A method and system for detecting transformer winding deformation based on a combination of short-circuit impedance and frequency response methods.

Abstract

This invention provides a method and system for detecting transformer winding deformation based on a combination of short-circuit impedance and frequency response methods, belonging to the field of transformer deformation detection technology. It includes: connecting the transformer winding pair under test to a detection circuit based on the combination of short-circuit impedance and frequency response methods; injecting a sweep frequency signal into the transformer winding pair under test through a signal source in the detection circuit; when the frequency of the sweep frequency signal is in a preset low-frequency band, determining whether the transformer winding pair under test is deformed based on the short-circuit impedance; when the frequency of the sweep frequency signal is in a preset high-frequency band, acquiring the winding response signal corresponding to the sweep frequency signal using the detection circuit; establishing a corresponding frequency response curve using an improved signal analysis method, and determining whether the transformer winding pair under test is deformed based on the frequency response curve; overcoming the limitations of traditional detection methods, improving the convenience, accuracy, and real-time performance of detection, and providing a more efficient and reliable solution for the maintenance and fault diagnosis of power transformers.

Description

Technical Field

[0001] This invention relates to the field of transformer winding detection technology, specifically to a transformer winding deformation detection method and system based on a combination of short-circuit impedance method and frequency response method. Background Technology

[0002] In the daily operation of power systems, power transformers, as core equipment for energy conversion and transmission, are directly related to the safety and stability of the power system. However, during operation, transformer windings often deform due to factors such as short-circuit faults, lightning strikes, and mechanical vibrations. Winding deformation not only reduces the transformer's operating efficiency but may also lead to more serious faults, such as inter-turn short circuits and phase-to-phase short circuits, posing a significant threat to the safety of the power system.

[0003] Currently, there are various methods for detecting transformer winding deformation. Among them, the short-circuit impedance method and the frequency response method are two of the more common methods. The short-circuit impedance method is simple to operate, and it determines whether the winding is deformed by measuring the impedance change of the transformer under different loads. The frequency response method determines whether the winding is deformed by acquiring the corresponding frequency response curve of the response signal and analyzing the frequency response curve. However, both of these methods have significant limitations when used alone. For example, they have limited sensitivity to detect minute deformations inside the winding, and the analysis of the response signal and the acquisition of the frequency response curve are affected by noise, resulting in poor accuracy and making them unsuitable for fault diagnosis. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a transformer winding deformation detection method and system based on the combination of short-circuit impedance method and frequency response method, which can overcome the limitations of traditional detection methods in the background art, improve the convenience, accuracy and real-time performance of detection, and provide a more efficient and reliable solution for the maintenance and fault diagnosis of power transformers.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A transformer winding deformation detection method based on a combination of short-circuit impedance and frequency response methods includes the following steps: The transformer winding under test is connected to a pre-set detection circuit based on a combination of short-circuit impedance method and frequency response method. A sweep frequency signal is injected into the transformer winding under test through the signal source in the detection circuit. When the frequency of the sweep signal is a preset low frequency band, the short-circuit impedance is calculated based on the detection circuit and the electrical parameters obtained by the detection circuit, and the short-circuit impedance is used to determine whether the winding of the transformer under test is deformed. When the frequency of the sweep signal is a preset high frequency band, the winding response signal corresponding to the sweep signal is obtained by the detection circuit. Based on the winding response signal, an improved signal analysis method is used to establish the corresponding frequency response curve, and the deformation of the transformer winding under test is determined based on the frequency response curve.

[0006] The aforementioned preset low frequency band is the frequency band where the sweep frequency signal frequency is less than 1000Hz, and the preset high frequency band is the frequency band where the sweep frequency signal frequency is greater than 1000Hz. Within a preset low-frequency band, the short-circuit impedance value calculated at 50Hz is compared with the short-circuit impedance value under normal conditions. When the rate of change of the short-circuit impedance value exceeds the specified range, it is determined that the winding of the transformer under test has deformed.

[0007] The improved signal analysis method described above includes Fourier transform based on windowing to establish the corresponding frequency response curve, specifically including: The continuous sweep frequency signal and the winding response signal are subjected to Fourier transform based on windowing processing, respectively; The frequency response curve is calculated using the swept frequency signal and the winding response signal after Fourier transform based on windowing.

[0008] The expression for the Fourier transform based on windowing is as follows: ; In the formula, For input, it is a continuous time signal; for input, it is the time-domain waveform of the excitation voltage in the frequency sweep signal or the response current in the winding response signal. Let τ represent the window function, τ be the translation amount of the window function, and ω be the angular frequency; is a complex exponential function, and is the kernel function that converts a time-domain signal to the frequency domain to perform the Fourier transform; It is the swept frequency signal or winding response signal after Fourier transform based on windowing processing.

[0009] The expression for the frequency response curve mentioned above is: ; In the formula, This represents the signal after the time-domain response voltage or current signal in the winding response signal has undergone a Fourier transform based on windowing. This represents the signal after the time-domain excitation voltage signal in the frequency sweep signal has undergone a Fourier transform based on windowing. This represents the frequency response curve.

[0010] The improved signal analysis method described above includes establishing the corresponding frequency response curve based on multi-scale complex continuous wavelet transform, specifically including: Multi-scale complex continuous wavelet transform is performed on the continuous sweep frequency signal and the winding response signal respectively; The frequency response curve is calculated from the swept frequency signal and the winding response signal after multi-scale complex continuous wavelet transform.

[0011] The above expression based on multi-scale complex continuous wavelet transform is as follows: ; In the formula, The voltage or current signal in the continuous sweep frequency signal or the winding response signal is the input. wavelet basis functions The complex conjugate of is used to calculate the local correlation between the signal and the wavelet; These are wavelet coefficients, representing the signal at different scales. and time The time-frequency energy distribution under [the specified conditions].

[0012] The expression for the frequency response curve mentioned above is: ; In the formula, This represents the wavelet coefficients of the excitation voltage signal in the swept frequency signal. This represents the wavelet coefficients of the current or voltage signal in the winding response signal. The time axis is ω, and the angular frequency is ω. This represents the frequency response curve.

[0013] The method also includes a step of diagnosing the deformation of the transformer winding under test by using the frequency response curve corresponding to the sweep frequency signal of any frequency band, specifically: An improved signal analysis method is used to establish the frequency response curve of the swept signal of the transformer winding under test in any frequency band, and the distance characteristic value between the frequency response curve and the normal curve in that frequency band is calculated. The distance feature value is input into the trained GA-SVM classification model for processing to obtain the deformation type of the transformer winding under test.

[0014] The formula for calculating the distance characteristic value mentioned above is: ; in: The distance feature value is obtained by normalizing the distance value. This is the distance value; ; In the formula, This represents the discrete amplitude data of the normal curve within the corresponding frequency band. This represents the discrete amplitude data of the frequency response curve within the corresponding frequency band. This indicates the phase angle corresponding to the normal curve. This indicates the phase angle corresponding to the frequency response curve. This indicates the total number of signal sampling points.

[0015] The aforementioned deformation types include at least axial deformation, radial deformation, and axis misalignment; The training process of the GA-SVM classification model includes: Under the same frequency segmentation, obtain the deformation frequency response curves of transformer windings with various deformation types, as well as the normal frequency response curves of normal transformer windings; Each segment of the same frequency is divided into multiple frequency intervals, and multiple deformed frequency response curves and multiple normal frequency response curves are obtained under each frequency interval. Based on the deformed frequency response curves and normal frequency response curves of the same frequency division interval, the distance characteristic value of the deformed frequency response curve of each deformation type is calculated, and a frequency division training set is established through the distance characteristic value of the same frequency division interval. The model optimization parameters of the preset SVM model are obtained by optimizing the GA genetic algorithm, and the SVM model based on the model optimization parameters is trained by the frequency division training set of each frequency division interval. The SVM model is defined as a GA-SVM classification model once it reaches the preset training termination condition.

[0016] The training termination condition mentioned above is that the loss function of the preset SVM model does not exceed a threshold. The expression for the loss function of the preset SVM model is: ; ; ; In the formula, This indicates the number of frequency division intervals. This represents the number of samples in the frequency division training set. For the first In the frequency division interval, the first The true distance feature value of each sample Indicates the first In the frequency division interval, the first The predicted distance feature value of each sample. To preset the first loss function, A second loss function is preset; Indicator variable representing the true label, when the first... The true class of each sample is hour, ,otherwise ; Indicates the first The predicted classification result for each sample is The probability, This represents the total number of samples across all frequency division intervals. The total number of categories; After normalization , After normalization , λ is the loss function of the preset SVM model, and λ is the weight.

[0017] The above-mentioned detection circuit construction includes: amplifying the signal output from the signal source through a power amplifier and applying it to one phase winding of the transformer under test; short-circuiting the opposite winding of the transformer under test; connecting sampling resistors at both ends of the winding of the transformer under test; acquiring the voltage signal at both ends of the winding under test and the current signal of the test circuit through the sampling resistors; and the winding response signal including the voltage signal at both ends of the winding under test and the current signal of the test circuit.

[0018] The system using the above-mentioned transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method includes: The signal injection module is used to connect the transformer winding under test to a pre-set detection circuit based on a combination of short-circuit impedance method and frequency response method, and inject a sweep frequency signal into the transformer winding under test through the signal source in the detection circuit. The short-circuit impedance module is used to calculate the short-circuit impedance based on the detection circuit and the electrical parameters obtained by the detection circuit when the frequency of the sweep signal is in the preset low frequency band, and to determine whether the winding of the transformer under test is deformed based on the short-circuit impedance. The sweep frequency response module is used to obtain the winding response signal corresponding to the sweep frequency signal by means of a detection circuit when the frequency of the sweep frequency signal is a preset high frequency band. The curve judgment module is used to establish the corresponding frequency response curve based on the winding response signal using an improved signal analysis method, and to determine whether the winding of the transformer under test is deformed based on the frequency response curve.

[0019] The system also includes a deformation diagnosis module, which uses an improved signal analysis method to establish the frequency response curve of the transformer winding under test in any frequency band for the sweep signal, calculates the distance feature value between the frequency response curve and the normal curve in that frequency band, and inputs the distance feature value into the trained GA-SVM classification model for processing to obtain the deformation type of the transformer winding under test.

[0020] The transformer winding deformation detection method and system based on the combination of short-circuit impedance method and frequency response method mentioned in this invention have the following beneficial effects: 1. By combining the short-circuit impedance method with the frequency response method for transformer deformation detection, both the short-circuit impedance value and the frequency response curve can be obtained in a single test, achieving complementary advantages of the two detection methods and improving the comprehensiveness and accuracy of the detection. This overcomes the limitations of the short-circuit impedance method when used alone, which has limited sensitivity in detecting minute deformations inside the windings, and the frequency response method when used alone, which suffers from poor accuracy in analyzing the response signal and obtaining the frequency response curve due to noise, thus hindering fault diagnosis. While overcoming the limitations of traditional detection methods, this approach improves the convenience, accuracy, and real-time performance of detection, providing a more efficient and reliable solution for the maintenance and fault diagnosis of power transformers.

[0021] In step 2, by calculating the distance feature value between the frequency response curve and the normal curve in this frequency band, and inputting the distance feature value into the trained GA-SVM classification model for processing, the corresponding deformation type is obtained. This solves the problem of high misjudgment rate when using the frequency response method to diagnose winding deformation by relying on human experience, and realizes accurate diagnosis of transformer winding deformation. At the same time, during the training process of the GA-SVM classification model, the entire frequency band is divided into intervals for training, so that the classification model can learn the feature differences of different frequency intervals, providing a good learning foundation for the accurate identification of subsequent classification models.

[0022] 3. By performing signal processing on the swept frequency signal and winding response signal using Fourier transform based on windowing and / or multi-scale complex continuous wavelet transform, the problem of poor accuracy caused by noise in current signal analysis methods for obtaining frequency response curves is solved, which is not conducive to fault diagnosis. The frequency response curves obtained by these two methods have high clarity and greatly reduce the impact of noise. Attached Figure Description

[0023] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the detection method in an embodiment of the present invention; Figure 2 This is a schematic diagram of the connection of the detection system in an embodiment of the present invention; Figure 3 This is a schematic diagram of the connection of the detection circuit in an embodiment of the present invention; Figure 4 This is a schematic diagram of the connection of an electronic device in an embodiment of the present invention. Detailed Implementation

[0024] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0025] Example 1: To address the limitations of current methods, such as the limited sensitivity of short-circuit impedance method in detecting minute deformations within windings when used alone, and the poor accuracy of frequency response method in analyzing response signals and obtaining frequency response curves due to noise, which hinders fault diagnosis, this embodiment provides a transformer winding deformation detection method based on a combination of short-circuit impedance and frequency response methods. Figure 1 As shown, it includes: S1, the transformer winding under test is connected to a pre-set detection circuit based on a combination of short-circuit impedance method and frequency response method, and a sweep frequency signal is injected into the transformer winding under test through the signal source in the detection circuit.

[0026] Combining the short-circuit impedance method with the frequency response method as a diagnostic method for transformer winding deformation can combine the advantages of both testing methods. In a single test, both the frequency response curve and the short-circuit impedance value can be obtained. By detecting the frequency response impedance curves of each transformer winding, and then comparing the obtained test results laterally or longitudinally, the differences in the frequency response impedance curves of each winding can be used to determine whether the winding may be deformed.

[0027] Specifically, see the connection diagram of the detection circuit when combining the short-circuit impedance method and the frequency response method. Figure 3 ,Depend on Figure 3 As can be seen, when using the combined method to detect transformer winding faults, it is necessary to continuously change the frequency of the external excitation power supply to obtain a continuous impedance-frequency response curve. During testing, the primary condition is a signal source that can continuously and stably output different frequencies. The signal output by the sweep frequency signal source is amplified by a power amplifier and then applied to one phase winding of the transformer. At the same time, the opposite winding of the transformer under test is short-circuited. In order to obtain an accurate voltage signal, sampling resistors R1 and R2 are connected to the two ends of the winding respectively to collect the voltage signal at both ends of the winding under test and the current signal of the test circuit. The winding response signal mentioned below includes the voltage signal at both ends of the winding under test and the current signal of the test circuit.

[0028] S2, when the frequency of the sweep signal is a preset low frequency band, the short-circuit impedance is calculated based on the detection circuit and the electrical parameters obtained by the detection circuit, and the short-circuit impedance is used to determine whether the windings of the transformer under test are deformed.

[0029] The preset low-frequency band can be a sweep signal with a frequency less than 1000 Hz. Within this frequency band, the equivalent circuit of the transformer winding is the same as that of the low-voltage impedance method. Impedance curves at different frequencies are plotted to observe whether they are linear. At 50 Hz, the impedance value is compared with the short-circuit impedance value under normal conditions. When the rate of change of the short-circuit impedance value exceeds the range specified in the regulations, it can be determined that the winding has deformed.

[0030] S3, when the frequency of the sweep signal is a preset high frequency band, the winding response signal corresponding to the sweep signal is obtained by the detection circuit.

[0031] Among them, the preset high frequency band can be a sweep frequency signal with a frequency greater than 1000 Hz. In this case, the frequency response curve is plotted by the winding response signal to analyze whether the transformer winding is deformed.

[0032] S4. Based on the winding response signal, an improved signal analysis method is used to establish the corresponding frequency response curve, and the frequency response curve is used to determine whether the winding of the transformer under test is deformed.

[0033] Specifically, the frequency response curves established using the improved signal analysis method described above are as follows: S411 performs Fourier transform based on windowing on the continuous sweep frequency signal and the winding response signal, respectively.

[0034] Specifically, the Fourier transform based on windowing is as follows: ; In the formula, The input is a continuous-time signal, representing the time-domain waveform of the excitation voltage in the sweep frequency signal or the response current in the winding response signal; Represents the window function. The translation amount of the window function. Angular frequency; The complex exponential function represents the kernel function that transforms a time-domain signal into the frequency domain to perform the Fourier transform. It is the swept frequency signal or winding response signal after Fourier transform based on windowing processing.

[0035] S412 obtains the frequency response curve through the swept frequency signal and the winding response signal based on the windowed Fourier transform.

[0036] Specifically, the frequency response curves mentioned above are as follows: ; In the formula, This represents the signal after the time-domain response voltage or current signal in the winding response signal has undergone a Fourier transform based on windowing. This represents the signal after the time-domain excitation voltage signal in the frequency sweep signal has undergone a Fourier transform based on windowing. This represents the frequency response curve obtained through the ratio.

[0037] Optionally, the frequency response curve established using the improved signal analysis method described above can also be: S421 performs multi-scale complex continuous wavelet transform on the continuous sweep frequency signal and the winding response signal, respectively.

[0038] Specifically, the above-mentioned multi-scale complex continuous wavelet transform is as follows: ; In the formula, The input is a continuous sweep frequency signal or a voltage or current signal from the winding response signal. wavelet basis functions The complex conjugate of is used to calculate the local correlation between the signal and the wavelet; These are wavelet coefficients, representing the signal at different scales. and time The time-frequency energy distribution under [the specified conditions].

[0039] Specifically, the inverse wavelet transform can be expressed as: ; The above equation is the inverse wavelet transform, where Represents the allowable constant. For wavelet basis functions, Indicates the translation factor. Represents the scaling factor; where: ,and ; In the formula, The constant represents the allowable constant, a necessary condition to ensure the invertibility of the wavelet transform, and reflects the wavelet mother function. Energy distribution characteristics; wavelet mother function Fourier transform; ω is the angular frequency, and ω represents the integral variable.

[0040] Furthermore, in the above formula: ; In the formula, Let be the wavelet basis functions, indicating that the wavelet transform is performed through the wavelet mother function. Translation and scaling yield a series of wavelet sequences; Indicates the translation factor. Represents the scaling factor; where: ; In the formula, and These represent the center frequency and bandwidth of the complex Morlet wavelet, respectively. For wavelet mother function, This is a complex exponential term used to provide the oscillatory properties of the wavelet; It is a Gaussian window function, which guarantees the local attenuation characteristics of the wavelet.

[0041] S422 obtains the frequency response curve using a sweep frequency signal and a winding response signal based on a multi-scale complex continuous wavelet transform.

[0042] Specifically, the frequency response curves mentioned above are as follows: ; In the formula, This represents the wavelet coefficients of the excitation voltage signal in the swept frequency signal. This represents the wavelet coefficients of the current or voltage signal in the winding response signal. For the timeline, ω is the angular frequency.

[0043] Optionally, the above methods also include: S5, using the frequency response curve corresponding to the sweep signal of any frequency band, performs deformation diagnosis of the transformer winding under test, specifically as follows: S51, using an improved signal analysis method, establishes the frequency response curve corresponding to the sweep signal of the transformer winding under test in any frequency band, and calculates the distance characteristic value between the frequency response curve and the normal curve in that frequency band.

[0044] Specifically, the aforementioned distance characteristic values ​​are as follows: ; in: This represents the distance feature value obtained by normalizing the distance value. Indicates the distance value;

[0045] In the formula, This represents the discrete amplitude data of the normal curve within the corresponding frequency band. This represents the discrete amplitude data of the frequency response curve within the corresponding frequency band. This represents the phase angle corresponding to the normal curve. This represents the phase angle corresponding to the frequency response curve. This represents the total number of signal sampling points.

[0046] S52, input the distance feature value into the trained GA-SVM classification model for processing to obtain the corresponding deformation type.

[0047] The GA-SVM classification model described above is trained in the following way: S521, under the same frequency segmentation, obtain the deformation frequency response curve of transformer windings with various deformation types, as well as the normal frequency response curve of normal transformer windings; the frequency segmentation can be a commonly used frequency range in practice, such as 1-2500 Hz.

[0048] S522 divides the frequency band into multiple frequency division intervals and obtains multiple deformed frequency response curves and multiple normal frequency response curves under each frequency division interval; it can be divided into 3 frequency division intervals, namely 1-800hz, 800-1700hz, and 1700-2500hz.

[0049] S523, based on the deformation frequency response curves and normal frequency response curves of the same frequency division interval, calculate the distance characteristic value of the deformation frequency response curve of each deformation type, and establish a frequency division training set through the distance characteristic values ​​of the same frequency division interval. The deformation types include at least axial deformation, radial deformation and axis offset.

[0050] S524 uses the GA genetic algorithm to optimize the model parameters of the preset SVM model, and trains the SVM model based on the optimized model parameters using the frequency division training sets of each frequency division interval.

[0051] S525, until the SVM model reaches the preset training termination condition, the SVM model that has reached the training termination condition is identified as the GA-SVM classification model.

[0052] The training termination condition mentioned above can be that the loss function of the preset SVM model does not exceed a threshold. The loss function of the preset SVM model can be: ; ; ; In the formula, Indicates the number of frequency division intervals. This indicates the number of samples in the frequency division training set. For the first i In the frequency division interval, the first j The true distance feature value of each sample Indicates the first i In the frequency division interval, the first j The predicted distance feature value of each sample. To preset the first loss function, To pre-define the second loss function, Indicator variable representing the true label, when the first... k The true class of each sample is c hour, ,otherwise ; Indicates the first k The predicted classification result for each sample is c The probability, This represents the total number of samples across all frequency intervals. The total number of categories, After normalization , After normalization , The loss function for the pre-defined SVM model, For weights.

[0053] Example 2: This application provides a transformer winding deformation detection system based on a combination of short-circuit impedance and frequency response methods, applied to the transformer winding deformation detection method based on the combination of short-circuit impedance and frequency response methods in Embodiment 1, such as... Figure 2 As shown, it includes: The signal injection module is used to connect the transformer winding under test to a pre-set detection circuit based on a combination of short-circuit impedance method and frequency response method, and inject a sweep frequency signal into the transformer winding under test through the signal source in the detection circuit. The short-circuit impedance module is used to calculate the short-circuit impedance based on the detection circuit and the electrical parameters obtained by the detection circuit when the frequency of the sweep signal is in the preset low frequency band, and to determine whether the windings of the transformer under test are deformed based on the short-circuit impedance. The sweep frequency response module is used to obtain the winding response signal corresponding to the sweep frequency signal by means of a detection circuit when the frequency of the sweep frequency signal is a preset high frequency band. The curve judgment module is used to establish the corresponding frequency response curve based on the winding response signal using an improved signal analysis method, and to determine whether the winding of the transformer under test is deformed based on the frequency response curve.

[0054] Example 3: This application provides an electronic device, such as... Figure 4 As shown, it includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method of Embodiment 1.

[0055] Example 4: This application provides a non-transitory computer-readable storage medium that stores computer instructions that cause a computer to execute the method of embodiment 1.

[0056] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0057] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0058] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0059] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0060] Those skilled in the art will understand that all or part of the steps in the above facts and methods can be implemented by a program instructing related hardware. The program or the program described therein can be stored in a computer-readable storage medium. When the program is executed, it includes the following steps: at this time, the corresponding method steps are introduced. The storage medium can be ROM / RAM, magnetic disk, optical disk, etc.

[0061] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting transformer winding deformation based on a combination of short-circuit impedance method and frequency response method, characterized in that, Includes the following steps: The transformer winding under test is connected to a pre-set detection circuit based on a combination of short-circuit impedance method and frequency response method. A sweep frequency signal is injected into the transformer winding under test through the signal source in the detection circuit. When the frequency of the sweep signal is a preset low frequency band, the short-circuit impedance is calculated based on the detection circuit and the electrical parameters obtained by the detection circuit, and the short-circuit impedance is used to determine whether the winding of the transformer under test is deformed. When the frequency of the sweep signal is a preset high frequency band, the winding response signal corresponding to the sweep signal is obtained by the detection circuit. Based on the winding response signal, an improved signal analysis method is used to establish the corresponding frequency response curve, and the deformation of the transformer winding under test is determined based on the frequency response curve.

2. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 1, characterized in that, The preset low frequency band is a frequency band where the sweep frequency signal frequency is less than 1000Hz, and the preset high frequency band is a frequency band where the sweep frequency signal frequency is greater than 1000Hz. Within a preset low-frequency band, the short-circuit impedance value calculated at 50Hz is compared with the short-circuit impedance value under normal conditions. When the rate of change of the short-circuit impedance value exceeds the specified range, it is determined that the winding of the transformer under test has deformed.

3. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 1, characterized in that, The improved signal analysis method includes Fourier transform based on windowing to establish the corresponding frequency response curve, specifically including: The continuous sweep frequency signal and the winding response signal are subjected to Fourier transform based on windowing processing, respectively; The frequency response curve is calculated using the swept frequency signal and the winding response signal after Fourier transform based on windowing.

4. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 3, characterized in that, The expression for the Fourier transform based on windowing is as follows: ； In the formula, For input, it is a continuous time signal; for input, it is the time-domain waveform of the excitation voltage in the frequency sweep signal or the response current in the winding response signal. Let τ represent the window function, τ be the translation amount of the window function, and ω be the angular frequency; is a complex exponential function, and is the kernel function that converts a time-domain signal to the frequency domain to perform the Fourier transform; It is the swept frequency signal or winding response signal after Fourier transform based on windowing processing.

5. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 3, characterized in that, The expression for the frequency response curve is: ； In the formula, This represents the signal after the time-domain response voltage or current signal in the winding response signal has undergone a Fourier transform based on windowing. This represents the signal after the time-domain excitation voltage signal in the frequency sweep signal has undergone a Fourier transform based on windowing. This represents the frequency response curve.

6. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 1, characterized in that, The improved signal analysis method includes establishing the corresponding frequency response curve based on multi-scale complex continuous wavelet transform, specifically including: Multi-scale complex continuous wavelet transform is performed on the continuous sweep frequency signal and the winding response signal, respectively. The frequency response curve is calculated from the swept frequency signal and the winding response signal after multi-scale complex continuous wavelet transform.

7. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 6, characterized in that, The expression based on multi-scale complex continuous wavelet transform is as follows: ； In the formula, The voltage or current signal in the continuous sweep frequency signal or the winding response signal is the input. wavelet basis functions The complex conjugate of is used to calculate the local correlation between the signal and the wavelet; These are wavelet coefficients, representing the signal at different scales. and time The time-frequency energy distribution under [the specified conditions].

8. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 6, characterized in that, The expression for the frequency response curve is: ； In the formula, This represents the wavelet coefficients of the excitation voltage signal in the swept frequency signal. This represents the wavelet coefficients of the current or voltage signal in the winding response signal. The time axis is ω, and the angular frequency is ω. This represents the frequency response curve.

9. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 1, characterized in that, The method also includes a step of diagnosing the deformation of the transformer winding under test by using the frequency response curve corresponding to the sweep frequency signal of any frequency band, specifically: An improved signal analysis method is used to establish the frequency response curve of the swept signal of the transformer winding under test in any frequency band, and the distance characteristic value between the frequency response curve and the normal curve in that frequency band is calculated. The distance feature value is input into the trained GA-SVM classification model for processing to obtain the deformation type of the transformer winding under test.

10. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 9, characterized in that, The formula for calculating the distance characteristic value is as follows: ； in: The distance feature value is obtained by normalizing the distance value. This is the distance value; ； In the formula, This represents the discrete amplitude data of the normal curve within the corresponding frequency band. This represents the discrete amplitude data of the frequency response curve within the corresponding frequency band. This indicates the phase angle corresponding to the normal curve. This indicates the phase angle corresponding to the frequency response curve. This indicates the total number of signal sampling points.

11. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 9, characterized in that, The deformation types mentioned include at least axial deformation, radial deformation, and axis offset; The training process of the GA-SVM classification model includes: Under the same frequency segmentation, obtain the deformation frequency response curves of transformer windings with various deformation types, as well as the normal frequency response curves of normal transformer windings; Each segment of the same frequency is divided into multiple frequency intervals, and multiple deformed frequency response curves and multiple normal frequency response curves are obtained under each frequency interval. Based on the deformed frequency response curves and normal frequency response curves of the same frequency division interval, the distance characteristic value of the deformed frequency response curve of each deformation type is calculated, and a frequency division training set is established through the distance characteristic value of the same frequency division interval. The model optimization parameters of the preset SVM model are obtained by optimizing the GA genetic algorithm, and the SVM model based on the model optimization parameters is trained by the frequency division training set of each frequency division interval. The SVM model is defined as a GA-SVM classification model once it reaches the preset training termination condition.

12. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 11, characterized in that, The training termination condition is that the loss function of the preset SVM model does not exceed a threshold. The expression for the loss function of the preset SVM model is: ； ； ； In the formula, This indicates the number of frequency division intervals. This represents the number of samples in the frequency division training set. For the first In the frequency division interval, the first The true distance feature value of each sample Indicates the first In the frequency division interval, the first The predicted distance feature value of each sample. To preset the first loss function, A second loss function is preset; Indicator variable representing the true label, when the first... The true class of each sample is hour, ,otherwise ; Indicates the first The predicted classification result for each sample is The probability, This represents the total number of samples across all frequency division intervals. The total number of categories; After normalization , After normalization , λ is the loss function of the preset SVM model, and λ is the weight.

13. The transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method according to claim 1, characterized in that, The construction of the detection circuit includes: amplifying the signal output from the signal source through a power amplifier and applying it to one phase winding of the transformer under test; short-circuiting the opposite winding of the transformer under test; connecting sampling resistors at both ends of the winding of the transformer under test; acquiring the voltage signal at both ends of the winding under test and the current signal of the test circuit through the sampling resistors; and the winding response signal including the voltage signal at both ends of the winding under test and the current signal of the test circuit.

14. A system using the transformer winding deformation detection method based on the combination of short-circuit impedance method and frequency response method as described in any one of claims 1-13, characterized in that, The system includes: The signal injection module is used to connect the transformer winding under test to a pre-set detection circuit based on a combination of short-circuit impedance method and frequency response method, and inject a sweep frequency signal into the transformer winding under test through the signal source in the detection circuit. The short-circuit impedance module is used to calculate the short-circuit impedance based on the detection circuit and the electrical parameters obtained by the detection circuit when the frequency of the sweep signal is in the preset low frequency band, and to determine whether the winding of the transformer under test is deformed based on the short-circuit impedance. The sweep frequency response module is used to obtain the winding response signal corresponding to the sweep frequency signal by means of a detection circuit when the frequency of the sweep frequency signal is a preset high frequency band. The curve judgment module is used to establish the corresponding frequency response curve based on the winding response signal using an improved signal analysis method, and to determine whether the winding of the transformer under test is deformed based on the frequency response curve.

15. The transformer winding deformation detection system based on the combination of short-circuit impedance method and frequency response method according to claim 14, characterized in that, The system also includes a deformation diagnosis module, which uses an improved signal analysis method to establish the frequency response curve of the transformer winding under test in any frequency band for the sweep signal, calculates the distance feature value between the frequency response curve and the normal curve in that frequency band, and inputs the distance feature value into the trained GA-SVM classification model for processing to obtain the deformation type of the transformer winding under test.

Images