Transformer excitation inrush current blocking and internal fault identification method and system

Abstract

The application provides a transformer excitation inrush current blocking and internal fault identification method and system, which synchronously collects three-phase currents of a primary side and a secondary side of the transformer, performs ratio conversion and phase compensation on the secondary side current; performs power-invariant Clarke transformation on the three-phase currents of the two sides of the transformer respectively, and constructs modal differential current and modal braking current in a unified alpha-beta modal coordinate; then, the same wavelet base and the same decomposition scale are adopted to synchronously decompose the differential channel and the braking channel, and the differential operation energy E op and the braking restraint energy E res are constructed in the same sliding time window; finally, a joint criterion is established according to the starting threshold E th and the dynamic braking boundary K·E res , so that the excitation inrush current and external transient blocking and internal fault identification are realized. Through the unified modal coordinate, the double-channel same-scale processing and the relative energy constraint, the application reduces the influence of three-phase coupling, common transient components and single-point noise on the discrimination result.

Description

Technical Field

[0001] This invention relates to the field of transformer transient signal processing technology, specifically to a method and system for transformer inrush current blocking and internal fault identification. Background Technology

[0002] Transformers are crucial equipment in the power transmission and distribution system. During transient processes such as no-load closing, external fault clearing, and sudden voltage changes in the system, the magnetization state of the transformer core may change drastically, leading to significant aperiodic components, high-order harmonic components, and high-frequency transient abrupt changes in the excitation current. These transient currents may be misinterpreted as internal faults by differential protection devices, causing malfunctions and affecting the reliability of both the transformer and the power grid.

[0003] Existing methods for identifying and blocking inrush current in transformers mainly include second harmonic restraint, waveform symmetry criteria, discontinuity angle criteria, Fourier transform spectrum analysis, and data-driven identification methods based on artificial intelligence. While these methods can distinguish between inrush current and internal fault current to some extent, they still have shortcomings.

[0004] First, traditional second harmonic braking methods rely on specific harmonic content. When modern transformer core materials are improved, the amplitude of inrush current decreases, or the transient process becomes more complex, the second harmonic content may decrease, leading to instability in the protection blocking criteria. Second, Fourier transform methods are suitable for steady-state or quasi-steady-state periodic signal analysis, while transformer inrush current and internal fault current are both strongly non-stationary transient signals. Simple frequency domain analysis cannot accurately reflect the timing and energy concentration of transient changes. Third, artificial intelligence recognition methods typically rely on training with a large number of samples, and the model's generalization ability is greatly affected by the coverage of the training data. When on-site operating conditions, transformer capacity, wiring method, sampling frequency, or current transformer characteristics change, the reliability of model recognition may decrease. Fourth, existing methods often directly process three-phase currents, failing to fully utilize the modal decoupling characteristics of the three-phase system. They are easily affected by zero-sequence components, three-phase coupling relationships, and imbalance factors, resulting in insufficient accuracy in transient feature extraction. In addition, existing schemes based on wavelet or wavelet packet analysis usually directly process the three-phase differential current of A, B, and C, and mainly rely on the high-frequency components and thresholds of a single differential channel for judgment, making it difficult to simultaneously characterize the transient intensity of internal fault differences and the common transient braking level under external impact.

[0005] Therefore, it is necessary to propose a transformer inrush current blocking and internal fault identification method that does not rely on sample training, can simultaneously construct differential transient energy and braking transient energy under unified modal coordinates, and uses the relative relationship between the two for protection judgment. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for transformer inrush current blocking and internal fault identification, which solves the problems of existing wavelet discrimination methods directly processing three-phase differential current in the phase coordinate system, the susceptibility of single differential channel characteristics to three-phase coupling and common transient impact, and the difficulty of adapting fixed discrimination boundaries to changes in operating conditions.

[0007] The technical solution to achieve the purpose of this invention is as follows:

[0008] A method for transformer inrush current blocking and internal fault identification includes:

[0009] Step S1: Synchronously collect the three-phase current on the primary side and the three-phase current on the secondary side of the transformer, and perform reduction and phase compensation on the three-phase current on the secondary side according to the transformer ratio, connection group and phase relationship.

[0010] Step S2: Perform power-invariant Clarke transformation on the primary three-phase current and the reduced secondary three-phase current respectively to obtain the corresponding α-mode current and β-mode current on both sides.

[0011] Step S3: Construct modal differential current and modal braking current under the same modal coordinates, and use the same wavelet basis, decomposition scale and boundary processing method to perform synchronous discrete wavelet decomposition on modal differential current and modal braking current to extract high-frequency detail coefficients;

[0012] Step S4: Within the same sliding time window, perform square integration on the α and β mode high-frequency detail coefficients of the modal differential current and modal braking current to obtain the differential operation energy E. op and braking constraint energy E res ;

[0013] Step S5, based on the energy activation threshold E th Construct a joint energy criterion with braking coefficient K, using K·E res As a dynamic braking boundary that varies with the current common transient level;

[0014] Step S6: Output a protection blocking signal or an internal fault action signal based on the joint energy criterion and the continuous sampling point confirmation condition.

[0015] Furthermore, the modal differential current and modal braking current are constructed as follows:

[0016] i μ,diff [n] = i μ,1 [n] - i ′ μ,2 [n];

[0017] i μ,res [n] = (|i μ,1 [n]| + |i′ μ,2 [n]|) / 2;

[0018] Among them, i μ,diff [n] represents the μ-mode differential current at the nth sampling point, i μ,1 [n] represents the primary side μ-mode current, i μ,res [n] represents the μ-mode braking current at the nth sampling point, i μ,2 ′[n] is the secondary side μ-mode current after reduction, phase compensation and Clarke transformation, μ∈{α,β}.

[0019] Furthermore, the wavelet basis is the Daubechies wavelet.

[0020] Furthermore, the Daubechies wavelet is a db4 wavelet.

[0021] Furthermore, the differential operation energy E op and braking constraint energy E res The α-β joint mode energies are calculated using the following formulas:

[0022] E op [n] = Σ k=n-Tw+1 n {(d α,diff [k]) 2 + (d β,diff [k]) 2};

[0023] E res [n] = Σ k=n-Tw+1 n {(d α,res [k]) 2 + (d β,res [k]) 2};

[0024] Among them, T w The sliding time window contains the number of sampling points, where n represents the discrete sampling point number of the original three-phase current or modal current, and d... α,diff [k]、d β,diff [k] represents the high-frequency detail coefficients obtained after the α and β mode differential currents are decomposed by the same wavelet decomposition, respectively, and d α,res [k]、d β,res [k] represents the high-frequency detail coefficients obtained after the α and β mode braking currents are decomposed by the same wavelet decomposition.

[0025] Furthermore, the energy activation threshold E th Based on the average differential operation energy μ within the background window before transient start-up bgand standard deviation σ bg Confirmed as:

[0026] E th = μ bg + λσ bg ;

[0027] Where λ is a set coefficient.

[0028] Furthermore, step S6 specifically includes: only when E op >E th And E op K·E res When M sampling points are satisfied consecutively, the current transient state is determined to have internal fault characteristics, and a protection action signal is output. When E op [n] is not greater than E th or not greater than K·E res [n], or when the number of sampling points that continuously meet the conditions is less than M, output a protection blocking signal.

[0029] Furthermore, the output signal also includes the differential operation energy E. op Braking constraint energy E res or effective energy margin .

[0030] Furthermore, the effective energy margin for:

[0031] ;

[0032] Where max represents the maximum value function.

[0033] A transformer inrush current blocking and internal fault identification system, executing the method, includes a three-phase current sampling unit, a reduction and phase compensation unit, a Clarke mode transform unit, a mode differential / braking construction unit, a dual-channel wavelet feature extraction unit, a joint energy calculation unit, a threshold and braking boundary unit, and a protection output unit, wherein:

[0034] The three-phase current sampling unit is used to synchronously acquire the three-phase current signals of the primary and secondary sides of the transformer.

[0035] The reduction and phase compensation unit is used to reduce and compensate the phase of the three-phase current on the secondary side according to the transformer ratio, connection group and phase relationship;

[0036] The Clarke mode transformation unit performs power-invariant Clarke transformation on the primary three-phase current and the reduced secondary three-phase current respectively to obtain the corresponding α-mode current and β-mode current on both sides.

[0037] Modal differential / braking construction unit, used to construct modal differential current and modal braking current in the same modal coordinates;

[0038] The dual-channel wavelet feature extraction unit is used to synchronously discretely decompose the modal differential current and modal braking current using the same wavelet basis, decomposition scale and boundary processing method to extract high-frequency detail coefficients;

[0039] The joint energy calculation unit is used to perform square integration on the α and β mode high-frequency detail coefficients of the modal differential current and modal braking current within the same sliding time window, to obtain the differential operation energy E. op and braking constraint energy E res ;

[0040] Threshold and braking boundary unit, used to initiate the action based on the energy threshold E th Construct a joint energy criterion with braking coefficient K, using K·E res As a dynamic braking boundary that varies with the current common transient level;

[0041] The protection output unit is used to output protection blocking signals or internal fault action signals based on the joint energy criterion and the continuous sampling point confirmation conditions.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] 1) This invention performs power-invariant Clarke transformation on the three-phase currents on both sides of the transformer, constructs differential and braking channels under unified modal coordinates, and reduces the impact of three-phase coupling, zero-sequence components and phase changes on transient feature extraction.

[0044] 2) The differential channel and the braking channel use the same wavelet basis, the same decomposition scale and the same sliding time window to form a common energy quantity with consistent modal basis, frequency band range and time scale, which can be directly constrained by relative energy.

[0045] 3) This invention uses α-β joint modal energy and sliding window square integral to reduce the impact of single phase, single mode, single point noise or instantaneous spikes on the discrimination results.

[0046] 4) This invention does not rely on a large number of training samples and does not require pre-training of artificial intelligence models, thus having good engineering adaptability and real-time performance.

[0047] 5) The present invention uses the energy start-up threshold and dynamic braking boundary as a joint criterion, which can maintain the sensitivity of action when the differential energy of internal faults increases significantly, and improve the reliability of blocking when the excitation inrush current, external faults or common transient components are enhanced.

[0048] 6) This invention can be embedded in existing microcomputer protection devices or intelligent substation protection and control devices, requiring only algorithm processing based on the sampled current signal, which facilitates engineering promotion. Attached Figure Description

[0049] Figure 1 This is a block diagram of the system structure of the present invention.

[0050] Figure 2 This is a flowchart of the method of the present invention.

[0051] Figure 3 A schematic diagram of the construction of dual-channel homogeneous transient energy.

[0052] Figure 4 This is a schematic diagram of the joint energy criterion. Detailed Implementation

[0053] The present application will be further described in detail below with reference to the accompanying drawings, so that those skilled in the art can better understand and implement the present invention, but the embodiments described are not intended to limit the present invention.

[0054] Example 1

[0055] This invention proposes a transformer protection method based on modal decoupling and differential-braking dual-channel energy constraint. The method first acquires the three-phase currents on the primary and secondary sides of the transformer, and performs reduction and phase compensation on the secondary current. Then, it performs power-invariant Clarke transform on the three-phase currents on both sides respectively, constructing modal differential current and modal braking current under a unified α-β modal coordinate system. Subsequently, it uses the same wavelet basis, the same decomposition scale, and the same sliding time window to synchronize the two channels, obtaining the differential operation energy E. op and braking constraint energy E res Finally, based on the energy activation threshold E th With dynamic braking boundary K·E res A joint criterion is established, and the condition for output protection blocking or action signal is confirmed using M consecutive sampling points. The above steps form an overall technical relationship of "unified modal coordinates—dual-channel homogeneous characteristics—joint transient energy—relative energy constraint," rather than a simple parallel arrangement of several signal processing methods. The key technical point of this invention is:

[0056] 1) How to map the three-phase currents on both sides of the transformer to a unified orthogonal modal coordinate system, while preserving the common transient information on both sides and reducing the influence of three-phase coupling and zero-sequence components;

[0057] 2) How to use the same frequency band and the same time scale for high-frequency transient feature extraction for modal differential channels and modal braking channels so that the two types of energy can be directly compared;

[0058] 3) How to construct α-β combined transient energy to reduce the sensitivity of single phase or single mode characteristics to closing angle, fault phase and coordinate axis direction;

[0059] 4) How to utilize energy start-up threshold, dynamic braking boundary and continuous sampling point confirmation conditions to achieve rapid identification of excitation inrush current, external transient reliable lockout and internal faults.

[0060] like Figures 2-3 The method of the present invention specifically includes the following steps.

[0061] Step S1: Acquire three-phase current signals and perform compensation.

[0062] The three-phase current on the primary side and the three-phase current on the secondary side of the transformer are synchronously acquired using a current transformer or an electronic current sensor, and are denoted as primary current i. A1 i B1 i C1 And the secondary side currents iA2, iB2, and iC2. Based on the transformer ratio, connection group, and phase compensation relationship, the secondary side currents are reduced and phase corrected to obtain i A2′ i B2′ i C2′ .

[0063] Step S2: Perform Clarke mode transformation on both sides of the current.

[0064] Perform power-invariant Clarke transforms on the primary three-phase current and the normalized secondary three-phase current respectively: [i α,j i β,j i 0,j ] T = T Clarke · [i A,j i B,j i C,j ] T , j ∈ {1, 2′}.

[0065] Among them, T Clarke This is the power-invariant Clarke transformation matrix. Using the same transformation matrix for both currents preserves both differential and common transient components in a unified modal coordinate system.

[0066] The reduction and Clarke transformation of the secondary current include specific steps.

[0067] Defined as "the secondary side after reduction and phase compensation" "Modal current" has the following formula:

[0068] Let the three-phase sampling current vector on the secondary side be:

[0069]

[0070] The secondary three-phase current vectors after reduction and phase compensation are:

[0071]

[0072] in: The equivalent reduction factor is determined by the transformer's rated current ratio, the current transformer ratios on both sides, and the selected reduction reference side. For wiring group compensation matrix; Indicates the transformer connection group; This represents the reduced and phase-compensated secondary-side three-phase current vectors. A power-invariant Clarke transform is then applied to the reduced and phase-compensated secondary-side three-phase currents:

[0073]

[0074] The power-invariant Clarke transformation matrix is:

[0075]

[0076] therefore:

[0077]

[0078] in, This indicates the secondary side three-phase current obtained after sequentially undergoing transformation ratio reduction, connection group phase compensation, and Clarke transformation. Modal current.

[0079] Step S3: Construct the modal differential channel current and the modal braking channel current.

[0080] For μ∈{α,β}, the modal differential current and modal braking current are respectively: i μ,diff [n] = i μ,1 [n] -i ′ μ,2 [n], i μ,res [n] = (|i μ,1 [n]| + |i ′ μ,2 [n]|) / 2.

[0081] Where n represents the discrete sampling point number; i μ,diff [n] represents the μ-mode differential current at the nth sampling point, used to characterize the degree of imbalance between the mode currents on both sides of the transformer; i μ,res[n] represents the μ-mode braking current at the nth sampling point, used to characterize the intensity of the common current and common transient component on both sides.

[0082] Modal differential current characterizes the degree of imbalance between the currents on both sides, while modal braking current characterizes the common transient level on both sides.

[0083] Step S4: Simultaneously extract high-frequency features of dual-channel wavelet.

[0084] The Daubechies 4 wavelet (db4 wavelet) was used to perform discrete wavelet decomposition on the α and β mode differential current and the α and β mode braking current using the same wavelet basis, the same number of decomposition levels, and the same boundary extension method, respectively, to obtain the high-frequency detail coefficients of the differential channel and the braking channel. The identical processing parameters ensure that the high-frequency characteristics of the two channels correspond to the same frequency band, providing a comparable basis for subsequent relative energy constraints.

[0085] Step S5: Construct the α-β joint transient energy.

[0086] Let the length of the sliding time window be T. w For each sampling point, the differential operation energy and braking constraint energy are calculated within the same sliding time window:

[0087] E op [n] = Σ k=n-Tw+1 n {(d α,diff [k]) 2 + (d β,diff [k]) 2};

[0088] E res [n] = Σ k=n-Tw+1 n {(d α,res [k]) 2 + (d β,res [k]) 2}

[0089] Where, d α,diff [k]、d β,diff [k] represent the high-frequency detail coefficients obtained after db4 discrete wavelet decomposition of the α and β mode differential currents, respectively; d α,res [k]、d β,res [k] represents the high-frequency detail coefficients obtained after wavelet decomposition of the α and β mode braking currents, respectively; k represents the wavelet detail coefficient index. The above wavelet coefficients are directly related to the modal differential current and the modal braking current. For example, the first-order discrete wavelet decomposition can be expressed as:

[0090] ;

[0091] ;

[0092] in, These are the coefficients of the db4 high-pass analysis filter. Because the standard discrete wavelet transform involves decimation by a factor of two, the original sampling point index n and the wavelet coefficient index k are not in one-to-one correspondence.

[0093] The combined energy of α and β orthogonal modes can reduce the sensitivity of a single mode to fault phase and coordinate axis direction changes.

[0094] Step S6: Establish joint energy criteria and continuity confirmation criteria.

[0095] Set energy start-up threshold E th And braking coefficient K, in terms of K·E res [n] serves as the dynamic braking boundary. When E op [n] M consecutive sampling points are simultaneously greater than E th and K·E res When [n], the current transient state is determined to have internal fault characteristics and a protection action signal is output; when E op [n] is not greater than E th or not greater than K·E res If [n], or if the number of sampling points that continuously meet the conditions is less than M, a protection blocking signal is output. The blocking signal can be used for inrush current, normal closing transients, or external disturbances.

[0096] Step S7: Output the identification and energy assessment results.

[0097] Based on the judgment result, the output includes a protection lockout signal, an internal fault identification signal, a protection action signal, and differential operation energy E. op Braking constraint energy E res Energy margin E op -K·E res At least one of them.

[0098] A schematic diagram of the joint energy criterion is shown below. Figure 4 The horizontal axis represents time, and the vertical axis represents transient energy. The graph includes the differential operation energy E. op Dynamic braking boundary K·E res Energy start-up threshold E th In addition to the protection of the locked area and the protection of the action area.

[0099] The internal fault identification signal can be clearly defined as a binary flag indicating whether an internal fault has been identified. When E op M consecutive sampling points are simultaneously greater than E th and K·E res At that time, the internal fault identification signal F int =1; otherwise, the internal fault identification signal Fint =0. This identification flag belongs to the algorithm's judgment result.

[0100] The protection action signal is the action control signal output to the protection execution circuit based on the internal fault identification signal.

[0101] like Figure 1 The method of the present invention can be implemented by a transformer inrush current blocking and internal fault identification system, the system comprising the following units:

[0102] 1) Three-phase current sampling unit 1, used to synchronously acquire the three-phase current signals of the primary and secondary sides of the transformer;

[0103] 2) Calculation and phase compensation unit 2, used to calculate and compensate the phase of the three-phase current on the secondary side according to the transformer ratio, connection group and phase relationship;

[0104] 3) Clarke mode conversion unit 3, used to convert the three-phase currents on both sides of the transformer into α-mode current, β-mode current and zero-sequence mode current respectively;

[0105] 4) Modal differential / braking construction unit 4, used to construct differential channel current and braking channel current under the same modal coordinates;

[0106] 5) Dual-channel wavelet feature extraction unit 5, used to extract high-frequency detail coefficients of differential channel and braking channel using the same wavelet basis, decomposition scale and boundary processing method;

[0107] 6) Joint energy calculation unit 6, used to calculate differential operation energy Eop and braking constraint energy Eres within the same sliding time window;

[0108] 7) Threshold and braking boundary unit 7, used to form dynamic braking boundary K·Eres based on energy start threshold Eth and braking coefficient K;

[0109] 8) Protection output unit 8, used to output protection blocking signal or internal fault action signal according to the joint energy criterion and continuous sampling point confirmation conditions.

[0110] This invention reduces the impact of three-phase coupling, common transient components, and single-point noise on the discrimination results by using unified modal coordinates, dual-channel same-scale processing, and relative energy constraints.

[0111] Example 2

[0112] This embodiment uses a three-phase transformer as an example. During the no-load closing process of the transformer, the protection device synchronously collects the three-phase current signals of the primary and secondary sides of the transformer. The protection device first performs reduction processing on the secondary side current according to the transformer ratio, connection group, and phase compensation relationship, so that the primary and secondary side currents are on the same reference side.

[0113] The protection device performs power-invariant Clarke transformation on the primary three-phase current and the reduced secondary three-phase current to obtain the corresponding α-mode, β-mode, and zero-sequence mode currents on both sides. Subsequently, α- and β-mode differential currents and mode braking currents are constructed under the same modal coordinates. This process gives the differential channel and braking channel a unified modal basis and retains common transient information on both sides.

[0114] Then, the db4 wavelet is used to perform a single-level discrete wavelet decomposition on the α and β mode differential currents and the modal braking current, extracting the first-level high-frequency detail coefficients of the two channels respectively. Within a preset sliding time window T... w Within this, the differential operation energy E is obtained by integrating the squares of the high-frequency detail coefficients of the α and β modes. op [n] and braking constraint energy E res [n].

[0115] Based on this, set the energy start-up threshold E. th And braking coefficient K, in terms of K·E res [n] forms a dynamic braking boundary. When the transformer is closed under no-load conditions and inrush current is generated or an external disturbance occurs, if E op [n] does not consecutively satisfy the condition that it is greater than E. th and K·E res If the condition is met, the protection device will output a protection lockout signal.

[0116] When an internal fault occurs in the transformer, the high-frequency detail coefficients of the modal differential channel increase significantly, and the α-β combined differential operation energy E op [n] increases rapidly. If E op [n] M consecutive sampling points are simultaneously greater than the energy threshold E th and dynamic braking boundary K·E res If [n] is detected, the protection device determines that there is an internal fault in the transformer and outputs a protection action signal.

[0117] In this embodiment, the sampling frequency can be set to 5 kHz, 10 kHz or higher; the wavelet basis is preferably the db4 wavelet; the sliding window length T w The sampling frequency and protection action time requirements can be set to half a cycle, one cycle, or several milliseconds window; the number of continuous confirmation points M can be adjusted according to the action speed and noise immunity requirements.

[0118] Energy start-up threshold E th The settings can be adjusted based on transformer capacity, rated current, sampling noise level, and protection sensitivity, or corrected online based on background energy before transient start-up; the braking coefficient K can be set according to the safety requirements of differential protection. For transformers of different capacity levels, voltage levels, and connection groups, the above parameters can be adjusted, but the core processing flow remains unchanged, namely: unified mode transformation of three-phase currents on both sides, dual-channel construction of mode differential / braking, extraction of wavelet features at the same scale, construction of α-β joint transient energy, and joint discrimination based on start-up threshold and dynamic braking boundary.

[0119] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and incorporate common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

Claims

1. A method for transformer inrush current blocking and internal fault identification, characterized in that, include: Step S1: Synchronously collect the three-phase current on the primary side and the three-phase current on the secondary side of the transformer, and perform reduction and phase compensation on the three-phase current on the secondary side according to the transformer ratio, connection group and phase relationship. Step S2: Perform power-invariant Clarke transformation on the primary three-phase current and the reduced secondary three-phase current respectively to obtain the corresponding α-mode current and β-mode current on both sides. Step S3: Construct modal differential current and modal braking current under the same modal coordinates, and use the same wavelet basis, decomposition scale and boundary processing method to perform synchronous discrete wavelet decomposition on modal differential current and modal braking current to extract high-frequency detail coefficients of α and β modes; Step S4: Within the same sliding time window, perform square integration on the α and β mode high-frequency detail coefficients of the modal differential current and modal braking current to obtain the differential operation energy E. op and braking constraint energy E res ; Step S5, based on the energy activation threshold E th Construct a joint energy criterion with braking coefficient K, using K·E res As a dynamic braking boundary that varies with the current common transient level; Step S6: Output a protection blocking signal or an internal fault action signal based on the joint energy criterion and the continuous sampling point confirmation condition.

2. The transformer inrush current blocking and internal fault identification method according to claim 1, characterized in that, The modal differential current and modal braking current are constructed as follows: i μ,diff [n] = i μ,1 [n] - i ′ μ,2 [n]; i μ,res [n] = (|i μ,1 [n]| + |i ′ μ,2 [n]|) / 2; Among them, i μ,diff [n] represents the μ-mode differential current at the nth sampling point, i μ,1 [n] represents the primary side μ-mode current, i μ,res [n] represents the μ-mode braking current at the nth sampling point, i μ,2 ′[n] is the secondary side μ-mode current after reduction, phase compensation and Clarke transformation, μ∈{α,β}.

3. The transformer inrush current blocking and internal fault identification method according to claim 2, characterized in that, The wavelet basis is the Dobercy wavelet.

4. The transformer inrush current blocking and internal fault identification method according to claim 3, characterized in that, The Dobesy wavelet is a db4 wavelet.

5. The transformer inrush current blocking and internal fault identification method according to claim 1, characterized in that, The differential operation energy E op and braking constraint energy E res The α-β joint mode energies are as follows: AND op [n] = Σ k=n-Tw+1 n {(d α,diff [k]) 2 + (d β,diff [k]) 2 }; AND res [n] = Σ k=n-Tw+1 n {(d α,res [k]) 2 + (d β,res [k]) 2 }; Among them, T w The sliding time window contains the number of sampling points, where n represents the discrete sampling point number of the original three-phase current or modal current, and d... α,diff [k]、d β,diff [k] represents the high-frequency detail coefficients obtained after the α and β mode differential currents are decomposed by the same wavelet decomposition, respectively, and d α,res [k]、d β,res [k] represents the high-frequency detail coefficients obtained after the α and β mode braking currents are decomposed by the same wavelet decomposition.

6. The transformer inrush current blocking and internal fault identification method according to claim 5, characterized in that, The energy activation threshold E th Based on the average differential operation energy μ within the background window before transient start-up bg and standard deviation σ bg Confirmed as: E th = m bg + ls bg ; Where λ is a set coefficient.

7. The transformer inrush current blocking and internal fault identification method according to claim 1, characterized in that, Step S6 specifically includes: only when E op >E th And E op K·E res When M sampling points are satisfied consecutively, the current transient state is determined to have internal fault characteristics, and a protection action signal is output. When E op [n] is not greater than E th or not greater than K·E res [n], or when the number of sampling points that continuously meet the conditions is less than M, output a protection blocking signal.

8. The transformer inrush current blocking and internal fault identification method according to claim 7, characterized in that, The output signal also includes differential operation energy E op Braking constraint energy E res or effective energy margin .

9. The transformer inrush current blocking and internal fault identification method according to claim 8, characterized in that, The effective energy margin for: ； Where max represents the maximum value function.

10. A transformer inrush current blocking and internal fault identification system, characterized in that, The method according to any one of claims 1 to 9 comprises a three-phase current sampling unit, a reduction and phase compensation unit, a Clarke mode transformation unit, a mode differential / braking construction unit, a dual-channel wavelet feature extraction unit, a joint energy calculation unit, a threshold and braking boundary unit, and a protection output unit, wherein: The three-phase current sampling unit is used to synchronously acquire the three-phase current signals of the primary and secondary sides of the transformer. The reduction and phase compensation unit is used to reduce and compensate the phase of the three-phase current on the secondary side according to the transformer ratio, connection group and phase relationship; The Clarke mode transformation unit performs power-invariant Clarke transformation on the primary three-phase current and the reduced secondary three-phase current respectively to obtain the corresponding α-mode current and β-mode current on both sides. Modal differential / braking construction unit, used to construct modal differential current and modal braking current in the same modal coordinates; The dual-channel wavelet feature extraction unit is used to synchronously discretely decompose the modal differential current and modal braking current using the same wavelet basis, decomposition scale and boundary processing method to extract high-frequency detail coefficients; The joint energy calculation unit is used to perform square integration on the α and β mode high-frequency detail coefficients of the modal differential current and modal braking current within the same sliding time window, to obtain the differential operation energy E. op and braking constraint energy E res ; Threshold and braking boundary unit, used to initiate the action based on the energy threshold E th Construct a joint energy criterion with braking coefficient K, using K·E res As a dynamic braking boundary that varies with the current common transient level; The protection output unit is used to output protection blocking signals or internal fault action signals based on the joint energy criterion and the continuous sampling point confirmation conditions.

Images