Three-phase transformer winding top deformation diagnosis method based on high-voltage oscillatory wave detection

Abstract

The invention discloses a three-phase transformer winding top deformation diagnosis method based on high-voltage oscillation wave detection, and relates to the field of transformer safety guarantee and fault diagnosis. Firstly, a mechanical device for controlling deformation enables the top of a winding of a transformer to generate deformation of different degrees, and deformation of the winding of the transformer is simulated. And the high-voltage oscillatory wave test device is further used to obtain oscillatory wave test curves under different fault degrees. Finally, according to the diagnosis method based on the high-voltage oscillatory waves, normalization processing is carried out on transformer winding oscillatory wave time sequence signals through a normal cumulative distribution function, a time sequence is mapped to a discrete symbol sequence, then the entropy value of the symbol sequence is calculated through a fuzzy membership function, and the high-voltage oscillatory waves are obtained. The method can effectively reflect the irregularity and complexity of the signal. The deformation fault of the three-phase power transformer can be safely and efficiently detected and diagnosed, and a guarantee is provided for power transmission safety.

Description

Technical Field

[0001] The present invention relates to the field of transformer safety assurance and fault diagnosis, and specifically to a diagnostic method for the deformation of the top of the three-phase transformer winding based on high-voltage oscillation wave detection. Background Art

[0002] With the continuous expansion of the scale of the power system and the increasing improvement of operation requirements, the transformer, as the core equipment of the power system, its safe and stable operation is crucial to the reliability of the entire power grid. However, during operation, the transformer winding is prone to being affected by factors such as short-circuit impact, mechanical vibration, and thermal stress, resulting in winding deformation. Winding deformation not only reduces the mechanical stability of the transformer but may also trigger faults such as insulation damage and partial discharge, ultimately leading to equipment outage.

[0003] Although traditional winding state detection methods such as short-circuit impedance method and frequency response analysis method have achieved certain results in practical applications, they have limitations such as low sensitivity, long test time, and insensitivity to minor faults. In recent years, detection methods based on oscillation waves have gradually received attention. Oscillation waves can effectively reflect changes in winding state and have advantages such as strong anti-interference ability and fast test speed. By analyzing the waveform characteristics, energy attenuation law, etc. of oscillation waves, rapid diagnosis of winding deformation can be achieved.

[0004] In a three-phase transformer, the deformation of the top of the winding is one of the common fault forms, which has a significant impact on the safe operation of the transformer. Therefore, studying the diagnostic method for the deformation of the top of the three-phase transformer winding based on oscillation waves has important theoretical and practical significance for improving the accuracy, timeliness, and reliability of transformer fault diagnosis.

[0005] The oscillation wave method has relatively high sensitivity to specific faults of the transformer winding and can relatively accurately identify the state of winding deformation of the transformer. The present invention proposes a diagnostic method for the deformation of the top of the power transformer winding based on high-voltage oscillation wave detection. Its core is based on oscillation wave signals. The time series signal of the transformer winding oscillation wave is normalized by using the normal cumulative distribution function, and the time series is mapped to a discrete symbol sequence. Then, the entropy value of the symbol sequence is calculated using the fuzzy membership function. This method can effectively reflect the irregularity and complexity of the signal. The present invention can more reliably and effectively evaluate the performance of the transformer winding.

[0006] The existing publication No. CN118914716A discloses a diagnostic method for the inward deformation of the transformer winding based on high-voltage oscillation wave detection, which is for the diagnosis of the inward deformation of the transformer winding, not the top winding of the transformer. In addition, for the inward deformation of the transformer winding, the oscillation wave signal is used as the time-domain excitation and response signal, which is converted into a frequency-domain signal through the H-W model, and the winding deformation situation is evaluated by calculating the frequency division comparison fault parameters. Summary of the Invention

[0007] (1) Technical Problem to be Solved

[0008] In view of the deficiencies of the prior art, the present invention provides a diagnostic method for the deformation of the top of the three-phase transformer winding based on high-voltage oscillation wave detection, which solves the technical problems mentioned in the background art.

[0009] (2) Technical Solution

[0010] To achieve the above object, the present invention is realized through the following technical solutions: A diagnostic method for the deformation of the top of the three-phase transformer winding based on high-voltage oscillation wave detection, including a fault simulation platform and a diagnostic method, wherein the fault simulation platform is used to provide the deformation of the top of the three-phase transformer winding;

[0011] The diagnostic method includes the following steps:

[0012] S1: Simulate the deformation fault of the power transformer winding, including:

[0013] S11: Fix the three-phase transformer winding on the winding placement platform;

[0014] S12: Start the motor of the front and rear moving device to adjust the winding placement platform to a suitable position; start the motor of the left and right moving device to adjust the left and right moving device to a suitable position; start the motors of the up and down moving device -1 and the up and down moving device -2 to adjust the up and down moving device -1 and the up and down moving device -2 to a suitable height;

[0015] S13: Start the motor of the extrusion device to control the telescopic rod of the extrusion device to extend downward, and at the same time drive the extrusion plate of the extrusion device to extrude the power transformer winding to cause deformation; after deformation, lift the extrusion device again;

[0016] S14: Then adjust the position of the extrusion device so that the extrusion plate of the extrusion device is located above other positions of the winding;

[0017] S15: Repeat the operations of S13 to S14 to simulate the top deformations of different positions and different degrees of the three-phase windings;

[0018] S2: Conduct the oscillation wave test on the power transformer winding and obtain data;

[0019] S21: Define the upper end of the high-voltage winding of the power transformer as connection point A and the lower end as connection point B.

[0020] S22: Connect connection point A and connection point B of the power transformer winding to the high-frequency high-voltage switch and the high-frequency high-voltage DC power supply. The high-frequency high-voltage switch performs periodic actions, thereby performing periodic charge and discharge on points A and B;

[0021] S23: Repeat the above process to obtain the transformer oscillation wave waveform data under different fault conditions and the oscillation wave waveform data under intact conditions through the signal acquisition device;

[0022] S3: Process data and handle winding deformation fault;

[0023] S31: The oscillation wave signal collected by the data acquisition system is expressed as a time series in the form of x = {x1, x2, x3, ..., x N}, use the normal cumulative distribution function to normalize the signal between 0 and 1 to reduce the impact of outliers, and obtain the normalized time series y = {y1, y2, y3, …, y N};

[0024]

[0025] Where σ and γ are the standard deviation and mean of the time series x respectively;

[0026] S32: Map yi (i=1:N) to the sequence zc, integer indexed from 1 to c, as follows:

[0027]

[0028] is the sequence z c The i-th member of , c is a class parameter, which means it can be assigned to The number of classes;

[0029] S33: In the dispersion entropy, the closer to the integer k belongs to the kth class, since there is a series z at the boundary between the two classes c The fuzziness of the membership of the members of the class, so the fuzzy membership function Mk is regarded as the class, and express The degree of membership relative to the kth class;

[0030] Each Assigned to one class or two. One class means that when the data point is located in the core area of a class, its membership is close to 1 in this class and close to 0 in other classes, without ambiguity. Two classes means that the data point is located in the boundary area of adjacent classes and needs to be fuzzily assigned to two adjacent classes (such as class k and class k+1). Since the membership function should have corresponding conditions, the trapezoidal membership function is used for class 1 and class c, and the triangular membership function is used for other classes (k≠1 and k≠c). The fuzzy function used is as follows:

[0031]

[0032] S34: Construct a time series with dimension m and time delay d according to the following formula

[0033]

[0034] j = 1, 2, ..., N-(m + 1)d

[0035] S35: Each vector is mapped to a dispersion pattern according to its membership degree

[0036]

[0037] When several states must be true simultaneously, when belongs to class v0, belongs to class v1, belongs to class belongs to class v m-1 i.e., in the case of an "AND" operation, use a t-norm;

[0038] The algebraic product operator is the most representative t-norm operator used in the above formula;

[0039]

[0040] where is For the membership degree of pattern is equal to the product of the membership degrees with respect to classes v0, v1, ..., v m-1 ;

[0041] S36: The probability of each dispersion pattern in the time series is calculated as follows:

[0042]

[0043] where represents the sum of the membership degrees of all sequences to the dispersion pattern , divided by the total number of embedded signals with embedding dimension m;

[0044] S37: Calculate the fuzzy dispersion entropy with embedding bedding dimension m and number of classes c:

[0045]

[0046] S38: Calculate the normalized allocation ratio according to the following formula: The allocation ratio is divided by the maximum possible allocation ratio. Since the fuzzy dispersion entropy ranges from 0 to ln, the normalized fuzzy dispersion entropy is calculated as follows:

[0047]

[0048] S39: Determine the deformation degree of the winding according to the magnitude of the fuzzy dispersion entropy under different deformation degrees of the winding:

[0049] If the fuzzy dispersion entropy < 0.1, it is determined that the deformation degree of the transformer winding is slight deformation;

[0050] If 0.1 < fuzzy dispersion entropy < 0.3, it is determined that the deformation degree of the transformer winding is obvious deformation;

[0051] If the fuzzy dispersion entropy > 0.3, it is determined that the deformation degree of the transformer winding is severe deformation.

[0052] (III) Beneficial effects

[0053] The present invention provides a method for diagnosing the deformation at the top of the three-phase transformer winding based on high-voltage oscillation wave detection. It has the following beneficial effects:

[0054] The method for diagnosing the deformation at the top of the three-phase transformer winding based on high-voltage oscillation wave detection uses the oscillation wave method and has relatively high sensitivity to specific faults of the transformer winding, and can more accurately identify the state of the deformation of the transformer winding. The present invention proposes a method for diagnosing the deformation at the top of the power transformer winding based on high-voltage oscillation wave detection. Its core is based on the oscillation wave signal. The normal cumulative distribution function is used to normalize the time series signal of the transformer winding oscillation wave, map the time series to a discrete symbol sequence, and then use the fuzzy membership function to calculate the entropy value of the symbol sequence. This method can effectively reflect the irregularity and complexity of the signal. The present invention can more reliably and effectively evaluate the performance of the transformer winding. Description of the drawings

[0055] Figure 1a It is a schematic diagram of the use state of the device for simulating the deformation at the top of the winding of the present invention;

[0056] Figure 1b It is the front view of the device for simulating the deformation at the top of the winding of the present invention;

[0057] Figure 1c It is the left view of the device for simulating the deformation at the top of the winding of the present invention;

[0058] Figure 1d It is the top view of the device for simulating the deformation at the top of the winding of the present invention;

[0059] Figure 2 It is the wiring diagram of the oscillation wave test of the power transformer of the present invention;

[0060] Figure 3 It is the flow block diagram adopted by the present invention. Detailed implementation manners

[0061] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] As Figure 1a , Figure 1b , Figure 1c , Figure 1d shown, the mechanical device of the fault simulation platform mainly includes: three-phase transformer winding 1, winding placement platform 2, bottom frame connection and bottom support -1 (3), bottom frame connection and bottom support -2 (4), bottom frame connection and bottom support -3 (5), bottom frame connection and bottom support -4 (6), bottom quadrilateral frame 7, triangular frame 8, top frame (9), up and down moving device motor base -1 (10), up and down moving device motor base -2 (11), up and down moving device control motor -1 (12), up and down moving device control motor -2 (13), up and down moving device track rod -1 (14), up and down moving device track rod -2 (15), up and down moving device moving screw rod -1 (16), up and down moving device moving screw rod -2 (17), up and down moving device -1 (18), up and down moving device -2 (19), left and right moving device control motor 20, left and right moving device track rod -1 (21), left and right moving device track rod -2 (22), left and right moving device 23, extrusion device 24, extrusion device control motor 25, extrusion device telescopic rod 26, extrusion device extrusion plate 27, front and back moving device track rod -1 (28), front and back moving device track rod -2 (29), front and back moving device moving power belt 30, moving power belt front fixing device 31, moving power belt rear fixing device 32, front and back moving device control motor 33;

[0063] Figure 3 A diagnostic method for the top deformation of a three-phase transformer winding based on high-voltage oscillating wave detection specifically includes the following steps:

[0064] Step 1: Simulate the deformation fault of the power transformer winding, including:

[0065] 1) Fix the three-phase transformer winding 1 on the winding placement platform 2.

[0066] 2) Start the front - rear moving device to control the motor 33 to adjust the winding placement platform 2 to a suitable position; start the left - right moving device to control the motor 20 to adjust the left - right moving device 23 to a suitable position; start the up - down moving device to control the motor - 1 (12) and the up - down moving device to control the motor - 2 (13) to adjust the up - down moving device - 1 (18) and the up - down moving device - 2 (19) to a suitable height.

[0067] 3) Start the extrusion device to control the motor 25 to control the extrusion device telescopic rod 26 to extend downward, and at the same time drive the extrusion device pressing plate 27 to extrude the power transformer winding 1, causing deformation. After deformation, then lift the extrusion device.

[0068] 4) Then adjust the position of the extrusion device 24 so that the extrusion device pressing plate 27 is located above other positions of the winding.

[0069] 5) Repeat the operations from step 3) to step 4) to simulate the top deformations at different positions and degrees of the three - phase winding.

[0070] Step 2: Conduct the oscillating wave test on the power transformer winding and obtain data

[0071] 1) Define the upper end of the high - voltage winding 36 of the power transformer as connection point A and the lower end as connection point B.

[0072] 2) Connect the connection points A and B of the power transformer winding to the high - frequency high - voltage switch 41 and the high - frequency high - voltage DC power supply 40. The high - frequency high - voltage switch 41 performs periodic actions, thereby performing periodic charge - discharge on points A and B;

[0073] 3) Repeat the above process to obtain the transformer oscillating wave waveform data under different fault conditions and the oscillating wave waveform data under normal conditions through the signal acquisition device 42.

[0074] Step 3: Process the data and handle the winding deformation fault.

[0075] 1) The oscillating wave signal collected by the data acquisition system is expressed in the form of a time series x = {x1, x2, x3, …, x N}, and the normal cumulative distribution function is used to normalize the signal between 0 - 1 to reduce the influence of outliers, obtaining the normalized time series y = {y1, y2, y3, …, y N}.

[0076]

[0077] Among them, σ and γ are the standard deviation and mean value of the time series x respectively.

[0078] 2) Map yi (i = 1:N) to the sequence zc, the integer index from 1 to c, as follows:

[0079]

[0080] is the i-th member of sequence z c The class parameter c represents the number of classes that can be assigned to the classes of

[0081] 3) In the dispersion entropy, the one closer to the integer k belongs to the k-th class. Due to the ambiguity of the membership of the members of the series z c at the boundary between two classes, the fuzzy membership function Mk is regarded as the class, and represents the membership degree relative to the k-th class. Each is assigned to one or two classes (integer index k = 1,..., c). Since the membership function should meet the corresponding conditions, we use the trapezoidal membership function for classes 1 and c, and the triangular membership function for other classes (k≠1 and k≠c). The fuzzy functions used are as follows:

[0082]

[0083] 4) Construct a time series with dimension m and time delay d according to the following formula

[0084]

[0085] j = 1, 2,..., N-(m + 1)d

[0086] 5) Each vector is mapped to the dispersion mode according to its membership degree

[0087]

[0088] When several states must be true at the same time, that is, for the "AND" operation, the t-norm is used. The algebraic product operator is the most representative t-norm operator used in the above formula

[0089]

[0090] where is the membership degree for the mode equals the product of the membership degrees with respect to the classes v0, v1,..., v m-1

[0091] 6) The probability of each dispersion mode in the time series is calculated as follows:​

[0092]

[0093] wherein represents the sum of the membership degrees belonging to all sequences of the dispersion mode , divided by the total number of embedded signals with an embedding dimension of m.

[0094] 7) Calculate the fuzzy dispersion entropy with an embedded bedding dimension of m and a number of classes of c:

[0095]

[0096] 8) Calculate the normalized allocation ratio according to the following formula: The allocation ratio is divided by the maximum possible allocation ratio. Since the fuzzy dispersion entropy changes from 0 to ln(cm), the normalized fuzzy dispersion entropy (NFuzzDispEn) is calculated as follows:

[0097]

[0098] 9) Judge the deformation degree of the winding according to the magnitude of the fuzzy dispersion entropy (NFuzzDispEn) under different deformation degrees of the winding.

[0099] It should be noted that in the description of the invention, the orientation or positional relationship indicated by terms such as "upper", "lower", "left", "right", "front", "rear", etc. is the description of the structure of the present invention based on the figures shown, only for the convenience of describing the present invention simply, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be construed as a limitation of the present invention.

[0100] For the "first" and "second" in this technical solution, they are only used to distinguish the names of corresponding structures with the same or similar structures or similar functions, not the arrangement of the importance of these structures, nor do they have an order, or compare sizes, or other meanings.

[0101] In addition, unless otherwise clearly specified and limited, the terms "installation" and "connection" should be understood in a broad sense. For example, the connection can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two structures. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the general idea of the present invention and in connection with the specific situation of this solution.

Claims

1. A three-phase transformer winding top deformation diagnosis method based on high-voltage oscillating wave detection, including a fault simulation platform and a diagnosis method, characterized in that: The fault simulation platform is used to provide deformation at the top of the three-phase transformer winding; The diagnostic method includes the following steps: S1: Simulate the deformation fault of the power transformer winding, including: S11: Fix the three-phase transformer winding (1) on the winding placement platform (2); S12: Start the front and rear moving device control motor (33) to adjust the winding placement platform (2) to a suitable position; start the left and right moving device control motor (20) to adjust the left and right moving device (23) to a suitable position; start the up and down moving device control motor-1 (12) and up and down moving device control motor-2 (13) to adjust the up and down moving device-1 (18) and up and down moving device-2 (19) to a suitable height; S13: Start the extrusion device control motor (25) to control the extrusion device telescopic rod (26) to extend downward, and at the same time drive the extrusion device extrusion plate (27) to extrude the power transformer winding (1) to cause deformation; after deformation, lift the extrusion device again; S14: Then adjust the position of the extrusion device (24) so that the extrusion device extrusion plate (27) is located above other positions of the winding; S15: Repeat the operations of S13 to S14 to simulate the top deformation of different positions and different degrees of the three-phase winding; S2: Conduct the oscillating wave test on the power transformer winding and obtain data; S21: Define the upper end of the high-voltage winding (36) of the power transformer as connection point A and the lower end as connection point B. S22: Connect the connection points A and B of the power transformer winding to the high-frequency high-voltage switch (41) and the high-frequency high-voltage DC power supply (40). The high-frequency high-voltage switch (41) performs periodic actions, so as to perform periodic charge and discharge on points A and B; S23: Repeat the above process to obtain the transformer oscillating wave waveform data under different fault conditions and the oscillating wave waveform data under normal conditions through the signal acquisition device (42); S3: Process the data to process the winding deformation fault; S31: The oscillation wave signal collected by the data acquisition system is expressed as a time series in the form of x = {x1, x2, x3, ..., x N }, use the normal cumulative distribution function to normalize the signal between 0 and 1 to reduce the influence of outliers, and obtain the normalized time series y = {y1, y2, y3, ..., y N }; where σ and γ are the standard deviation and mean value of the time series x respectively; S32: Map yi (i = 1:N) to the sequence zc, the integer index from 1 to c, as follows: is the i-th member of sequence z c and c is a class parameter representing the number of classes that can be assigned to ; S33: Among the dispersion entropies, those closer to the integer k belong to the k-th class. Due to the ambiguity of the membership degrees of the members of the series z at the boundary between two classes c , the fuzzy membership function Mk is regarded as this class, and denotes the membership degree relative to the k-th class; Each Assigned to one class or two classes. One class means that when the data point is located in the core area of a class, its membership is close to 1 in that class and close to 0 in other classes, without ambiguity. Two classes means that the data point is located in the boundary area of adjacent classes and needs to be fuzzily assigned to two adjacent classes (such as class k and class k+1). Since the membership function should meet the corresponding conditions, the trapezoidal membership function is used for class 1 and class c, and the triangular membership function is used for other classes (k≠1 and k≠c). The fuzzy function used is as follows: S34: Construct a time series with dimension m and time delay d according to the following formula S35: Each vector is mapped to a dispersion pattern according to its membership degree When belongs to category v0, belongs to category v1, belongs to category v2, ……, belongs to category v m-1 at that time, i.e., when performing the "AND" operation, use the t-norm; The algebraic product operator is the most representative t-norm operator used in the above formula; Among them is the membership degree for the pattern equals the product of the membership degrees m-1 with respect to classes v0, v1,..., v S36: The probability calculation of each dispersion mode in the time series is as follows: Among them represents the sum of the membership degrees of all sequences of the dispersion mode, divided by the total number of embedded signals with an embedding dimension of m; S37: Calculate the fuzzy dispersion entropy with the embedding bedding dimension of m and the number of classes of c: S38: Calculate the normalized distribution ratio according to the following formula: The distribution ratio is divided by the maximum possible distribution ratio. Since the fuzzy dispersion entropy changes from 0 to ln, the normalized fuzzy dispersion entropy is calculated as follows: S39: Judge the deformation degree of the winding according to the size of the fuzzy dispersion entropy under different deformation degrees of the winding: If the fuzzy dispersion entropy < 0.1, it is judged that the deformation degree of the transformer winding is slight deformation; If 0.1 < fuzzy dispersion entropy < 0.3, it is judged that the deformation degree of the transformer winding is obvious deformation; If the fuzzy dispersion entropy > 0.3, it is judged that the deformation degree of the transformer winding is severe deformation.

Images