parts among the various embodiments in this specification, reference can be made to each other. In particular, for the terminal embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for the relevant parts, reference can be made to the descriptions in the method embodiments.

[0069] Although the present invention has been described in detail by reference to the accompanying drawings and in conjunction with the preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, those of ordinary skill in the art can make various equivalent modifications or substitutions to the embodiments of the present invention, and all such modifications or substitutions should be within the scope of the present invention / Any person skilled in the art within the technical scope disclosed by the present invention can easily conceive of changes or substitutions, which should all be covered by the protection scope of the present invention.

Claims

1. A method for diagnosing transformer winding faults, characterized in that, Including: Step S1: Obtain the frequency response curves of the transformer winding in the normal state and the abnormal state through sweep frequency testing. During the test, the secondary winding is short-circuited and grounded, the head end of the primary winding is connected to the sweep frequency signal, and the tail end is suspended. The test frequency band is from 1 kHz to 2 MHz. Step S2: Determine the number of units of the equivalent network according to the number of characteristic frequencies of the frequency response curve of the normal state winding, and reconstruct the equivalent network of the normal winding composed of multiple RLC units. Step S3: Calculate the voltage and current parameters of all branches based on the reconstructed equivalent network, and construct a Jacobian matrix for characterizing the sensitivity of the frequency response function to the change of network parameters. Step S4: Calculate the frequency response residuals between the normal frequency response curve and the abnormal frequency response curve, decompose and reconstruct the Jacobian matrix and the frequency response residuals respectively, obtain the corresponding reconstructed real frequency response residuals and real Jacobian matrix, and construct a real relationship equation between the reconstructed real Jacobian matrix and the real frequency response residuals. Step S5: Analyze the Pearson correlation coefficients between the column vectors of the reconstructed real Jacobian matrix and the real frequency response residuals, sort the column vectors from high to low according to the correlation, and record the corresponding network element numbers. Step S6: Select the network elements corresponding to several column vectors with the highest correlation, construct a least squares equation to solve the corrected values of the fault element parameters, update the equivalent network parameters, and recalculate the frequency response curve. Step S7: Iteratively execute steps S3 to S6 until the residual between the updated frequency response curve and the measured abnormal curve is less than the preset first residual threshold, and determine the fault location, type and severity according to the finally corrected network element number and parameter change.

2. The transformer winding fault diagnosis method according to claim 1, wherein Step S1 specifically includes: Step S11: The test frequency band is from 1 kHz to 2 MHz. The sweep frequency signal is generated by a signal generator and input through the head end of the primary winding, and the tail end remains suspended. Step S12: The secondary winding of the transformer under test is grounded through a short-circuit wire to eliminate the interference of the secondary side distributed parameters on the frequency response curve. Step S13: Calculate the frequency response function by measuring the spectral data of the current at the first end and the voltage at the last end of the primary side , and the calculation formula is: ； Among them, is the frequency response function, which characterizes the impedance characteristics of the transformer winding under high-frequency signals; is the voltage amplitude at the end of the primary winding of the transformer at frequency f, with the unit of volt; is the current amplitude at the beginning of the primary winding of the transformer at frequency f, with the unit of ampere; is the phase factor, which reflects the phase difference between voltage and current ; is the imaginary unit, that is, it satisfies , which characterizes the phase shift characteristics of inductance and capacitance.

3. The transformer winding fault diagnosis method according to claim 2, wherein Step S2 includes: Step S21: Jointly analyze the amplitude-frequency characteristics and phase-frequency characteristics of the frequency response curve in the normal state, and statistically count the total number N of characteristic frequencies through extreme point detection and inflection point identification. Extreme points include peaks and valleys. Step S22: Based on the total number of characteristic frequencies N, compensate the mutual inductance effect in the high-frequency band according to a preset ratio, and dynamically determine the number of winding equivalent network units n. The calculation formula is: , where is rounding down; Step S23: Construct a winding equivalent network with a chain series structure from the RLC units with the determined number of units. Step S24: Based on the pre-stored least squares optimization algorithm, with the measured frequency response curve as the benchmark, iteratively adjust the resistance, inductance, and capacitance parameters of each unit so that the frequency response residual between the frequency response curve of the winding equivalent network and the measured curve converges to a preset second residual threshold; the objective function of the pre-stored least squares optimization algorithm is: ; where is the frequency response function measured through frequency sweep testing, characterizing the frequency response characteristics of the winding under normal conditions; is the voltage amplitude at the end of the primary winding of the transformer in the equivalent network at frequency , with the unit of volts; is the current amplitude at the start of the primary winding of the transformer in the equivalent network at frequency , with the unit of amperes; is the phase factor, reflecting the phase difference between the voltage and current in the equivalent network .

4. The transformer winding fault diagnosis method according to claim 3, characterized in that, Step S3 includes: Step S31: Based on the reconstructed equivalent network, solve the voltage values and current values of the RLC elements in all branches through circuit analysis. Step S32: For each RLC element, calculate the sensitivity of the frequency response function to the parameters according to its voltage or current value. Among them, the sensitivity of the inductance parameter is characterized by the product of the square of the branch current value and the angular frequency, the sensitivity of the capacitance parameter is characterized by the product of the square of the branch voltage value and the angular frequency, the sensitivity of the resistance parameter is directly characterized by the square of the branch current value, and the sensitivity of the mutual inductance parameter is characterized by the product of the current product of the associated inductance branches and the angular frequency. Step S33: Fill in all sensitivity values into the matrix according to the pre-stored test frequency points and parameter indexes to form the Jacobian matrix A. Step S34: Compare the measured sensitivity with the Jacobian matrix elements, and when the deviation between the measured sensitivity and the matrix elements exceeds a preset deviation threshold, re-execute Steps S31 to S33 until the deviation converges.

5. The transformer winding fault diagnosis method according to claim 4, wherein, Step S4 includes: Step S41: Calculate the frequency response residual between the normal frequency response curve and the abnormal frequency response curve, and decompose the real part and the imaginary part of the frequency response residual. Then, vertically splice the real part and the imaginary part of the frequency response residual to obtain a real residual vector ; Step S42: Decompose the Jacobian matrix A into its real and imaginary parts, and vertically concatenate the real and imaginary parts of the Jacobian matrix A to obtain a reconstructed real Jacobian matrix ; Step S43: Correlate the reconstructed real Jacobian matrix with the real residual vector and construct its real relationship equation: , where is the network parameter correction value vector, which is used to represent the parameter change caused by the fault.

6. The transformer winding fault diagnosis method according to claim 5, wherein Step S5 includes: Step S51: Calculate the correlation between each column vector of the real Jacobian matrix and the real frequency response residual ; the Pearson correlation coefficient formula is: where, is the number of test frequency points; is the value of the i-th element in the column vector ; is the value of the i-th element in the real frequency response residual ; ​ Step S52: Perform a traversal operation on all column vectors of the real Jacobian matrix , and successively substitute each column vector into the Pearson correlation coefficient formula to calculate its correlation coefficient value with the real frequency response residual ; during the calculation process, record the correlation coefficient value corresponding to each column vector and the column index g of this column vector in the real Jacobian matrix , and at the same time establish a mapping relationship between the column index g and the equivalent network parameter number of the transformer winding, that is, each column index g corresponds to one or a group of component parameters in the winding equivalent network; Step S53: According to the calculated correlation coefficient values, sort all column vectors of the real Jacobian matrix in descending order; while sorting, according to the recorded column indexes and the mapping relationship between the column indexes and the transformer winding equivalent network parameter numbers, output the transformer winding equivalent network parameter number information corresponding to each column vector in sequence to form a parameter number sequence.

7. The transformer winding fault diagnosis method according to claim 6, wherein Step S6 includes: Step S61. Based on the Pearson correlation coefficient sorting result, select the real Jacobian matrix column vectors with the top 5% correlation with the real frequency response residual, record their corresponding network element numbers, and form a set S of faulty element numbers; ​ Step S62: Construct a subset of the Jacobian matrix from the column vectors extracted from the real Jacobian matrix ; Step S63: Extract the elements corresponding to the faulty component number set S from the network parameter correction value vector to form a network parameter correction value sub-vector ; Step S64, based on the Jacobian matrix subset and the subvector of the network parameter correction value to construct a least squares equation: ; Step S65: Solve the constructed least - square equation to obtain the corrected values of the network element parameters , and its calculation formula is: , where is the transposed matrix of the Jacobian matrix subset ; Step S66: Based on the corrected values of the network element parameters obtained by solving Update the parameters of the winding equivalent network; and recalculate the frequency response curve based on the updated network parameters.

8. The transformer winding fault diagnosis method according to claim 7, wherein Step S7 includes: After the iteration terminates, based on the finally corrected network element numbers and parameter changes, combined with the pre-stored fault feature library, determine the location, type, and severity of the fault; the fault feature library contains the corresponding relationships between fault types, fault locations, and parameter changes and fault severities for different element numbers. Based on the parameter changes of the network elements, judge the fault type and severity; the fault types include short circuit, open circuit, and parameter drift, and the severity levels include mild fault, moderate fault, and severe fault.

9. A terminal, characterized in that, Includes: A processor; A memory for storing the execution instructions of the processor; Wherein, the processor is configured to execute the method according to any one of claims 1-8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the program is executed by the processor, it implements the method according to any one of claims 1-8.

Images