State evaluation method and system of transformer on-load tap-changer, electronic equipment and medium

Abstract

The invention provides a transformer on-load tap-changer state evaluation method and system, electronic equipment and a medium, and belongs to the field of fault diagnosis of transformer on-load tap-changers. The method comprises the following steps: acquiring electrical and mechanical performance parameters of the on-load tap-changer of the transformer, and generating an oil chromatographic data characteristic matrix according to oil chromatographic data; for the vibration signals, generating a vibration signal feature matrix; generating a driving motor current signal characteristic matrix according to the driving motor current signal; fusing the three feature matrixes to generate a multi-dimensional information feature matrix; and calculating correlation among data by using an attention mechanism, endowing a multi-dimensional information feature matrix with a weight, and inputting the multi-dimensional information feature matrix into a BIGRU model for state evaluation. The characteristic data of the electrical characteristics and the mechanical characteristics of the on-load tap-changer are comprehensively selected, and the evaluation model is used for fault diagnosis, so that the diagnosis accuracy and adaptability are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault diagnosis of transformer on-load tap changer devices, and in particular relates to a state evaluation method, system, electronic equipment and medium for transformer on-load tap changer devices. Background Art

[0002] With the rapid development of my country's power industry, the equipment capacity and scale of the power grid are getting larger and larger, which leads to higher economic losses caused by the failure and maintenance of power equipment, and the social effects will become more and more serious, which puts forward higher technical requirements for the guarantee of power supply safety and quality. Therefore, effective safeguard measures must be taken to improve the reliability of normal operation of power equipment and the stability of system operation, so as to effectively ensure the quality of power supply.

[0003] As the only movable part in the transformer, the on-load tap changer (OLTC) can not only reduce and avoid large voltage fluctuations with its accurate and timely action, but also force the distribution of load flow, tap the reactive and active output of equipment, and ensure the safe and reliable operation of the power system. Statistics show that OLTC failures account for about 30% of the total transformer failures, and the failure types are basically mechanical failures, such as contact failure, loose components, spring fatigue, abnormal switching sequence, etc. Once a mechanical failure occurs in OLTC, it may cause the line to trip and lose transmission power, or even cause the converter transformer to catch fire and burn, causing huge economic losses and adverse social impacts.

[0004] However, traditional OLTC status detection methods are mainly based on electrical performance measurements, including testing transition resistance impedance and judging the switching time based on the current on-off conditions, and the existing electrical measurement methods are often powerless against the hidden dangers of mechanical failures during the OLTC switching process. In recent years, a mechanical status vibration detection method has been proposed to detect mechanical performance by arranging vibration sensors on the surface of the equipment to collect vibration signals and extracting the characteristics of vibration signals under different working conditions to achieve status detection. However, the current traditional mechanical fault diagnosis methods mostly rely on a single feature, and have problems such as low diagnostic accuracy and poor adaptability. Summary of the invention

[0005] The purpose of the embodiments of the present invention is to provide a method, system, electronic device and medium for evaluating the status of an on-load tap changer of a transformer, so as to completely or at least partially solve the technical problems existing in the above-mentioned prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for evaluating the state of an on-load tap changer of a transformer, comprising: Acquire electrical performance parameters and mechanical performance parameters of the transformer on-load tap changer, wherein the electrical performance parameters include oil chromatogram data, and the mechanical performance parameters include vibration signals and drive motor current signals; Generate an oil chromatogram data feature matrix for the oil chromatogram data, wherein the oil chromatogram data feature matrix includes the volume fraction of gas generated by decomposition of insulating oil of the on-load tap changer under abnormal conditions, the total hydrocarbon value, the temperature of the on-load decomposition switch overheating fault point, the temperature of the discharge fault point, and the acetylene growth rate; For the vibration signal, multi-scale wave propagation entropy is used to extract the vibration signal characteristics of the hydropower unit, and energy entropy, kurtosis and spectral kurtosis are introduced to generate a vibration signal feature matrix; Generate a driving motor current signal characteristic matrix for the driving motor current signal, wherein the driving motor current signal characteristic matrix includes the duration of the driving motor current and the sum of the absolute values ​​of the driving motor current in the motor starting stage, the motor stable operation stage and the motor stop operation stage; Fusion of the oil chromatogram data feature matrix, the vibration signal feature matrix and the drive motor current signal feature matrix to generate a multi-dimensional information feature matrix; The attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, which are then input into a pre-built BIGRU model to realize the state evaluation of the transformer on-load tap changer.

[0007] Optionally, the gas generated by decomposition of the insulating oil of the on-load tap changer under abnormal conditions includes at least hydrogen, methane, ethane, ethylene and acetylene.

[0008] Optionally, the temperature of the on-load switch overheating fault point is calculated according to the following formula:

[0009] Where, T G is the overheating fault point temperature, is the volume fraction of ethylene, is the volume fraction of ethane.

[0010] Optionally, the temperature of the discharge fault point is calculated according to the following formula:

[0011] In the formula, is the discharge fault point temperature, is the hydrogen gas volume fraction, is the volume fraction of acetylene.

[0012] Optionally, calculate the acetylene growth rate according to the following formula:

[0013] In the formula, is the acetylene growth rate, is the change in acetylene volume fraction, is the time interval.

[0014] Optionally, the calculation process of the multi-scale fluctuation spread entropy includes: Map the coarse-grained time series and perform linear transformation on the mapping result to obtain Z; Calculate embedding vectors according to embedding dimension and time delay, and map each embedding vector to a fluctuation-based distribution pattern; Calculate the probability of each dispersion model and calculate the dispersion entropy based on fluctuations based on the Shannon entropy calculation method; The fluctuation-based spread entropy under the scale factor is calculated to obtain the multi-scale fluctuation spread entropy function.

[0015] Optionally, calculate the energy entropy according to the following formula:

[0016] In the formula, represents the vibration amplitude, Represents the ratio of the current position energy to the total energy, Represents a time series.

[0017] Optionally, calculate kurtosis according to the following formula:

[0018] In the formula, N is the data sample size, is the i-th data sample, is the sample mean.

[0019] Optionally, calculate the spectral kurtosis according to the following formula: ; ; In the formula, Indicates frequency, represents the power spectral density of the current frequency, represents the average power spectral density.

[0020] Optionally, an attention mechanism is used to calculate the correlation between data in a multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, including an additive attention scoring mechanism:

[0021] In the formula, It is the attention scoring mechanism after the hidden layer undergoes a full connection operation. u s , w , b are the randomly initialized time series, attention weight matrix and bias matrix, respectively. for i Characteristics j Hidden layer states of group data.

[0022] Optionally, the attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, and also includes an attention distribution mechanism:

[0023] In the formula, for i Characteristics j The degree of attention that the group data receives for the target prediction quantity.

[0024] Optionally, the attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, and also includes an information weighted summation mechanism:

[0025] In the formula, is the score vector after weighted summation, for i Characteristics j The degree of attention paid to the target prediction quantity by the group data, for i Characteristics j Hidden layer states of group data.

[0026] Optionally, the construction process of the BIGRU model includes: Use the forward GRU network and the reverse GRU network to build the initial BIGRU model; The multiverse algorithm is used to optimize the hyperparameters of the initial BIGRU model to obtain the target BIGRU model, wherein the hyperparameters of the initial BIGRU model include learning rate, number of iterations, number of forward hidden layer nodes, and number of backward hidden layer nodes.

[0027] Optionally, the process of optimizing the hyperparameters of the initial BIGRU model using a multiverse algorithm includes: In the population initialization process, Circle chaos mapping is added to generate a chaotic sequence to initialize the universe through the chaotic sequence; Calculate the expansion rate of each universe, and introduce an adaptive mechanism to adjust the wormhole existence probability E and the travel distance rate D parameters to update the universe expansion rate; Construct a fitness function, calculate the fitness, and determine whether the maximum number of iterations has been reached. If so, obtain the optimized hyperparameters.

[0028] Optionally, an adaptive mechanism is introduced to adjust the wormhole existence probability E and the travel distance rate D parameters, and the formula for updating the universe expansion rate is:

[0029] In the formula, represents the jth parameter of the current optimal universe, , , represents a random number that follows a uniform distribution and is between [0, 1]. represents the probability of wormhole existence in each generation of the universe, D represents the travel distance rate, represents the minimum value of the jth parameter, represents the maximum value of the jth parameter, represents the jth parameter of the i-th initialized universe.

[0030] Optionally, the fitness function formula is:

[0031] In the formula, is the number of correctly classified samples; is the total number of samples.

[0032] In a second aspect, an embodiment of the present application further provides a transformer on-load tap changer status assessment system, comprising: An acquisition unit, used to acquire electrical performance parameters and mechanical performance parameters of the transformer on-load tap changer, wherein the electrical performance parameters include oil chromatogram data, and the mechanical performance parameters include vibration signals and drive motor current signals; A first generating unit is used to generate an oil chromatogram data feature matrix for the oil chromatogram data, wherein the oil chromatogram data feature matrix includes the volume fraction of gas generated by decomposition of insulating oil of the on-load tap changer under abnormal conditions, the total hydrocarbon value, the temperature of the overheating fault point of the on-load decomposition switch, the temperature of the discharge fault point, and the acetylene growth rate; The second generating unit is used to extract the vibration signal characteristics of the hydropower unit by using multi-scale wave propagation entropy for the vibration signal, and introduce energy entropy, kurtosis and spectral kurtosis to generate a vibration signal characteristic matrix; a third generating unit, configured to generate a driving motor current signal characteristic matrix for the driving motor current signal, wherein the driving motor current signal characteristic matrix includes a duration of the driving motor current and a sum of absolute values ​​of the driving motor current in a motor startup phase, a motor stable operation phase, and a motor stop operation phase; A fusion unit, used for fusing the oil chromatogram data feature matrix, the vibration signal feature matrix and the drive motor current signal feature matrix to generate a multi-dimensional information feature matrix; An evaluation unit is used to calculate the correlation between data in a multidimensional information feature matrix using an attention mechanism and assign weights to the multidimensional information feature matrix, obtain data samples with weights, and input the data samples with weights into a pre-built BIGRU model to achieve status evaluation of the transformer on-load tap changer.

[0033] In a third aspect, an embodiment of the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for assessing the status of an on-load tap changer of a transformer when executing the program.

[0034] In a fourth aspect, an embodiment of the present application further provides a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-mentioned method for assessing the state of an on-load tap changer of a transformer.

[0035] It can be seen from the above technical solutions that the present invention has the following advantages: In the transformer on-load tap changer state assessment method, system, electronic device and medium provided in the present application, characteristic data that can reflect the electrical characteristics and mechanical characteristics of the on-load tap changer are comprehensively selected to form a characteristic matrix, and the transformer on-load tap changer state assessment model based on multi-dimensional information fusion is used for fault diagnosis, thereby improving the diagnostic accuracy and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings required for use in the description will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0037] Figure 1 A flow chart of a method for evaluating the status of an on-load tap changer of a transformer provided by an embodiment of the present invention; Figure 2 A timing diagram of a vibration signal of an on-load tap changer provided by an embodiment of the present invention; Figure 3 A coarse-grained time series graph provided by an embodiment of the present invention; Figure 4 A flowchart of a MFDE method provided by an embodiment of the present invention; Figure 5 A schematic diagram of a motor current waveform division result provided by an embodiment of the present invention; Figure 6 A structural diagram of an attention mechanism provided by an embodiment of the present invention; Figure 7 A structural diagram of a GRU provided in an embodiment of the present invention; Figure 8 A structural diagram of a BIGRU provided in an embodiment of the present invention; Fig. 9 An IMVO algorithm optimization flow chart provided by an embodiment of the present invention; Fig.10 A detailed flow chart of a method for evaluating the state of an on-load tap changer of a transformer provided by an embodiment of the present invention; Fig.11 A schematic structural diagram of a transformer on-load tap changer status assessment system provided by an embodiment of the present invention; Fig.12 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In the detailed description below, various embodiments of the present disclosure will be described more fully. The present disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but rather the present disclosure should be understood to cover all adjustments, equivalents and / or alternatives that fall within the spirit and scope of the various embodiments of the present disclosure.

[0039] Hereinafter, the terms "include" or "may include" that may be used in various embodiments of the present disclosure indicate the presence of the disclosed functions or operations, and do not limit the addition of one or more functions or operations. In addition, as used in various embodiments of the present disclosure, the terms "include", "have" and their cognates are intended only to indicate specific features, numbers, steps, operations, or combinations of the foregoing items, and should not be understood as first excluding the presence of one or more other features, numbers, steps, operations, or combinations of the foregoing items or the possibility of adding one or more features, numbers, steps, operations, or combinations of the foregoing items.

[0040] In various embodiments of the present disclosure, the expression "or" or "at least one of A or / and B" includes any combination or all combinations of the words listed at the same time. For example, the expression "A or B" or "at least one of A or / and B" may include A, may include B, or may include both A and B.

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] See also Figure 1 The flowchart of a method for evaluating the status of an on-load tap changer of a transformer in a specific embodiment is shown, and includes the following execution steps: Step 100: Acquire electrical performance parameters and mechanical performance parameters of the transformer on-load tap changer, wherein the electrical performance parameters include oil chromatogram data, and the mechanical performance parameters include vibration signals and drive motor current signals.

[0043] Specifically, the performance indicators of OLTC mainly include electrical performance and mechanical performance. In terms of electrical performance evaluation, when the vacuum on-load tap changer of the transformer operates normally, the arc extinguishing process is generally carried out in the vacuum tube during the switching process. The insulating oil of the tap changer does not directly bear the arc extinguishing function, and the insulating oil will not deteriorate or produce gas. When an electrical fault occurs in the vacuum on-load tap changer, characteristic gas will be produced in the insulating oil. Oil chromatography analysis can effectively detect latent electrical faults inside the vacuum on-load tap changer; in terms of mechanical performance evaluation, on the one hand, during the OLTC switching process, the vibration signal generated by the multiple collisions between the moving and static contacts contains the mechanical state information of each component, and the vibration analysis method can effectively evaluate the mechanical state of the OLTC. On the other hand, the OLTC uses the drive motor to provide driving force for the spring energy storage, and completes the gear switching through a series of mechanical actions of the connecting rod, gear box, fast mechanism, contact and other components. The motor current signal also contains rich mechanical state information. In summary, the oil chromatography data, vibration signal and drive motor current signal of the on-load tap changer are selected.

[0044] Step 101: Generate an oil chromatogram data feature matrix for the oil chromatogram data, wherein the oil chromatogram data feature matrix includes the volume fraction of gas generated by decomposition of insulating oil of the on-load tap changer under abnormal conditions, the total hydrocarbon value, the temperature of the on-load decomposition switch overheating fault point, the temperature of the discharge fault point, and the acetylene growth rate.

[0045] In some embodiments, the insulating oil of the on-load tap changer will not decompose to produce gas under normal circumstances. When the vacuum tube leaks or the insulation performance deteriorates, causing the arc to fail to extinguish normally, the isolating switch connected in series with the vacuum tube will break the arc and decompose to produce gas. This embodiment intends to select five characteristic gases As the main research object, the corresponding defects of different characteristic gas contents are analyzed and calculated. Therefore, the first thing introduced is The volume fractions of the five gases and the total hydrocarbon value are used as characteristic values. It should be noted that for the same type of switches, due to different manufacturers, different models, and different structural designs, the production of characteristic gases in defective conditions will also vary greatly.

[0046]

[0047] In the formula, is the volume fraction of gas 𝑖, 10 -6 ; is the volume of gas 𝑖, and its units are the same as the total volume; is the total volume of the mixed gas.

[0048] Specifically, under abnormal conditions, the gases generated by the decomposition of the insulating oil of the on-load tap changer include at least hydrogen, methane, ethane, ethylene and acetylene.

[0049] In some embodiments, the overheating fault of the on-load tap changer is generally caused by poor contact of the current-carrying contacts, long-term overload, or poor heat dissipation of the medium around the contacts. Overheating of the contacts will inevitably cause the insulating oil to decompose, and H2 is usually the first characteristic gas to appear; as the temperature rises, CH4, C2H6 and C2H4 will be generated. At this time, the fault characteristic gases are CH4 and C2H4, and as the temperature of the fault point increases, the proportion of C2H4 gradually increases. The temperature of the overheating fault point of the on-load decomposition switch is calculated according to the following formula:

[0050] Where, T G is the overheating fault point temperature, is the volume fraction of ethylene, is the volume fraction of ethane.

[0051] The electrical faults of the vacuum on-load tap-changer of the transformer are mainly discharge faults. The discharge faults of the on-load tap-changer can be divided into three types: partial discharge, spark discharge and arc discharge according to the different energy density of the discharge. Partial discharge and spark discharge are low-energy discharge faults, and arc discharge is a malignant fault of high-energy discharge. It is generally believed that partial discharge and spark discharge will not quickly cause insulation breakdown of the vacuum on-load tap-changer, which is mainly reflected in abnormal oil chromatography analysis and light gas action of the on-load tap-changer. The fault is relatively easy to find and handle; arc discharge usually causes heavy gas action of the on-load tap-changer and the transformer trips. H2 and C2H2 are usually used as the main characteristic gases of discharge faults. Arc discharge also means that the temperature of the faulty part is extremely high. The temperature of the discharge fault point is calculated according to the following formula:

[0052] In the formula, is the discharge fault point temperature, is the hydrogen gas volume fraction, is the volume fraction of acetylene.

[0053] At the same time, during actual operation, due to the main contact recovery voltage, heating of the transition resistor wire and other reasons, discharge characteristic gases such as acetylene and overheating characteristic gases such as ethylene and methane are often detected in the insulating oil of the normally operating vacuum on-load tap changer, but the acetylene gas production rate is generally slow. Therefore, after the characteristic gas is detected, in order to effectively distinguish whether the gas production of the on-load tap changer is normal gas production or fault gas production, this application introduces the acetylene growth rate as a new characteristic value. The state of the tap changer is comprehensively judged by combining the volume fraction content of different characteristic gases of the on-load tap changer, the fault point temperature and the acetylene growth rate. Calculate the acetylene growth rate according to the following formula:

[0054] In the formula, is the acetylene growth rate, is the change in acetylene volume fraction, is the time interval.

[0055] In summary, a feature matrix is ​​formed for the oil chromatography data: .

[0056] Step 102: for the vibration signal, use multi-scale wave propagation entropy to extract the vibration signal characteristics of the hydropower unit, and introduce energy entropy, kurtosis and spectral kurtosis to generate a vibration signal feature matrix.

[0057] In recent years, a mechanical state vibration detection method has been proposed. Vibration sensors are arranged on the surface of the equipment to collect vibration signals, and the characteristics of vibration signals under different working conditions are extracted to achieve state detection. Since vibration signals and their related characteristics can often directly reflect the internal faults of the equipment, the vibration detection method has high sensitivity and accuracy for mechanical state detection. At the same time, since the entire measurement is non-invasive, the detection is easy to operate and has strong engineering practice value. However, the fault vibration signal contains a lot of noise interference, and its dynamic law is generally very nonlinear. Traditional nonlinear system analysis indicators such as mean, variance, kurtosis, skewness, peak, waveform factor, etc. are difficult to effectively extract fault feature information. Therefore, new indicators are needed to analyze the system properties or dynamic laws of the signal. This application introduces multi-scale fluctuation spread entropy (MFDE) for the vibration signal of the on-load tap changer to characterize the complexity of the time series at different scales, so as to enhance the fault characteristics of the vibration signal. The timing diagram of the vibration signal of the on-load tap changer is shown in the attached figure. Figure 2 shown.

[0058] Specifically, the calculation process of multi-scale fluctuation spread entropy includes the following steps: S1: Map the coarse-grained time series and perform a linear transformation on the mapping result to obtain Z.

[0059] For example, the coarse-grained time series Map to ,Right now: ; Among them, μ and are the expectation and variance of x respectively.

[0060] Then Linear transformation to , The elements are Integer:

[0061] Among them, round() is the rounding method, and c is the number of categories in the symbolization process.

[0062] S2: Calculate embedding vectors based on embedding dimension and time delay, and map each embedding vector to a fluctuation-based distribution pattern.

[0063] Exemplarily, the embedding vector is calculated according to the embedding dimension m and the delay d. :

[0064] Each embedding vector Mapping to a wave-based dispersion pattern :

[0065] each The number of all possible walking patterns is .

[0066] S3: Calculate the probability of each diffusion model and calculate the dispersion entropy based on fluctuations based on the Shannon entropy calculation method.

[0067] For example, each possible scatter pattern is calculated The probability of :

[0068] Where count() means Map to The number of equal Map to The number of is divided by the total number of embedding vectors corresponding to the embedding dimension m.

[0069] S4: Calculate the fluctuation-based spread entropy under the scale factor and obtain the multi-scale fluctuation spread entropy function.

[0070] Exemplarily, based on the calculation method of Shannon entropy, the dispersion entropy based on fluctuation is calculated:

[0071] The difference between adjacent elements of the scatter pattern is considered in the FDE calculation, which is called the fluctuation-based scatter pattern. In this algorithm, a pattern vector with a dimension of m-1 can be obtained, and each element of the pattern vector ranges from -c+1 to c-1. Therefore, there are a total of When all the scatter patterns have equal probability values, the entropy value is the largest, which is , at this time the signal is completely random.

[0072] In MFDE calculations, see Figure 3 , Figure 4 As shown, for a given time series , first use floor to divide it into non-overlapping segments of, τ is the scale factor, is the floor operation. Then the average value of each segment is calculated. The coarse-grained process of MFDE can be simply understood as averaging the original time series in a window of length τ to obtain the time series:

[0073] Calculate the FDE under the scale factor τ and obtain the function MFDE of the scale factor τ:

[0074] The parameters of MFDE include embedding dimension m, number of categories n, time delay s, and scale factor τ. If the embedding dimension m is too large, it will be difficult to detect small changes in the signal. If m is too small, it will be difficult to observe dynamic changes in the signal. The usual selection range is 2≤m≤5, and m is 3; the number of categories is usually 3≤n≤8, and n=3; the time delay s is generally set to 1 to avoid losing important frequency information due to excessive delay. When τ=1, the coarse-grained time series is equal to the original signal. Figure 3 The coarse-grained time series for τ=2 and τ=3 are shown. Figure 4 Flowchart illustrating the MFDE method, where τ m The maximum scale factor to set.

[0075] At the same time, this application still calculates some traditional time-frequency domain eigenvalues ​​to enhance the fault characterization capability, and mainly extracts three types of eigenvalues: energy entropy, kurtosis, and spectral kurtosis.

[0076] In some embodiments, energy entropy describes the uncertainty of energy distribution of a signal in different frequency bands or time windows, which can be used to measure the complexity and randomness of the signal. Energy entropy is calculated according to the following formula:

[0077] In the formula, represents the vibration amplitude, Represents the ratio of the current position energy to the total energy, Represents a time series.

[0078] In some embodiments, kurtosis is a measure of the fourth-order standard moment, which is used to describe the peak degree of a signal, can be used to capture non-periodic mutations in a signal, and is used to detect instantaneous faults of a signal in fault diagnosis. The calculation formula is as follows:

[0079] In the formula, N is the data sample size, is the i-th data sample, is the sample mean.

[0080] Kurtosis is a feature that is sensitive to outliers. A single extreme value will affect the kurtosis of the entire segment. Spectral kurtosis is an extended concept of time domain feature kurtosis, which can be used to extract transient vibration frequency bands and envelope feature analysis. Spectral kurtosis can represent the sharpness of the spectrum and can be approximately realized as the kurtosis of the power spectrum density. The envelope spectrum of the signal is obtained by using the fast Fourier transform. The power spectrum density of the signal can be approximately obtained by normalizing the square of the spectrum. The power spectrum density is calculated as follows:

[0081] In the formula, Indicates frequency, Represents the power spectral density of the current frequency, and the spectral kurtosis is calculated according to the following formula: ; In the formula, Indicates frequency, represents the power spectral density of the current frequency, represents the average power spectral density.

[0082] Step 103: Generate a drive motor current signal characteristic matrix for the drive motor current signal, wherein the drive motor current signal characteristic matrix includes the duration of the drive motor current and the sum of the absolute values ​​of the drive motor current in the motor startup phase, the motor stable operation phase and the motor stop operation phase.

[0083] In some embodiments, the drive motor is the power source of the on-load tap changer, and its function is to adjust the working position of the on-load tap changer to the position required for operation during on-load voltage regulation. By observing the collected current waveform, it can be found that the motor current waveform can be divided into three stages: Phase 1: Motor start-up phase. Its characteristic is that there is a large inrush current at the moment the motor starts with load, and the current signal stabilizes after about 0.3s; Stage 2: The motor is in stable operation. Its characteristics are that the motor is working stably, the current signal amplitude is basically unchanged, and the OLTC switching action is completed in this stage; Phase 3: Motor stop phase. After a period of time after the OLTC switching action is completed, the motor current is cut off and drops to 0.

[0084] The drive motor is the power source for OLTC operation. Its output torque is closely related to the current signal. The duration T of the drive motor current is c It is equivalent to the total switching time of the on-load tap changer, so this application uses Tc as the first feature of the drive motor current signal.

[0085] When the on-load tap changer is stuck or the spring is fatigued, the torque of the drive motor will change. The common on-load tap changer drive motor is a three-phase asynchronous motor. The following will take the three-phase asynchronous motor as an example to discuss the relationship between the output torque T2 of the drive motor and the stator current I1. According to the power balance relationship, when the three-phase asynchronous motor is running stably, the electromagnetic torque P em The relationship with the stator phase current I1 is as follows:

[0086] Where: P1 is the input power; PCu1 is the stator copper loss; P Fe is the stator iron loss, which is ignored here; m1 is the number of phases; U1 is the stator phase voltage; is the stator power factor; R1 is the equivalent resistance on the stator side.

[0087] Ignoring mechanical losses and additional losses, the output power P2 of the three-phase asynchronous motor is:

[0088]

[0089] Where: P m is the mechanical power of the motor; s is the slip rate of the motor; n1 is the synchronous speed; n is the rotor speed.

[0090] From the power-torque relationship, we can know that the output torque T2 of the three-phase asynchronous motor is: ; ; Where Ω is the mechanical angular velocity of the rotor.

[0091] Combining the above formula, we can get:

[0092] because , and n1, U1, The output torque T2 of the three-phase asynchronous motor is approximately proportional to the stator current I1, which is a constant value in steady state. Therefore, this application uses the sum of the absolute values ​​of the drive motor current as a new feature of the drive motor current signal. At the same time, when different mechanical faults occur in the OLTC, the current characteristic parameters of each stage change differently, so it is necessary to divide these three stages. Based on the previous analysis, the intervals of these three stages are: , , The division results are shown in the attached Figure 5 Therefore, the sum of the absolute values ​​of each segment is calculated as the eigenvalue.

[0093]

[0094] In the formula, It is a certain stage of the current signal of the driving motor. The three stages are represented by Ⅰ, Ⅱ and Ⅲ.

[0095] Step 104: The oil chromatography data feature matrix, the vibration signal feature matrix and the drive motor current signal feature matrix are integrated to generate a multi-dimensional information feature matrix.

[0096] In summary, a characteristic matrix is ​​formed for the driving motor current signal: .

[0097] By analyzing the characteristics of the on-load tapchanger oil chromatographic data, vibration signal and drive motor current signal, a large multi-dimensional information feature matrix is ​​finally formed. .

[0098] In some embodiments, different feature data have large differences in magnitude. If such data with large differences in magnitude are directly used as the input of the evaluation model, it is easy to cause the small data to be swallowed by the network when it is approximately equal to zero relative to the large data, and the input data with lower magnitude is regarded as zero input by default. Most of the time, these data with smaller values ​​play a vital role in network training. Therefore, the different feature value data of the oil chromatogram data, vibration signal and drive motor current signal obtained in the above text are subjected to minimum and maximum (Min-Max) standardization processing to obtain data between 0 and 1, thereby reducing the interference of the data magnitude.

[0099] Step 105: Utilize the attention mechanism to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain data samples with weights, and input the data samples with weights into a pre-built BIGRU model to achieve status evaluation of the transformer on-load tap changer.

[0100] In order to better find useful information between input feature data and target output, highlight the main features related to fault diagnosis, and improve the accuracy of fault diagnosis, this application applies the attention mechanism (AM) to multi-dimensional feature data analysis, calculates the correlation between data and assigns different weights to different feature data, thereby replacing the original data. The following is an introduction to the principle of the attention mechanism. The attention mechanism is essentially a weighted probability distribution mechanism that can quickly filter out high-value information from a large amount of information, assign high weights to important content, and assign corresponding low weights to less important content. The structure of the attention mechanism can be regarded as two modules, Encode and Decode: Encode is an encoder that converts input data into binary data; Decode is a decoder that converts binary data into output data type. Its structure is shown in the attached figure. Figure 6 , in the figure X 1— X 10 Indicates the input data type.

[0101] Specifically, the attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, including an additive attention scoring mechanism:

[0102] In the formula, It is the attention scoring mechanism after the hidden layer undergoes a full connection operation. u s , w , b are the randomly initialized time series, attention weight matrix and bias matrix, respectively. for i Characteristics j Hidden layer states of group data.

[0103] Specifically, the attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, which also includes an attention distribution mechanism:

[0104] In the formula, for i Characteristics j The degree of attention that the group data receives for the target prediction quantity.

[0105] The attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, which also includes an information weighted summation mechanism:

[0106] In the formula, is the score vector after weighted summation, for i Characteristics j The degree of attention paid to the target prediction quantity by the group data, for i Characteristics j Hidden layer states of group data.

[0107] In a specific implementation, the construction process of the BIGRU model includes: Use the forward GRU network and the reverse GRU network to build the initial BIGRU model; The multiverse algorithm is used to optimize the hyperparameters of the initial BIGRU model to obtain the target BIGRU model, wherein the hyperparameters of the initial BIGRU model include learning rate, number of iterations, number of forward hidden layer nodes, and number of backward hidden layer nodes.

[0108] In some embodiments, the structure diagram of GRU is as shown in the attached Figure 7 GRU replaces the forget gate and input gate of the LSTM network with an update gate and merges the forget gate and input gate. Figure 7 middle, x t is the current moment (i.e.t time) input; h t-1 and h t They are the state outputs of GRU at the previous moment (i.e., moment t-1) and the current moment respectively; is the candidate state at the current moment; r t and z t They are the outputs of the reset gate and update gate at the current moment. The output calculation formula of GRU is as follows.

[0109]

[0110]

[0111] Where: W r and b r Reset the weight and bias of the gate respectively; W z and b z They are the weight and bias of the update gate respectively; W d and b d are the weight and bias of the gated unit respectively; σ(·) and tanh(·) are the sigmoid function and the hyperbolic tangent activation function respectively.

[0112] BIGRU is an improvement on GRU. It contains two GRUs with opposite directions (i.e., forward GRU and reverse GRU). The two GRUs do not interfere with each other. They both serve as input and control the output together. Figure 8 Compared with GRU, BIGRU can mine the information contained in the monitoring data to a greater extent, so the algorithm has the characteristics of few parameters and strong optimization ability. The BiGRU network is constructed by the forward GRU network and the reverse GRU network. These two layers of GRU networks are responsible for capturing historical information and future information respectively. Finally, the outputs of the two layers of the network are integrated according to their respective positions, which can improve memory ability and prediction accuracy. The hyperparameter optimization of the BIGRU model mainly includes four parameters: learning rate (LR), number of iterations (NI), number of nodes in the forward hidden layer (NF), and number of nodes in the backward hidden layer (NB).

[0113] This application introduces the multiverse algorithm (MVO) to optimize the hyperparameters of the BIGRU model. The MVO algorithm is a meta-inspiration algorithm developed based on the laws of physical motion between universes, and has good global optimization capabilities. In the cyclic model of the multiverse theory, some physicists believe that white holes are produced where collisions occur between parallel universes. White holes repel all objects, including light beams, with extremely high repulsive forces. The behavior of frequently observed black holes is completely opposite to that of white holes. Black holes attract all objects, including light beams, with extremely high attraction. Wormholes connect different universes together and play the role of time and space travel tunnels in algorithmic theory, allowing objects to instantly reach another universe from one universe. The MVO algorithm will be introduced in detail below.

[0114] The MVO algorithm assumes that each set of candidate solutions to the problem is a universe, each parameter in the candidate solution is an object in the universe, and the jth parameter of the i-th universe is ,in represents the minimum value of the jth parameter, represents the maximum value of the jth parameter. Assume that there are I universes, each containing J parameters, and each parameter in the universe is initialized according to a random number:

[0115] in, represents the i-th universe, represents the jth parameter of the i-th initialized universe.

[0116] The MVO algorithm believes that each universe has an expansion rate that characterizes the degree of expansion of the universe. The expansion rate is proportional to the corresponding fitness function value of the solution. The expansion rate determines the formation of black holes, white holes, and wormholes.

[0117] The expansion rate of each universe According to the objective function f Get, as shown below:

[0118] Then each cosmic expansion rate is normalized to obtain the standardized cosmic expansion rate , as shown in the following formula:

[0119] The MVO algorithm is divided into an exploration phase and a development phase. The exploration phase first uses the standardized expansion rate of the universe to Calculate the probability of each universe producing a white hole , as shown below:

[0120] Then calculate the cumulative probability of each universe producing a white hole , as shown below:

[0121] In the exploration phase, the parameters of black holes and white holes are updated according to the roulette mechanism as shown in the formula:

[0122] in, represents the jth parameter of the kth universe that produces a white hole, Represents a random number that follows a uniform distribution and is between [0, 1].

[0123] In the development phase, wormholes are used to perform refined local searches. The parameter update based on wormholes in the development phase is shown in the formula:

[0124] in, represents the jth parameter of the current optimal universe, D represents the travel distance rate, and E represents the probability of the existence of a wormhole. , , Represents a random number that follows a uniform distribution and is between [0, 1].

[0125] There are two important coefficients in the development stage, namely the probability of wormhole existence E and the travel distance rate D. The probability of wormhole existence is shown in the formula:

[0126] in, Indicates the minimum probability of the existence of a wormhole (the MVO algorithm takes 0.2), It represents the maximum value of the maximum existence probability of the wormhole (the MVO algorithm takes 1), q represents the current number of iterations, and Q represents the maximum number of iterations.

[0127] The distance traveled rate is shown in the formula:

[0128] h is an adjustment parameter of D. The higher h is, the faster the local search is. In this embodiment, h is set to 8.

[0129] In the MVO algorithm, the optimization process starts with initializing I random universes. Then in each iteration, objects in the high expansion rate universe will always tend to move toward the low expansion rate universe through black hole and white hole tunnels, and each universe may be randomly teleported to the current optimal universe through wormholes. This process will continue to iterate until the final iteration stop condition is met, which is generally a predefined maximum number of iterations.

[0130] For example, the traditional MVO algorithm uses a random generation method to generate the initial population, and the distribution uniformity of particles is poor. Fig. 9 As shown, the present application adds Circle chaos mapping to generate a chaotic sequence during population initialization. Using a chaotic sequence to initialize the universe improves the optimization ability and search efficiency of the MVO algorithm. The process of optimizing the hyperparameters of the initial BIGRU model using the multiverse algorithm includes the following steps: S1: Circle chaos mapping is added during the initialization process to generate a chaotic sequence to initialize the universe through the chaotic sequence.

[0131] For example, the mathematical representation of the Circle chaos map is as follows:

[0132] Among them, mod represents the modulus operation, Represents the generated chaotic sequence corresponding to the jth parameter of the ith universe.

[0133] Compared with random numbers, Circle mapping can generate uniformly distributed chaotic sequences. It avoids the possibility of generating blank spaces. Using Circle mapping to initialize the universe can provide a high-quality search space for the algorithm, which is conducive to improving the convergence accuracy of the algorithm.

[0134] S2: Calculate the expansion rate of each universe, and introduce an adaptive mechanism to adjust the wormhole existence probability E and the travel distance rate D parameters to update the universe expansion rate.

[0135] Exemplarily, based on the above analysis of the MVO algorithm, it can be seen that in this algorithm, adjusting the two important parameters of the wormhole existence probability E and the travel distance rate D can play a key role in balancing the two stages of development and search. The imbalance of the traditional multiverse optimization algorithm in the two stages will cause the algorithm to fall into a local optimal solution or even fail. In order to ensure the effectiveness of the algorithm, this application adopts an adaptive mechanism to adjust E and D on the basis of the conventional algorithm, so that the values ​​of these two parameters are automatically adjusted as the fitness and the number of iterations change, so that the optimization effect of the algorithm is better. The mathematical expression of the adaptive adjustment mechanism of E is as follows:

[0136] In the formula Represents the probability of wormhole existence in each generation of the universe; , and They represent the average, maximum and minimum values ​​of the fitness function of each generation of the universe during the iteration process; Represents the fitness function value of the i-th generation universe object.

[0137] The adaptive adjustment mechanism for D can be expressed as follows:

[0138] Specifically, an adaptive mechanism is introduced to adjust the wormhole existence probability E and the travel distance rate D parameters, and the formula for updating the universe expansion rate is:

[0139] In the formula, represents the jth parameter of the current optimal universe, , , represents a random number that follows a uniform distribution and is between [0, 1]. represents the probability of wormhole existence in each generation of the universe, D represents the travel distance rate, represents the minimum value of the jth parameter, represents the maximum value of the jth parameter, represents the jth parameter of the i-th initialized universe.

[0140] S3: Construct a fitness function, calculate the fitness, and determine whether the maximum number of iterations has been reached. If so, obtain the optimized hyperparameters.

[0141] Specifically, this application uses the IMVO algorithm to optimize the four key parameters of the BIGRU model, namely, the learning rate (LR), the number of iterations (NI), the number of forward hidden layer nodes (NF), and the number of backward hidden layer nodes (NB). The fitness function formula is:

[0142] In the formula, is the number of correctly classified samples; is the total number of samples.

[0143] In one embodiment, Fig.10 This is a detailed flow chart of a method for evaluating the status of an on-load tap changer of a transformer provided according to an embodiment of the present invention. This embodiment is further optimized and expanded on the basis of the above embodiments.

[0144] Step 1: Extract features from oil chromatography data. The five characteristic gases are taken as the main research objects. The volume fraction content of different characteristic gases in the on-load tap changer, the fault point temperature and the acetylene growth rate are used to form the oil chromatography data characteristic matrix (Xoil) to comprehensively judge the state of the tap changer.

[0145] Step 2: Extraction of vibration signal features of on-load tap changer. Multiscale fluctuation distribution entropy (MFDE) is introduced to characterize the complexity of time series at different scales, so as to enhance the fault characteristics of vibration signals. At the same time, three types of traditional time-frequency domain eigenvalues, energy entropy, kurtosis, and spectral kurtosis are calculated to enhance the fault characterization capability and form a vibration signal feature matrix.

[0146] Step 3: Extraction of the feature of the on-load tap changer drive motor current signal. First, the duration Tc of the drive motor current is taken as the first feature of the drive motor current signal. Secondly, according to the change process of the on-load tap changer, the drive motor current signal is divided into three stages, and the sum of the absolute values ​​of the drive motor current in different stages is taken as the new feature of the drive motor current signal. Thus, a drive motor current signal feature matrix is ​​formed.

[0147] Step 4: Data normalization. Perform Min-Max normalization on the obtained data samples.

[0148] Step 5: Attention mechanism correlation calculation. The attention mechanism (AM) is used to calculate the correlation between data and assign different weights to different feature data, thereby replacing the original data and obtaining data samples with specific weights.

[0149] Step 6: Optimize the parameters related to the BIGRU network model. IMVO is used to optimize the key parameters that affect the accuracy of BIGRU fault diagnosis, such as learning rate, number of iterations, number of nodes in the forward hidden layer, and number of nodes in the backward hidden layer, to obtain the optimal parameters of the BIGRU diagnosis model.

[0150] Step 7: Model training and on-load tap changer status assessment. The data samples with specific weights are input into the BIGRU model, and the BIGRU network model under the optimal parameter group is used to train and analyze the data, and finally the on-load tap changer status assessment based on multi-dimensional information parameters is realized, and multi-dimensional information fusion diagnosis is realized.

[0151] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0152] like Fig.11 As shown, the following is an embodiment of the transformer on-load tap changer status assessment system provided by the embodiment of the present disclosure, which belongs to the same inventive concept as the transformer on-load tap changer status assessment method of the above-mentioned embodiments. For details not fully described in the embodiment of the new energy power supply photovoltaic panel power generation and storage data management and control system, reference can be made to the embodiment of the transformer on-load tap changer status assessment method.

[0153] An acquisition unit 110 is used to acquire electrical performance parameters and mechanical performance parameters of the transformer on-load tap changer, wherein the electrical performance parameters include oil chromatogram data, and the mechanical performance parameters include vibration signals and drive motor current signals; The first generating unit 111 is used to generate an oil chromatogram data feature matrix for the oil chromatogram data, wherein the oil chromatogram data feature matrix includes the volume fraction of gas generated by decomposition of insulating oil of the on-load tap changer under abnormal conditions, the total hydrocarbon value, the temperature of the overheating fault point of the on-load decomposition switch, the temperature of the discharge fault point, and the acetylene growth rate; The second generating unit 112 is used to extract the vibration signal characteristics of the hydropower unit by using multi-scale wave propagation entropy for the vibration signal, and introduce energy entropy, kurtosis and spectral kurtosis to generate a vibration signal characteristic matrix; A third generating unit 113 is used to generate a driving motor current signal characteristic matrix for the driving motor current signal, wherein the driving motor current signal characteristic matrix includes the duration of the driving motor current and the sum of the absolute values ​​of the driving motor current in the motor starting stage, the motor stable operation stage and the motor stop operation stage; A fusion unit 114 is used to fuse the oil chromatogram data feature matrix, the vibration signal feature matrix and the drive motor current signal feature matrix to generate a multi-dimensional information feature matrix; The evaluation unit 115 is used to calculate the correlation between the data in the multidimensional information feature matrix by using the attention mechanism and assign weights to the multidimensional information feature matrix, obtain data samples with weights, and input the data samples with weights into a pre-built BIGRU model to realize the state evaluation of the transformer on-load tap changer.

[0154] Fig.12 It is a schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0155] The state assessment method of the transformer on-load tap changer provided in the embodiment of the present application can be applied to electronic devices. Those skilled in the art will appreciate that the electronic device structure involved in the embodiment of the present invention does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown, or combine certain components, or arrange different components. In the embodiment of the present invention, the electronic device includes but is not limited to a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described and / or required herein.

[0156] The electronic device may include a processor, an external memory interface, an internal memory, a universal serial bus (USB) interface, a charging management module, a power management module, a battery, a wireless communication module, an audio module, a speaker, a microphone, a sensor module, buttons, a camera, a display, and a SIM card interface, etc.

[0157] It is to be understood that the structure illustrated in the embodiments of the present application does not constitute a specific limitation on the electronic device. In other embodiments of the present application, the electronic device may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0158] The processor may include one or more processing units, for example, the processor may include a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU), etc. Among them, different processing units may be independent devices or integrated into one or more processors.

[0159] The processor can be the nerve center and command center of the electronic device. The controller can generate an operation control signal according to the instruction operation code and timing signal to complete the control of fetching and executing instructions.

[0160] A memory may also be provided in the processor for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. The memory may store instructions or data that the processor has just used or is cyclically used. If the processor needs to use the instruction or data again, it may be directly called from the memory. This avoids repeated access, reduces the waiting time of the processor, and thus improves system efficiency.

[0161] The external memory interface can be used to connect an external memory card, such as a MicroSD card, to expand the storage capacity of the electronic device. The external memory card communicates with the processor through the external memory interface to implement data storage functions. For example, files such as music and videos can be saved in the external memory card.

[0162] The internal memory can be used to store computer executable program codes, which include instructions. The processor executes various functional applications and data processing of the electronic device by running the instructions stored in the internal memory. The internal memory may include a program storage area and a data storage area. The internal memory may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, a universal flash storage (UFS), etc.

[0163] The wireless communication function of an electronic device can be realized through an antenna, a wireless communication module, a modem processor, and a baseband processor.

[0164] The wireless communication module can provide wireless communication solutions for electronic devices, including wireless local area networks (WLAN) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication (NFC), infrared technology (IR), etc.

[0165] Electronic devices can implement audio functions, etc. through audio modules, speakers, receivers, microphones, headphone jacks, and application processors.

[0166] Electronic devices can achieve shooting functions through ISP, camera, video codec, GPU, display and application processor.

[0167] Electronic devices can achieve display functions through GPU, display screen and application processor.

[0168] The GPU is a microprocessor for image processing that connects the display screen and the application processor. The GPU is used to perform mathematical and geometric calculations for graphics rendering. The processor may include one or more GPUs that execute program instructions to generate or change display information.

[0169] The display screen is used to display images, videos, etc. The display screen includes a display panel.

[0170] The storage medium provided in the present application stores a program product capable of implementing a method for evaluating the state of an on-load tap changer of a transformer.

[0171] In some possible implementations, the subject matter of the present disclosure, namely, method and system for assessing the condition of an on-load tap changer of a transformer, can be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of various exemplary implementations of the present disclosure described in the above “Exemplary Method” section of this specification.

[0172] The storage medium of the present disclosure can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0173] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for evaluating the state of a transformer on-load tap changer, characterized in that: include: Acquire electrical performance parameters and mechanical performance parameters of the transformer on-load tap changer, wherein the electrical performance parameters include oil chromatogram data, and the mechanical performance parameters include vibration signals and drive motor current signals; Generate an oil chromatogram data feature matrix for the oil chromatogram data, wherein the oil chromatogram data feature matrix includes the volume fraction of gas generated by decomposition of insulating oil of the on-load tap changer under abnormal conditions, the total hydrocarbon value, the temperature of the on-load decomposition switch overheating fault point, the temperature of the discharge fault point, and the acetylene growth rate; For the vibration signal, multi-scale wave propagation entropy is used to extract the vibration signal characteristics of the hydropower unit, and energy entropy, kurtosis and spectral kurtosis are introduced to generate a vibration signal feature matrix; Generate a driving motor current signal characteristic matrix for the driving motor current signal, wherein the driving motor current signal characteristic matrix includes the duration of the driving motor current and the sum of the absolute values ​​of the driving motor current in the motor starting stage, the motor stable operation stage and the motor stop operation stage; Fusion of the oil chromatogram data feature matrix, the vibration signal feature matrix and the drive motor current signal feature matrix to generate a multi-dimensional information feature matrix; The attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, which are then input into a pre-built BIGRU model to realize the state evaluation of the transformer on-load tap changer.

2. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: Under abnormal conditions, the gases produced by the decomposition of the insulating oil of the on-load tap-changer include at least hydrogen, methane, ethane, ethylene and acetylene.

3. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: The temperature of the on-load switch overheating fault point is calculated according to the following formula: Where, T G is the overheating fault point temperature, is the volume fraction of ethylene, is the volume fraction of ethane.

4. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: The temperature of the discharge fault point is calculated according to the following formula: In the formula, is the discharge fault point temperature, is the hydrogen gas volume fraction, is the volume fraction of acetylene.

5. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: The acetylene growth rate was calculated according to the following formula: In the formula, is the acetylene growth rate, is the change in acetylene volume fraction, is the time interval.

6. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: The calculation process of multi-scale fluctuation spread entropy includes: Map the coarse-grained time series and perform linear transformation on the mapping result to obtain Z; Calculate embedding vectors according to embedding dimension and time delay, and map each embedding vector to a fluctuation-based distribution pattern; Calculate the probability of each dispersion model and calculate the dispersion entropy based on fluctuations based on the Shannon entropy calculation method; The fluctuation-based spread entropy under the scale factor is calculated to obtain the multi-scale fluctuation spread entropy function.

7. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: Energy entropy is calculated according to the following formula: In the formula, represents the vibration amplitude, Represents the ratio of the current position energy to the total energy, Represents a time series.

8. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: Kurtosis is calculated according to the following formula: In the formula, N is the data sample size, is the i-th data sample, is the sample mean.

9. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: The spectral kurtosis is calculated according to the following formula: ； ； In the formula, Indicates frequency, represents the power spectral density of the current frequency, represents the average power spectral density.

10. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: The attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, including the additive attention scoring mechanism: In the formula, It is the attention scoring mechanism after the hidden layer undergoes a full connection operation. u s , w , b are the randomly initialized time series, attention weight matrix and bias matrix, respectively. for i Characteristics j Hidden layer states of group data.

11. The method for evaluating the status of the transformer on-load tap changer according to claim 10, characterized in that: The attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, which also includes the attention distribution mechanism: In the formula, for i Characteristics j The degree of attention that the group data receives for the target prediction quantity.

12. The method for evaluating the status of the transformer on-load tap changer according to claim 11, characterized in that: The attention mechanism is used to calculate the correlation between data in the multidimensional information feature matrix and assign weights to the multidimensional information feature matrix to obtain weighted data samples, which also includes an information weighted summation mechanism: In the formula, is the score vector after weighted summation, for i Characteristics j The degree of attention paid to the target prediction quantity by the group data, for i Characteristics j Hidden layer states of group data.

13. The method for evaluating the status of a transformer on-load tap changer according to claim 1, characterized in that: The construction process of the BIGRU model includes: Use the forward GRU network and the reverse GRU network to build the initial BIGRU model; The multiverse algorithm is used to optimize the hyperparameters of the initial BIGRU model to obtain the target BIGRU model, wherein the hyperparameters of the initial BIGRU model include learning rate, number of iterations, number of forward hidden layer nodes, and number of backward hidden layer nodes.

14. The method for evaluating the status of an on-load tap changer of a transformer according to claim 13, characterized in that: The process of optimizing the hyperparameters of the initial BIGRU model using the multiverse algorithm includes: In the population initialization process, Circle chaos mapping is added to generate a chaotic sequence to initialize the universe through the chaotic sequence; Calculate the expansion rate of each universe, and introduce an adaptive mechanism to adjust the wormhole existence probability E and the travel distance rate D parameters to update the universe expansion rate; Construct a fitness function, calculate the fitness, and determine whether the maximum number of iterations has been reached. If so, obtain the optimized hyperparameters.

15. The method for evaluating the status of the transformer on-load tap changer according to claim 14, characterized in that: An adaptive mechanism is introduced to adjust the wormhole existence probability E and the travel distance rate D parameters, and the formula for updating the universe expansion rate is: In the formula, represents the jth parameter of the current optimal universe, , , represents a random number that follows a uniform distribution and is between [0, 1]. represents the probability of wormhole existence in each generation of the universe, D represents the travel distance rate, represents the minimum value of the jth parameter, represents the maximum value of the jth parameter, represents the jth parameter of the i-th initialized universe.

16. The method for evaluating the status of an on-load tap changer of a transformer according to claim 14, characterized in that: The fitness function formula is: In the formula, is the number of correctly classified samples; is the total number of samples.

17. A transformer on-load tap changer status assessment system, characterized in that: include: An acquisition unit, used to acquire electrical performance parameters and mechanical performance parameters of the transformer on-load tap changer, wherein the electrical performance parameters include oil chromatogram data, and the mechanical performance parameters include vibration signals and drive motor current signals; A first generating unit is used to generate an oil chromatogram data feature matrix for the oil chromatogram data, wherein the oil chromatogram data feature matrix includes the volume fraction of gas generated by decomposition of insulating oil of the on-load tap changer under abnormal conditions, the total hydrocarbon value, the temperature of the overheating fault point of the on-load decomposition switch, the temperature of the discharge fault point, and the acetylene growth rate; The second generating unit is used to extract the vibration signal characteristics of the hydropower unit by using multi-scale wave propagation entropy for the vibration signal, and introduce energy entropy, kurtosis and spectral kurtosis to generate a vibration signal characteristic matrix; a third generating unit, configured to generate a driving motor current signal characteristic matrix for the driving motor current signal, wherein the driving motor current signal characteristic matrix includes a duration of the driving motor current and a sum of absolute values ​​of the driving motor current in a motor startup phase, a motor stable operation phase, and a motor stop operation phase; A fusion unit, used for fusing the oil chromatogram data feature matrix, the vibration signal feature matrix and the drive motor current signal feature matrix to generate a multi-dimensional information feature matrix; An evaluation unit is used to calculate the correlation between data in a multidimensional information feature matrix using an attention mechanism and assign weights to the multidimensional information feature matrix, obtain data samples with weights, and input the data samples with weights into a pre-built BIGRU model to achieve status evaluation of the transformer on-load tap changer.

18. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method for evaluating the condition of an on-load tap changer of a transformer according to any one of claims 1 to 16 are implemented.

19. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for assessing the condition of an on-load tap changer of a transformer according to any one of claims 1 to 16 are implemented.

Images