Multi-sensor-based power transformer fault automatic early warning device and early warning method

Abstract

The application relates to a multi-sensor-based power transformer fault automatic early warning device and method. The device comprises a determination module, which determines power weight values corresponding to IMF components of each frequency band of the same type according to the energy change correlation between the IMF components of the same frequency band and the energy intensity of the IMF components of each frequency band for decomposed data of the same power type; an analysis module, which analyzes the amplitude fluctuation characteristics of each IMF component in the time domain and the power spectrum density distribution thereof in the frequency domain to obtain power filter values corresponding to each IMF component; a reconstruction module, which reconstructs multiple IMF components in each kind of decomposed data after filtering to obtain each kind of reconstruction data corresponding to each kind of decomposed data; and a warning module, which determines to issue a warning if the average of the differential values of the amplitudes of each kind of reconstruction data is greater than a corresponding preset threshold. The application solves the limitation problem of the traditional filtering method in a complex signal environment and improves the accuracy and reliability of fault warning.

Description

Technical Field

[0001] This application relates to the field of electrical data processing technology, and in particular to an automatic early warning device and method for power transformer faults based on multiple sensors. Background Technology

[0002] In modern society, electricity has become an indispensable energy source for people's production and daily life. The stable operation of the power system is crucial to the normal functioning of society and the sustainable development of the economy. Transformers, as key equipment in the power system that realizes voltage transformation, power distribution, and transmission, are of paramount importance in terms of performance and reliability. Prolonged operation of transformers can easily lead to failures, which can not only affect the power supply in local areas but also cause large-scale power outages, resulting in huge economic losses and impacts on society.

[0003] During transformer operation, the voltage and current signals on the primary and secondary sides are electrical data reflecting the equipment's status. When a transformer malfunctions, abnormal amplitude and waveform distortion will appear in the voltage and current signals. Self-excited oscillation is a common fault phenomenon, where the electromagnetic field inside the transformer undergoes strong nonlinear oscillations, causing abnormal fluctuations in the output voltage and current. However, in real-world scenarios, transformers are often affected by external environmental mechanical vibrations and electromagnetic interference, which can interfere with the waveform period, amplitude, and different frequency bands of the voltage and current signals. When the interference frequency is close to the self-excited oscillation frequency or the interference amplitude is greater than the self-excited oscillation signal, the self-excited characteristics are easily masked, making it difficult to accurately identify self-excited oscillation faults. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide an automatic early warning device and method for power transformer faults based on multiple sensors. The specific technical solution adopted is as follows: In a first aspect, an automatic early warning device for power transformer faults based on multiple sensors is provided, the device comprising: The determination module is used to determine the power weight value corresponding to the IMF components of the same type in each frequency band based on the correlation of energy changes between IMF components in the same frequency band and the energy intensity of IMF components in each frequency band for the decomposed data of the same power type. Multiple decomposed data are obtained by variational mode decomposition of multiple power data of the transformer, including primary side voltage data, secondary side voltage data, primary side current data and secondary side current data. The analysis module is used to analyze the amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain, and obtain the power filtering value corresponding to each IMF component. The reconstruction module is used to reconstruct multiple IMF components in each type of filtered decomposition data according to the corresponding power weight value, so as to obtain each type of reconstructed data corresponding to each type of decomposition data; wherein, the multiple IMF components after filtering are obtained by performing Gaussian kernel standard deviation filtering on each of the multiple IMF components, and the Gaussian kernel standard deviation is determined according to the power filtering value corresponding to each IMF component. The early warning module is used to issue an early warning if the mean difference of the amplitude of each type of reconstructed data is greater than the corresponding preset threshold; the preset threshold corresponding to each type of reconstructed data is determined based on the corresponding historical power data.

[0005] Optionally, the determining module is also used for: For decomposed data of the same power type, calculate the cross power spectral density of the IMF components of the primary and secondary sides in the same frequency band, the auto-power spectral density of the IMF components of the primary side in the same frequency band, and the auto-power spectral density of the IMF components of the secondary side in the same frequency band. The power weight value corresponding to the IMF component of this frequency band in this power type is determined based on the ratio of the square of the cross power spectral density to the density product; the density product indicates the product of the self-power spectral density of the IMF component of the primary side in this frequency band and the self-power spectral density of the IMF component of the secondary side in this frequency band.

[0006] Optionally, the analysis module is also used for: For each type of decomposed data, perform Hilbert transform on each IMF component to determine the amplitude of each IMF component at each sampling point. Based on the mean and variance of the absolute values ​​of the differences of all amplitudes corresponding to each IMF component, determine the amplitude fluctuation characteristic value corresponding to each IMF component of this decomposed data. Based on the amplitude fluctuation characteristic value of each IMF component corresponding to each type of decomposed data and the power spectral density distribution of the corresponding IMF component at each frequency point in the frequency domain, the power filtering value corresponding to the IMF component is obtained.

[0007] Optionally, the analysis module is also used for: For each type of decomposed data, perform a Fourier transform on each IMF component to determine the power of each frequency point in each IMF component, and calculate the total power of all frequency points in each IMF component. Based on the amplitude fluctuation characteristic value of each IMF component corresponding to each type of decomposed data and the proportion of power at each frequency point in the corresponding IMF component to the total power, the power filtering value corresponding to that IMF component is obtained.

[0008] Optionally, the refactoring module is also used for: The Gaussian kernel standard deviation corresponding to each IMF component is determined by multiplying the power filter value corresponding to each IMF component in each decomposition data with the preset coefficient. Each IMF is filtered based on the Gaussian kernel standard deviation corresponding to each IMF component, resulting in multiple filtered IMF components.

[0009] Optionally, the refactoring module is also used for: Based on the power weight value corresponding to each IMF component, determine the reconstruction weight of each filtered IMF component; Based on the reconstruction weights of multiple IMF components and the corresponding IMF components in each type of decomposed data after filtering, each type of reconstructed data is obtained.

[0010] Optionally, the refactoring module is also used for: The reconstruction weight of each IMF component after filtering is determined by the ratio of the power weight value corresponding to each IMF component to the sum of the power weight values ​​corresponding to multiple IMFs.

[0011] Optionally, the refactoring module is also used for: The reconstructed data for each type of decomposed data is obtained by summing the products of the reconstruction weights of the multiple IMF components in each type of decomposed data and the corresponding IMF components.

[0012] Optionally, the early warning module is also used for: The signal amplitude difference value of each historical power data at each time moment is calculated using the first-order backward difference method, and the mean and variance of the signal amplitude difference value of each historical power data are calculated. Based on the mean and variance of the signal amplitude difference for each type of historical power data, a preset threshold is determined for each type of historical power data.

[0013] Secondly, a method for automatic early warning of power transformer faults based on multiple sensors is provided, the method comprising: For decomposed data of the same type of electrical quantity, the power weight value corresponding to the IMF component of each frequency band of the same type is determined based on the correlation of energy changes between IMF components of the same frequency band and the energy intensity of IMF components of each frequency band. Multiple decomposed data are obtained by variational mode decomposition of multiple power data of transformers, including primary voltage data, secondary voltage data, primary current data and secondary current data. The amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain are analyzed to obtain the power filtering value corresponding to each IMF component. Based on the corresponding power weight value, the multiple IMF components in each decomposed data after filtering are reconstructed to obtain each reconstructed data corresponding to each decomposed data. Among them, the multiple IMF components after filtering are obtained by performing Gaussian kernel standard deviation filtering on each of the multiple IMF components. The Gaussian kernel standard deviation is determined according to the power filtering value corresponding to each IMF component. If the mean difference of the amplitude of each type of reconstructed data is greater than the corresponding preset threshold, an early warning will be issued; the preset threshold for each type of reconstructed data is determined based on the corresponding historical power data.

[0014] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this application.

[0015] This application offers the following advantages: it overcomes the limitations of traditional filtering methods in complex signal environments by adaptively assigning reconstruction weights to IMF components in different frequency bands, and then reconstructing the denoised components accordingly. The weight allocation is based on the correlation of energy changes between the two-sided components and the energy intensity of the components themselves. This ensures that signal components effectively characterizing the electromagnetic coupling properties within the transformer are enhanced during reconstruction, while signal components caused by local external interference that disrupt the correlation between the two sides are suppressed, thereby optimizing the output signal.

[0016] Intelligent adaptive adjustment of filter parameters is achieved: the filter adjustment value is obtained by analyzing the amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain, and the kernel standard deviation of the Gaussian filter is determined independently for each IMF component. This allows the filter to adaptively adjust the filter intensity according to the unique fluctuation pattern of each component: strong smoothing is applied to components with drastic fluctuations and complex frequency components to suppress noise; weak smoothing is applied to components with stable fluctuations and concentrated frequency components to preserve potential fault characteristics, thereby avoiding the loss of effective signal while denoising.

[0017] The accuracy and reliability of fault early warning are improved: the reconstructed data obtained through the above adaptive reconstruction and adaptive filtering process can more accurately reflect the real changes in the transformer's operating status. By monitoring whether the average value of the amplitude change of this reconstructed data signal exceeds a threshold set based on historical normal operating data, early warning is issued, significantly reducing the missed and false alarm rates for early and weak fault characteristics, and improving the overall performance of the early warning device. Attached Figure Description

[0018] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the structure of an automatic early warning device for power transformer faults based on multiple sensors in one embodiment; Figure 2 This is a flowchart illustrating an automatic early warning system for power transformer faults based on multiple sensors, as described in one embodiment. Figure 3 This is a schematic diagram of the structure of an electronic device in one embodiment. Detailed Implementation

[0020] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an automatic early warning device and method for power transformer faults based on multiple sensors proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0022] The following description, in conjunction with the accompanying drawings, details a specific scheme for an automatic early warning device for power transformer faults based on multiple sensors, as provided in this application. For example... Figure 1 As shown, the device includes: The determination module 11 is used to determine the power weight value corresponding to the IMF component of each frequency band based on the correlation of energy changes between IMF components in the same frequency band and the energy intensity of IMF components in each frequency band for the decomposed data of the same power type.

[0023] Among them, multiple decomposed data are obtained by variational mode decomposition of multiple power data of the transformer. The multiple power data include primary side voltage data, secondary side voltage data, primary side current data and secondary side current data. Each decomposed data corresponds one-to-one with each power data. Each decomposed data contains multiple IMF (Intrinsic Mode Function) components of the corresponding voltage data. Multiple IMF components correspond one-to-one with multiple frequency bands.

[0024] Decomposition data of the same electrical quantity type indicates decomposition data with the same physical properties. For example, the current data on the primary side and the current data on the secondary side are of the same electrical quantity type, and the voltage data on the primary side and the voltage data on the secondary side are of the same electrical quantity type.

[0025] The multi-sensor-based automatic fault early warning device for power transformers provided in this application embodiment is used for fault detection and warning of power transformers. The device also includes a data acquisition unit module, a host module, a communication and storage module, an early warning module, and a power supply module. The data acquisition module contains multiple data acquisition units: a primary side voltage and current acquisition unit and a secondary side voltage and current acquisition unit, respectively installed on the primary side terminals (usually the high-voltage side) and secondary side terminals (usually the low-voltage side) of the transformer via voltage and current transformers, for real-time acquisition of voltage and current waveforms on the high-voltage and low-voltage sides. The host module uses an industrial-grade embedded processor to receive and process the data acquired by the data acquisition module. It can perform synchronous sampling, signal decomposition, and data analysis, detecting the transformer's operating status through changes in the transformer's power data. The communication and storage module is responsible for uploading the detection results and historical data to the power management center via a wireless network and storing the transformer's operating data locally. The early warning module consists of a buzzer and a warning light; when a transformer fault occurs, the host module controls the early warning module to issue a warning. The power supply module is a 220V power supply independent of the transformer. The enclosure of the multi-sensor-based automatic early warning device for power transformer faults must have an IP54 rating (dustproof level 5, waterproof level 4) or IP65 rating (dustproof level 6, waterproof level 5) and electromagnetic interference resistance to prevent electromagnetic interference from the transformer and external environmental interference from affecting the normal operation of the early warning device. The control flow of the multi-sensor-based automatic early warning device for power transformer faults is as follows: device power-on → multi-channel synchronous sampling → signal preprocessing → data feature analysis and fault index calculation → fault determination → device alarm issuance → data transmission, realizing full online monitoring and fault early warning of the transformer's operating status.

[0026] The primary-side voltage and current acquisition unit is connected to the primary-side terminals of the transformer (usually the high-voltage side) via voltage and current transformers, while the secondary-side voltage and current acquisition unit is connected to the secondary-side terminals (usually the low-voltage side) via voltage and current transformers. Voltage data from both the primary and secondary sides is acquired using the voltage transformers, and current data from both sides is acquired using the current transformers, yielding the transformer's power data. Then, the Sinc interpolation method (Sinc interpolation) is used to reconstruct a continuous time-series signal from the various power data signals. The Sinc interpolation method is an existing technology, and its specific process will not be elaborated upon. The sampling frequency for the power data is 10kHz, and the sampling duration is 10 minutes. The sampling frequency and sampling duration can be set according to actual conditions, but for the reliability of the acquired signal, the sampling frequency setting should satisfy the Nyquist sampling theorem.

[0027] Transformer faults (such as self-excited oscillation) can cause abnormal distortions in power data waveforms. Monitoring changes in the transformer's voltage and current signals can be used to detect these faults. However, when a transformer is affected by external environmental factors such as mechanical vibration and electromagnetic interference, electromagnetic coupling effects can cause abnormal amplitudes and waveforms in the voltage and current signals, generating complex harmonic effects at different frequency bands. When the interference frequency is close to the self-excited oscillation frequency or the interference amplitude is greater than the self-excited oscillation signal, the data characteristics of the self-excited oscillation in the voltage and current signals will be masked by the interference signal, making it difficult to accurately identify the self-excited oscillation fault and thus affecting the accuracy of power transformer fault detection.

[0028] This application considers the correlation and data characteristic changes between the voltage and current data on the primary and secondary sides of the transformer during operation, respectively, to determine the power weight values ​​corresponding to the IMF components in each frequency band. The collected power data set can be represented as Y, where... , For the first Data from sampling points, with a total of sampling points. ,in , This represents the primary side voltage data. This represents the primary side current data. Represents secondary side voltage data. This represents the secondary current data.

[0029] Considering that transformer self-oscillation and external interference can cause complex multi-frequency interference to power data signals, resulting in the superposition of fluctuations and disturbances in different frequency bands on the stable power data signal, it is difficult to accurately analyze the signal variation characteristics under the entire power data signal. Firstly, variational mode decomposition is performed on each type of power data to obtain the corresponding... Each IMF component, that is, each decomposition data includes One IMF component, The size can be set according to the actual situation without special restrictions. In this embodiment, The size is set to 6. This embodiment decomposes complex power data signals into multiple IMF components in different frequency bands, which facilitates the analysis of the variation characteristics of self-excited oscillation and external interference noise from each component level.

[0030] In a power system, electrical energy input to the primary side is converted from high-voltage to low-voltage electricity via magnetic field coupling and transmitted to the secondary side. Electromagnetic induction makes the voltage and current on the secondary side proportional to the voltage and current on the primary side, respectively. Therefore, the voltage and current signals on the primary and secondary sides are highly correlated in terms of data changes. However, the power data signals can fluctuate significantly due to self-excited oscillations and external interference. Self-excited oscillations, caused by the continuous oscillation phenomenon resulting from the electromagnetic coupling of the windings inside the transformer, cause both the primary and secondary power data to be affected by the same fluctuations under the influence of internal electromagnetic coupling. External interference, through transformer vibration or electromagnetic disturbances, interferes with the power data but does not affect the internal electromagnetic coupling effect of the transformer. It only significantly affects the power data on the side closer to the interference source; the primary and secondary power data will not exhibit the same fluctuations under external interference.

[0031] Based on the correlation of data changes among primary-side voltage data, secondary-side voltage data, primary-side current data, and secondary-side current data, the power weight value corresponding to the IMF component of each frequency band in the decomposed data of the same power type is determined, which is used to characterize the correlation between the IMF components of power data signals under different frequency bands. Therefore, in one embodiment, the determining module is further used to: For decomposed data of the same power type, calculate the cross power spectral density of the IMF components of the primary and secondary sides in the same frequency band, the auto-power spectral density of the IMF components of the primary side in the same frequency band, and the auto-power spectral density of the IMF components of the secondary side in the same frequency band. The power weight value corresponding to the IMF component of this frequency band in this power type is determined based on the ratio of the square of the cross power spectral density to the density product; the density product indicates the product of the self-power spectral density of the IMF component of the primary side in this frequency band and the self-power spectral density of the IMF component of the secondary side in this frequency band.

[0032] The first side voltage data One IMF component (i.e., the kth frequency band) and the second-side voltage data IMF components For example, first calculate... and Self-power spectral density and cross-power spectral density The self-power spectral density represents the energy distribution of a signal at different frequencies, reflecting the strength of each frequency component of the signal in the frequency domain. The cross-power spectral density, on the other hand, represents how the energy of two signals changes simultaneously at the same frequency, reflecting the correlation between the changes in the two signals. and The Electricity weight values ​​of each IMF component It can be represented as: ; in, The first indicator of voltage data The power weight values ​​corresponding to each frequency band (i.e., the kth IMF component), and the voltage data include primary-side voltage data and secondary-side voltage data. The first indicator of primary side voltage data One IMF component, The first indicator of secondary voltage data One IMF component, instruct and cross power spectral density, instruct The self-power spectral density, instruct The self-power spectral density.

[0033] The power weight value reflects the correlation between IMF components of the same type of power data on different sides. It can be used as an adaptive reconstruction weight for each component during subsequent IMF component reconstruction, controlling the importance of different IMF components during reconstruction. Self-excited oscillations and external interference can cause disturbances to voltage data signals across multiple frequency bands, which are reflected in the corresponding IMF components at different frequency bands. Since self-excited oscillations cause voltage data on both sides to fluctuate to the same degree, the IMF components on these two sides at the same frequency band exhibit high correlation. The larger the value, the more likely the IMF component in that frequency band exhibits normal data characteristics of transformer operation, with the primary and secondary voltages maintaining their original linear relationship. Noise disturbances are also less in the IMF component within this frequency band. Conversely, under external interference, the voltage data on both sides show asynchronous changes in signal fluctuations. The smaller the value, the more noise disturbances are present in the IMF component of that frequency band.

[0034] Similarly, based on the ratio of the square of the cross-power spectral density to the product of the self-power spectral density of the k-th IMF component of the primary and secondary current data, the power weight value corresponding to the k-th IMF component of the current data is determined, serving as the adaptive weight for subsequent reconstruction of the IMF components. The self-power spectral density product indicates the power weight of the k-th IMF component of the primary current data. The self-power spectral density of the IMF component and the second-side current data of the first IMF component The product of the self-power spectral densities of each IMF component.

[0035] Analysis module 12 is used to analyze the amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain, and obtain the power filtering value corresponding to each IMF component.

[0036] The power spectral density distribution of each IMF component in the frequency domain indicates the curve of power versus frequency obtained after performing a time-frequency transformation on the component (usually using a Welch periodogram or FFT).

[0037] The power filtering value corresponding to each IMF component can be used as an adaptive weight of the Gaussian kernel standard deviation when the Gaussian filter built into the early warning device filters the IMF component, so as to adjust the filtering and noise reduction effect of the IMF component.

[0038] The amplitude of the IMF component in transformer power data can measure the signal strength within that component. Under stable operating conditions, the amplitude variation is relatively stable. However, when the transformer experiences self-oscillation faults or is subjected to external interference, the amplitude fluctuations in the power data become significant. This makes it difficult for the Gaussian filter built into the early warning device to accurately denoise the power data signal based solely on the signal amplitude at each data sampling point. In the early stages of a fault, small fluctuations in the amplitude of each component of the power data can be mistaken for noise interference, leading to the distortion of the true power data signal through filtering and denoising. The changes in power data caused by self-oscillation are continuous, and the accumulation of abnormal power data over a period of time damages the transformer circuit. Furthermore, the changes in power data caused by external interference exhibit transient spike disturbance characteristics across various frequency bands.

[0039] The power filtering value corresponding to each IMF component can be determined by considering the amplitude fluctuation characteristics and spectral entropy of each IMF component under different conditions. Therefore, in one embodiment, the analysis module is also used for: For each type of decomposed data, perform Hilbert transform on each IMF component to determine the amplitude of each IMF component at each sampling point. Based on the mean and variance of the absolute values ​​of the differences of all amplitudes corresponding to each IMF component, determine the amplitude fluctuation characteristic value corresponding to each IMF component of this decomposed data. Based on the amplitude fluctuation characteristic value of each IMF component corresponding to each type of decomposed data and the power spectral density distribution of the corresponding IMF component at each frequency point in the frequency domain, the power filtering value corresponding to the IMF component is obtained.

[0040] Furthermore, the analysis module is also used for: For each type of decomposed data, perform a Fourier transform on each IMF component to determine the power of each frequency point in each IMF component, and calculate the total power of all frequency points in each IMF component. Based on the amplitude fluctuation characteristic value of each IMF component corresponding to each type of decomposed data and the proportion of power at each frequency point in the corresponding IMF component to the total power, the power filtering value corresponding to that IMF component is obtained.

[0041] It can be obtained through Hilbert transform Amplitude at each data sampling point Then, the first-order backward difference method is used to calculate... The absolute value of the difference between all amplitudes The absolute value of the difference between all amplitudes The calculation formula is: t represents the t-th sampling point, and T represents the number of sampling points. for The amplitude at the t-th sampling point. Then calculate the mean of the absolute values ​​of the amplitude differences. With variance The Hilbert transform and the first-order backward difference method are existing techniques, and their specific processes will not be elaborated here. The corresponding amplitude fluctuation characteristic value is , The calculation formula is , To avoid the denominator being zero, a small constant is used in this embodiment. Take 0.001, It can represent The magnitude of the fluctuations in the amplitude can provide a preliminary measure of anomalies in IMF components. A larger value indicates that the IMF component in that frequency band has experienced significant fluctuations in signal strength, but noise interference in that frequency band is difficult to distinguish.

[0042] Furthermore, construct Spectral entropy index ,in For the first The proportion of the energy value (power value) at each frequency point to the total energy (total power) of that IMF component is determined by... After performing a Fourier transform to the frequency domain, find the first... The energy of each frequency point in all The energy percentage at each frequency point is calculated. Spectral entropy reflects the complexity of the signal's frequency components. External interference causes transient spikes in the amplitude of the IMF component, resulting in energy dispersion and thus higher spectral entropy. Conversely, persistent amplitude changes caused by faults result in lower spectral entropy. Fourier transform is a current technique, and its specific process will not be elaborated further. Therefore… Corresponding power filter value for: ;in, for The corresponding amplitude fluctuation characteristic value, for The corresponding spectral entropy index.

[0043] Constructed Considering the variations in the amplitude of the IMF component signal in the power data and the signal complexity of the IMF component in this frequency band, the IMF signal changes caused by external interference exhibit significant amplitude and spectral entropy. If there is a lot of noise interference in the IMF component at a certain frequency band, then its... If the value is relatively large, then the IMF component in that frequency band may be a signal fluctuation caused by a fault, which can more clearly characterize the different fluctuations of the IMF component.

[0044] The reconstruction module 13 is used to reconstruct multiple IMF components in each type of filtered decomposition data according to the corresponding power weight value, so as to obtain each type of reconstructed data corresponding to each type of decomposition data.

[0045] The filtered IMF components are obtained by performing Gaussian kernel standard deviation filtering on each IMF component. The Gaussian kernel standard deviation is determined based on the power filtering value corresponding to each IMF component.

[0046] In one embodiment, the refactoring module is further configured to: The Gaussian kernel standard deviation corresponding to each IMF component is determined by multiplying the power filter value corresponding to each IMF component in each decomposition data with the preset coefficient. Each IMF is filtered based on the Gaussian kernel standard deviation corresponding to each IMF component, resulting in multiple filtered IMF components.

[0047] The standard deviation of the Gaussian kernel of the Gaussian filter built into the early warning device can be expressed as: ; in, for The corresponding Gaussian kernel standard deviation, These are the maximum and minimum values ​​of the Gaussian kernel standard deviation, respectively. for The corresponding power filtering value, , , The size can be set according to the actual situation. In this embodiment, The values ​​are 0.5 and 2 respectively.

[0048] The function expression is: ; The Gaussian kernel standard deviation designed in this way can be adaptively adjusted according to the signal fluctuation characteristics of different IMF components, avoiding the situation where relying solely on the signal amplitude of each data sampling point to denoise the IMF components of the power data signal can smooth out small signal fluctuations caused in the early stages of a fault, thus affecting the denoising effect on the data.

[0049] The adaptive Gaussian kernel standard deviation is set as a parameter in the Gaussian filter of the early warning device. The process of filtering and denoising, specifically Gaussian filtering, is an existing technique, and its details will not be elaborated here.

[0050] Similarly, the analysis module and the reconstruction module repeat all the above steps to denoise all IMF components of the primary side voltage data, all IMF components of the primary side current data, all IMF components of the secondary side voltage data, and all IMF components of the secondary side current data, respectively.

[0051] In one embodiment, the refactoring module is further configured to: Based on the power weight value corresponding to each IMF component, determine the reconstruction weight of each filtered IMF component; Based on the reconstruction weights of multiple IMF components and the corresponding IMF components in each type of decomposed data after filtering, each type of reconstructed data is obtained.

[0052] Furthermore, the refactoring module is also used for: The reconstruction weight of each IMF component after filtering is determined by the ratio of the power weight value corresponding to each IMF component to the sum of the power weight values ​​corresponding to multiple IMFs.

[0053] The refactoring module is also used for: The reconstructed data for each type of decomposed data is obtained by summing the products of the reconstruction weights of the multiple IMF components in each type of decomposed data and the corresponding IMF components.

[0054] Based on the power weight value corresponding to each IMF component, the IMFs of the denoised primary and secondary voltage and current data are reconstructed to obtain the denoised reconstructed data. Taking primary and secondary voltage data as examples, the reconstruction weights of the primary and secondary voltage data are... for: ;in, The reconstructed weights for the voltage data, which includes primary-side voltage data and secondary-side voltage data, For the voltage data of the first The electricity weight values ​​corresponding to each IMF component For the first There are 1 IMF component, where K is the number of IMF components.

[0055] Therefore, the reconstructed data of the primary and secondary voltage data after denoising are as follows: ; in, This is the reconstructed data of the primary side voltage after denoising. This is the reconstructed data of the secondary side voltage after denoising. For the first There are multiple IMF components, where K is the number of IMF components. For the reconstruction weights of voltage data, This is the primary side voltage data after noise reduction. This is the noise-reduced secondary voltage data.

[0056] Similarly, based on the reconstruction weights of the primary and secondary current data, the denoised primary and secondary current data are reconstructed to obtain the reconstructed primary current data and the reconstructed secondary current data after denoising.

[0057] Because the changes in power data of the same quantity type on the primary and secondary sides of a transformer are correlated, the power weight values ​​of the corresponding quantity type are used when reconstructing the IMF components. The first... The higher the correlation of an IMF component, the less external noise interference it is affected by, and the better it reflects the transformer's operating state. Therefore, its weight should be adjusted during IMF component reconstruction. A higher weight indicates more noise interference in the IMF component of that frequency band, and vice versa, thus reducing its weight during reconstruction. The smaller.

[0058] The early warning module 14 is used to determine and issue an early warning if the mean difference value of the amplitude of each reconstructed data is greater than the corresponding preset threshold.

[0059] The preset threshold for each type of reconstructed data is determined based on the historical power data of the corresponding transformer.

[0060] In one embodiment, the warning module is further used for: The signal amplitude difference value of each historical power data at each time moment is calculated using the first-order backward difference method, and the mean and variance of the signal amplitude difference value of each historical power data are calculated. Based on the mean and variance of the signal amplitude difference for each type of historical power data, a preset threshold is determined for each type of historical power data.

[0061] Historical power data can be obtained through the communication and storage unit of the early warning device. One hour of historical power data from a transformer during which no faults occurred is selected as the standard operating data for a normal transformer. Then, the first-order backward difference method is used to calculate the signal amplitude difference values ​​for the primary side voltage data, secondary side voltage data, primary side current data, and secondary side current data at each moment in the standard operating data for the normal transformer, with a difference interval of 1 second. The mean and variance of the signal amplitude difference values ​​for each type of historical power data are then obtained. A threshold is set according to the three-fold variance criterion based on the mean and variance of the amplitude difference values ​​for each signal. , These are the threshold values ​​for primary side voltage, primary side current, secondary side voltage, and secondary side current, respectively. The three-fold variance criterion is existing technology, and the specific process will not be elaborated here.

[0062] The mean of the signal amplitude difference values ​​of the various reconstructed data after denoising is then calculated within the sampling time. If the mean difference value of each signal is greater than the corresponding preset threshold, the signal segment is determined to be an abnormal signal, indicating a fault. The early warning device controls the early warning unit to issue a warning through the host unit, and simultaneously synchronizes the fault status to the power management center.

[0063] This application addresses the limitations of traditional filtering methods in complex signal environments by adaptively assigning reconstruction weights to IMF components in different frequency bands and reconstructing the denoised components accordingly. The weight allocation is based on the correlation of energy changes between the two-sided components and the energy intensity of the components themselves. This ensures that signal components effectively characterizing the electromagnetic coupling properties within the transformer are enhanced during reconstruction, while signal components caused by local external interference that disrupt the correlation between the two sides are suppressed, thereby optimizing the output signal.

[0064] Intelligent adaptive adjustment of filter parameters is achieved: the filter adjustment value is obtained by analyzing the amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain, and the kernel standard deviation of the Gaussian filter is determined independently for each IMF component. This allows the filter to adaptively adjust the filter intensity according to the unique fluctuation pattern of each component: strong smoothing is applied to components with drastic fluctuations and complex frequency components to suppress noise; weak smoothing is applied to components with stable fluctuations and concentrated frequency components to preserve potential fault characteristics, thereby avoiding the loss of effective signal while denoising.

[0065] The accuracy and reliability of fault early warning are improved: the reconstructed data obtained through the above adaptive reconstruction and adaptive filtering process can more accurately reflect the real changes in the transformer's operating status. By monitoring whether the average value of the amplitude change of this reconstructed data signal exceeds a threshold set based on historical normal operating data, early warning is issued, significantly reducing the missed and false alarm rates for early and weak fault characteristics, and improving the overall performance of the early warning device.

[0066] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs.

[0067] This application also provides an automatic early warning method for power transformer faults based on multiple sensors, such as... Figure 2 As shown, the method includes: S21. For the decomposed data of the same type of electrical quantity, the power weight value corresponding to the IMF component of each frequency band of the same type is determined according to the correlation of energy change between IMF components of the same frequency band and the energy intensity of IMF components of each frequency band. Multiple decomposed data are obtained by variational mode decomposition of multiple power data of the transformer. Multiple power data include primary side voltage data, secondary side voltage data, primary side current data and secondary side current data. S22. Analyze the amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain to obtain the power filtering value corresponding to each IMF component. S23. Reconstruct the multiple IMF components in each decomposed data after filtering according to the corresponding power weight value to obtain each reconstructed data corresponding to each decomposed data; wherein, the multiple IMF components after filtering are obtained by performing Gaussian kernel standard deviation filtering on each of the multiple IMF components, and the Gaussian kernel standard deviation is determined according to the power filtering value corresponding to each IMF component. S24. If the mean difference of the amplitude of each type of reconstructed data is greater than the corresponding preset threshold, an early warning is issued; the preset threshold corresponding to each type of reconstructed data is determined based on the corresponding historical power data.

[0068] For the method embodiments, since they are basically corresponding to the apparatus embodiments, the relevant parts can be referred to in the description of the apparatus embodiments.

[0069] It should be understood that, although Figure 2 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 2 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0070] Figure 3 This is a schematic diagram of the structure of an electronic device according to an example embodiment of this application. The electronic device includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the method described in any of the above embodiments. Figure 3 The electronic device 30 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0071] like Figure 3 As shown, the electronic device 30 can be manifested as a general-purpose computing device, such as a server device. The components of the electronic device 30 may include, but are not limited to: at least one processor 31, at least one memory 32, and a bus 33 connecting different system components (including memory 32 and processor 31).

[0072] Bus 33 includes a data bus, an address bus, and a control bus.

[0073] The memory 32 may include volatile memory, such as random access memory (RAM) 321 and / or cache memory 322, and may further include read-only memory (ROM) 323.

[0074] The memory 32 may also include a program tool 325 (or utility) having a set (at least one) program module 324, such program module 324 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0075] The processor 31 executes various functional applications and data processing, such as the methods provided in any of the above embodiments, by running computer programs stored in the memory 32.

[0076] Electronic device 30 can also communicate with one or more external devices 34 (e.g., keyboard, pointing device, etc.). This communication can be performed via input / output (I / O) interface 35. Furthermore, electronic device 30 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public network, such as the Internet) via network adapter 36. As shown, network adapter 36 communicates with other modules of electronic device 30 via bus 33. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with electronic device 30, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0077] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0078] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method provided in any of the above embodiments.

[0079] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0080] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0081] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the above embodiments.

[0082] The program code for executing the computer program product of this application can be written in any combination of one or more programming languages. The program code can be executed entirely on the user device, partially on the user device, as a standalone software package, partially on the user device and partially on a remote device, or entirely on a remote device.

[0083] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0084] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application.

[0085] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An automatic early warning device for power transformer faults based on multiple sensors, characterized in that, The device includes: The determination module is used to determine the power weight value corresponding to the IMF components of the same type in each frequency band based on the correlation of energy changes between IMF components in the same frequency band and the energy intensity of IMF components in each frequency band for the decomposed data of the same power type. Multiple decomposed data are obtained by variational mode decomposition of multiple power data of the transformer, including primary side voltage data, secondary side voltage data, primary side current data and secondary side current data. The analysis module is used to analyze the amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain, and obtain the power filtering value corresponding to each IMF component. The reconstruction module is used to reconstruct multiple IMF components in each type of filtered decomposition data according to the corresponding power weight value, so as to obtain each type of reconstructed data corresponding to each type of decomposition data; wherein, the multiple IMF components after filtering are obtained by performing Gaussian kernel standard deviation filtering on each of the multiple IMF components, and the Gaussian kernel standard deviation is determined according to the power filtering value corresponding to each IMF component. The early warning module is used to issue an early warning if the mean difference value of the amplitude of each type of reconstructed data is greater than the corresponding preset threshold; the preset threshold corresponding to each type of reconstructed data is determined based on the corresponding historical power data. The determination module is also used for: For decomposed data of the same power type, calculate the cross power spectral density of the IMF components of the primary and secondary sides in the same frequency band, the auto-power spectral density of the IMF components of the primary side in the same frequency band, and the auto-power spectral density of the IMF components of the secondary side in the same frequency band. The power weight value corresponding to the IMF component of this frequency band in this power type is determined based on the ratio of the square of the cross power spectral density to the density product; the density product indicates the product of the self-power spectral density of the IMF component of the primary side in this frequency band and the self-power spectral density of the IMF component of the secondary side in this frequency band. The analysis module is also used for: For each type of decomposed data, perform Hilbert transform on each IMF component to determine the amplitude of each IMF component at each sampling point. Based on the mean and variance of the absolute values ​​of the differences of all amplitudes corresponding to each IMF component, determine the amplitude fluctuation characteristic value corresponding to each IMF component of this decomposed data. Based on the amplitude fluctuation characteristic value of each IMF component corresponding to each type of decomposed data and the power spectral density distribution of the corresponding IMF component at each frequency point in the frequency domain, the power filtering value corresponding to the IMF component is obtained.

2. The automatic early warning device for power transformer faults based on multiple sensors as described in claim 1, characterized in that, The analysis module is also used for: For each type of decomposed data, perform a Fourier transform on each IMF component to determine the power of each frequency point in each IMF component, and calculate the total power of all frequency points in each IMF component. Based on the amplitude fluctuation characteristic value of each IMF component corresponding to each type of decomposed data and the proportion of power at each frequency point in the corresponding IMF component to the total power, the power filtering value corresponding to that IMF component is obtained.

3. The automatic early warning device for power transformer faults based on multiple sensors as described in claim 1 or 2, characterized in that, The reconstruction module is also used for: The Gaussian kernel standard deviation corresponding to each IMF component is determined by multiplying the power filter value corresponding to each IMF component in each decomposition data with the preset coefficient. Each IMF is filtered based on the Gaussian kernel standard deviation corresponding to each IMF component, resulting in multiple filtered IMF components.

4. The automatic early warning device for power transformer faults based on multiple sensors as described in claim 3, characterized in that, The reconstruction module is also used for: Based on the power weight value corresponding to each IMF component, determine the reconstruction weight of each filtered IMF component; Based on the reconstruction weights of multiple IMF components and the corresponding IMF components in each type of decomposed data after filtering, each type of reconstructed data is obtained.

5. The automatic early warning device for power transformer faults based on multiple sensors as described in claim 4, characterized in that, The reconstruction module is also used for: The reconstruction weight of each IMF component after filtering is determined by the ratio of the power weight value corresponding to each IMF component to the sum of the power weight values ​​corresponding to multiple IMFs.

6. The automatic early warning device for power transformer faults based on multiple sensors as described in claim 4, characterized in that, The reconstruction module is also used for: The reconstructed data for each type of decomposed data is obtained by summing the products of the reconstruction weights of the multiple IMF components in each type of decomposed data and the corresponding IMF components.

7. The automatic early warning device for power transformer faults based on multiple sensors as described in claim 1, characterized in that, The early warning module is also used for: The signal amplitude difference value of each historical power data at each time moment is calculated using the first-order backward difference method, and the mean and variance of the signal amplitude difference value of each historical power data are calculated. Based on the mean and variance of the signal amplitude difference for each type of historical power data, a preset threshold is determined for each type of historical power data.

8. A method for automatic early warning of power transformer faults based on multiple sensors, characterized in that, The method is applied to the multi-sensor-based automatic early warning device for power transformer faults as described in any one of claims 1-6, and the method includes: For decomposed data of the same type of electrical quantity, the power weight value corresponding to the IMF component of each frequency band of the same type is determined based on the correlation of energy changes between IMF components of the same frequency band and the energy intensity of IMF components of each frequency band. Multiple decomposed data are obtained by variational mode decomposition of multiple power data of transformers, including primary voltage data, secondary voltage data, primary current data and secondary current data. The amplitude fluctuation characteristics of each IMF component in the time domain and its power spectral density distribution in the frequency domain are analyzed to obtain the power filtering value corresponding to each IMF component. Based on the corresponding power weight value, the multiple IMF components in each decomposed data after filtering are reconstructed to obtain each reconstructed data corresponding to each decomposed data. Among them, the multiple IMF components after filtering are obtained by performing Gaussian kernel standard deviation filtering on each of the multiple IMF components. The Gaussian kernel standard deviation is determined according to the power filtering value corresponding to each IMF component. If the mean difference of the amplitude of each type of reconstructed data is greater than the corresponding preset threshold, an early warning will be issued; the preset threshold for each type of reconstructed data is determined based on the corresponding historical power data.

Images