Intelligent Comprehensive Testing Method and System for Power Transformers

Abstract

This application relates to the technical field of power transformer testing, specifically to an intelligent comprehensive testing method and system for power transformers. The method includes: determining an estimated value of the number of vibration sources of the power transformer at the current moment, analyzing the complexity of each IMF component in the vibration signals at each measurement point, and separating all vibration separation signals from the reconstructed virtual multi-channel observation signals; respectively comparing the fundamental frequencies of each vibration separation signal with the voltage frequency and the current frequency, determining the core vibration signal and the winding vibration signal, analyzing the complexity of the core vibration signal and the winding vibration signal in the vibration signals at each measurement point, determining the core detection coefficient and the winding detection coefficient, and testing the conditions of the core and winding of the power transformer. This application aims to improve the accuracy of the comprehensive test results of power transformers.

Description

Technical Field

[0001] This application relates to the technical field of power transformer testing, and specifically to an intelligent comprehensive testing method and system for power transformers. Background Art

[0002] Existing comprehensive testing methods for power transformers usually perform various tests on power transformers to evaluate abnormal conditions of different components in the power transformer. Among them, the vibration test of the power transformer is an important test method, which is used to detect whether the operating states of the windings and iron cores in the transformer are abnormal. However, the vibration generated on the surface of the transformer tank during its load is mainly caused by the mixed vibration signals generated by the windings and iron cores. Therefore, in order to effectively evaluate the operating states of the windings and iron cores of the transformer, it is necessary to separate the winding vibration signal and the iron core vibration signal of the transformer from the vibration signal on the surface of the tank.

[0003] However, the existing methods do not consider that the vibrations of on-load tap changers, fan rotations, etc. in the transformer will also cause tank vibrations, as well as the noise influence in the collected transformer vibration signals, resulting in the separated signals being unable to accurately reflect the operating conditions of the windings and iron cores of the transformer. Furthermore, it is difficult to accurately evaluate the operating states of the windings and iron cores in the transformer through vibration testing, reducing the accuracy of the comprehensive test results of the transformer. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide an intelligent comprehensive testing method and system for power transformers. The specific technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of this application provides an intelligent comprehensive testing method for power transformers. The method includes the following steps:

[0006] Obtain the vibration signals at each measurement point on the power transformer at the current moment, as well as the load voltage signal and load current signal of the power transformer;

[0007] Use the modal decomposition algorithm to decompose the vibration signals at each measurement point at the current moment into multiple IMF components and a residual component. By analyzing the vibration signals at each measurement point, the discrete degree of the correlation between the residual components of each vibration signal and all IMF components, determine the vibration characteristic values of the vibration signals at each measurement point to obtain the number of vibration sources of the vibration signals at each measurement point; Based on the number of vibration sources of the vibration signals at all measurement points, determine the estimated value of the number of vibration sources of the power transformer at the current moment;

[0008] By analyzing the complexity of each IMF component in the vibration signals at each measurement point, determining the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point, and combining the estimated value of the number of vibration sources, obtaining all vibration separation signals from the reconstructed virtual multi-channel observation signals; obtaining the voltage frequency of the load voltage signal, the current frequency of the load current signal, and the fundamental frequency of each vibration separation signal at the current moment, and respectively comparing the difference between the fundamental frequency of each vibration separation signal and the voltage frequency and the current frequency to determine the core vibration signal and the winding vibration signal of the vibration signals at each measurement point at the current moment;

[0009] Respectively analyze the complexity of the core vibration signal and the winding vibration signal of the vibration signals at each measurement point, determine the core detection coefficient and the winding detection coefficient of the power transformer at the current moment, and test the conditions of the core and the winding of the power transformer at the current moment.

[0010] Preferably, the method for determining the vibration characteristic value of the vibration signal at each measurement point is:

[0011] Taking the set composed of the vibration signals at each measurement point, the remaining components of the vibration signal, and all IMF components as the virtual multi-channel observation signals of the vibration signals at each measurement point;

[0012] Obtaining the covariance matrix of the virtual multi-channel observation signals of the vibration signals at each measurement point, and calculating all the eigenvalues of the covariance matrix as the vibration characteristic values of the vibration signals at each measurement point.

[0013] Preferably, the method for obtaining the number of vibration sources of the vibration signal at each measurement point is:

[0014] Taking all the vibration characteristic values of the vibration signals at each measurement point as the input of the threshold segmentation algorithm, outputting the segmentation threshold, and taking the number of all vibration characteristic values greater than the segmentation threshold as the number of vibration sources of the vibration signals at each measurement point.

[0015] Preferably, the estimated value of the number of vibration sources of the power transformer at the current moment is the mode among the numbers of vibration sources of the vibration signals at all measurement points on the power transformer at the current moment.

[0016] Preferably, the method for determining the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point is:

[0017] In the vibration signals at each measurement point, calculating the approximate entropy of each IMF component, and taking the set composed of the first preset number of IMF components and the corresponding vibration signals in the ascending order of the approximate entropy of all IMF components as the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point.

[0018] Preferably, obtaining all vibration separation signals from the reconstructed virtual multi-channel observation signals includes:

[0019] Use the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point as the input of the independent component analysis algorithm. Among them, use the estimated number of vibration sources of the power transformer as the number of independent components in the independent component analysis algorithm, and output all separated signals as the vibration separation signals of the vibration signals at each measurement point.

[0020] Preferably, the method for determining the core vibration signal and the winding vibration signal of the vibration signals at each measurement point at the current moment is as follows:

[0021] Among the vibration signals at each measurement point, calculate the absolute value of the difference between the fundamental frequency of each vibration separation signal and twice the voltage frequency, and record it as the first absolute value of each vibration separation signal. Among the first absolute values of all vibration separation signals, use the vibration separation signal corresponding to the smallest first absolute value as the core vibration signal at each measurement point;

[0022] Calculate the absolute value of the difference between the fundamental frequency of each vibration separation signal and twice the voltage frequency, and record it as the second absolute value of each vibration separation signal. Among the second absolute values of all vibration separation signals, use the vibration separation signal corresponding to the smallest second absolute value as the winding vibration signal at each measurement point.

[0023] Preferably, the method for determining the core detection coefficient and the winding detection coefficient of the power transformer at the current moment is as follows:

[0024] Use the core vibration signals and winding vibration signals at each measurement point in the power transformer as the input of the Lempel-Ziv complexity algorithm respectively, and output the frequency complexity of the core vibration signal and the frequency complexity of the winding vibration signal at each measurement point;

[0025] Use the mean value of the frequency complexity of the core vibration signals in the vibration signals at all measurement points in the power transformer and the mean value of the frequency complexity of the winding vibration signals in the vibration signals at all measurement points as the core detection coefficient and the winding detection coefficient of the power transformer respectively.

[0026] Preferably, the test on the core and winding conditions of the power transformer at the current moment includes:

[0027] If the core detection coefficient of the power transformer at the current moment is less than the preset first threshold, the core is abnormal, otherwise the core is healthy; if the winding detection coefficient of the power transformer at the current moment is less than the preset second threshold, the winding is abnormal, otherwise the winding is healthy.

[0028] In a second aspect, an embodiment of the present application further provides an intelligent comprehensive test system for a power transformer, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the intelligent comprehensive test method for the power transformer described in any one of the above are implemented.

[0029] The present application has at least the following beneficial effects:

[0030] By successively performing preliminary denoising processing on the collected vibration signals and filtering the remaining noise in the vibration signals by using the estimated value of the number of vibration sources and approximate entropy obtained, the present application can effectively reduce the influence of the noise in the vibration signals on the vibration signals generated by the vibration sources separated in the subsequent vibration test process, thereby improving the accuracy of the winding vibration signals and core vibration signals separated subsequently; further, by performing blind source signal separation on the virtual multi-channel observation signals formed by recombining the IMF components selected from the vibration signals, and using the fundamental frequency of the obtained vibration separation signals to separate the winding vibration signals and core vibration signals in the collected vibration signals, the present application can effectively reduce the influence of the vibrations of components such as on-load tap changers and fans in the transformer on the subsequent separation of winding vibration signals and core vibration signals, improve the accuracy of the separated winding vibration signals and core vibration signals, thereby improving the accuracy of the assessment of the operating states of the windings and cores in the vibration test of the power transformer, and further improving the accuracy of the comprehensive test results of the power transformer. Description of the Drawings

[0031] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0032] Figure 1 It is a flowchart of the steps of an intelligent comprehensive test method for a power transformer provided by an embodiment of the present application;

[0033] Figure 2 It is a schematic diagram of the process of constructing a virtual multi-channel observation signal provided by an embodiment of the present application;

[0034] Figure 3 It is a schematic diagram of the process of obtaining core vibration signals and winding vibration signals provided by an embodiment of the present application. Detailed Embodiments

[0035] To further elaborate on the technical means and effects adopted by this application to achieve the intended invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, detail the specific implementation manner, structure, features, and effects of the intelligent comprehensive testing method and system for power transformers proposed according to this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0036] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs.

[0037] The following will specifically describe the specific solutions of the intelligent comprehensive testing method and system for power transformers provided by this application in conjunction with the accompanying drawings.

[0038] Please refer to Figure 1 , which shows the step flowchart of the intelligent comprehensive testing method for power transformers provided by one embodiment of this application. The method includes the following steps:

[0039] Step S1: Obtain the vibration signals at each measuring point on the power transformer at the current moment, as well as the load voltage signal and load current signal of the power transformer.

[0040] The intelligent comprehensive testing system for power transformers in this application consists of multiple testing modules. The testing modules include insulation resistance testing, direct current resistance testing, turns ratio testing, temperature rise testing, and vibration testing of the power transformer. Among them, insulation resistance testing evaluates the insulation performance of the power transformer winding by measuring the insulation resistance between the high-voltage winding and the low-voltage winding of the power transformer to ensure that there is no leakage or short circuit during operation; direct current resistance testing measures the resistance value of the power transformer winding to ensure that there is no open circuit or poor contact; turns ratio testing measures the voltage ratio of the high- and low-voltage sides of the power transformer to verify whether the turns ratio of the power transformer meets the design requirements; temperature rise testing monitors the temperature of the winding and oil of the power transformer operating under full load conditions to ensure that it does not overheat under normal working conditions; vibration testing measures the vibration signal on the surface of the fuel tank of the power transformer under the load operation state to detect whether there are any abnormalities in the operating state of the winding and iron core of the power transformer.

[0041] During the vibration test, multiple measuring points are set on the surface of the oil tank of the power transformer. A vibration acceleration sensor is installed at each measuring point to collect the vibration signals generated on the surface of the oil tank during the vibration test of the transformer to be tested. The load voltage signal and load current signal of the power transformer within a preset duration before and adjacent to the current moment are respectively obtained, as well as the vibration signals at each measuring point on the power transformer, which are respectively used as the load voltage signal, load current signal of the power transformer at the current moment, and the vibration signals at each measuring point on the power transformer. Among them, the positions of the measuring points can be set by the implementer according to specific circumstances. In this embodiment, the vertical surfaces at 1 / 4, 1 / 2, and 3 / 4 of the height on the high-voltage side of the transformer to be tested and facing the winding position of the transformer to be tested are selected as the measuring points of the power transformer.

[0042] Furthermore, the vibration signals at all measuring points are used as the input of the filtering algorithm to perform noise reduction processing on the vibration signals. It should be noted that there are many commonly used filtering algorithms. In this embodiment, the Gaussian filtering algorithm is used to perform noise reduction processing on the vibration signals. In actual application processes, as other implementation methods, the implementer can also use wavelet denoising algorithms and Kalman filtering algorithms. There are no special restrictions on the selection of filtering algorithms in this embodiment.

[0043] Among them, the Gaussian filtering algorithm is a well-known technology, and the specific process of performing noise reduction processing on the signal will not be elaborated here.

[0044] Step S2: Use the modal decomposition algorithm to decompose the vibration signals at each measuring point at the current moment into multiple IMF components and a residual component. By analyzing the discrete degree of the correlation between the vibration signals, the residual components of each vibration signal, and all IMF components at each measuring point, the vibration characteristic values of the vibration signals at each measuring point are determined to obtain the number of vibration sources of the vibration signals at each measuring point. Based on the number of vibration sources of the vibration signals at all measuring points, the estimated value of the number of vibration sources of the power transformer at the current moment is determined.

[0045] Since the vibration signals of the power transformer are composed of the vibration signals generated by multiple vibration sources in the vibration test, and for the existing problem of separating each independent source signal in the multi-mixed signal, it is mainly solved by using blind source separation technology, such as the FastICA algorithm based on multi-channel blind source separation and the SSA-ICA algorithm based on single-channel blind source separation. Among them, the multi-channel blind source separation method needs to use the signal matrix formed by multiple observation signals to achieve signal separation, while the single-channel blind source separation method needs to decompose the mixed signal collected by a single channel into multiple virtual observation signals. For example, the single-channel signal is decomposed and converted into multi-channel observation signals by using a signal decomposition algorithm, and then the multi-channel blind source separation method is used to solve it.

[0046] By decomposing the vibration signal into multiple signal components and combining the decomposed signal components into a virtual multi-channel observation signal, the separation of the vibration signals generated by different vibration sources in the vibration signal can be achieved. However, the directly decomposed signal components will not only contain the vibration components generated by the vibration sources, but also contain the noise components caused by the noise interference suffered by the vibration signal during its data acquisition. The virtual multi-channel observation signal synthesized using these noise components will affect the separation accuracy of the independent vibration signals generated by different vibration sources in the vibration signal. The vibration sources include, for example, the windings, iron cores, on-load tap changers, and fan rotations of power transformers.

[0047] Therefore, in order to reduce the influence of the residual noise in the vibration signal on the separation of the vibration signals of subsequent different vibration sources, the following processing is carried out, specifically:

[0048] Taking each vibration signal as the input of the improved complete ensemble empirical mode decomposition with adaptive noise algorithm, multiple IMF components and a residual component of each vibration signal are output, where the IMF components correspond to the vibration components or noise components in the vibration signal.

[0049] Among them, the improved complete ensemble empirical mode decomposition with adaptive noise algorithm is a well-known technology, and its specific principle will not be elaborated here.

[0050] Furthermore, combining each vibration signal, all the IMF components of each vibration signal, and the residual to form a virtual multi-channel observation signal, which is specifically expressed as follows:

[0051] B(i)= , where B(i) represents the virtual multi-channel observation signal of the vibration signal at the i-th measuring point; represents the vibration signal at the i-th measuring point; represents the k-th IMF component of the vibration signal at the i-th measuring point; res(i) represents the residual component of the vibration signal at the i-th measuring point; T represents the transpose of the matrix.

[0052] Preferably, the schematic diagram of the construction process of the virtual multi-channel observation signal provided in this embodiment is as Figure 2 shown.

[0053] Since the number of vibration sources that appear in the vibration test of the power transformer is fixed, therefore, by estimating the number of vibration sources and retaining all the IMF components in the vibration signal that contain the vibration information of the vibration sources and the number of retained IMF components is equal to the number of vibration sources, the influence of the residual noise in the vibration signal on the separation of the vibration signals of subsequent different vibration sources can be reduced.

[0054] Specifically, since the vibration signals and noise signals generated by the windings, iron cores, on-load tap changers, and fan rotations of power transformers are usually uncorrelated, and the signal power of the vibration signals is greater than the noise power, there are obvious differences between the vibration signal eigenvalues and noise eigenvalues in the covariance matrix of the virtual multi-channel observation signals. Specifically, the differences between the noise eigenvalues are relatively small, while the differences between the vibration signal eigenvalues and the noise eigenvalues are relatively large.

[0055] Based on the above analysis, obtain the covariance matrix of the virtual multi-channel observation signals of the vibration signals at each measurement point, and calculate all the eigenvalues of the covariance matrix as the vibration eigenvalues of the vibration signals at each measurement point.

[0056] It should be noted that there are many methods for calculating the eigenvalues of a matrix. In this embodiment, the singular value decomposition algorithm is used to obtain all the eigenvalues in the covariance matrix; in actual application processes, as other implementation manners, implementers can also use other eigenvalue decomposition algorithms such as the QR algorithm. Regarding the selection of the method for calculating the eigenvalues of a matrix, this embodiment does not make special restrictions.

[0057] Among them, the method for obtaining the covariance matrix of the signal and the singular value decomposition algorithm are both well-known technologies, and their specific principles will not be elaborated here.

[0058] Furthermore, take all the vibration eigenvalues of the vibration signals at each measurement point as the input of the threshold segmentation algorithm, output the segmentation threshold, and take the number of all vibration eigenvalues greater than the segmentation threshold as the number of vibration sources of the vibration signals at each measurement point.

[0059] It should be noted that there are many commonly used threshold segmentation algorithms. In this embodiment, the maximum between-class variance algorithm is used. In actual application processes, as other implementation manners, implementers can also use other threshold segmentation algorithms. Regarding the selection of the threshold segmentation algorithm, this embodiment does not make special restrictions.

[0060] Among them, the maximum between-class variance algorithm is a well-known technology, and its specific principle process will not be elaborated here.

[0061] Furthermore, take the mode of the number of vibration sources of the vibration signals at all measurement points on the power transformer at the current moment as the estimated value of the number of vibration sources of the power transformer at the current moment, which is used to represent the estimated result of the number of vibration sources that appear in the vibration test of the power transformer.

[0062] Step S3: By analyzing the complexity of each IMF component in the vibration signals at each measurement point, determine the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point, and combine the estimated number of vibration sources to separate all vibration separation signals from the reconstructed virtual multi-channel observation signals; obtain the voltage frequency of the load voltage signal, the current frequency of the load current signal, and the fundamental frequency of each vibration separation signal at the current moment, and respectively compare the differences between the fundamental frequency of each vibration separation signal and the voltage frequency and the current frequency to determine the core vibration signal and the winding vibration signal of the vibration signals at each measurement point at the current moment.

[0063] Since the noise signal usually has more obvious disorder compared to the vibration signals generated by the vibration sources such as windings, cores, on-load tap-changers, and fan rotations in a power transformer, the noise components in the vibration signals have greater signal complexity than the vibration components.

[0064] Therefore, calculate the approximate entropy of each IMF component in the vibration signals at each measurement point respectively to characterize the complexity of the signals in the IMF component. The smaller the approximate entropy of the IMF component, the more vibration information of the vibration sources of the power transformer the IMF component contains, and the less likely it is to be a noise component.

[0065] Among them, the calculation of the approximate entropy of the signal is a well-known technology, and the specific process will not be elaborated.

[0066] Furthermore, in the vibration signals at each measurement point, take the set composed of the first preset number of IMF components and the corresponding vibration signals in the ascending order arrangement result of the approximate entropies of all IMF components as the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point.

[0067] Among them, in this embodiment, the value of the preset number is numerically equal to the estimated number of vibration sources of the power transformer.

[0068] Furthermore, take the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point as the input of the independent component analysis algorithm (FastICA). Among them, take the estimated number of vibration sources of the power transformer as the number of independent components in the independent component analysis algorithm, and output all separation signals as the vibration separation signals of the vibration signals at each measurement point.

[0069] Among them, the independent component analysis algorithm is a well-known technology, and its specific principle will not be elaborated.

[0070] Since the winding vibration of a power transformer is caused by the electromagnetic force generated by the current flowing through the winding, and the amplitude of this vibration is proportional to the square of the winding current, the fundamental frequency of the vibration signal generated by the winding is twice the current frequency; while the iron core of a power transformer is composed of laminated silicon steel sheets, and the magnetostriction of the silicon steel sheets under a strong magnetic field causes the vibration of the iron core, and the amplitude of this vibration is proportional to the square of the voltage, so the fundamental frequency of the vibration signal generated by the iron core is twice the voltage frequency.

[0071] Based on the above analysis, obtain the voltage frequency of the load voltage signal and the current frequency in the load current signal, and extract the fundamental frequency of each vibration separation signal in each vibration signal. Among the vibration signals at each measuring point, calculate the absolute value of the difference between the fundamental frequency of each vibration separation signal and twice the voltage frequency, and record it as the first absolute value of each vibration separation signal. Among the first absolute values of all vibration separation signals, take the vibration separation signal corresponding to the smallest first absolute value as the iron core vibration signal at each measuring point;

[0072] Calculate the absolute value of the difference between the fundamental frequency of each vibration separation signal and twice the voltage frequency, and record it as the second absolute value of each vibration separation signal. Among the second absolute values of all vibration separation signals, take the vibration separation signal corresponding to the smallest second absolute value as the winding vibration signal at each measuring point.

[0073] Preferably, the schematic diagram of the process for obtaining the iron core vibration signal and the winding vibration signal provided in this embodiment is as Figure 3 shown.

[0074] Step S4: Analyze the complexity of the iron core vibration signal and the winding vibration signal of the vibration signal at each measuring point respectively, determine the iron core detection coefficient and the winding detection coefficient of the power transformer at the current moment, and conduct a comprehensive test on the power transformer at the current moment.

[0075] Since the vibration signals generated by the windings and iron cores of power transformers under normal pressure mainly consist of fundamental frequency, 1st, 3rd, and 5th harmonic components, and when the windings and iron cores are abnormal, 50Hz and its odd multiples of frequency components will be generated, making the vibration signals generated by the abnormal windings and iron cores have more complex signal frequency components compared to the vibration signals generated by the normal windings and iron cores.

[0076] Based on the above analysis, take the iron core vibration signal and the winding vibration signal at each measuring point in the power transformer as the input of the Lempel-Ziv complexity algorithm respectively, and output the frequency complexity of the iron core vibration signal and the frequency complexity of the winding vibration signal at each measuring point;

[0077] The average value of the frequency complexity of the core vibration signal at all measurement points in the power transformer and the average value of the frequency complexity of the winding vibration signal at all measurement points are respectively used as the core detection coefficient and the winding detection coefficient of the power transformer.

[0078] Among them, the Lempel-Ziv complexity algorithm is a well-known technology, and the specific process of obtaining the frequency complexity using the Lempel-Ziv complexity algorithm will not be elaborated here.

[0079] During the vibration test process, if the core detection coefficient of the power transformer at the current moment is less than the preset first threshold, the core is abnormal; otherwise, the core is healthy. If the winding detection coefficient of the power transformer at the current moment is less than the preset second threshold, the winding is abnormal; otherwise, the winding is healthy.

[0080] Among them, there is no limitation on the magnitude relationship between the preset first threshold and the preset second threshold. In this embodiment, the values of the preset first threshold and the preset second threshold are both 0.5. Implementers can also set them according to specific situations, and this embodiment does not make special restrictions.

[0081] In addition, the test results of the remaining test modules of the power transformer are obtained respectively. In the insulation resistance test module, if the measured insulation resistance should meet the insulation resistance value specified by the standard, usually required to be at least 1 MΩ / kV, then the output is normal; otherwise, the output is abnormal. In the direct current resistance test module, if the measured winding resistance is within the error range specified by the standard (usually ±5%), and the deviation between the winding resistances of each phase shall not exceed 2%, then the output is normal; otherwise, the output is abnormal. In the turns ratio test module, if the turns ratio error should meet the requirements specified in the standard, usually ±0.5%, then the output is normal; otherwise, the output is abnormal. In the temperature rise test module, if the result meets the temperature rise limit value in the standard (such as the winding temperature rise does not exceed 65 K), then the output is normal; otherwise, the output is abnormal.

[0082] Based on the same inventive concept as the above method, the embodiment of the present application also provides an intelligent comprehensive test system for a power transformer, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned intelligent comprehensive test methods for the power transformer.

[0083] It should be noted that: the above sequence of the embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0084] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the differences between each embodiment and other embodiments are emphasized.

[0085] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principle of the present application shall be included within the protection scope of the present application.

Claims

1. An intelligent comprehensive testing method for power transformers, characterized in that The method includes the following steps: Obtain the vibration signals at each measurement point on the power transformer at the current moment, as well as the load voltage signal and load current signal of the power transformer; Adopt a modal decomposition algorithm to decompose the vibration signals at each measurement point at the current moment into multiple IMF components and a residual component. By analyzing the discrete degree of the correlation between the vibration signals, the residual components of each vibration signal, and all IMF components at each measurement point, determine the vibration characteristic values of the vibration signals at each measurement point to obtain the number of vibration sources of the vibration signals at each measurement point; Based on the number of vibration sources of the vibration signals at all measurement points, determine the estimated value of the number of vibration sources of the power transformer at the current moment; By analyzing the complexity of each IMF component in the vibration signals at each measurement point, determine the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point, and combine the estimated value of the number of vibration sources to obtain all vibration separation signals from the reconstructed virtual multi-channel observation signals; Obtain the voltage frequency of the load voltage signal, the current frequency of the load current signal, and the fundamental frequency of each vibration separation signal at the current moment, and respectively compare the difference between the fundamental frequency of each vibration separation signal and the voltage frequency and the current frequency to determine the core vibration signal and winding vibration signal of the vibration signals at each measurement point at the current moment; Analyze the complexity of the core vibration signal and winding vibration signal of the vibration signals at each measurement point respectively, determine the core detection coefficient and winding detection coefficient of the power transformer at the current moment, and test the conditions of the core and winding of the power transformer at the current moment.

2. The intelligent comprehensive testing method for the power transformer according to claim 1, wherein The method for determining the vibration characteristic values of the vibration signals at each measurement point is as follows: Take the set composed of the vibration signals at each measurement point, the residual component of this vibration signal, and all IMF components as the virtual multi-channel observation signal of the vibration signals at each measurement point; Obtain the covariance matrix of the virtual multi-channel observation signal of the vibration signals at each measurement point, and calculate all the eigenvalues of the covariance matrix as the vibration characteristic values of the vibration signals at each measurement point.

3. The intelligent comprehensive testing method for the power transformer according to claim 1, characterized in that The method for obtaining the number of vibration sources of the vibration signals at each measurement point is as follows: Take all the vibration characteristic values of the vibration signals at each measurement point as the input of the threshold segmentation algorithm, output the segmentation threshold, and take the number of all vibration characteristic values greater than the segmentation threshold as the number of vibration sources of the vibration signals at each measurement point.

4. The intelligent comprehensive testing method for a power transformer according to claim 1, characterized in that, The estimated value of the number of vibration sources of the power transformer at the current moment is the mode among the numbers of vibration sources of the vibration signals at all measurement points on the power transformer at the current moment.

5. The intelligent comprehensive testing method for the power transformer according to claim 1, characterized in that The method for determining the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point is as follows: In the vibration signals at each measurement point, calculate the approximate entropy of each IMF component, and take the set composed of the first preset number of IMF components and the corresponding vibration signals in the ascending order of the approximate entropy of all IMF components as the reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point.

6. The intelligent comprehensive testing method for a power transformer according to claim 1, characterized in that, Obtaining all vibration separation signals from the reconstructed virtual multi-channel observation signals includes: The reconstructed virtual multi-channel observation signals of the vibration signals at each measurement point are used as the input of the independent component analysis algorithm. Among them, the estimated number of vibration sources of the power transformer is used as the number of independent components in the independent component analysis algorithm, and all separated signals are output as the vibration separation signals of the vibration signals at each measurement point.

7. The intelligent comprehensive testing method for a power transformer according to claim 1, characterized in that, The method for determining the core vibration signal and the winding vibration signal of the vibration signals at each measurement point at the current moment is as follows: In the vibration signals at each measurement point, calculate the absolute value of the difference between the fundamental frequency of each vibration separation signal and twice the voltage frequency, which is denoted as the first absolute value of each vibration separation signal. Among the first absolute values of all vibration separation signals, the vibration separation signal corresponding to the smallest first absolute value is used as the core vibration signal at each measurement point; Calculate the absolute value of the difference between the fundamental frequency of each vibration separation signal and twice the voltage frequency, which is denoted as the second absolute value of each vibration separation signal. Among the second absolute values of all vibration separation signals, the vibration separation signal corresponding to the smallest second absolute value is used as the winding vibration signal at each measurement point.

8. The intelligent comprehensive testing method for the power transformer according to claim 1, characterized in that, The method for determining the core detection coefficient and the winding detection coefficient of the power transformer at the current moment is as follows: The core vibration signals and the winding vibration signals at each measurement point in the power transformer are respectively used as the input of the Lempel-Ziv complexity algorithm, and the frequency complexity of the core vibration signal and the frequency complexity of the winding vibration signal at each measurement point are output; The mean value of the frequency complexity of the core vibration signals in the vibration signals at all measurement points in the power transformer and the mean value of the frequency complexity of the winding vibration signals in the vibration signals at all measurement points are respectively used as the core detection coefficient and the winding detection coefficient of the power transformer.

9. The intelligent comprehensive testing method for a power transformer according to claim 1, wherein, The test on the core and winding conditions of the power transformer at the current moment includes: If the core detection coefficient of the power transformer at the current moment is less than the preset first threshold, the core is abnormal, otherwise the core is healthy; if the winding detection coefficient of the power transformer at the current moment is less than the preset second threshold, the winding is abnormal, otherwise the winding is healthy.

10. An intelligent integrated test system for a power transformer, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent comprehensive test method of the power transformer as described in any one of claims 1-9.

Images