Oil paper insulation state evaluation method based on harmonic-discharge coupling analysis

Abstract

The invention provides an oil paper insulation state evaluation method and system based on harmonic-discharge coupling analysis, and relates to the technical field of transformer oil paper insulation, and the method comprises the steps: configuring harmonic parameters, employing a Morlet wavelet convolution kernel to extract the time-frequency energy distribution of each harmonic, constructing a harmonic fingerprint matrix, and calculating the harmonic fingerprint matrix; oil paper insulation data are collected to construct a nonlinear cross feature to generate an enhanced feature vector, then the dynamic attention weight of each harmonic component to an insulation state is calculated, and then dual-channel degradation evaluation is performed through an equivalent circuit equation of a physical channel and an attention LSTM network of a data channel to obtain a harmonic influence factor and a state evaluation matrix; and finally, calculating an HIF index through weighted fusion of the harmonic frequency deviation and the amplitude exceeding degree, and outputting an evaluation result according to the HIF index. According to the method, the problems of insufficient precision and poor interpretability of a traditional method in a harmonic complex coupling scene are solved, and the accuracy and engineering applicability of oil paper insulation state evaluation are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer oil-paper insulation, and in particular, to an oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis. Background Art

[0002] With the rapid development of China's power industry, power transmission and transformation equipment is developing towards larger capacity and higher voltage. 500 kV AC / DC electrical equipment has been widely used and has become the core part of the power grid system. Oil-immersed high-voltage transformers are one of the most important equipment among them, so they must have high reliability to ensure the safe and efficient operation of the power grid. This insulation system is mainly composed of a combination of oil and paper, forming a composite insulating medium. Traditional oil-immersed high-voltage transformers use cellulose insulating paper. Therefore, more and more high-voltage transformers are beginning to use new insulating papers to ensure the best operating state.

[0003] However, traditional oil-paper insulation state evaluation methods have significant limitations: First, existing technologies are mostly designed based on power frequency conditions and fail to effectively quantify the non-linear coupling effect of harmonic frequencies and contents on discharge characteristics, resulting in an evaluation error of more than 35% under complex working conditions with high-order harmonics (such as new energy grid connection scenarios); Second, although pure data-driven models (such as traditional LSTM, CNN) can extract discharge characteristics, they cannot establish the correlation between characteristic parameters and physical aging mechanisms such as electric field distortion and thermal stress, violating the requirement of IEC 60599 standard for the interpretability of insulation state evaluation; Finally, existing methods rely on single indicators such as partial discharge inception voltage (PDIV) or dielectric loss tangent value, and it is difficult to characterize the synergistic degradation effect caused by the superposition of multiple harmonics (such as the coupling effect of high-frequency harmonics accelerating cellulose hydrolysis and low-frequency harmonics inducing local overheating). Summary of the Invention

[0004] The purpose of the present invention is to provide an oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis, which can accurately evaluate the oil-paper insulation state of transformers, improve the insulation level of water high-voltage transformers, and maintain the safe and stable operation of high-voltage transformers.

[0005] The technical solution of the present invention is as follows: The present application provides an oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis, which includes the following steps: S1. Configure harmonic parameters, and use the Morlet wavelet convolution kernel to extract the time-frequency energy distribution of each harmonic according to the configured harmonic parameters to construct a harmonic fingerprint matrix; S2. Collect oil-paper insulation data to construct non-linear cross features to generate enhanced feature vectors; S3. Calculate the dynamic attention weights of each harmonic component on the insulation state based on the harmonic fingerprint matrix and the enhanced feature vector; S4. Based on the dynamic attention weights of each harmonic component on the insulation state, perform dual-channel degradation assessment through the equivalent circuit equation of the physical channel and the attention LSTM network of the data channel to obtain the harmonic influence factor and the state evaluation matrix; S5. Based on the harmonic influence factor and the state evaluation matrix, calculate the HIF index by weighted fusion of the harmonic frequency offset and the degree of amplitude exceeding the standard, and output the evaluation result according to the HIF index.

[0006] Further, in step S1, the above harmonic parameters include harmonic frequency and harmonic content.

[0007] Further, in step S1, the extraction of the time-frequency energy distribution of each harmonic by the Morlet wavelet convolution kernel satisfies: , where g represents the time-frequency joint distribution function, t represents time, f k represents the k-th frequency component, j represents the imaginary unit, σ k represents the bandwidth parameter.

[0008] Further, in step S2, the above oil-paper insulation data includes partial discharge inception voltage, average discharge amount, maximum discharge amount, and discharge repetition rate.

[0009] Further, in step S2, the calculation process of collecting the oil-paper insulation data to construct non-linear cross features to generate an enhanced feature vector includes: Extract partial discharge feature quantities and normalize them: , , Generate non-linear cross terms: , Generate an enhanced feature vector: , where X is the partial discharge feature vector, PDIV is the partial discharge inception voltage, Q max is the maximum discharge amount, Q avg is the average discharge amount, R discharge is the discharge repetition rate, X norm is the normalized partial discharge feature vector, is the feature mean of the unaged specimen, is the standard deviation of the unaged specimen, α is the non-linear correlation term between the discharge intensity and voltage, β is the non-linear characteristic term of the discharge quantity distribution, T represents the transpose of the matrix, is the enhanced eigenvector.

[0010] Further, in step S3, the calculation process of calculating the dynamic attention weights of each harmonic component on the insulation state according to the harmonic fingerprint matrix and the enhanced eigenvector includes: Calculation of electric field distortion constraint: , Calculation of PDIV decay rate: , Calculation of thermal stress constraint: , , Calculation of dynamic attention weights: , In the formula, E eff is the effective electric field distortion quantity, n is the upper limit of the harmonic order, h is the harmonic order, A h is the amplitude of the h-th harmonic, f h is the frequency of the h-th harmonic, ΔPDIV is the decay amount of the partial discharge inception voltage, k 1 is the decay proportionality coefficient, k 2 is the thermal stress proportionality coefficient, E 0 is the reference electric field strength, ΔT is the temperature rise, τ aging is the aging time constant, A and B are the Arrhenius equation coefficients, T amb is the ambient temperature, w h is the dynamic attention weight of the h-th harmonic, σ is the Sigmoid function, A i is the amplitude of the i-th harmonic, f i is the frequency of the i-th harmonic, Q max is the maximum discharge quantity.

[0011] Further, in step S4, the equivalent circuit equation of the physical channel is: , In the formula, Q theory is the theoretical charge quantity,C is the equivalent capacitance, V harmontic is the harmonic voltage, t is the time, R insul is the time-varying parameter of the insulation resistance.

[0012] Further, in step S4, the calculation formula of the attention LSTM network of the above data channel includes: , , , In the formula, is the attention hidden state, w h is the dynamic attention weight of the h-th harmonic, n is the upper limit of the harmonic order, h is the harmonic order, h t is the h-th hidden state, L is the total loss function, L MSE is the mean square error loss, L SSIM is the structural similarity loss, L Phys is the physical constraint loss, is the fluctuation of the maximum discharge amount, V is the voltage, t is the time.

[0013] Further, in step S5, the calculation formula of the above state evaluation matrix includes: , In the formula, S is the state evaluation matrix, ΔPDIV / PDIV is the discharge starting voltage change rate, is the fluctuation of the maximum discharge amount, is the fluctuation of the average discharge amount, R is the current resistance value, R 0 is the reference resistance value, H is the harmonic distortion rate, f is the frequency, A is the Arrhenius equation coefficient, τ aging is the aging time constant, HIF is the harmonic influence factor.

[0014] Further, in step S5, the calculation formula of the above HIF index calculated by weighted fusion of harmonic frequency offset and amplitude overrun degree includes: , In the formula, HIF is the harmonic influence factor, n is the upper limit of the harmonic order, h is the harmonic order, w h is the dynamic attention weight of the h-th harmonic, fh is the frequency of the h-th harmonic, f 0 is the reference frequency, A h is the amplitude of the h-th harmonic, A 0 is the reference amplitude.

[0015] Compared with the prior art, the present invention has at least the following advantages or beneficial effects: A method for evaluating the state of oil-paper insulation based on harmonic-discharge coupling analysis according to the present invention quantifies the harmonic frequency-domain characteristics by constructing a harmonic fingerprint matrix, combines a dynamic physical attention mechanism and a dual-channel degradation evaluation model to realize the non-linear mapping between harmonic parameters and partial discharge characteristics, and innovatively introduces a differentiable physical constraint layer to embed the electric field distortion model and the thermal stress equation into the deep learning framework, and designs a harmonic influence factor (HIF) quantization evaluation index, which solves the problems of insufficient accuracy and poor interpretability of traditional methods in complex harmonic coupling scenarios, and significantly improves the accuracy and engineering applicability of the evaluation of the state of oil-paper insulation; BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 is a step diagram of a method for evaluating the state of oil-paper insulation based on harmonic-discharge coupling analysis according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Usually, the components of the embodiments of the present application described and illustrated herein can be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application claimed, but merely represents the selected embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts fall within the scope of protection of the present application.

[0020] It should be noted that like reference numerals and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.

[0021] It should be noted that, in this text, the term "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a series of elements includes not only those elements but also other elements not expressly listed, or elements that are inherent to such process, method, article, or apparatus. Without further limitation, elements defined by the statement "comprising..." do not preclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the said elements.

[0022] The following will, with reference to the accompanying drawings, elaborate on some embodiments of the present application. Without conflict, the following embodiments and the various features in the embodiments can be combined with each other.

[0023] Embodiment 1 Please refer to Figure 1 , Figure 1 which shows a step diagram of a method for evaluating the oil-paper insulation state based on harmonic-discharge coupling analysis provided by an embodiment of the present application.

[0024] The present application provides a method for evaluating the oil-paper insulation state based on harmonic-discharge coupling analysis, which includes the following steps: S1. Configure harmonic parameters, and extract the time-frequency energy distribution of each harmonic using a Morlet wavelet convolution kernel according to the configured harmonic parameters to construct a harmonic fingerprint matrix; S2. Collect oil-paper insulation data to construct non-linear cross features to generate an enhanced feature vector; S3. Calculate the dynamic attention weights of each harmonic component on the insulation state according to the harmonic fingerprint matrix and the enhanced feature vector; S4. Based on the dynamic attention weights of each harmonic component on the insulation state, perform dual-channel degradation evaluation through the equivalent circuit equation of the physical channel and the attention LSTM network of the data channel to obtain the harmonic influence factor and the state evaluation matrix; S5. Based on the harmonic influence factor and the state evaluation matrix, calculate the HIF index by weighted fusion of the harmonic frequency offset and the amplitude over-standard degree, and output the evaluation result according to the HIF index.

[0025] As a preferred embodiment, in step S1, the harmonic parameters include harmonic frequency and harmonic content.

[0026] As a preferred embodiment, in step S1, the extraction of the time-frequency energy distribution of each harmonic by the Morlet wavelet convolution kernel satisfies: , wherein, g represents the time-frequency joint distribution function, t represents time, f k represents the k-th frequency component, j represents the imaginary unit, σ k represents the bandwidth parameter.

[0027] As a preferred embodiment, in step S2, the oil-paper insulation data includes the partial discharge inception voltage, average discharge amount, maximum discharge amount, and discharge repetition rate.

[0028] As a preferred embodiment, in step S2, the calculation process of collecting oil-paper insulation data to construct non-linear cross features to generate enhanced feature vectors includes: Extracting partial discharge feature quantities and normalizing them: , , Generating non-linear cross terms: , Generating enhanced feature vectors: , wherein, X is the partial discharge feature vector, PDIV is the partial discharge inception voltage, Q max is the maximum discharge amount, Q avg is the average discharge amount, R discharge is the discharge repetition rate, X norm is the normalized partial discharge feature vector, is the feature mean of the unaged specimen, is the standard deviation of the unaged specimen, α is the non-linear correlation term between discharge intensity and voltage, β is the non-linear feature term of discharge amount distribution, T represents the transpose of the matrix, is the enhanced feature vector.

[0029] As a preferred embodiment, in step S3, the calculation process of calculating the dynamic attention weights of each harmonic component on the insulation state according to the harmonic fingerprint matrix and the enhanced feature vector includes: Calculating the electric field distortion constraint: , Calculating the PDIV attenuation rate: , Calculating the thermal stress constraint: , , Dynamic attention weight calculation: , In the formula, E eff is the effective electric field distortion variable, n is the upper limit of the harmonic order, h is the harmonic order, A h is the amplitude of the h-th harmonic, f h is the frequency of the h-th harmonic, ΔPDIV is the partial discharge inception voltage attenuation, k 1 is the attenuation proportionality coefficient, k 2 is the thermal stress proportionality coefficient, E 0 is the reference electric field strength, ΔT is the temperature rise, τ aging is the aging time constant, A and B are Arrhenius equation coefficients, T amb is the ambient temperature, w h is the dynamic attention weight of the h-th harmonic, σ is the Sigmoid function, A i is the amplitude of the i-th harmonic, f i is the frequency of the i-th harmonic, Q max is the maximum discharge amount.

[0030] As a preferred implementation manner, in step S4, the equivalent circuit equation of the physical channel is: , In the formula, Q theory is the theoretical charge amount, C is the equivalent capacitance, V harmontic is the harmonic voltage, t is the time, R insul is the time-varying parameter of the insulation resistance.

[0031] As a preferred implementation manner, in step S4, the calculation formula of the attention LSTM network of the data channel includes: , , , In the formula, is the attention hidden state, w h is the dynamic attention weight of the h-th harmonic, n is the upper limit of the harmonic order,h is the harmonic order, h t is the h-th hidden state, L is the total loss function, L MSE is the mean squared error loss, L SSIM is the structural similarity loss, L Phys is the physical constraint loss, is the fluctuation of the maximum discharge capacity, V is the voltage, t is the time.

[0032] As a preferred implementation manner, in step S5, the calculation formula of the state evaluation matrix includes: , In the formula, S is the state evaluation matrix, ΔPDIV / PDIV is the rate of change of the discharge starting voltage, is the fluctuation of the maximum discharge capacity, is the fluctuation of the average discharge capacity, R is the current resistance value, R 0 is the reference resistance value, H is the harmonic distortion rate, f is the frequency, A is the Arrhenius equation coefficient, τ aging is the aging time constant, HIF is the harmonic influence factor.

[0033] As a preferred implementation manner, in step S5, the calculation formula for calculating the HIF index by weighted fusion of the harmonic frequency offset and the amplitude exceeding degree includes: ,

[0034] In the formula, HIF is the harmonic influence factor, n is the upper limit of the harmonic order, h is the harmonic order, w h is the dynamic attention weight of the h-th harmonic, f h is the frequency of the h-th harmonic, f 0 is the reference frequency, A h is the amplitude of the h-th harmonic, A 0 is the reference amplitude.

[0035] It can be understood that the structure shown in the figure is only schematic. An oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis may also include more or fewer components than those shown in the figure, or have a different configuration from that shown in the figure. Each component shown in the figure can be implemented by hardware, software, or a combination thereof.

[0036] In the embodiments provided in this application, it should be understood that the disclosed method can also be implemented in other ways. The embodiments described above are merely illustrative. For example, the flowcharts or block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the methods and computer program products according to multiple embodiments of this application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0037] In addition, each functional module in various embodiments of this application can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0038] If the above functions are implemented in the form of software functional modules and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disks, or optical discs that can store program code.

[0039] The above is only the preferred embodiment of this application and is not used to limit this application. For those skilled in the art, this application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included in the protection scope of this application.

[0040] It is obvious to those skilled in the art that the present application is not limited to the details of the above-mentioned exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or basic characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present application is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present application. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. A method for evaluating the state of oil-paper insulation based on harmonic-discharge coupling analysis, characterized in that: The following steps are involved: S1. Configure harmonic parameters, and use Morlet wavelet convolution kernel to extract the time-frequency energy distribution of each harmonic according to the configured harmonic parameters to construct a harmonic fingerprint matrix; S2, collecting oil-paper insulation data to construct nonlinear cross features to generate enhanced feature vectors; S3, calculating the dynamic attention weight of each harmonic component on the insulation state according to the harmonic fingerprint matrix and the enhanced eigenvector; S4. Based on the dynamic attention weight of each harmonic component on the insulation state, dual-channel degradation evaluation is performed through the equivalent circuit equation of the physical channel and the attention LSTM network of the data channel to obtain the harmonic impact factor and state evaluation matrix; S5. Based on the harmonic impact factor and the state assessment matrix, the HIF index is calculated by weighted fusion of the harmonic frequency offset and the amplitude exceeding the standard, and the assessment result is output according to the HIF index.

2. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1, characterized in that: In step S1, the harmonic parameters include harmonic frequency and harmonic content.

3. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1 is characterized in that: In step S1, the time-frequency energy distribution of each harmonic extracted by the Morlet wavelet convolution kernel satisfies: ， In the formula, g represents the time-frequency joint distribution function, t Indicates time, f k represents the kth frequency component, j represents the imaginary unit, σ k Indicates the bandwidth parameter.

4. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1, characterized in that: In step S2, the oil-paper insulation data includes partial discharge starting voltage, average discharge capacity, maximum discharge capacity and discharge repetition rate.

5. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 4 is characterized in that: In step S2, the calculation process of collecting oil-paper insulation data to construct nonlinear cross features to generate enhanced feature vectors includes: Extract the local discharge characteristics and normalize them: ， ， Generate nonlinear cross terms: ， Generate an augmented feature vector: ， Where X is the local discharge characteristic vector, PDIV is the local discharge inception voltage, Q max is the maximum discharge capacity, Q avg is the average discharge capacity, R discharge is the discharge repetition rate, X norm is the normalized partial discharge feature vector, is the characteristic mean of the unaged sample, is the standard deviation of the unaged sample, α is the nonlinear correlation term between discharge intensity and voltage, β is the nonlinear characteristic term of discharge quantity distribution, T represents the transpose of the matrix, is the enhanced feature vector.

6. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1, characterized in that: In step S3, the process of calculating the dynamic attention weight of each harmonic component to the insulation state according to the harmonic fingerprint matrix and the enhanced eigenvector includes: Electric field distortion constraint calculation: ， PDIV decay rate calculation: ， Thermal stress constraint calculation: ， ， Dynamic attention weight calculation: ， In the formula, E eff is the effective electric field distortion, n is the upper limit of harmonic order, h is the harmonic order, A h is the amplitude of the hth harmonic, f h is the frequency of the hth harmonic, ΔPDIV is the attenuation of the partial discharge inception voltage, k 1 is the attenuation proportional coefficient, k 2 is the thermal stress proportionality coefficient, E 0 is the reference electric field strength, ΔT is the temperature rise, τ aging is the aging time constant, A and B are the coefficients of the Arrhenius equation, T amb is the ambient temperature, w h is the dynamic attention weight of the h-th harmonic, σ is the Sigmoid function, A i is the amplitude of the ith harmonic, f i is the frequency of the ith harmonic, Q max is the maximum discharge capacity.

7. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1, characterized in that: In step S4, the equivalent circuit equation of the physical channel is: ， In the formula, Q theory is the theoretical charge, C is the equivalent capacitance, V harmontic is the harmonic voltage, t For time, R insul is the time-varying parameter of insulation resistance.

8. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1, characterized in that: In step S4, the calculation formula of the attention LSTM network of the data channel is include: ， ， ， In the formula, is the attention hidden state, w h is the dynamic attention weight of the hth harmonic, n is the upper limit of the harmonic order, h is the harmonic order, h t is the hth hidden state, L is the total loss function, L MSE is the mean square error loss, L SSIM is the structural similarity loss, L Phys is the physical constraint loss, is the fluctuation of the maximum discharge amount, V is the voltage, t For time.

9. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1, characterized in that: In step S5, the calculation formula of the state evaluation matrix includes: ， In the formula, S is the state evaluation matrix, ΔPDIV / PDIV is the discharge starting voltage change rate, is the fluctuation of the maximum discharge capacity, is the fluctuation of the average discharge amount, R is the current resistance value, R0 is the reference resistance value, H is the harmonic distortion rate, f is the frequency, A is the coefficient of the Arrhenius equation, τ aging is the aging time constant, and HIF is the harmonic influence factor.

10. The oil-paper insulation state evaluation method based on harmonic-discharge coupling analysis according to claim 1, characterized in that: In step S5, the calculation formula for calculating the HIF index by weighted fusion of harmonic frequency offset and amplitude exceeding the standard degree includes: ， In the formula, HIF is the harmonic impact factor, n is the upper limit of the harmonic order, h is the harmonic order, w h is the dynamic attention weight of the h-th harmonic, f h is the frequency of the hth harmonic, f 0 is the reference frequency, A h is the amplitude of the hth harmonic, A 0 is the reference amplitude.

Images