Method for detecting anti-short-circuit capability of power transformer based on digital twinning technology

Abstract

The invention discloses a method for detecting the short-circuit resistance of a power transformer based on a digital twinning technology, and the method comprises the steps: building a three-dimensional transformer vibration simulation model, and carrying out the coupling calculation of a circuit, a magnetic field, a structural force field and a flow field. Calculating current and voltage curves of the winding and magnetic field distribution of the transformer in electromagnetic field simulation; solving a deformation trend and a fault type of the winding in a structural force field; the vibration characteristics of the measuring points on the surface of the oil tank are calculated in the flow field, so that the relationship between transformer winding deformation and vibration and reactance characteristics is obtained, the short-circuit damage process of the transformer can be analyzed, the method is of great significance in mastering the damage mechanism of the transformer, the anti-short-circuit capability of the power transformer is detected from multiple aspects, the detection accuracy is improved, and the detection cost is reduced. Related tests can be carried out when transformer equipment factories or related scientific research institutions are combined with transformer short-circuit tests, transformer short-circuit laws are mastered, and the anti-short-circuit capacity of the transformer is optimized and improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power transformers, and in particular to a method for detecting the short-circuit resistance of a power transformer based on digital twin technology. Background Art

[0002] With the development of the national economy, power grid construction is developing towards ultra-high voltage and large capacity, and the scale and voltage level of the power grid are also growing. As an important equipment in the power transmission system, the capacity of a single large transformer is getting larger and larger, and its reliability has correspondingly higher requirements. However, due to the limitations of transportation conditions, the volume of large transformers cannot be very large, so the electromagnetic load they bear is very high, resulting in a significant increase in the leakage magnetic field, which brings a series of problems, one of which is the problem of coil short-circuit strength. During operation, the transformer is very likely to suffer from various sudden extreme conditions, such as short-circuit faults, lightning strikes, etc. In this case, a huge short-circuit force will be generated in the coil. If the equipment is not able to resist these extreme conditions, it may cause further serious failures, which may cause damage to the insulation and structural parts of the coil, affecting the insulation performance of the equipment, or even cause the coil to loosen, twist, deform, break the wire, or even collapse the entire coil, or cause the coil to burn due to inter-turn short circuit caused by insulation damage, which will then cause a large-scale power outage in the power grid.

[0003] There are studies on short-circuit electromagnetic forces. Many existing documents have conducted in-depth analysis and research on the short-circuit electromagnetic forces and strength issues of transformer windings. The finite element method was used to perform dynamic strength calculations and static strength calculations. The short-circuit electromagnetic forces were calculated based on the existing large-scale transformer short-circuit electromagnetic force calculation software, and the SuperSAP software was used to solve the short-circuit strength and stability issues of the windings. At the same time, the large displacement geometric nonlinear theory was used to study the radial stability of the transformer low-voltage windings, and the changing law of the winding stability critical load was calculated and analyzed, and a method for improving the winding's ability to resist short circuits was given.

[0004] However, in the existing detection method of the short-circuit resistance of the power transformer, the detection is limited to the static aspect, that is, the moment when the winding short-circuit current reaches the peak value is checked, and the detection process is relatively simple, resulting in the inaccuracy of the detection result.

[0005] Therefore, the prior art still needs to be improved and enhanced. Summary of the invention

[0006] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a method for detecting the short-circuit resistance of a power transformer based on digital twin technology, aiming to solve the problem that in the existing method for detecting the short-circuit resistance of a power transformer, the detection is limited to the static aspect, that is, the moment when the short-circuit current of the winding reaches the peak value is checked, and the detection process is relatively simple, resulting in the problem that the detection result is inaccurate.

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

[0008] In the first aspect, a method for detecting the short-circuit resistance of a power transformer based on digital twin technology includes:

[0009] Step 1: Establish a multi-physics field simulation model of the transformer short-circuit dynamic process, taking into account the leakage magnetic field, electromotive force and displacement deformation during short-circuit and reclosing multiple impacts;

[0010] Step 2: Improve the transformer model, including the construction of radial, axial and clamp models, and carry out simulation calculation and analysis of the electromagnetic field, electrodynamic force, displacement deviation and other aspects of the transformer winding model;

[0011] Step 3: Analyze the electrical and vibration data of the transformer short-circuit test, establish a three-dimensional transformer simulation model based on the electrical and vibration data of the transformer short-circuit test, and calculate the transformer short-circuit reactance value using the measured values ​​of the voltage and current of the primary and secondary windings of the winding;

[0012] Step 4: Based on the above theoretical and mathematical model analysis, a transformer is selected for winding test, relevant parameters of the winding when various faults occur are collected, and the relationship between short-circuit reactance and winding deformation is analyzed;

[0013] Step 5: Use digital twin technology to perform high-fidelity modeling of the transformer, and use the digital twin transformer model to perform simulation analysis under various combinations of deformations and various conditions; organize the simulation data to form a database, and detect the power transformer entity based on the database to achieve the detection of the power transformer's short-circuit resistance.

[0014] Furthermore, the step 1 includes:

[0015] Step 1.1: Establishment of circuit-magnetic field coupling mathematical model:

[0016] The accurate calculation of transformer electromagnetic field, short-circuit current and electromagnetic force is the basis of winding stress field. In practical applications, most electromagnetic devices are excited by voltage sources. Therefore, in the calculation of transformer transient electromagnetic field, the constraint equation of the voltage connecting the winding and the external circuit is considered.

[0017] The transient electromagnetic field equation is:

[0018]

[0019] In the formula, J is the current density in the coil, and there is a relationship J = nI / S, n is the number of turns of the coil, S is the total cross-sectional area of ​​the coil, and I is the branch current;

[0020] Winding circuit equation:

[0021]

[0022] Where, {u}={-u1 u2} T , {e}={e1 e2} T ,{i}={i1 i2} T , resistance array [R] = diag [R1 R2], leakage impedance array [L] = diag [L 1σ L 2σ ];

[0023] If the effective length of each coil turn is Lef, the number of turns of the winding is nc, and the winding area is divided into ne units, then the phase potential of the winding is:

[0024]

[0025] The spatial discrete equation of the transient process solved by coupling the transient eddy current field and the winding external circuit equation is:

[0026]

[0027] Step 1.2: Simulation of core hysteresis characteristics and establishment of circuit coupling mathematical model:

[0028] Under multiple no-load closing conditions, the synthetic flux Φ of the transient process n The expression is:

[0029]

[0030] After half a cycle of closing (ωt=π), the synthetic magnetic flux reaches its maximum value. When the initial phase angle α=0 at the moment of closing, the core magnetic flux reaches its maximum value.

[0031] The JA model is used to describe the nonlinear magnetization characteristics of the core, and a "circuit-hysteresis" coupling calculation model is established to calculate and analyze the impact current. When the JA theory is used to analyze and calculate the transformer excitation current, it is necessary to combine the magnetic circuit equation of the no-load closing dynamic process and the circuit equation describing its external characteristics with the JA model. Taking the equivalent circuit of rated operating conditions as an example, it is written as follows:

[0032]

[0033] Its matrix form expression is:

[0034]

[0035] The magnetic circuit equation of the short-circuit reclosing dynamic process, the circuit equation of the external characteristics and the JA model are combined and written as follows:

[0036]

[0037] In the formula, H is the magnetic field intensity; M is the magnetization intensity; l is the magnetic path length; N1 is the number of turns of the primary winding of the transformer; i1 is the current of the primary winding of the transformer; N2 is the number of turns of the secondary winding of the transformer; i2 is the current of the secondary winding of the transformer; s is the cross-sectional area of ​​the transformer core; B is the core magnetic flux density; R2 is the sum of the secondary resistance of the transformer and the load resistance; L2 is the sum of the secondary winding inductance and the load inductance;

[0038] In the above formula Convert to In the form of:

[0039]

[0040] Furthermore, the step 2 includes:

[0041] When the transformer is subjected to a short-circuit shock, the inner winding is subjected to inward pressure, the outer winding is subjected to outward tensile force, and the radial compression force on the inner winding is much greater than the radial tension force on the outer winding. Under multiple short-circuit conditions, the distribution of radial compression force on the inner winding in the circumferential direction is different from that under single short-circuit conditions, and is unevenly distributed. The radial force on the winding is the largest in the middle of the winding axial height and between the two stays near the 0° position in the window, and plastic deformation is most likely to occur. After the winding undergoes elastic-plastic deformation, multiple short circuits of the same type will inevitably cause cumulative deformation on the winding;

[0042] Elastic limit bending moment M re The expression is:

[0043]

[0044] In the formula, the span length of the winding between the two struts is l, the cross-sectional thickness is d, and the cross-sectional height is h. Since the inner winding radius is much larger than the span length, the winding between the two struts can be equivalent to a straight beam, and the axial force of the winding is equivalent to a uniformly distributed load P r , let the deflection of the winding in the y direction be w, and the downward direction be positive;

[0045] The bending moment corresponding to the curvature k of the winding is:

[0046]

[0047] When ξ→0, the entire winding cross section enters the plastic stage, and |k|→∞, it can be deduced that:

[0048] M rs =1.5M re

[0049] Where M re is the plastic limit bending moment;

[0050] When M r >M rs When , the winding is in the completely plastic stage. As the bending moment increases further, the elastic area of ​​the cross section becomes zero, the winding cross section enters the plastic state, and the stress of the entire cross section reaches the limit value.

[0051] Under multiple short-circuit shocks, plastic deformation will accumulate, and it is necessary to establish an elastic-plastic mechanical model of the winding. During the modeling process, the stress-strain curve is nonlinear, and the cumulative deformation of the winding and the displacement of the coil are calculated.

[0052] Furthermore, step 3 includes:

[0053] The dynamic stability of the transformer is not only related to the various support positions, but also to its own structure and temperature. Each short circuit will cause different residual deformations in the transformer, which will accumulate in multiple short circuits and affect the original structure of the winding.

[0054] Step 3.1: Dynamic stability of winding without cumulative deformation:

[0055] When the short-circuit electromotive force acts on the winding coil, the winding can be analyzed as an elastically constrained arc arch model between two stays, as shown below. Its dimensionless kinetic energy T and total potential energy U are:

[0056]

[0057] Where: ρ is the material density, A is the cross-sectional area, R is the radius of the arch, Ex is the elastic modulus of the arch, I is the second-order moment of the arch section, α is the half-span central angle, g is the radial motion velocity, w is the axial motion velocity, rx is the radius of gyration of the arch section, εm is the membrane strain of the arch, G is the radial displacement, q is the uniform electromagnetic force, and γ is the constraint flexibility parameter of the arch foot;

[0058] Step 3.2: Dynamic stability of windings with accumulated deformation:

[0059] After multiple short-circuit shocks, the transformer windings in operation may produce cumulative deformation, which changes the geometric configuration of the arch axis and forms initial defects. The mechanical properties of the arch are significantly affected by defects, and serious defects will affect the dynamic stability of the transformer. The dimensionless kinetic energy T and total potential energy U of the possible structures of several windings after cumulative deformation are:

[0060]

[0061] Where: W is the axial displacement, V is the volume, N is the axial pressure of the arch, and M is the bending moment of the arch.

[0062] Furthermore, the step 3 also includes:

[0063] Step 3.3, transformer short circuit test:

[0064] The scheme of testing leakage magnetic field using the principle of electromagnetic induction, the magnetic density is calculated according to the following formula:

[0065]

[0066] Where U is the voltage peak; W is the number of turns of the sensor coil (take 600 turns); S is the area of ​​the sensor coil;

[0067] Since the potential of the winding in the short-circuit electrodynamic test is as high as tens of thousands of volts, the winding force test is as small as mm level, and the short-circuit electrodynamic force has a dynamic process acquisition, through theoretical calculation and experimental verification, it is determined that the piezoelectric film short-circuit electrodynamic test system is used. Under the action of the huge short-circuit electromagnetic force, the winding will deform, causing the piezoelectric film sensor to be under pressure, and then generate a corresponding electrical signal;

[0068] When the piezoelectric film is subjected to force in the x direction, thickness deformation will occur and polarization will occur. The charge qx generated in the x direction is proportional to the force Fx:

[0069] q x =d 11 F x

[0070] Where, qx is the piezoelectric coefficient; d11 is the charge on the plane of the x-axis; Fx is the force applied along the x-axis;

[0071] Then, a laser sensor is used to directly measure the radial vibration of the middle coil of the winding in the model.

[0072] Furthermore, the step 3 also includes:

[0073] Step 3.4: Establishment of mathematical model of winding axial vibration:

[0074] The transformer winding can be regarded as a plurality of coils, which are separated by insulating pads and pressed together by the upper and lower pressure plates and iron yoke pads. Under the action of radial electromagnetic force, the radial vibration dynamics model of the winding is established;

[0075] The acceleration of the winding vibration is:

[0076]

[0077] Step 3.5: Extraction of transformer vibration signal features based on wavelet packet transform

[0078] The vibration signal is decomposed by wavelet packet decomposition, and then the frequency band energy feature vector of the vibration signal is constructed to perform transformer fault diagnosis. The specific method is as follows:

[0079] 1) Perform three-layer wavelet packet decomposition on the transformer vibration signal collected in the experiment and simulation;

[0080] 2) Reconstruct the sub-band sequences within the 8 frequency bands obtained in step 1) to obtain 8 wavelet packet reconstructed signals, each of which contains the information of each frequency band of the original signal from low to high, and fully reflects the operating status of the transformer system;

[0081] 3) Obtain the energy entropy of 8 wavelet packets, and construct a feature vector from the energy entropy obtained after decomposing the 8 wavelet packets, denoted as T, then:

[0082] T=[H 3,0 ,H 3,1 ,H 3，2 ,H 3,3 ,…H 3，7 ]

[0083] Normalize the eigenvectors:

[0084]

[0085] Furthermore, the step 4 includes:

[0086] Transformer fault diagnosis analysis based on KELM, the extreme learning machine is a typical single hidden layer feedforward neural network. Since the traditional feedforward neural network needs to adjust all network parameters, its learning efficiency is far from meeting the needs of practical applications; the kernel extreme learning process can be described as: given a training set and kernel function K(u, v), calculate the output function:

[0087]

[0088] The calculated wavelet packet decomposition energy entropy corresponding to the winding vibration signal is used as the input of KELM and divided into a training group and a test group in a ratio of 2:1;

[0089] According to the above principles, the population size, maximum evolutionary generation, inertia weight, acceleration factors c1 and c2 of the particle swarm algorithm are set to perform transformer fault diagnosis.

[0090] Furthermore, the step 5 includes:

[0091] A digital twin model of the power transformer is established through a self-compiled program for electromagnetic, short-circuit force, short-circuit impedance, strain and winding parameter identification; the twin database is the real-time simulation data of the virtual transformer, which is embedded in the experimental platform through software writing to calculate and analyze the transformer simulation model data in real time; the data flow is a dynamic connection between the physical space and the digital space, optimizing the simulation equipment to be closer to the actual operation and maintenance situation.

[0092] The technical solution adopted by the present invention has the following beneficial effects:

[0093] (1) In this application, a three-dimensional transformer vibration simulation model is established to perform circuit, magnetic field, structural force field and flow field coupling calculations. In the electromagnetic field simulation, the current and voltage curves of the winding and the magnetic field distribution of the transformer are calculated; the deformation trend and fault type of the winding are solved in the structural force field; the vibration characteristics of the measuring points on the surface of the oil tank are calculated in the flow field, so as to obtain the relationship between the deformation and vibration of the transformer winding and the reactance characteristics. The short-circuit damage process of the transformer can be analyzed, which is of great significance for understanding the damage mechanism of the transformer. The short-circuit resistance of the power transformer can be tested from multiple aspects to improve the accuracy of the test. Relevant tests can be carried out in the transformer equipment factory or relevant scientific research institutions in conjunction with the transformer short-circuit test to understand the transformer short-circuit law, which is of great guiding significance for optimizing the design and improving the product's short-circuit resistance.

[0094] (2) Conduct field tests on transformers to measure the short-circuit reactance values ​​under various winding deformation conditions, and place vibration sensors to measure the vibration of transformer windings under various deformation conditions. Combined with the results of model simulation, analyze the relationship between transformer winding deformation and short-circuit reactance and vibration. Combined with the analysis of real transformer short-circuit tests, obtain the transformer short-circuit damage process and related mechanisms. On this basis, carry out optimization research on the transformer short-circuit resistance verification method to achieve lean verification of transformer short-circuit resistance. Transformer manufacturers can use this software to guide design and manufacturing to ensure that the products produced meet the requirements of power grid operation.

[0095] (3) The digital twin model of the power transformer is established through a self-written program with independent intellectual property rights for electromagnetic, short-circuit force, short-circuit impedance, strain and winding parameter identification. The twin database is the real-time simulation data of the virtual transformer. It is embedded in the experimental platform through software writing, and the transformer simulation model data is calculated and analyzed in real time; the data flow is a dynamic connection between the physical space and the digital space, and the optimized simulation equipment is closer to the actual operation and maintenance situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 A transformer equivalent circuit model for a detection method of short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0097] Figure 2A hysteresis loop family based on a JA model for a detection method of the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0098] Figure 3 A transformer excitation inrush current waveform of a detection method for short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0099] Figure 4 A coupling calculation transformer overall model for a detection method of a power transformer short-circuit resistance capability based on digital twin technology provided by the present invention;

[0100] Figure 5 A winding coupling calculation model for a detection method of the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0101] Figure 6 The invention provides a method for detecting the short-circuit resistance of a power transformer based on digital twin technology, which is a winding cumulative deformation;

[0102] Figure 7 A schematic diagram of winding compression deformation of a method for detecting the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0103] Figure 8 A schematic diagram of the plastic expansion of the cross section of a winding when it is under pressure in a method for detecting the short-circuit resistance of a power transformer based on the digital twin technology provided by the present invention;

[0104] Fig. 9 A schematic diagram of the nonlinearity of copper conductor materials in a method for detecting the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0105] Fig.10 A loading and unloading development path for a detection method for the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0106] Fig.11 A diagram of an elastically constrained arc arch model of a winding coil under a short-circuit impact in a detection method for the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0107] Fig.12 A defect model after cumulative deformation of windings in a detection method for short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0108] Fig.13 A laser ranging working principle diagram of a method for detecting the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0109] Fig.14 A transformer geometric model for a method for detecting short-circuit resistance of a power transformer based on digital twin technology provided by the present invention;

[0110] Fig.15 A schematic diagram of the steps of a method for detecting the short-circuit resistance of a power transformer based on digital twin technology provided by the present invention. DETAILED DESCRIPTION

[0111] In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0112] First, as Fig.15 As shown, a method for detecting the short-circuit resistance of a power transformer based on digital twin technology includes:

[0113] Step 1: Establish a multi-physics field simulation model of the transformer short-circuit dynamic process, taking into account the leakage magnetic field, electromotive force and displacement deformation during short-circuit and reclosing multiple impacts;

[0114] (1) Circuit-magnetic field coupling mathematical model:

[0115] The accurate calculation of transformer electromagnetic field, short-circuit current and electromagnetic force is the basis of winding stress field. In practical applications, most electromagnetic devices are excited by voltage sources. Therefore, in the calculation of transformer transient electromagnetic field, for such problems that are actually excited by voltage sources, the constraint equation of the voltage connecting the winding to the external circuit must be considered.

[0116] The transient electromagnetic field equation is:

[0117]

[0118] Where J is the current density in the coil, and there is a relationship J = nI / S, n is the number of coil turns, S is the total cross-sectional area of ​​the coil, and I is the branch current.

[0119] The transformer equivalent circuit is as follows Figure 1 As shown. According to the reference direction specified in the figure, the following winding circuit equation can be obtained:

[0120]

[0121] Where, {u}={-u1 u2} T , {e}={e1 e2} T ,{i}={i1 i2} T , resistance array [R] = diag [R1 R2], leakage impedance array [L] = diag [L1σ L 2σ ].

[0122] If the effective length of each coil turn is Lef, the number of turns of the winding is nc, and the winding area is divided into ne units, then the phase potential of the winding is:

[0123]

[0124] The spatial discrete equation of the transient process solved by coupling the transient eddy current field and the winding external circuit equation is:

[0125]

[0126] (2) Mathematical model of core hysteresis characteristics simulation and circuit coupling:

[0127] After a transformer short circuit occurs, the circuit is disconnected and reclosed. The operating conditions where multiple impact faults occur may be no-load or light-load closing current impact, or more seriously, the secondary reclosing short-circuit current impact when the short-circuit fault is not eliminated. The circuit and magnetic field coupling analysis of this transient process requires simulation of the core hysteresis characteristics and the influence of residual magnetism on the secondary impact current, magnetic field and electromotive force. The current results mainly focus on the study of the influence of no-load closing on the impact current. The electromagnetic coupling of the secondary impact process after reclosing is more complex and more difficult.

[0128] Under multiple no-load closing conditions, the synthetic flux Φ of the transient process n The expression is:

[0129]

[0130] After half a cycle of closing (ωt=π), the synthetic magnetic flux reaches its maximum value. When the initial phase angle α=0 at the moment of closing, the core magnetic flux reaches its maximum value.

[0131] This project uses the JA model to describe the nonlinear magnetization characteristics of the core, establishes a "circuit-hysteresis" coupling calculation model, and calculates and analyzes the impact current.

[0132] Applying a series of sinusoidal alternating magnetic field strengths H to the transformer core can obtain the corresponding magnetization intensity M. According to the relationship between the magnetic flux density B and the magnetization intensity M of the ferromagnetic material, a series of hysteresis loop family BH curves are obtained, such as Figure 2 shown.

[0133] When applying JA theory to analyze and calculate the transformer excitation current, it is necessary to combine the magnetic circuit equation of the no-load closing dynamic process and the circuit equation describing its external characteristics with the JA model. Taking the equivalent circuit of rated operating conditions as an example, it is written as follows:

[0134]

[0135] Its matrix form expression is:

[0136]

[0137] When the transformer is short-circuited multiple times, in a cycle interval, when the core enters the saturation zone, the instantaneous value of the excitation current is very large, such as Figure 1 In the section ωt=θ1-θ2, when the core exits the saturation zone, there is only normal excitation current, and its instantaneous value is very small and can be ignored. Figure 3 In the segments ωt=0-θ1 and ωt=θ2-2π.

[0138] The magnetic circuit equation of the short-circuit reclosing dynamic process, the circuit equation of the external characteristics and the JA model are combined and written as follows:

[0139]

[0140] In the formula, H is the magnetic field intensity; M is the magnetization intensity; l is the magnetic path length; N1 is the number of turns of the primary winding of the transformer; i1 is the current of the primary winding of the transformer; N2 is the number of turns of the secondary winding of the transformer; i2 is the current of the secondary winding of the transformer; s is the cross-sectional area of ​​the transformer core; B is the core magnetic flux density; R2 is the sum of the secondary resistance of the transformer and the load resistance; L2 is the sum of the secondary winding inductance and the load inductance.

[0141] In the above formula Convert to In the form of:

[0142]

[0143] (3) Calculation of coupling between electromagnetic field and structural field (stress field):

[0144] When applying the finite element electromagnetic field and structural field coupling method, it is necessary to share the transformer solution model when calculating the two fields separately. Therefore, when establishing the coupled field solution model, it is necessary to comprehensively consider the solution accuracy of the electromagnetic field and the structural field and the coupling calculation accuracy.

[0145] The magnetic field-structure field direct coupling finite element model for short-circuit response calculation of three-phase three-column oil-immersed power transformer, the overall model and the detailed model of the inner winding are as follows: Figure 4 and 5 shown.

[0146] Step 2: Improve the transformer model, including the construction of radial, axial, and clamp models, conduct simulation calculations and analysis of the electromagnetic field, electrodynamic force, displacement deviation, etc. of the transformer winding model, analyze the deformation mechanism of the transformer winding, and classify and establish the deformation model of the winding. Use the finite element method to perform electromagnetic simulation on various deformed windings; through simulation and analysis of various deformed windings, obtain the distribution law of leakage magnetic field, leakage magnetic field energy, leakage inductance, unbalance coefficient and other related parameters, and obtain the relationship between impedance and winding deformation, so as to provide a criterion for judging transformer winding deformation;

[0147] Study on elastic-plastic deformation of windings under multiple short-circuit conditions:

[0148] When the transformer is subjected to a short-circuit shock, the inner winding is subjected to inward pressure, the outer winding is subjected to outward tensile force, and the radial compression force on the inner winding is much greater than the radial tension force on the outer winding. Under multiple short-circuit conditions, the distribution of radial compression force on the inner winding in the circumferential direction is different from that under single short-circuit conditions. It is unevenly distributed. The radial force on the winding is the largest in the middle of the winding axial height and between the two stays near the 0° position in the window, and plastic deformation is most likely to occur. After the winding undergoes elastic-plastic deformation, it is bound to form cumulative deformation on the winding after experiencing multiple similar short circuits, such as Figure 6 shown.

[0149] Assume that the span length of the winding between the two struts is l, the cross-sectional thickness is d, and the cross-sectional height is h. Since the inner winding radius is much larger than the span length, the winding between the two struts can be equivalent to a straight beam, and the axial force of the winding is equivalent to a uniformly distributed load P. r Assume that the deflection of the winding in the y direction is w, and the downward direction is positive. The winding is compressed and deformed as follows Figure 7 shown.

[0150] Elastic limit bending moment M re The expression is:

[0151]

[0152] The plastic expansion process of the winding compression section is as follows: Figure 8 At the initial stage of compression, the winding undergoes elastic deformation, which is an elastic state for the entire winding cross section. The maximum stress at the outer edge of the winding cross section is less than the yield stress, and the strain is less than the yield strain. After the short-circuit force acting on the winding is unloaded, the elastic deformation will completely disappear, and the winding will completely return to its initial state.

[0153] The bending moment corresponding to the curvature k of the winding is:

[0154]

[0155] When ξ→0, the entire winding cross section enters the plastic stage, and |k|→∞, it can be deduced that:

[0156] M rs =1.5M re (12)

[0157] Where M re is the plastic limit bending moment.

[0158] When M r >M rs When , the winding is in the completely plastic stage. As the bending moment increases further, the elastic area of ​​the cross section becomes zero, the entire winding cross section enters the plastic state, and the stress of the entire cross section reaches the limit value.

[0159] The mechanical properties of the copper wire of the winding material are important factors affecting the deformation of the winding. The copper wire is a typical elastic-plastic material. When the impact load on the winding exceeds its yield limit, the winding does not completely recover after the load is unloaded, but produces irreversible plastic deformation. Under multiple short-circuit impacts, the plastic deformation will accumulate, and it is necessary to establish an elastic-plastic mechanical model of the winding. During the modeling process, the stress-strain curve is nonlinear, such as Fig. 9 shown.

[0160] Under multiple short-circuit conditions, the winding cross section will experience multiple loading and unloading processes. Therefore, in the process of nonlinear analysis of the winding structure, it is necessary to determine the winding cross section. The judgment criteria for loading and unloading. In a certain iteration, points A and B are in the increment variable Δε i Different stress development paths under the action of Fig.10 shown.

[0161] The yield function distinguishes the state of the material in the stress space by the spatial surface it describes. When the stress is within the yield surface (f < 0), the material is in an elastic state; when the stress is on the yield surface (f = 0), given the stress increment dσ, the isotropically strengthened material will have three different reactions, and the judgment criteria for the strengthening material loading and unloading are:

[0162] (1) New plastic strain appears, satisfying is plastic loading;

[0163] (2) No new plastic strain occurs, and the yield surface is returned to the surface, satisfying for elastic unloading;

[0164] (3) No new plastic strain occurs, and the system is still on the yield surface, satisfying Neutral load change.

[0165] The large deflection elastic-plastic deformation-rebound anti-coupling mathematical model and its related boundary conditions are adopted, and the equilibrium condition is established on the displacement and deformation after deformation. The residual deformation of each coil of the winding under multiple short-circuit impact loads is calculated, and the distribution law of residual deformation under different span lengths, cross-sectional heights and thicknesses is analyzed. The cumulative deformation mechanism of the transformer winding in operation is studied.

[0166] The total rebound potential energy of the large deflection elastic-plastic deformation-rebound anti-coupling system is:

[0167]

[0168] In the formula, Π' sp is the total rebound potential energy of the straight beam; A(K') is the strain energy density of the rebound straight beam; K' is the bending curvature of the rebound straight beam; ε' represents the radial strain of the rebound straight beam; N is the radial force; M is the bending moment; and V is the equivalent shear force.

[0169] The total rebound complementary energy of the large deflection elastic-plastic deformation-rebound anti-coupling system is:

[0170]

[0171] The total rebound potential energy functional of the large deflection elastic-plastic deformation-rebound anti-coupling system can be further expressed as:

[0172]

[0173] Where l is the winding span length; E is the winding elastic modulus; I is the winding section inertia moment and cross-sectional area; M(x) is the winding section bending moment; w′ is the winding rebound deformation deflection; u′ is the winding rebound deformation displacement.

[0174] Considering the continuity of rotation angles between unit nodes and the need to calculate deformation, the displacement f within the unit includes the interpolation of deflection w and displacement u as follows:

[0175]

[0176] It can be seen that the stiffness matrix is ​​a function of displacement. This equation has a nonlinear solution and belongs to a geometric nonlinear solution problem. The Newton-Raphson iteration method can be used to solve it.

[0177] The cumulative deformation of the winding and the displacement of the coil are calculated by the above method.

[0178] Step 3: Analyze the electrical and vibration data of the transformer short-circuit test, establish a three-dimensional transformer simulation model based on the electrical and vibration data of the transformer short-circuit test, and calculate the transformer short-circuit reactance value using the measured values ​​of the voltage and current of the primary and secondary windings of the winding;

[0179] The dynamic stability of the transformer is not only related to the various support positions, but also to its own structure and temperature. Each short circuit will cause different residual deformations in the transformer, which will accumulate in multiple short circuits and affect the original structure of the winding.

[0180] (1) Dynamic stability of winding without cumulative deformation:

[0181] When the short-circuit electromotive force acts on the winding coil, the winding can be analyzed as an elastically constrained arc arch model between two stays, such as Fig.11 As shown, its dimensionless kinetic energy T and total potential energy U are:

[0182]

[0183] Where: ρ is the material density, A is the cross-sectional area, R is the radius of the arch, Ex is the elastic modulus of the arch, I is the second-order moment of the arch section, α is the central angle of the half span, g is the radial motion velocity, w is the axial motion velocity, rx is the radius of gyration of the arch section, εm is the membrane strain of the arch, G is the radial displacement, q is the uniformly distributed electromagnetic force, and γ is the constraint flexibility parameter of the arch foot.

[0184] (2) Dynamic stability of windings with accumulated deformation:

[0185] After multiple short-circuit shocks, the windings of the transformer in operation may produce cumulative deformation, which changes the geometric configuration of the arch axis and forms initial defects. The mechanical properties of the arch are significantly affected by defects, and serious defects will affect the dynamic stability of the transformer. Fig.12 The possible structures of several windings after cumulative deformation, their dimensionless kinetic energy T and total potential energy U are:

[0186]

[0187] Where: W is the axial displacement, V is the volume, N is the axial pressure of the arch, and M is the bending moment of the arch.

[0188] Transformer short circuit test experimental study:

[0189] The scheme of testing leakage magnetic field using the principle of electromagnetic induction, the magnetic density is calculated according to the following formula:

[0190]

[0191] Where U is the voltage peak; W is the number of turns of the sensor coil (take 600 turns); S is the area of ​​the sensor coil.

[0192] Since the potential of the short-circuit electrodynamic test winding is as high as tens of thousands of volts, the test of the winding force is as small as mm level, and the short-circuit electrodynamics has a dynamic process collection, it is determined through theoretical calculations and experimental verification that the piezoelectric film short-circuit electrodynamic test system is used. Under the action of the huge short-circuit electromagnetic force, the winding will deform, causing the piezoelectric film sensor to be under pressure, and then generate a corresponding electrical signal.

[0193] When the piezoelectric film is subjected to force in the x direction, thickness deformation will occur and polarization will occur. The charge qx generated in the x direction is proportional to the force Fx.

[0194] q x =d 11 F x (twenty two)

[0195] Where qx ​​is the piezoelectric coefficient; d11 is the charge on the plane of the x-axis; and Fx is the force applied along the x-axis.

[0196] Laser sensors are used to directly measure the radial vibration of the middle coil of the winding in the model. The measurement system consists of four laser sensors, a vibration measuring instrument, a data cable and a microcomputer terminal.

[0197] like Fig.13 As shown in the figure, the light emitted by the laser is focused by the converging lens and vertically incident on the surface of the object to be measured. The movement of the object or the change of the surface causes the incident point to move along the incident light axis. The scattered light at the incident point is incident on the photodetector through the receiving lens, and the scattered light is imaged after being converged by the receiving lens. The two images collected before and after the moving object are processed by software to calculate its displacement.

[0198]

[0199] sinα=n / a

[0200] cosβ=n / x=cos(180°-θ-(90°-α))=sinθcosα-cosθsinα (23)

[0201] From similar triangles we can get:

[0202]

[0203] Study on vibration characteristics of power transformer windings:

[0204] (1) Mathematical model of winding axial vibration:

[0205] The transformer winding can be regarded as a plurality of coils, which are separated by insulating pads and pressed together by the upper and lower pressure plates and iron yoke pads. Under the action of radial electromagnetic force, the radial vibration dynamics model of the winding is as follows: Fig.14 shown.

[0206] The acceleration of the winding vibration is:

[0207]

[0208] (2) Transformer vibration signal feature extraction based on wavelet packet transform

[0209] The vibration signal is decomposed by wavelet packet decomposition, and then the frequency band energy feature vector of the vibration signal is constructed to perform transformer fault diagnosis. The specific method is as follows:

[0210] 1) Perform three-layer wavelet packet decomposition on the transformer vibration signal collected in the experiment and simulation;

[0211] 2) Reconstruct the sub-band sequences within the 8 frequency bands obtained in step 1) to obtain 8 wavelet packet reconstructed signals, each of which contains the information of each frequency band of the original signal from low to high, and fully reflects the operating status of the transformer system;

[0212] 3) Obtain the energy entropy of 8 wavelet packets, and construct a feature vector from the energy entropy obtained after decomposing the 8 wavelet packets, denoted as T, then:

[0213] T=[H 3，0 , H 3，1 , H 3，2 , H 3，3 ,····H 3，7 ] (26)

[0214] Normalize the eigenvectors:

[0215]

[0216] Step 4: Based on the above theoretical and mathematical model analysis, a transformer is selected for winding test, relevant parameters of the winding when various faults occur are collected, and the relationship between short-circuit reactance and winding deformation is analyzed;

[0217] Based on the above theoretical and mathematical model analysis, a transformer was selected for winding test. The specific transformer winding was set with faults such as loose axial preload, pad fall-off, change in winding radial force, and radial deformation of the coil. The relevant parameters of the winding when various faults occurred were collected, and the relationship between short-circuit reactance and winding deformation was analyzed. In addition, the vibration data of the transformer in operation was collected online and compared with the vibration signal estimated by applying the transformer vibration model analysis to determine the operating status of the transformer winding.

[0218] Transformer fault diagnosis and analysis based on KELM;

[0219] The extreme learning machine is a typical single hidden layer feedforward neural network. Since traditional feedforward neural networks need to adjust all network parameters, their learning efficiency is far from meeting the needs of practical applications. The kernel extreme learning process can be described as: given a training set and kernel function K(u, v), calculate the output function:

[0220]

[0221] The calculated wavelet packet decomposition energy entropy corresponding to the winding vibration signal is used as the input of KELM and divided into a training group and a test group in a ratio of 2:1.

[0222] According to the above principles, in the process of transformer fault diagnosis, the population size, maximum evolutionary generation, inertia weight, and acceleration factors c1 and c2 of the particle swarm algorithm can be set to perform fault diagnosis.

[0223] Step 5: Use digital twin technology to perform high-fidelity modeling of the transformer, and use the digital twin transformer model to perform simulation analysis under various combinations of deformations and various conditions; organize the simulation data to form a database, and detect the power transformer entity based on the database to achieve the detection of the power transformer's short-circuit resistance.

[0224] Digital twins are technologies that use digital technologies such as artificial intelligence, sensors, and simulation tools to achieve parallel interaction, precise mapping, and virtual-real iteration from physical atoms to data bits. They are the process and method of transforming and connecting the physical world and the digital world. Digital twins have the following typical characteristics:

[0225] (1) High fidelity: The digital world holographically replicates the physical world in terms of multiple dimensions, angles, and attributes, including its composition, morphological behavior, and operating rules;

[0226] (2) Scalability: Digital models can be disassembled, integrated, copied, modified, deleted, and so on according to the self-deduction of the digital world or the changes in the physical world;

[0227] (3) Interoperability: Both the digital model and the physical world have standard interfaces and specification definitions. Information can be exchanged between different mathematical models, between different physical terminals, and between mathematical models and physical terminals.

[0228] A digital twin is a digital model of an existing or future physical entity object. It uses actual measurement, simulation, and data analysis to perceive, diagnose, and predict the state of the physical entity object in real time. It regulates the behavior of the physical entity object through optimization and instructions, and evolves itself through mutual learning between related digital models. It also improves the decision-making of stakeholders during the life cycle of the physical entity.

[0229] The digital twin system consists of five levels: user domain, digital twin, measurement and control entity, real domain and cross-domain functional entity.

[0230] The first layer is the user domain that uses digital twins, including humans, human-machine interfaces, application software, and other digital twins;

[0231] The second layer is the digital twin corresponding to the physical entity target object. It is a digital model that reflects the characteristics of a physical object from a certain perspective, and provides three types of functions: modeling management, simulation services, and twin co-intelligence. Modeling management involves the digital model and display of physical objects, synchronization with physical object models, and operation management. Simulation services include model simulation, analysis services, report generation, and platform support. Twin co-intelligence involves the interface, interoperability, and secure access to resources such as co-intelligence twins. Modeling management, simulation services, and twin co-intelligence transmit the information required to realize the state perception, diagnosis, and prediction of physical objects.

[0232] The third layer is the measurement and control entity that connects the digital twin and the physical entity from the measurement and control domain to realize the state perception and control function of the physical object;

[0233] The fourth layer is the real physical domain where the physical entity target object corresponding to the digital twin is located. There is a transmission of measurement data flow and control information flow between the measurement and control entity and the real physical domain.

[0234] The data flow and information flow transmission between measurement and control entities, digital twins, and user domains requires the support of cross-domain functional entities such as information exchange, data assurance, and security assurance. Information exchange enables the exchange of information between digital twins through appropriate protocols. Security assurance is responsible for the authentication, authorization, confidentiality, and integrity related to the security of the digital twin system. Data assurance and security assurance are together responsible for the accuracy and integrity of the digital twin system data.

[0235] In the present invention, the JA theory is applied to analyze the residual magnetic flux in the core and the excitation surge current in the winding at different closing angles, and the influence of the phase-to-phase magnetic field on the leakage magnetic field is analyzed at the same time, and the magnetic field, current and dynamic force under multiple short-circuit impacts are accurately calculated. The source of the vibration signal of the power transformer is analyzed. Among them, the core vibration caused by magnetostriction, the core vibration caused by leakage magnetic field and the winding vibration caused by the electromotive force are the main causes of transformer vibration.

[0236] A numerical calculation model for the nonlinear dynamic buckling of elastic restrained arches under dynamic short-circuit current loads is proposed, and the mechanical properties and buckling behavior of the windings under conditions of no cumulative deformation and cumulative deformation are analyzed. Finally, a dynamic stability impact test of the windings is designed to test the dynamic buckling load of the winding coils under the action of electromagnetic force, and the correctness of the theoretical results is verified by the experimental results. A relatively complete method for solving the coupling problem after the windings produce cumulative deformation is established, and the spatial distribution and time-varying laws of the electromagnetic field, structural field, flow field and temperature field involved in the cumulative deformation of the windings under multiple short-circuit impacts are analyzed.

[0237] A three-dimensional model of the transformer including insulating pads and clamps was established using three-dimensional drawing software. Based on the coupled field theory of electromagnetic and mechanical fields, the vibration characteristics of the transformer winding under steady-state operating conditions were analyzed in terms of electromagnetic, modal, and harmonic responses. The vibration mode of the winding under typical orders was obtained, and the symmetry was good. By comparing the natural frequency and frequency response characteristics of the winding in normal and loose states, it was found that as the preload increased, the natural frequency increased, and the vibration acceleration amplitude increased. Combined with the winding vibration test, the correctness of the simulation calculation was verified, providing a basis for the vibration detection method of transformer winding faults.

[0238] A vibration signal acquisition system was built. The main function of the system is to collect vibration signals on the surface of the transformer. Piezoelectric accelerometers and vibration signal acquisition and processing equipment were selected to complete the construction of the vibration signal acquisition system. Research on the state identification algorithm of power transformer windings. The extreme learning machine neural network model is applied to winding fault diagnosis. For the sensitive problem of parameter selection, the extreme learning machine is used for optimization. The results of the example analysis show that the proposed method can effectively diagnose the winding identification of the transformer, and has better performance in efficiency and accuracy.

[0239] The digital twin model of the power transformer is established through a self-written program with independent intellectual property rights for electromagnetic, short-circuit force, short-circuit impedance, strain and winding parameter identification. The twin database is the real-time simulation data of the virtual transformer, which is embedded in the experimental platform through software writing to calculate and analyze the transformer simulation model data in real time; the data flow is a dynamic connection between the physical space and the digital space, optimizing the simulation equipment to be closer to the actual operation and maintenance situation.

[0240] The technical solution adopted by the present invention has the following beneficial effects:

[0241] (1) In this application, a three-dimensional transformer vibration simulation model is established to perform circuit, magnetic field, structural force field and flow field coupling calculations. In the electromagnetic field simulation, the current and voltage curves of the winding and the magnetic field distribution of the transformer are calculated; the deformation trend and fault type of the winding are solved in the structural force field; the vibration characteristics of the measuring points on the surface of the oil tank are calculated in the flow field, so as to obtain the relationship between the deformation and vibration of the transformer winding and the reactance characteristics. The short-circuit damage process of the transformer can be analyzed, which is of great significance for understanding the damage mechanism of the transformer. The short-circuit resistance of the power transformer can be tested from multiple aspects to improve the accuracy of the test. Relevant tests can be carried out in the transformer equipment factory or relevant scientific research institutions in conjunction with the transformer short-circuit test to understand the transformer short-circuit law, which is of great guiding significance for optimizing the design and improving the product's short-circuit resistance.

[0242] (2) Conduct field tests on transformers to measure the short-circuit reactance values ​​under various winding deformation conditions, and place vibration sensors to measure the vibration of transformer windings under various deformation conditions. Combined with the results of model simulation, analyze the relationship between transformer winding deformation and short-circuit reactance and vibration. Combined with the analysis of real transformer short-circuit tests, obtain the transformer short-circuit damage process and related mechanisms. On this basis, carry out optimization research on the transformer short-circuit resistance verification method to achieve lean verification of transformer short-circuit resistance. Transformer manufacturers can use this software to guide design and manufacturing to ensure that the products produced meet the requirements of power grid operation.

[0243] (3) The digital twin model of the power transformer is established through a self-written program with independent intellectual property rights for electromagnetic, short-circuit force, short-circuit impedance, strain and winding parameter identification. The twin database is the real-time simulation data of the virtual transformer. It is embedded in the experimental platform through software writing, and the transformer simulation model data is calculated and analyzed in real time; the data flow is a dynamic connection between the physical space and the digital space, and the optimized simulation equipment is closer to the actual operation and maintenance situation.

[0244] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the schemes disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed in this disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the claims.

Claims

1. A method for detecting the short-circuit resistance of a power transformer based on digital twin technology, characterized in that: include: Step 1: Establish a multi-physics field simulation model of the transformer short-circuit dynamic process, taking into account the leakage magnetic field, electromotive force and displacement deformation during short-circuit and reclosing multiple impacts; Step 2: Improve the transformer model, including the construction of radial, axial and clamp models, and carry out simulation calculation and analysis of the electromagnetic field, electrodynamic force and displacement deviation of the transformer winding model; Step 3: Analyze the electrical and vibration data of the transformer short-circuit test, establish a three-dimensional transformer simulation model based on the electrical and vibration data of the transformer short-circuit test, and calculate the transformer short-circuit reactance value using the measured values ​​of the voltage and current of the primary and secondary windings of the winding; Step 4: Based on the above theoretical and mathematical model analysis, a transformer is selected for winding test, relevant parameters of the winding when various faults occur are collected, and the relationship between short-circuit reactance and winding deformation is analyzed; Step 5: Use digital twin technology to perform high-fidelity modeling of the transformer, and use the digital twin transformer model to perform simulation analysis under various combinations of deformations and various conditions; organize the simulation data to form a database, and detect the power transformer entity based on the database to achieve the detection of the power transformer's short-circuit resistance.

2. The method for detecting the short-circuit resistance of a power transformer based on digital twin technology according to claim 1 is characterized in that: The step 1 includes: Step 1.1: Establishment of circuit-magnetic field coupling mathematical model: The accurate calculation of transformer electromagnetic field, short-circuit current and electromagnetic force is the basis of winding stress field. In practical applications, most electromagnetic devices are excited by voltage sources. Therefore, in the calculation of transformer transient electromagnetic field, the constraint equation of the voltage connecting the winding and the external circuit is considered. The transient electromagnetic field equation is: In the formula, J is the current density in the coil, and there is a relationship J = nI / S, n is the number of turns of the coil, S is the total cross-sectional area of ​​the coil, and I is the branch current; Winding circuit equation: Where, {u}={-u1 u2} T , {e}={e1 e2} T ,{i}={i1 i2} T , resistance array [R] = diag [R1 R2], leakage impedance array [L] = diag [L 1σ L 2σ ]; If the effective length of each coil turn is Lef, the number of turns of the winding is nc, and the winding area is divided into ne units, then the phase potential of the winding is: The spatial discrete equation of the transient process solved by coupling the transient eddy current field and the winding external circuit equation is: Step 1.2: Simulation of core hysteresis characteristics and establishment of circuit coupling mathematical model: Under multiple no-load closing conditions, the synthetic flux Φ of the transient process n The expression is: After half a cycle of closing (ωt=π), the synthetic magnetic flux reaches its maximum value. When the initial phase angle α=0 at the moment of closing, the core magnetic flux reaches its maximum value. The JA model is used to describe the nonlinear magnetization characteristics of the core, and a "circuit-hysteresis" coupling calculation model is established to calculate and analyze the impact current. When the JA theory is used to analyze and calculate the transformer excitation current, it is necessary to combine the magnetic circuit equation of the no-load closing dynamic process and the circuit equation describing its external characteristics with the JA model. Taking the equivalent circuit of rated operating conditions as an example, it is written as follows: Its matrix form expression is: The magnetic circuit equation of the short-circuit reclosing dynamic process, the circuit equation of the external characteristics and the JA model are combined and written as follows: In the formula, H is the magnetic field intensity; M is the magnetization intensity; l is the magnetic path length; N1 is the number of turns of the primary winding of the transformer; i1 is the current of the primary winding of the transformer; N2 is the number of turns of the secondary winding of the transformer; i2 is the current of the secondary winding of the transformer; s is the cross-sectional area of ​​the transformer core; B is the core magnetic flux density; R2 is the sum of the secondary resistance of the transformer and the load resistance; L2 is the sum of the secondary winding inductance and the load inductance; In the above formula Convert to In the form of:

3. The method for detecting the short-circuit resistance of a power transformer based on digital twin technology according to claim 1 is characterized in that: The step 2 includes: When the transformer is subjected to short-circuit impact, the inner winding is subjected to inward pressure, the outer winding is subjected to outward tensile force, and the radial compression force on the inner winding is much greater than the radial tension force on the outer winding; under multiple short-circuit conditions, the distribution of radial compression force on the inner winding in the circumferential direction is different from that under single short-circuit conditions, and is unevenly distributed. The radial force on the winding is the largest in the middle of the winding axial height and between the two stays near the 0° position in the window, and plastic deformation is most likely to occur; after the winding undergoes elastic-plastic deformation, cumulative deformation is bound to occur on the winding after multiple short circuits of the same type; Elastic limit bending moment M re The expression is: In the formula, the span length of the winding between the two struts is l, the cross-sectional thickness is d, and the cross-sectional height is h. Since the inner winding radius is much larger than the span length, the winding between the two struts can be equivalent to a straight beam, and the axial force of the winding is equivalent to a uniformly distributed load P r , let the deflection of the winding in the y direction be w, and the downward direction be positive; The bending moment corresponding to the curvature k of the winding is: When ξ→0, the entire winding cross section enters the plastic stage, and |k|→∞, it can be deduced that: M rs =1.5M re Where M re is the plastic limit bending moment; When M r >M rs When , the winding is in the completely plastic stage. As the bending moment increases further, the elastic area of ​​the cross section becomes zero, the winding cross section enters the plastic state, and the stress of the entire cross section reaches the limit value. Under multiple short-circuit shocks, plastic deformation will accumulate, and it is necessary to establish an elastic-plastic mechanical model of the winding. During the modeling process, the stress-strain curve is nonlinear, and the cumulative deformation of the winding and the displacement of the coil are calculated.

4. The method for detecting the short-circuit resistance of a power transformer based on digital twin technology according to claim 1 is characterized in that: Step 3 includes: The dynamic stability of the transformer is not only related to the various support positions, but also to its own structure and temperature. Each short circuit will cause different residual deformations in the transformer, which will accumulate in multiple short circuits and affect the original structure of the winding. Step 3.1: Dynamic stability of winding without cumulative deformation: When the short-circuit electromotive force acts on the winding coil, the winding can be analyzed as an elastically constrained arc arch model between two stays, as shown below. Its dimensionless kinetic energy T and total potential energy U are: Where: ρ is the material density, A is the cross-sectional area, R is the radius of the arch, Ex is the elastic modulus of the arch, I is the second-order moment of the arch section, α is the half-span central angle, g is the radial motion velocity, w is the axial motion velocity, rx is the radius of gyration of the arch section, εm is the membrane strain of the arch, G is the radial displacement, q is the uniform electromagnetic force, and γ is the constraint flexibility parameter of the arch foot; Step 3.2: Dynamic stability of windings with accumulated deformation: After multiple short-circuit shocks, the transformer windings in operation may produce cumulative deformation, which changes the geometric configuration of the arch axis and forms initial defects. The mechanical properties of the arch are significantly affected by defects, and serious defects will affect the dynamic stability of the transformer. The dimensionless kinetic energy T and total potential energy U of the possible structures of several windings after cumulative deformation are: Where: W is the axial displacement, V is the volume, N is the axial pressure of the arch, and M is the bending moment of the arch.

5. The method for detecting the short-circuit resistance of a power transformer based on digital twin technology according to claim 1 is characterized in that: The step 3 also includes: Step 3.3, transformer short circuit test: The scheme of testing leakage magnetic field using the principle of electromagnetic induction, the magnetic density is calculated according to the following formula: Where U is the voltage peak; W is the number of turns of the sensor coil (take 600 turns); S is the area of ​​the sensor coil; Since the potential of the winding in the short-circuit electrodynamic test is as high as tens of thousands of volts, the winding force test is as small as mm level, and the short-circuit electrodynamic force has a dynamic process acquisition, through theoretical calculation and experimental verification, it is determined that the piezoelectric film short-circuit electrodynamic test system is used. Under the action of the huge short-circuit electromagnetic force, the winding will deform, causing the piezoelectric film sensor to be under pressure, and then generate a corresponding electrical signal; When the piezoelectric film is subjected to force in the x direction, thickness deformation will occur and polarization will occur. The charge qx generated in the x direction is proportional to the force Fx: q x =d 11 F x Where, qx is the piezoelectric coefficient; d11 is the charge on the plane of the x-axis; Fx is the force applied along the x-axis; Then, a laser sensor is used to directly measure the radial vibration of the middle coil of the winding in the model.

6. The method for detecting the short-circuit resistance of a power transformer based on digital twin technology according to claim 1 is characterized in that: The step 3 also includes: Step 3.4: Establishment of mathematical model of winding axial vibration: The transformer winding can be regarded as a plurality of coils, which are separated by insulating pads and pressed together by the upper and lower pressure plates and iron yoke pads. Under the action of radial electromagnetic force, the radial vibration dynamics model of the winding is established. The acceleration of the winding vibration is: Step 3.5: Extraction of transformer vibration signal features based on wavelet packet transform: The vibration signal is decomposed by wavelet packet decomposition, and then the frequency band energy feature vector of the vibration signal is constructed to perform transformer fault diagnosis. The specific method is as follows: 1) Perform three-layer wavelet packet decomposition on the transformer vibration signal collected in the experiment and simulation; 2) Reconstruct the sub-band sequences within the 8 frequency bands obtained in step 1) to obtain 8 wavelet packet reconstructed signals, each of which contains the information of each frequency band of the original signal from low to high, and fully reflects the operating status of the transformer system; 3) Obtain the energy entropy of 8 wavelet packets, and construct a feature vector from the energy entropy obtained after decomposing the 8 wavelet packets, denoted as T, then: T=[H 3，0 ,H 3，1 ,H 3，2 ,H 3，3 ,····H 3，7 ] Normalize the eigenvectors:

7. The method for detecting the short-circuit resistance of a power transformer based on digital twin technology according to claim 1 is characterized in that: The step 4 includes: Transformer fault diagnosis analysis based on KELM. The extreme learning machine is a typical single hidden layer feedforward neural network. Since the traditional feedforward neural network needs to adjust all network parameters, its learning efficiency is far from meeting the needs of practical applications. The kernel extreme learning process can be described as: given a training set and kernel function K(u,υ), calculate the output function: The calculated wavelet packet decomposition energy entropy corresponding to the winding vibration signal is used as the input of KELM and divided into a training group and a test group in a ratio of 2:1; According to the above principles, the population size, maximum evolutionary generation, inertia weight, acceleration factors c1 and c2 of the particle swarm algorithm are set to perform transformer fault diagnosis.

8. The method for detecting the short-circuit resistance of a power transformer based on digital twin technology according to claim 1 is characterized in that: The step 5 includes: A digital twin model of the power transformer is established through a self-compiled program for electromagnetic, short-circuit force, short-circuit impedance, strain and winding parameter identification; the twin database is the real-time simulation data of the virtual transformer, which is embedded in the experimental platform through software writing to calculate and analyze the transformer simulation model data in real time; the data flow is a dynamic connection between the physical space and the digital space, optimizing the simulation equipment to be closer to the actual operation and maintenance situation.

Images