Ultra-wideband transformer PD source intelligent detection platform and method

Abstract

The invention discloses an ultra-wideband transformer PD source intelligent detection platform and method, and relates to the related field of electrical variable measurement, and the platform comprises a configuration module which is used for configuring a sensor group, an initial solution construction module which is used for carrying out the wavelet decomposition of a pulse signal set, obtaining a wavelet decomposition coefficient under each scale, and carrying out the calculation of the wavelet decomposition coefficient, the correlation analysis module is used for carrying out particle similarity calculation on m initial particles, the updating iteration module is used for carrying out correlation selection on particle correlation and then generating correlation optimization through the current iteration frequency and the iteration state, and the de-noising processing module is used for dynamically adjusting the inertia weight. The feature extraction module is used for establishing a feature extraction result for the de-noised signal set, and the anomaly analysis module is used for carrying out transformer anomaly analysis according to the feature extraction result and generating an anomaly report. The technical problem that the distinguishing accuracy of ultra-wideband PD signal detection and a noise source is low is solved, and the technical effects of extracting PD pulse characteristic parameters and achieving intelligent clustering and separation are achieved.

Description

Technical Field

[0001] The present application relates to the field of electrical variable measurement, and in particular to an ultra-wideband transformer PD source intelligent detection platform and method. Background Art

[0002] With the rapid construction and development of smart grids in my country, the reliability and safety of grid operation have gradually improved. As key equipment for power transmission and transformation in AC and DC systems, the operating status of power transformers and converter transformers directly affects the stable operation of the entire power grid. At present, oil-paper composite insulation is a common insulation method for large transformers, which is mainly composed of insulating oil, insulating paperboard and other solid insulating materials. Although oil-paper insulation is required to have sufficient electrical strength and mechanical properties during design, it is inevitable that accidental factors in the production, equipment and operation process will cause defects in the insulation system, which will lead to equipment failure. Partial discharge (PD) is the main cause of transformer oil-paper insulation defect failure and can be used as a means to assess the degree of insulation degradation. Therefore, conducting PD detection (monitoring) under AC and DC voltages on power transformers and converter transformers to obtain discharge information will help to understand the insulation condition of the operating transformer and ensure its safe and stable operation. Summary of the invention

[0003] The present application solves the technical problem of low accuracy in ultra-wideband PD signal detection and noise source differentiation by providing an ultra-wideband transformer PD source intelligent detection platform and method, and achieves the technical effect of quickly extracting PD pulse characteristic parameters and realizing intelligent clustering and separation.

[0004] The present application provides an ultra-wideband transformer PD source intelligent detection platform, and the platform is applied to an ultra-wideband transformer PD source intelligent detection method, including: a configuration module, used to configure a sensor group, the sensor group includes an HFCT, a bushing end screen sensor, and a detection impedance, and the sensor group is used to monitor the partial discharge pulse signal of the transformer to establish a pulse signal set; an initial solution construction module, used to perform wavelet decomposition on the pulse signal set, obtain the wavelet decomposition coefficients at each scale, and generate m initial particles in the solution space using the wavelet decomposition coefficients; an association analysis module, used to calculate the particle similarity of the m initial particles, establish particle similarity association, and use the particle similarity association to perform particle combination to establish particle association; an update iteration module, which is used to generate association optimization through the current iteration number and iteration state after association selection of the particle association, and use the association optimization and the particle individual extreme value and particle group extreme value to perform m initial particle update iterations; a denoising processing module, which is used to dynamically adjust the inertia weight, perform iterative optimization, establish an optimal wavelet threshold according to the iterative optimization result, and use the optimal wavelet threshold to perform denoising processing on the pulse signal set to establish a denoised signal set; a feature extraction module, which is used to extract first-order shape parameter features, kurtosis features, skewness features, and frequency domain features of the denoised signal set to establish a feature extraction result; an abnormality analysis module, which is used to perform transformer abnormality analysis according to the feature extraction result and generate an abnormality report.

[0005] In a possible implementation, the association analysis module performs the following processing: configuring a similarity calculation formula as follows: ;in, Characterizing particles and particles The particle similarity correlation, and Characterize and measure particles separately and particles The fitness function of the individual quality, is the particle fitness deviation coefficient, is the particle's own deviation coefficient; configure a combined isolation strategy, sort the particle similar associations in order, and then use the combined isolation strategy to perform association screening from the maximum association to complete the particle combination and establish particle association.

[0006] In a possible implementation, the update iteration module performs the following processing: after performing association normalization processing on the particle association, performing downward association ratio clustering from the highest association value to establish a downward association ratio clustering result; when the downward association ratio clustering result meets a preset threshold, the clustering ends, and the particles corresponding to the particle associations in the cluster are used as association selection particles; the current number of iterations is obtained, and a first optimization weight is generated based on the number of iterations; the current iteration state is obtained, and a second optimization weight is generated based on the iteration state; after comprehensive calculation of the first optimization weight and the second optimization weight, an optimization reward is generated using low-fitness particles in the association selection particles and the comprehensive calculation result, and the high-fitness particles in the association selection particles are updated using the optimization reward, and the update of the low-fitness particles is stopped, and the update iteration is completed according to the update result.

[0007] In a possible implementation, the update iteration module performs the following processing: remove the low fitness particles that stop updating from the m initial particles to establish updated initial particles; perform individual updates on the non-associated selected particles in the updated initial particles based on the individual extreme values ​​of the particles and the extreme values ​​of the particle group to generate a first update result; perform individual updates on the associated selected particles in the updated initial particles based on the optimization reward, the individual extreme values ​​of the particles, and the extreme values ​​of the particle group to generate a second update result; and complete the update iteration after adding new particles to the first update result and the second update result.

[0008] In a possible implementation, the feature extraction module performs the following processing: Calculate the kurtosis feature by the formula as follows: ;in, Characterizes the kurtosis, Characterize the random variable A value, represents the mean of a random variable, is the standard deviation of the random variable, is the total number of random variables, The probability density function that characterizes a random variable is The value at; the skewness characteristic is calculated by the formula as follows: ;in, Characterize the skewness; use the kurtosis feature and the skewness feature as feature extraction results.

[0009] In a possible implementation, the abnormality analysis module performs the following processing: collects equipment status parameters of the transformer and establishes an equipment status data set, wherein the equipment status data set includes load data, voltage level data, and operating time data; collects environmental parameters of the transformer and establishes an environmental data set, wherein the environmental data set includes temperature data and humidity data; uses the equipment status data set, the environmental data set, and the feature extraction results to identify transformer abnormalities and establish an abnormality report.

[0010] In a possible implementation, the following processing is performed: a monitoring module is used to establish a focus window according to the abnormal report; a feedback monitoring module is used to perform feedback update on the feature extraction module based on the focus window, and transformer monitoring management is performed according to the feature extraction module after feedback update.

[0011] The present application also provides an ultra-wideband transformer PD source intelligent detection method, comprising: configuring a sensor group, the sensor group comprising an HFCT, a bushing end screen sensor, and a detection impedance, using the sensor group to monitor the partial discharge pulse signal of the transformer, and establishing a pulse signal set; performing wavelet decomposition on the pulse signal set, obtaining the wavelet decomposition coefficients at each scale, and using the wavelet decomposition coefficients to generate m initial particles in the solution space; performing particle similarity calculation on the m initial particles, establishing particle similarity association, and using the particle similarity association to perform particle combination and establish particle association; After the sub-association performs association selection, an association optimization is generated through the current iteration number and iteration state, and the association optimization and the particle individual extreme value and the particle group extreme value are used to perform m initial particle update iterations; the inertia weight is dynamically adjusted, and iterative optimization is performed. An optimal wavelet threshold is established according to the iterative optimization result, and a denoised signal set is established after the pulse signal set is denoised using the optimal wavelet threshold; first-order shape parameter features, kurtosis features, skewness features, and frequency domain features are extracted from the denoised signal set to establish a feature extraction result; transformer abnormality analysis is performed according to the feature extraction result, and an abnormality report is generated.

[0012] One or more technical solutions provided in this application have at least the following technical effects or advantages: The present application provides an ultra-wideband transformer PD source intelligent detection platform and method, which relate to the technical field of electrical variable measurement, solves the technical problem of low accuracy in ultra-wideband PD signal detection and noise source distinction, and achieves the technical effect of quickly extracting PD pulse characteristic parameters and realizing intelligent clustering and separation. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical scheme of the embodiment of the present application, the accompanying drawings of the embodiment of the present application will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or a certain step or several steps of operations can be removed from these processes.

[0014] Figure 1 A schematic diagram of the structure of an ultra-wideband transformer PD source intelligent detection platform provided in an embodiment of the present application.

[0015] Figure 2 A schematic diagram of time domain characteristic parameters of a single discharge waveform of an ultra-wideband transformer PD source intelligent detection platform provided in an embodiment of the present application.

[0016] Figure 3 A flowchart of an ultra-wideband transformer PD source intelligent detection method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0018] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0019] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0020] The embodiment of the present application provides an ultra-wideband transformer PD source intelligent detection platform, such as Figure 1 As shown, the method includes: According to an ultra-wideband transformer PD source intelligent detection platform of the embodiment of the present application, it is used to solve the technical problem of low accuracy of ultra-wideband PD signal detection and noise source distinction, and achieve the technical effect of quickly extracting PD pulse characteristic parameters and realizing intelligent clustering and separation. An ultra-wideband transformer PD source intelligent detection platform includes: a configuration module 10, an initial solution construction module 20, an association analysis module 30, an update iteration module 40, a denoising processing module 50, a feature extraction module 60, and an abnormality analysis module 70.

[0021] Configuration module 10 is used to configure a sensor group, the sensor group includes HFCT, bushing end screen sensor, and detection impedance. The sensor group is used to monitor the partial discharge pulse signal of the transformer and establish a pulse signal set. Specifically, in the partial discharge monitoring of the transformer, the sensor group is configured to collect and analyze signals. The configured sensor group may include HFCT, bushing end screen sensor, detection impedance, etc. The HFCT is used to capture the high-frequency current signal outside the transformer, especially the high-frequency current signal generated by partial discharge. The bushing end screen sensor is used to monitor the partial discharge signal, especially the electric field and local current outside the transformer, and can more accurately detect the electric field changes caused by partial discharge. The detection impedance can judge the partial discharge situation by detecting the insulation impedance change of the transformer. Partial discharge is usually accompanied by aging and deterioration of the insulating material, and the impedance of the transformer will change with the occurrence of discharge.

[0022] Furthermore, the HFCT can be installed outside the transformer, especially at the input or output end of the current conductor, so as to monitor the changes in the external current. The bushing end screen sensor is installed at the end of the transformer bushing to facilitate the detection of signals caused by the local current or electric field of the transformer. The detection impedance sensor is usually installed on the winding or grounding part of the transformer to monitor the state of the insulating material.

[0023] At the same time, using the sensor group to monitor the partial discharge pulse signal of the transformer means starting the sensor group to monitor the signal. Each sensor will collect signals from the target area at the set time interval and transmit the collected signals to the data acquisition system. It can include real-time collection of high-frequency current signals through HFCT sensors, and sending the data to the acquisition system for further analysis and processing, continuous monitoring of local electric field changes through the bushing end screen sensor, recording the changes and transmitting them to the data processing unit, and then monitoring the insulation state of the transformer in real time through the impedance sensor, recording the impedance change data, and transmitting it to the data acquisition system through signals, and ensuring that the data acquisition process of all sensors is synchronized to ensure accurate capture of partial discharge signals and avoid timing errors affecting data quality.

[0024] Finally, based on the processed signal data, the frequency, time and space characteristics of the partial discharge signal can be extracted using signal processing algorithms including fast Fourier transform FFT and time-frequency analysis. By classifying the partial discharge signals with different characteristics, different types of discharge signal sets can be generated. For example, the discharge signal can be divided into categories such as slight discharge, normal discharge and severe discharge according to its intensity and frequency. The integration of different sensor data can form a comprehensive partial discharge pulse signal set.

[0025] The initial solution construction module 20 is used to perform wavelet decomposition on the pulse signal set, obtain the wavelet decomposition coefficients at each scale, and generate m initial particles in the solution space using the wavelet decomposition coefficients; specifically, multi-layer wavelet decomposition is performed on the pulse signal. Each layer of decomposition generates a low-frequency part (approximate coefficient) and a high-frequency part (detail coefficient). Usually, N layers of wavelet decomposition are performed to obtain signal components of different scales, and an appropriate wavelet decomposition layer number N is selected. The wavelet decomposition layer number N is related to the frequency range and the desired resolution of the signal, wherein a higher number of decomposition layers can extract more detailed signal features.

[0026] Furthermore, wavelet decomposition will generate coefficients of different scales, mainly including low-frequency parts (approximate coefficients) and high-frequency parts (detail coefficients). The low-frequency coefficients (approximate coefficients) reflect the long-term trend of the signal. The low-frequency part is the stable part of the signal, which usually contains the global characteristics of the signal. The high-frequency coefficients (detail coefficients) reflect the local changes or instantaneous characteristics of the signal. The high-frequency part can capture the characteristics of short-term changes such as sharp changes and pulses in the signal. Each layer of wavelet decomposition will generate a set of coefficients, and each set of coefficients will be merged in turn to obtain the characteristics of the signal at different scales. After obtaining the wavelet decomposition coefficients at each scale, particles can be generated through the solution space. It means that the number of wavelet decomposition coefficients and the characteristic dimension determine the dimension of the solution space. In the solution space, each particle corresponds to a potential solution, which consists of multiple dimensions, and each dimension represents a feature or coefficient value. For example, each wavelet decomposition coefficient can be regarded as a dimension of the particle, and each particle is initialized by the wavelet decomposition coefficient. Specifically, for the coefficients obtained by decomposition at each layer, the position and velocity of the initial particle are generated. Further, the position of the initial particle can be directly associated with the wavelet decomposition coefficients of each layer. The dimension of the initial particle is equal to the number of decomposition layers multiplied by the number of wavelet coefficients in each layer, and the particle speed is adjusted by the current coefficient value and the historical best solution to determine the initial particle speed; according to the speed and position of the initialized particle, a search is performed in the solution space to determine m initial particles.

[0027] The association analysis module 30 is used to calculate the particle similarity of the m initial particles, establish the particle similarity association, combine the particles using the particle similarity association, and establish the particle association; the specific configuration of the association analysis module 30 will be described in detail below. The association analysis module 30 may further include: a first configuration unit, used to configure a similarity calculation formula, as follows: ; in, Characterizing particles and particles The particle similarity correlation, and Characterize and measure particles separately and particles The fitness function of the individual quality, is the particle fitness deviation coefficient, is the particle's own deviation coefficient; the second configuration unit is used to configure a combined isolation strategy, after sorting the particle similar associations in order, perform association screening from the maximum association through the combined isolation strategy to complete the particle combination and establish particle association.

[0028] Specifically, in the particle swarm optimization (PSO) algorithm, the process of calculating particle similarity and establishing particle similarity associations can help improve optimization efficiency and enhance the exploration ability of the particle swarm in the solution space. By calculating the similarity between particles, particles with similar characteristics can be found and combined together for optimization to improve the search effect. The particle similarity is a measure of the similarity between particles. The position or speed difference of the particles is usually used as the basis for similarity calculation. Similarity calculation can be performed by configuring the similarity calculation formula. The similarity calculation formula is configured as follows: ; in, Characterizing particles and particles The particle similarity correlation, and Characterize and measure particles separately and particles The fitness function of the individual quality, is the particle fitness deviation coefficient, is the particle's own deviation coefficient.

[0029] Traverse the similarities between m particles and store them in the similarity matrix, further divide the particles into several groups according to the similarity, and establish a particle group structure. By setting a similarity threshold, similar particles can be grouped together. The threshold of similarity can be used to set a fixed threshold or dynamically adjust it. For example, when the Euclidean distance between two particles is less than a certain set value, they are considered similar and can be grouped together. At the same time, all particles are divided into multiple subgroups according to the similarity matrix and the set threshold T. The particles in each subgroup are similar, and it can be considered that they explore similar areas in the solution space. Through the similarity matrix, particles with high similarity are grouped together to form multiple particle subgroups. Each subgroup represents an exploration direction or solution space area.

[0030] Further configuration of the combination isolation strategy, the combination isolation strategy refers to selecting multiple particles from a particle group with high similarity for combination to maintain the diversity and effectiveness of the particle group, and the particles that have completed the combination pairing can no longer be paired with other particles to achieve the purpose of isolation, and after sequential screening from the maximum association, the pairwise combination of particles can be constructed, and the particles that cannot be combined will be eliminated. After the particle combination is completed, the solution space position of the particle group is updated, the new fitness value is calculated, and the updated particle group and the similarity association information are synchronized to the next round of iteration of the particle swarm optimization process to establish particle association, thereby improving the optimization efficiency and the quality of the final solution.

[0031] The update iteration module 40 is used to generate the association optimization through the current iteration number and iteration state after the particle association is selected, and perform the m initial particle update iterations by using the association optimization and the particle individual extreme value and the particle group extreme value; the specific configuration of the update iteration module 40 will be described in detail below. The update iteration module 40 may further include: a clustering unit, which is used to perform downward association proportion clustering from the highest association value after the particle association is normalized, and establish a downward association proportion clustering result; a particle association unit, which is used to terminate the clustering when the downward association proportion clustering result meets the preset threshold, and use the particle corresponding to the particle association in the cluster as the association selection particle; a first iteration unit, which is used to obtain the current iteration number and generate the first optimization weight based on the iteration number; a second iteration unit, which is used to obtain the current iteration state and generate the second optimization weight based on the iteration state; an update unit, which is used to generate the optimization reward by using the low fitness particles in the association selection particles and the comprehensive calculation result after the comprehensive calculation of the first optimization weight and the second optimization weight, and update the high fitness particles in the association selection particles by using the optimization reward, and stop the update of the low fitness particles, and complete the update iteration according to the update result.

[0032] Specifically, firstly, correlation normalization processing is performed based on particle correlation data, which means that correlation normalization processing is performed. The particle correlation data may include correlation proportion clustering based on pulse waveform characteristic parameters (such as time domain and frequency domain characteristics), starting from the highest correlation value and gradually downward. This means that the particles with the highest association value are classified into one category, and then the particles with the second highest association value are considered in turn. If their association with the existing category exceeds a certain threshold, they are classified into this category, otherwise a new category is formed, and then a downward association ratio clustering result is established, that is, the particles contained in each cluster and their association ratio, so as to establish the downward association ratio clustering result. Exemplarily, the association ratio clustering result can be set with a preset threshold, such as 70%, and clustering starts from the maximum and stops when 70% is met. When the downward association ratio clustering result meets the preset threshold, the clustering ends, and the particles corresponding to the particle association in the cluster are used as the association selection particles, which means checking whether the clustering result meets the preset clustering quality threshold (such as the average association of particles in the cluster, the difference in association between clusters, etc.). If the clustering result meets the preset clustering quality threshold, the clustering process ends, and the particles with the highest association are selected from the clustering results as the association selection particles. These particles will be used in the subsequent iterative optimization process.

[0033] The current number of iterations is further recorded, and the first optimization weight is calculated based on the number of iterations and the preset weight adjustment strategy (such as linear decrease, exponential decay, etc.). This weight will be used in the subsequent particle update process to balance the ability of global search and local search, and at the same time evaluate the current iteration state, which may include the distribution of particles, convergence speed, changes in the optimal solution, etc. According to the iteration state and the preset weight adjustment strategy, the second optimization weight is calculated. This weight may be used to adjust the exploration ability of particles in the search space to cope with different situations in the iteration process.

[0034] Finally, the first optimization weight and the second optimization weight are calculated comprehensively to obtain the comprehensive optimization weight. Among the associated selected particles, low fitness particles and high fitness particles are identified, and the optimization reward is generated according to the comprehensive optimization weight and the information of the low fitness particles. The optimization reward can be an adjustment amount, a velocity vector or a new position coordinate, and the optimization reward is used to guide the high fitness particles to move to a better solution. The high fitness particles are updated using the optimization reward, and the update of the low fitness particles is stopped (or a smaller update step is given to them). According to the update results, the new position and fitness value of the particles are evaluated to determine whether the iteration termination conditions are met (such as reaching the maximum number of iterations, the optimal solution is stable and unchanged, etc.). If the conditions are met, the iterative process ends; otherwise, return to continue the iteration, and find the optimal PD detection solution by continuously iterating and improving the position and association of the particles.

[0035] The specific configuration of the update iteration module 40 will be described in detail below. The update iteration module 40 may further include: a removal unit, used to remove the low fitness particles that stop updating from the m initial particles, and establish the updated initial particles; a first individual update unit, used to perform individual updates on the non-associated selected particles in the updated initial particles based on the individual extreme values ​​of the particles and the extreme values ​​of the particle group, and generate a first update result; a second individual update unit, used to perform individual updates on the associated selected particles in the updated initial particles based on the optimization reward, the individual extreme values ​​of the particles, and the extreme values ​​of the particle group, and generate a second update result; a particle supplement unit, used to complete the update iteration after supplementing the first update result and the second update result with new particles.

[0036] Specifically, the low-fitness particles that stop updating are eliminated from the m initial particles. This can be achieved by monitoring the particle fitness, identifying the low-fitness particles, and eliminating the low-fitness particles. The process can be to continuously monitor the fitness value of each particle during the iteration of the hybrid particle swarm optimization algorithm. The fitness value is usually determined according to the objective function value of the particle, reflecting the degree of pros and cons of the particle in the solution space. At the same time, a fitness threshold is set or the standard of low-fitness particles is determined according to the fitness distribution of particles in the current iteration. Particles with fitness lower than the standard are identified, and these particles are considered to be low-fitness particles that stop updating or fall into local optimality. These low-fitness particles are eliminated from the current m initial particles to leave space for subsequent particle updates and supplements, thereby establishing updated initial particles.

[0037] The non-associated selected particles in the updated initial particles are further individually updated based on the individual extreme values ​​of the particles and the extreme values ​​of the particle group. The non-associated selected particles in the updated initial particles refer to the part of the remaining particles that are not selected as the associated selected particles after the low fitness particles are eliminated. Then, for each non-associated selected particle, its speed and position are updated according to its individual extreme value (i.e., the best position found so far by the particle) and the group extreme value (i.e., the best position found so far in the entire particle group). On this basis, the first update result of the non-associated selected particles is obtained. The positions of these particles in the solution space have changed, and they may be closer to the global optimal solution.

[0038] The associated selected particles in the updated initial particles are further individually updated based on the optimization reward, the individual extreme value of the particles, and the extreme value of the particle group. The associated selected particles in the updated initial particles refer to selecting particles that are closely related to the highest association value as the associated selected particles based on the downward association ratio clustering result and the preset threshold. For the associated selected particles, in addition to considering their individual extreme values ​​and group extreme values, an optimization reward is introduced to further adjust their speed and position. This optimization reward can be determined based on factors such as the contribution of the particles in the cluster and the degree of improvement of the association value. The speed and position of the associated selected particles are updated by combining the individual extreme value, the group extreme value and the optimization reward to obtain a second update result. The positions of these particles in the solution space can be more superior due to the introduction of the optimization reward.

[0039] Finally, the first update result and the second update result are supplemented with new particles, which means that the number of particles that need to be supplemented is determined according to the number of the current particle swarm and the preset particle swarm size. New particles are randomly generated in the solution space. The positions and velocities of these particles can be initialized to random values ​​or determined according to a certain heuristic strategy. The new particles are supplemented to the first update result and the second update result to form an updated particle swarm. At the same time, the total number of particle swarms is kept unchanged to maintain the search ability and diversity of the algorithm. Thus, a complete iterative process is completed. The updated particle swarm will be used as the input of the next iteration to continue the search and optimization process to find the global optimal solution or the approximate optimal solution.

[0040] A denoising processing module 50 is used to dynamically adjust the inertia weight, perform iterative optimization, establish an optimal wavelet threshold according to the iterative optimization result, and establish a denoised signal set after denoising the pulse signal set using the optimal wavelet threshold; Specifically, the inertia weight is first calculated using a linear decreasing weight algorithm: ; in, is the inertia weight, is the maximum number of iterations, and are the minimum and maximum values ​​of the weight coefficient respectively.

[0041] Further, each particle is randomly assigned an initial position and velocity, which represent possible wavelet threshold candidates. The fitness function value is calculated for the current position of each particle (i.e., the wavelet threshold candidate). The fitness function can be defined based on the quality of the denoised signal (such as the signal-to-noise ratio SNR), and its inertia weight is dynamically adjusted according to the particle's historical performance (such as the improvement of the fitness value). For example, if the particle has performed well in the past few iterations (the fitness value has significantly improved), the inertia weight is reduced to strengthen the local search; conversely, if the performance is poor, the inertia weight is increased to promote the global search.

[0042] Further, the particle speed and position are updated according to the particle's current speed, inertia weight, individual optimal position and global optimal position, and it is determined whether the maximum number of iterations is reached or the change in fitness value is less than the preset threshold. If so, the iteration is stopped; otherwise, the next step is continued to compare the current fitness value of each particle with the individual optimal and global optimal fitness values. If better, the corresponding optimal position and fitness value are updated. When the iteration is completed, the global optimal position is the optimal wavelet threshold required.

[0043] The processed wavelet coefficients are further subjected to inverse wavelet transform to obtain a denoised signal set, and based on the optimal wavelet threshold determined above, the wavelet coefficients are subjected to threshold processing, that is, coefficients greater than the threshold are retained, and coefficients less than or equal to the threshold are set to zero or shrunk, and the processed wavelet coefficients are subjected to inverse wavelet transform to obtain a denoised signal set.

[0044] The feature extraction module 60 is used to extract the first-order shape parameter feature, kurtosis feature, skewness feature, and frequency domain feature of the denoised signal set to establish a feature extraction result; Figure 2 As shown, the specific configuration of the feature extraction module 60 will be described in detail. The feature extraction module 60 may further include: a first feature calculation unit, which is used to calculate the kurtosis feature by a formula as follows: ; in, Characterizes the kurtosis, Characterize the random variable A value, represents the mean of a random variable, is the standard deviation of the random variable, is the total number of random variables, The probability density function that characterizes a random variable is The value of ; The second feature calculation unit is used to calculate the skewness feature through the formula as follows: ; in, Characterize skewness; The feature extraction unit is used to take the kurtosis feature and the skewness feature as feature extraction results.

[0045] Specifically, if Figure 2 Shown is the maximum amplitude of the discharge pulse; pulse rising edge time It is the time from 10% of the maximum amplitude to 90% of the maximum amplitude, that is: - ; Pulse falling edge time It is the time from 90% of the maximum amplitude to 10% of the maximum amplitude, that is: - . 50% amplitude pulse duration It is the pulse waveform duration from 50% of the maximum amplitude of the rising edge to 50% of the maximum amplitude of the falling edge, that is: - ; The definition of 10% maximum amplitude pulse duration is similar to, - , thereby determining the first-order shape parameter characteristics.

[0046] Kurtosis Describes the degree to which the probability distribution of a random variable is concentrated on the mean, or the speed at which the random variable increases, that is, the steepness of the change of the distribution function. Assume that the probability distribution function of the random variable X is f(x), the mean is µ, and the variance is ,but ; in, Characterizes the kurtosis, Characterize the random variable A value, represents the mean of a random variable, is the standard deviation of the random variable, is the total number of random variables, The probability density function that characterizes a random variable is The value of ; Each sampling point on the single discharge waveform is treated as a random variable, and the sampling value of the local discharge signal at each sampling moment is used to calculate the And the value of this discharge probability Replace and The above formula can be expressed as: ; The above formula is the kurtosis of single discharge waveform of ultra-wideband partial discharge. The larger the value is, the greater the intensity of a discharge pulse in a single discharge waveform. Indicates the "height, shortness, fatness and thinness" of the spectrum. The larger it is, the sharper the data information is, the more concentrated the distribution is. The smaller it is, the more dispersed the data distribution is, thus obtaining the kurtosis feature.

[0047] Skewness Describe the symmetry of the probability distribution of random variables. Take each sample value on the partial discharge time domain signal waveform as a random variable and assume that the random variable The probability distribution function of , with mean µ and variance σ2, then: ; in, Characterize skewness; Similarly, some distribution parameter and its probability of occurrence are used instead of and , we can get the distribution of this kind From the above formula, we can see that It is the degree of symmetry of a distribution about the mean value of a parameter corresponding to a single discharge waveform, or the deviation of the discharge distribution from the normal distribution. If > 0, it means that its distribution is right-skewed relative to the normal distribution. When <0, it means that its distribution is left-skewed relative to the normal distribution. Indicates the difference in the positive and negative half-cycle discharge distribution, that is, the symmetry of the upper and lower amplitude distribution of a single partial discharge: ; Among them, N1 and N2 are the number of values ​​greater than zero and less than zero at a single partial discharge sampling point, respectively. and Represents the value less than zero and the value greater than zero of a single partial discharge sampling point. It can be seen from the formula that, >1 means the average value of the positive half cycle is greater than the average value of the negative half cycle; =1 means that the top and bottom are completely symmetrical; <1 means that the average value of the positive half cycle is smaller than the average value of the negative half cycle. The pulse factor I and margin factor L are both conventional mathematical definitions to obtain the skewness characteristics.

[0048] In addition to collecting the time domain waveform data of a single discharge pulse of partial discharge, the waveform frequency domain data is also collected. The waveform after denoising and removing redundant data is transformed in the frequency domain using fast Fourier transform, and the spectrum diagram is analyzed. The main concentration range of partial discharge and the frequency corresponding to the maximum amplitude are found from the transformed spectrum diagram; as the defect model changes, the frequency at the maximum amplitude will also change. According to the analysis of the partial discharge amplitude-frequency characteristic diagram of different discharge types of transformer oil-paper insulation, characteristic parameters are extracted from the amplitude-frequency characteristic diagram, including: the frequency at the maximum amplitude, the average amplitude in the range of 0 to 30 MHz, and the average amplitude in the range of 30 to 50 MHz, so as to obtain frequency domain characteristics.

[0049] The first-order shape parameter features, kurtosis features, skewness features, and frequency domain features obtained above are integrated to establish feature extraction results.

[0050] The abnormality analysis module 70 is used to perform transformer abnormality analysis based on the feature extraction result and generate an abnormality report.

[0051] The specific configuration of the abnormality analysis module 70 will be described in detail below. The abnormality analysis module 70 may further include: a first parameter acquisition unit, used to acquire equipment status parameters of the transformer and establish an equipment status data set, wherein the equipment status data set includes load data, voltage level data, and operating time data; a second parameter acquisition unit, used to acquire environmental parameters of the transformer and establish an environmental data set, wherein the environmental data set includes temperature data and humidity data; an abnormality identification unit, used to identify transformer abnormalities using the equipment status data set, the environmental data set, and the feature extraction result, and establish an abnormality report.

[0052] Specifically, the initial parameters of the data are collected by using a data collection device (such as a sensor, a monitor, etc.) connected to the transformer, which may include load data, voltage level data, and operating time data. The load data is obtained by collecting the load current, active power, reactive power and other data of the transformer in real time through a current transformer or a power monitoring device and storing them. The voltage level data is obtained by collecting the voltage level data of the input and output of the transformer using a voltage transformer or a voltage monitoring device, and recording and storing them. The operating time data is obtained by recording the operating time of the transformer through a time recording device (such as a clock module) or a data management system, including continuous operating time, cumulative operating time, etc. Finally, the load data, voltage level data, and operating time data are integrated to form an equipment status data set containing load data, voltage level data, and operating time data.

[0053] Furthermore, based on the temperature sensors and humidity sensors deployed around the transformer, the status of the transformer operating environment is accurately collected, which may include temperature data and humidity data. The temperature data is obtained by collecting the temperature data of the transformer and its surrounding environment in real time through the temperature sensor and storing it. The humidity data is obtained by collecting the humidity data of the environment in real time through the humidity sensor and storing it. The temperature data and humidity data are integrated to form an environmental data set including temperature data and humidity data.

[0054] Finally, the equipment status data set and the environmental data set are preprocessed to extract key features that can reflect the operating status of the transformer, such as load change rate, voltage fluctuation rate, temperature anomaly points, humidity change trend, etc. Machine learning algorithms such as support vector machine, decision tree, neural network, etc. are used to train the model using historical data including normal and abnormal state data, so that it can identify the abnormal state of the transformer. The real-time collected equipment status data and environmental data are input into the trained abnormal recognition model, and the model will output the abnormal recognition result according to the input data. When the model recognizes the abnormal state, an abnormal report is automatically generated. The report should include information such as the time when the abnormality occurred, the type of abnormality (such as overload, overheating, high humidity, etc.), the degree of abnormality, and possible processing suggestions. According to the content of the abnormal report, take corresponding processing measures, such as adjusting the load, reducing the ambient temperature, increasing ventilation, etc., and feed back the processing results to the abnormal recognition model to optimize the recognition performance of the model, generate an abnormal report, and provide strong protection for the safe operation of the transformer.

[0055] Below, an ultra-wideband transformer PD source intelligent detection platform may further include: an attention monitoring module, used to establish an attention window according to the abnormal report; a feedback monitoring module, used to perform feedback update on the feature extraction module based on the attention window, and perform transformer monitoring management according to the feature extraction module after feedback update.

[0056] Specifically, the monitoring module first receives the abnormality report from the abnormality identification unit. These reports contain information such as the time, type, degree and possible processing suggestions of the transformer abnormality. According to the content of the abnormality report, the monitoring module will establish an attention window. The attention window can be a time interval, which is used to cover the time point when the abnormality occurs, and can extend forward and backward for a certain period of time, so as to conduct a comprehensive analysis of the situation before and after the abnormality. Within the attention window, the monitoring module will set the corresponding monitoring parameters according to the abnormality type. Exemplarily, if the abnormality is caused by overload, the monitoring parameters may include load current, active power, etc.; if the abnormality is caused by overheating, the monitoring parameters may include temperature, heat dissipation, etc. After setting the monitoring parameters, the monitoring module will start real-time monitoring and continuously monitor the status of the transformer in the attention window.

[0057] Furthermore, the feedback monitoring module receives real-time monitoring data from the attention monitoring module. These data are obtained by real-time monitoring of the transformer within the attention window, and the feedback monitoring module will analyze the received real-time monitoring data. If the data shows that the state of the transformer is still abnormal, or the abnormal situation has intensified, the feedback mechanism will be triggered. After the feedback is triggered, the feedback monitoring module will update the feature extraction module according to the real-time monitoring data and the content of the abnormal report. The update content may include adjusting the feature extraction algorithm, adding new feature dimensions, etc., in order to more accurately reflect the current state of the transformer. The updated feature extraction module will re-extract features from the real-time monitoring data. The extracted features will be used for subsequent abnormality identification and analysis. If the state of the transformer is still identified as abnormal after re-extraction of the features, a new abnormality report will be generated and sent to relevant personnel for processing.

[0058] Finally, the monitoring data collected by the attention monitoring module and the feedback monitoring module will be integrated to form a comprehensive data set. This data set contains various parameters and characteristics of the transformer in different time periods and different states, and the integrated monitoring data will be deeply analyzed and mined. Through data analysis, the laws and trends of transformer state changes can be discovered, providing a scientific basis for subsequent monitoring and management, and formulating more reasonable monitoring strategies based on the results of data analysis. For example, the monitoring frequency can be adjusted, more sensitive thresholds can be set, etc., so as to more timely discover and handle abnormal conditions of the transformer, and continuously optimize the monitoring system according to the actual monitoring effect. This includes updating hardware equipment, improving software algorithms, improving the stability and reliability of the system, etc., and continuously monitoring and managing the transformer according to the formulated monitoring strategy and the optimized monitoring system. Through real-time monitoring, abnormality identification, feedback update and other means, the safe and stable operation of the transformer can be ensured, thereby realizing comprehensive, accurate and real-time monitoring and management of the transformer.

[0059] The embodiments of the present application solve the technical problem of low accuracy in distinguishing between ultra-wideband PD signal detection and noise sources, and achieve the technical effect of quickly extracting PD pulse characteristic parameters and realizing intelligent clustering and separation.

[0060] In the above, refer to Figure 1 The present invention describes in detail an ultra-wideband transformer PD source intelligent detection platform according to an embodiment of the present application. Figure 3 An ultra-wideband transformer PD source intelligent detection method according to an embodiment of the present application is described.

[0061] Step A100, configuring a sensor group, the sensor group including an HFCT, a bushing end screen sensor, and a detection impedance, using the sensor group to monitor the partial discharge pulse signal of the transformer, and establishing a pulse signal set; Execute step A200, after performing wavelet decomposition on the pulse signal set, obtain wavelet decomposition coefficients at each scale, and generate m initial particles in the solution space using the wavelet decomposition coefficients; Execute step A300, calculate particle similarity for the m initial particles, establish particle similarity association, combine particles using the particle similarity association, and establish particle association; Configure similar calculation formula as follows: ; in, Characterizing particles and particles The particle similarity correlation, and Characterize and measure particles separately and particles The fitness function of the individual quality, is the particle fitness deviation coefficient, is the particle's own deviation coefficient; A combination isolation strategy is configured, and after the particle similar associations are sorted in association order, association screening is performed from the maximum association through the combination isolation strategy to complete the particle combination and establish the particle association.

[0062] Execute step A400, after performing association selection on the particle association, generate association optimization according to the current iteration number and iteration state, and perform m initial particle update iterations using the association optimization and the particle individual extreme value and the particle group extreme value; After performing association normalization processing on the particle associations, performing downward association ratio clustering from the highest association value to establish a downward association ratio clustering result; When the downward association ratio clustering result meets the preset threshold, the clustering ends, and the particles corresponding to the particle associations in the cluster are used as the association selection particles; Obtaining a current number of iterations, and generating a first optimization weight based on the number of iterations; Acquire a current iteration state, and generate a second optimization weight based on the iteration state; After comprehensively calculating the first optimization weight and the second optimization weight, an optimization reward is generated by using the low fitness particles in the associated selected particles and the comprehensive calculation result, and the high fitness particles in the associated selected particles are updated by using the optimization reward, and the update of the low fitness particles is stopped, and the update iteration is completed according to the update result.

[0063] Remove the low fitness particles that stop updating from the m initial particles, and establish the updated initial particles; Performing individual update on non-associated selected particles in the updated initial particles based on individual particle extreme values ​​and particle group extreme values ​​to generate a first update result; Performing individual update on the associated selected particles in the updated initial particles based on the optimization reward, the individual extreme value of the particles, and the extreme value of the particle group to generate a second update result; After the first update result and the second update result are supplemented with new particles, the update iteration is completed.

[0064] Execute step A500, dynamically adjust the inertia weight, perform iterative optimization, establish an optimal wavelet threshold according to the iterative optimization result, perform denoising on the pulse signal set using the optimal wavelet threshold, and establish a denoised signal set; Next, step A600 is performed to extract first-order shape parameter features, kurtosis features, skewness features, and frequency domain features from the denoised signal set to establish a feature extraction result; The kurtosis feature is calculated by the formula as follows: ; in, Characterizes the kurtosis, Characterize the random variable A value, represents the mean of a random variable, is the standard deviation of the random variable, is the total number of random variables, The probability density function that characterizes a random variable is The value of ; The skewness feature is calculated by the formula as follows: ; in, Characterize skewness; The kurtosis feature and the skewness feature are taken as the feature extraction results.

[0065] Finally, step A700 is executed to perform transformer abnormality analysis based on the feature extraction result and generate an abnormality report.

[0066] Collecting equipment status parameters of the transformer to establish an equipment status data set, wherein the equipment status data set includes load data, voltage level data, and operating time data; Execute environmental parameter collection of the transformer and establish an environmental data set, wherein the environmental data set includes temperature data and humidity data; The device status data set, the environmental data set, and the feature extraction result are used to identify transformer anomalies and create an anomaly report.

[0067] The ultra-wideband transformer PD source intelligent detection platform also includes: A focus monitoring module, used for establishing a focus window according to the abnormal report; A feedback monitoring module is used to perform feedback update on the feature extraction module based on the focus window, and perform transformer monitoring management according to the feature extraction module after feedback update.

[0068] An ultra-wideband transformer PD source intelligent detection platform provided in an embodiment of the present application can execute an ultra-wideband transformer PD source intelligent detection method provided in any embodiment of the present application, and has functional modules and beneficial effects corresponding to the execution method.

[0069] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present application.

[0070] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. An ultra-wideband transformer PD source intelligent detection platform, characterized in that: The platform includes: A configuration module is used to configure a sensor group, wherein the sensor group includes an HFCT, a bushing end screen sensor, and a detection impedance, and the sensor group is used to monitor the partial discharge pulse signal of the transformer to establish a pulse signal set; An initial solution construction module, used to perform wavelet decomposition on the pulse signal set, obtain wavelet decomposition coefficients at each scale, and generate m initial particles in the solution space using the wavelet decomposition coefficients; An association analysis module is used to calculate the particle similarity of the m initial particles, establish particle similarity association, combine particles using the particle similarity association, and establish particle association; An update iteration module is used to generate an association optimization through the current iteration number and iteration state after performing association selection on the particle association, and perform m initial particle update iterations using the association optimization and the particle individual extreme value and the particle group extreme value; A denoising processing module is used to dynamically adjust the inertia weight, perform iterative optimization, establish an optimal wavelet threshold according to the iterative optimization result, and establish a denoised signal set after denoising the pulse signal set using the optimal wavelet threshold; A feature extraction module, used to extract first-order shape parameter features, kurtosis features, skewness features, and frequency domain features from the denoised signal set, and establish a feature extraction result; The abnormality analysis module is used to perform transformer abnormality analysis based on the feature extraction results and generate an abnormality report.

2. The ultra-wideband transformer PD source intelligent detection platform according to claim 1, characterized in that: The association analysis module is used for: The first configuration unit is used to configure a similarity calculation formula as follows: ； in, Characterizing particles and particles The particle similarity correlation, and Characterize and measure particles separately and particles The fitness function of the individual quality, is the particle fitness deviation coefficient, is the particle's own deviation coefficient; The second configuration unit is used to configure a combination isolation strategy, after sorting the particle similar associations in association order, perform association screening from the maximum association through the combination isolation strategy to complete the particle combination and establish the particle association.

3. The ultra-wideband transformer PD source intelligent detection platform as claimed in claim 2, characterized in that: The update iteration module is used for: A clustering unit, configured to perform association normalization processing on the particle associations, and then perform downward association ratio clustering from the highest association value to establish a downward association ratio clustering result; The particle association unit is used to terminate clustering when the downward association ratio clustering result meets a preset threshold, and to associate the particles in the cluster with the particles corresponding to the particles as the association selection particles; A first iteration unit, used to obtain a current number of iterations, and generate a first optimization weight based on the number of iterations; A second iteration unit, used to obtain a current iteration state, and generate a second optimization weight based on the iteration state; An updating unit is used to comprehensively calculate the first optimization weight and the second optimization weight, generate an optimization reward by using the low fitness particles in the associated selected particles and the comprehensive calculation result, update the high fitness particles in the associated selected particles by using the optimization reward, stop updating the low fitness particles, and complete the update iteration according to the update result.

4. The ultra-wideband transformer PD source intelligent detection platform as claimed in claim 3, characterized in that: The update iteration module is also used for: A removal unit is used to remove low-fitness particles that have stopped updating from the m initial particles, and establish updated initial particles; A first individual updating unit, used for performing individual updating on non-associated selected particles in the updated initial particles based on particle individual extreme values ​​and particle group extreme values, to generate a first updating result; A second individual updating unit is used to perform individual updating on the associated selected particles in the updated initial particles based on the optimization reward, the individual extreme value of the particles, and the extreme value of the particle group to generate a second update result; The particle supplement unit is used to supplement the first update result and the second update result with new particles to complete the update iteration.

5. The ultra-wideband transformer PD source intelligent detection platform as claimed in claim 1, characterized in that: The feature extraction module is used for: The first feature calculation unit is used to calculate the kurtosis feature by the formula as follows: ； in, Characterizes the kurtosis, Characterize the random variable A value, represents the mean of a random variable, is the standard deviation of the random variable, is the total number of random variables, The probability density function that characterizes a random variable is The value of ; The second feature calculation unit is used to calculate the skewness feature through the formula as follows: ； in, Characterize skewness; The feature extraction unit is used to take the kurtosis feature and the skewness feature as feature extraction results.

6. The ultra-wideband transformer PD source intelligent detection platform as claimed in claim 1, characterized in that: The abnormality analysis module is also used for: A first parameter acquisition unit is used to collect equipment status parameters of the transformer and establish an equipment status data set, wherein the equipment status data set includes load data, voltage level data, and operating time data; A second parameter collection unit is used to collect environmental parameters of the transformer and establish an environmental data set, wherein the environmental data set includes temperature data and humidity data; The abnormality identification unit is used to identify transformer abnormalities using the equipment status data set, the environmental data set, and the feature extraction result, and to establish an abnormality report.

7. The ultra-wideband transformer PD source intelligent detection platform according to claim 1, characterized in that: The platform also includes: A focus monitoring module, used for establishing a focus window according to the abnormal report; A feedback monitoring module is used to perform feedback update on the feature extraction module based on the focus window, and perform transformer monitoring management according to the feature extraction module after feedback update.

8. An ultra-wideband transformer PD source intelligent detection method, characterized in that: The method is performed by an ultra-wideband transformer PD source intelligent detection platform according to any one of claims 1 to 7, wherein the ultra-wideband transformer PD source intelligent detection method comprises: A sensor group is configured, wherein the sensor group includes an HFCT, a bushing end screen sensor, and a detection impedance, and the sensor group is used to monitor the partial discharge pulse signal of the transformer to establish a pulse signal set; After performing wavelet decomposition on the pulse signal set, wavelet decomposition coefficients at each scale are obtained, and m initial particles are generated in the solution space using the wavelet decomposition coefficients; Calculating particle similarity for the m initial particles, establishing particle similarity association, combining particles using the particle similarity association, and establishing particle association; After the particle association is selected, an association optimization is generated according to the current iteration number and iteration state, and the association optimization and the particle individual extreme value and the particle group extreme value are used to perform m initial particle update iterations; Dynamically adjust the inertia weight, perform iterative optimization, establish an optimal wavelet threshold according to the iterative optimization result, and establish a denoised signal set after denoising the pulse signal set using the optimal wavelet threshold; Extracting first-order shape parameter features, kurtosis features, skewness features, and frequency domain features from the denoised signal set to establish a feature extraction result; Transformer abnormality analysis is performed based on the feature extraction results to generate an abnormality report.

Images