A transformer fault diagnosis method, device, equipment and medium

Abstract

The application discloses a transformer fault diagnosis method, device and equipment and a medium, relates to the technical field of transformer fault diagnosis, and adopts a terahertz time domain spectrum technology to acquire terahertz spectrum information of transformer fault insulating oil, decomposes terahertz spectrum sequence data by using a complementary ensemble empirical mode decomposition (CEEMD), calculates the fuzzy entropy of each subsequence after decomposition, so as to form characteristic quantities reflecting the transformer fault state. This process only extracts and processes the terahertz spectrum information of the transformer insulating oil, ignores the influence of the insulating oil, the internal and external environment and the transformer equipment itself, and then a new transformer fault diagnosis model is established according to a regularized extreme learning machine (RELM) and a hill climbing optimization method (HOA). The fault type of the transformer is diagnosed according to the transformer fault diagnosis model, and the accuracy of the transformer fault diagnosis is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of transformer fault diagnosis technology, and in particular to a transformer fault diagnosis method, apparatus, equipment and medium. Background Technology

[0002] As one of the important electrical devices in the power system, the transformer undertakes the important tasks of power transmission and voltage transformation. It is a static electrical device used to convert an AC voltage of a certain value into another or several voltages of different values ​​at the same frequency. It usually has two or more windings. In order to transmit electrical energy, at the same frequency, it converts the AC voltage and current of one system into the voltage and current of another system through electromagnetic induction. Usually, the values ​​of these currents and voltages are different.

[0003] Transformers serve multiple functions, not only raising voltage to deliver electricity to user areas but also lowering voltage to various operating levels to meet electricity needs. A transformer malfunction can cause localized or even widespread power outages, inevitably resulting in significant economic losses. Common transformer faults include oil leaks at welded joints, seals, flange connections, bolts or pipe threads, cast iron components, radiators, and porcelain insulators and glass oil gauges. Therefore, establishing efficient and accurate transformer fault diagnosis methods and understanding their operational status are crucial for improving transformer performance and ensuring the safe and reliable operation of the power system.

[0004] Currently, commonly used transformer fault diagnosis methods include the IEC ratio method, Rogers ratio method, and David's triangle method, as well as intelligent diagnostic methods based on the characteristics of dissolved gases in oil. These methods all rely on chromatography to obtain data on dissolved gases in oil and use this data to determine the type of transformer fault. However, the gas content data is easily affected by the insulating oil in the transformer, the internal and external environment, and the transformer equipment itself, making it difficult for the measured gas content data to reflect the true condition of the transformer. Therefore, the accuracy of transformer fault diagnosis is low. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and medium for diagnosing transformer faults. It can solve the problem that existing methods are easily affected by the insulating oil in the transformer, the internal and external environment, and the transformer equipment itself, making it difficult for the measured gas content data to reflect the true state of the transformer, thus resulting in low accuracy in diagnosing transformer faults.

[0006] This invention provides a transformer fault diagnosis method, comprising the following steps:

[0007] Terahertz time-domain spectral signal data of transformer insulating oil were collected and formed into time-domain sequence data;

[0008] The time-domain sequence data is decomposed using complementary set empirical mode decomposition (CEEMD) to obtain intrinsic mode components (IMFs) subsequences with different center frequencies. The fuzzy entropy (FE) method is used to obtain the fuzzy entropy value of each IMF subsequence, and the fuzzy entropy values ​​of each IMF subsequence are combined into multiple feature column vectors.

[0009] A transformer fault diagnosis model based on Regularized Extreme Learning Machine (RELM) is constructed. The transformer fault diagnosis model is trained using multiple feature column vectors. During the training process, the HOA (Hike-Based Optimization) method is used to simulate constraints on the transformer fault diagnosis model in order to update the model parameters and obtain the trained transformer fault diagnosis model.

[0010] The feature column vector of the transformer to be tested is input into the trained transformer fault diagnosis model to obtain the fault type of the transformer after diagnosis.

[0011] Preferably, obtaining the intrinsic mode components (IMFs) subsequences with different center frequencies includes:

[0012] Different noise instances are added to the original time-domain sequence data s(t) to obtain a new set of time-domain sequence data. The equation for obtaining this new set of data is as follows:

[0013] s i (t)=s(t)+noise i

[0014] Where: s(t) represents the original signal; s i (t) represents a signal with added noise; noise i Represents an instance of random noise;

[0015] For each signal s with added noise i (t) Empirical Mode Decomposition (EMD) is performed to obtain a set of intrinsic mode components, whose decomposition equations are as follows:

[0016]

[0017] Among them: IMF i，k (t) represents the k-th eigenmode function of the i-th signal; N represents the number of eigenmode functions; residue i (t) represents the residual of the i-th signal;

[0018] After performing Empirical Mode Decomposition (EMD), a set of intrinsic mode components is obtained, and then ensemble averaging is performed. The ensemble averaging equation is as follows:

[0019]

[0020] in: Let represent the final k-th intrinsic mode function; M represents the number of noise instances; represents the k-th mode of the i-th noise instance;

[0021] From the set after ensemble averaging, select the intrinsic mode components (IMFs) that have high consistency across different noise instances, and use these IMFs as the IMF subsequences with different center frequencies.

[0022] Preferably, the step of obtaining the fuzzy entropy value of each intrinsic mode component (IMF) subsequence using the fuzzy entropy (FE) method includes:

[0023] Obtaining the subsequence of intrinsic mode components (IMFs) and The similarity is calculated using the following equation:

[0024]

[0025] Where: n represents the exponent of the fuzzy function; r represents the similarity tolerance threshold; Represents a sequence and Chebyshev distance; This represents a sequence of length m;

[0026] The average similarity of each IMF subsequence is obtained based on the similarity between each intrinsic mode component (IMF) subsequence. The equation for obtaining this average similarity is as follows:

[0027]

[0028] Where: N represents the total length of the sequence;

[0029] The fuzzy entropy value of each IMFs subsequence is obtained based on each intrinsic mode component (IMFs) subsequence, and the equation for obtaining it is as follows:

[0030] FuzzyEn(m,n,r)=lnφ m,n,r -lnφ m+1,r,n

[0031] Where m+1 represents the dimension.

[0032] Preferably, the model parameters for updating the transformer fault diagnosis model include:

[0033] When training the transformer fault diagnosis model, the parameters of the Regularized Extreme Learning Machine (RELM) and the HOA (Hike Optimization) method are first set. The parameters include population size, maximum number of iterations, number of hidden layer nodes, and regularization parameters.

[0034] Based on the step size and fitness threshold formed by the hiking optimization method HOA in continuously iterating and optimizing the speed and position of hikers within the Tobler hiking function, the parameters of the regularized extreme learning machine RELM are continuously updated; and the input layer weights and hidden layer biases obtained at the end of the iteration are set as the input parameters of the regularized extreme learning machine RELM to complete the optimization and update of the model parameters of the transformer fault diagnosis model.

[0035] Preferably, the iterative optimization of the hiker's speed and position within the Tobler hiking function includes:

[0036] The Tobler hiking function is expressed as:

[0037]

[0038] Where: V i，t S represents the speed of hiker i at iteration or time t; i，t Indicates the slope of the terrain;

[0039] The hiker's speed within the Tobler hiking function is represented as:

[0040] V i,t =V i,t-1 +γ i,t (β best -α i,t β i,t )

[0041] Where: V i，t V represents the current speed. i，t-1 Indicates the initial velocity; γ i,t β represents a random number in the range [0,1]; best Indicates the position of the lead hiker; α i，t The sweep factor SF of the hiker is in the range [1,3].

[0042] The hiker's position within the Tobler hiking function is represented as follows:

[0043] β i,t+1 =β i,t +V i,t

[0044] Where: β i,t+1 Indicates the current position;

[0045] Initialize the positions of the hikers, with each hiker corresponding to a result that includes the input layer weights and hidden layer biases;

[0046] The walking speed of each hiker is obtained according to the Tobler hiking function, and the real-time position of each hiker is updated. At the same time, the current fitness and terrain are updated to adjust the stride length.

[0047] The walking speed and real-time location of each hiker are continuously iterated and optimized to obtain the optimal fitness threshold and stride length.

[0048] This invention also provides a transformer fault diagnosis device, comprising:

[0049] The data acquisition module is used to acquire terahertz time-domain spectral signal data of transformer insulating oil and form time-domain sequence data;

[0050] The data processing module is used to decompose the time-domain sequence data using complementary set empirical mode decomposition (CEEMD) to obtain intrinsic mode components (IMFs) subsequences with different center frequencies; the fuzzy entropy (FE) method is used to obtain the fuzzy entropy value of each IMFs subsequence, and the fuzzy entropy value of each IMFs subsequence is used to form multiple feature column vectors.

[0051] The model module is used to construct a transformer fault diagnosis model based on Regularized Extreme Learning Machine (RELM). The transformer fault diagnosis model is trained using multiple feature column vectors. During the training process, the HOA (Hike-Based Optimization) method is used to simulate constraints on the transformer fault diagnosis model in order to update the model parameters and obtain the trained transformer fault diagnosis model.

[0052] The fault identification module is used to input the feature column vector of the transformer under test into the trained transformer fault diagnosis model to obtain the fault type of the transformer after diagnosis.

[0053] This invention also provides an electronic device, including a memory and a processor;

[0054] The memory is used to store computer programs;

[0055] When the processor executes the computer program stored in the memory, it implements the steps of the transformer fault diagnosis method described above.

[0056] This invention also provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the steps of a transformer fault diagnosis method as described above.

[0057] This invention provides a transformer fault diagnosis method, apparatus, equipment, and medium, which have the following advantages compared with the prior art:

[0058] This invention employs terahertz time-domain spectroscopy to obtain the terahertz spectral information of transformer fault insulating oil. It then utilizes Complementary Set Empirical Mode Decomposition (CEEMD) to decompose the terahertz spectral sequence data, calculating the fuzzy entropy of each subsequence to form characteristic quantities reflecting the transformer fault state. This process only extracts and processes the terahertz spectral information of the transformer insulating oil, i.e., only extracting characteristic quantities reflecting the transformer fault state, ignoring the influence of insulating oil, internal and external environment, and the transformer equipment itself on the terahertz spectral information. Subsequently, a novel transformer fault diagnosis model is established based on Regularized Extreme Learning Machine (RELM) and Hovering Optimization Method (HOA). The established model can efficiently search for more accurate fault information from the characteristic quantities. Finally, the transformer fault diagnosis model is used to diagnose the transformer fault type, significantly improving the accuracy of transformer fault diagnosis. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the overall process of a transformer fault diagnosis method provided in an embodiment of the present invention. Detailed Implementation

[0060] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0061] See Figure 1 This invention provides a transformer fault diagnosis method, comprising the following steps:

[0062] Step 1: Use a terahertz spectrometer to collect the terahertz time-domain spectral signal of the insulating oil under transformer fault conditions.

[0063] Step 2: Process the collected terahertz time-domain spectral signal of transformer fault insulating oil.

[0064] 1. First, the acquired terahertz time-domain spectral signal of transformer fault insulating oil is decomposed using complementary ensemble empirical mode decomposition (CIMD) to obtain multiple intrinsic mode function (IMF) components. The specific steps of CIMD include:

[0065] ① Add noise.

[0066] By adding different noise instances to the original signal s(t), a new set of signals is generated, as shown in the formula:

[0067] s i (t)=s(t)+noise i

[0068] Where: s(t) represents the original signal; s i (t) represents a signal with added noise; noise i This represents an instance of random noise.

[0069] ②EMD decomposition.

[0070] For each signal s with added noise i (t) Perform Empirical Mode Decomposition (EMD) to obtain a set of intrinsic mode functions, the formula of which is:

[0071]

[0072] Among them: IMF i，k (t) represents the k-th eigenmode function of the i-th signal; N represents the number of eigenmode functions; residue i (t) represents the residual of the i-th signal.

[0073] ③ The set average, its formula is:

[0074]

[0075] in: Let represent the final k-th intrinsic mode function; M represents the number of noise instances; represents the k-th mode of the i-th noise instance.

[0076] ④ Modal selection.

[0077] Select stable IMFs components that show high consistency across different noise instances from the set.

[0078] 2. Then, calculate the fuzzy entropy for each IMF component. Fuzzy entropy is an indicator that measures the complexity and uncertainty of a signal, and can effectively characterize the randomness and regularity of a signal. Its specific calculation steps include:

[0079] ① Sequence definition.

[0080] Divide the time series data into sequences of length m, denoted as... in And subtract its mean from it. The mean-removed sequence is obtained.

[0081] ② Definition of distance.

[0082] Define two sequences and Distance between Chebyshev distance is the absolute value of the maximum difference between corresponding elements in two sequences.

[0083] ③ Definition of similarity.

[0084] Introducing fuzzy membership To measure and The similarity is defined using an exponential function, and its formula is:

[0085]

[0086] Where: n represents the exponent of the fuzzy function; r represents the similarity tolerance threshold, which is usually taken as a certain proportion of the standard deviation of the sequence (such as 0.2 times the standard deviation).

[0087] ④ Calculation of average similarity.

[0088] Calculate the average similarity φ for each sequence m，r，n The calculation formula is as follows:

[0089]

[0090] Where: N represents the total length of the sequence.

[0091] ⑤ Fuzzy entropy calculation.

[0092] Increase the dimension to m+1 and repeat the above steps to calculate φ. m+1，r，n Then, the fuzzy entropy is calculated, and its formula is:

[0093] FuzzyEn(m,n,r)=lnφ m,n,r -lnφ m+1,r,n .

[0094] 3. Finally, the fuzzy entropy of each IMF component is used to form a column vector, which constitutes the feature sample of the transformer fault insulating oil terahertz spectral signal.

[0095] Step 3: Segmentation of terahertz time-domain spectral characteristics of transformer fault insulating oil.

[0096] The terahertz time-domain spectral characteristic samples of the preprocessed transformer fault insulating oil were divided into a training set of 70% and a test set of 30% using a random sampling method.

[0097] Step 4: Establish a transformer fault diagnosis model based on terahertz time-domain spectral characteristics.

[0098] The training set is used to train the transformer fault diagnosis model, establishing a transformer fault diagnosis model based on terahertz time-domain spectral characteristics. Then, the test set is used to test the diagnostic effect of the model to verify its performance.

[0099] This invention selects a regularized extreme learning machine (RELM) to establish a transformer fault diagnosis model and uses the hiking optimization algorithm (HOA) to determine the RELM model parameters.

[0100] 1. Regularized Extreme Learning Machine.

[0101] Regularized Extreme Learning Machine (RELM) is based on structural risk minimization. It introduces regularization into Extreme Learning Machine (ELM) and considers a term that controls the complexity of machine learning in the optimization objective function, which improves the problem of overfitting in ELM and enhances the generalization and robustness of the algorithm.

[0102] ELM is a single-hidden-layer feedforward neural network where the input layer weights and hidden layer biases are randomly generated. During execution, only the number of hidden layer nodes needs to be set, and the weight matrix between the hidden layer and the output layer is obtained through learning or training.

[0103] Given any N samples Where x i = [x1,x2,…,x n ] T ∈R n ,y i =[y1,y2,…,y m ] T ∈R m Then the number of input and output neurons in the model are n and m respectively. An ELM with S hidden layer nodes should, as far as possible, satisfy the following formula:

[0104]

[0105] Where: β i =[β1,β2,···,β n ] T The output weight vector between the output node and the i-th hidden layer node is represented by g(·); g(·) represents the activation function; w i =[w1,w2,…,w n ] T b represents the input weight vector between the input node and the i-th hidden layer node; i This represents the bias of the i-th hidden layer node.

[0106] set up The above formula can then be rewritten as: Hβ = Y, where Therefore, the learning objective function of ELM is:

[0107] min||Hβ-Y||

[0108] When g(·) is infinitely differentiable, the output parameters of the extreme learning machine do not need to be adjusted and can be obtained directly through calculation; therefore, the smallest and unique solution β satisfying the above equation can be calculated as:

[0109] β=H + Y

[0110] Wherein: H + Let H represent the Moore-Penrose generalized inverse of the hidden layer output matrix H.

[0111] RELM introduces a regularization term to control the norm of the output weight matrix (β), and its learning objective function is:

[0112]

[0113] Where: λ>0 indicates and Compromise parameters.

[0114] 2. Hiking optimization algorithm.

[0115] Since the hyperparameters of the RELM model have a significant impact on its training and learning performance, HOA is introduced to optimize the input layer weights and hidden layer biases of RELM to enhance its generalization ability. HOA is a metaheuristic optimization algorithm inspired by hiking experience. The mathematical model of HOA is based on the well-known Tobler's Hiking Function (THF), which determines the hiker's speed by considering the terrain elevation and walking distance. The specific implementation process of the HOA algorithm includes:

[0116] ①The Tobler Hiking Function (THF) has the following formula:

[0117]

[0118] Where: V i，t S represents the speed of hiker i at iteration or time t (unit: km / h); i，t Indicates the slope of the terrain.

[0119] ② The slope calculation formula is as follows:

[0120]

[0121] Where: dh represents the change in altitude; dt represents the distance covered by the hiker; θ i，t This represents the angle of inclination of the hiker at time t.

[0122] ③ The formula for updating hiker speed is:

[0123] V i,t =V i,t-1 +γ i,t (β best -α i,t β i,t )

[0124] Where: V i，t V represents the current speed. i，t-1 Indicates the initial velocity; γ i,t β represents a random number in the range [0,1]; best Indicates the position of the lead hiker; α i，t The sweep factor (SF) of the hiker is in the range [1,3].

[0125] ④ The formula for updating hiker locations is:

[0126] β i,t+1 =β i,t +V i,t

[0127] Where: β i,t+1 This indicates the updated position.

[0128] 3. The specific steps for optimizing RELM based on HOA are as follows:

[0129] ① Input data and initialization operation: Input preprocessed terahertz time-domain spectral data of transformer fault insulating oil.

[0130] ② Set the parameters for HOA and RELM, including: population size, maximum number of iterations, number of hidden layer nodes, regularization parameters, etc.

[0131] ③ Initialize the positions of the hikers, with each hiker corresponding to a potential solution (a combination of input layer weights and hidden layer biases).

[0132] ④ Calculate the fitness level for each hiker, i.e., the value of the objective function.

[0133] ⑤ Calculate the walking speed of each hiker according to the Tobler hiking function, update the position, and adjust the step length according to the current fitness and terrain (search space).

[0134] ⑥ Iterate continuously to find the optimal hiker position until the termination condition is met (reaching the maximum number of iterations or the fitness threshold).

[0135] ⑦ Use the input layer weights and hidden layer biases obtained after the iteration as input parameters of RELM to establish the optimal HOA-RELM model for transformer fault diagnosis.

[0136] Terahertz time-domain spectroscopy, as a novel, rapid, reliable, and non-destructive testing technology, possesses excellent coherence, extremely high signal-to-noise ratio, and good penetration. It has been widely applied in fields such as physical and chemical testing, non-destructive testing, materials science, and biomedicine. This invention uses terahertz time-domain spectroscopy to extract new characteristic quantities reflecting the fault state of transformers as input to the transformer fault diagnosis model, thereby achieving accurate diagnosis of transformer faults.

[0137] This invention employs terahertz time-domain spectroscopy to obtain terahertz spectral information of transformer fault insulating oil and proposes a novel HOA-RELM transformer fault diagnosis method. This method addresses the shortcomings of existing transformer fault diagnosis methods based on chromatography to obtain dissolved gases in oil. This method enables rapid, non-contact detection of transformer faults, greatly improving the accuracy and efficiency of fault diagnosis.

[0138] This invention also provides a transformer fault diagnosis device, comprising:

[0139] The data acquisition module is used to acquire terahertz time-domain spectral signal data of transformer insulating oil and form time-domain sequence data.

[0140] The data processing module is used to decompose the time-domain sequence data using complementary set empirical mode decomposition (CEEMD) to obtain intrinsic mode components (IMFs) subsequences with different center frequencies; the fuzzy entropy (FE) method is used to obtain the fuzzy entropy value of each IMFs subsequence, and the fuzzy entropy value of each IMFs subsequence is used to form multiple feature column vectors.

[0141] The model module is used to construct a transformer fault diagnosis model based on Regularized Extreme Learning Machine (RELM). The transformer fault diagnosis model is trained using multiple feature column vectors. During the training process, the HOA (Hike-Based Optimization) method is used to simulate constraints on the transformer fault diagnosis model in order to update the model parameters and obtain the trained transformer fault diagnosis model.

[0142] The fault identification module is used to input the feature column vector of the transformer under test into the trained transformer fault diagnosis model to obtain the fault type of the transformer after diagnosis.

[0143] This invention also provides an electronic device, including a memory and a processor.

[0144] Memory is used to store computer programs.

[0145] When the processor executes a computer program stored in memory, it implements the steps of the above-described transformer fault diagnosis method.

[0146] This invention also provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the steps of the transformer fault diagnosis method described above.

[0147] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A transformer fault diagnosis method characterized by, The method comprises the following steps: Collecting terahertz time-domain spectrum signal data of transformer insulating oil and forming time-domain sequence data; Decomposing the time-domain sequence data by using complementary ensemble empirical mode decomposition (CEEMD) to obtain intrinsic mode component (IMF) sub-sequences of different center frequencies; obtaining fuzzy entropy values of each intrinsic mode component (IMF) sub-sequence by using a fuzzy entropy (FE) method, and forming multiple feature column vectors from the fuzzy entropy values of each intrinsic mode component (IMF) sub-sequence; Constructing a transformer fault diagnosis model based on a regularized extreme learning machine (RELM); training the transformer fault diagnosis model by using the multiple feature column vectors, and simulating constraints on the transformer fault diagnosis model by using a hill climbing optimization algorithm (HOA) during the training process to update model parameters of the transformer fault diagnosis model, thereby obtaining a trained transformer fault diagnosis model; Inputting a feature column vector of a transformer to be tested into the trained transformer fault diagnosis model to obtain a fault type of the transformer after diagnosis.

2. The transformer fault diagnostic method of claim 1, wherein The intrinsic mode component (IMF) sub-sequences of different center frequencies are obtained by: Adding different noise instances to the original time-domain sequence data s(t) to obtain a group of new time-domain sequence data, and the equation for obtaining the new time-domain sequence data is: s i (t) = s(t) + noise i wherein: s(t) represents the original signal; s i (t) represents the signal with added noise; noise i represents a random noise instance; for each noise-added signal s i (t) performing empirical mode decomposition (EMD) to obtain a set of intrinsic mode components, and a decomposition equation thereof is: where: IMF i，k (t) represents the kth intrinsic mode function of the ith signal; N represents the number of intrinsic mode functions; residue i (t) represents the residue of the ith signal; Performing ensemble average on a group of intrinsic mode components obtained after empirical mode decomposition (EMD), and the ensemble average equation is: wherein: denotes the final kth eigenmodal function; M denotes the number of noise instances; denotes the kth mode of the ith noise instance; Selecting intrinsic mode components (IMFs) with high consistency in different noise instances from the ensemble average set as intrinsic mode components (IMFs) of different center frequencies.

3. The transformer fault diagnostic method of claim 1, wherein The fuzzy entropy values of each intrinsic mode component (IMF) sub-sequence are obtained by: obtaining the similarity of the intrinsic modal component IMF subsequence and The similarity obtaining equation is: where: n represents the exponent of the fuzzy function; r represents the similarity tolerance threshold; denotes a sequence and the Chebyshev distance of denotes a sequence of length m; According to the similarity between each intrinsic mode component (IMF) sub-sequence, obtaining the average similarity of each intrinsic mode component (IMF) sub-sequence, and the equation for obtaining the average similarity is: Wherein: N represents the total length of the sequence; According to each intrinsic mode component (IMF) sub-sequence, obtaining the fuzzy entropy value of each intrinsic mode component (IMF) sub-sequence, and the equation for obtaining the fuzzy entropy value is: FuzzyEn(m, n, r) = lnφ m,n,r - lnφ m+1,r,n Wherein: m+1 represents the dimension.

4. The transformer fault diagnostic method of claim 1, wherein The model parameters of the transformer fault diagnosis model are updated by: When training the transformer fault diagnosis model, first, setting parameters of the regularized extreme learning machine (RELM) and the hill climbing optimization algorithm (HOA), including population size, maximum number of iterations, number of hidden layer nodes, and regularization parameter; According to the step length and fitness threshold formed when the hill climbing optimization algorithm (HOA) continuously iteratively optimizes the speed and position of the walker in the Tobler walker function, the parameters of the regularized extreme learning machine (RELM) are continuously updated; and setting the input layer weight and hidden layer bias obtained when the iteration is completed as the input parameters of the regularized extreme learning machine (RELM), thereby completing the optimization and update of the model parameters of the transformer fault diagnosis model.

5. The transformer fault diagnostic method of claim 4, wherein, The speed and position of the walker in the Tobler walker function are iteratively optimized by: The Tobler walker function is represented as: where: V i，t represents the speed of the hiker i at iteration or time t; S i，t represents the slope of the terrain; The speed of the walker in the Tobler walker function is represented as: V i,t = V i,t-1 + γ i,t (β best - α i,t β i,t ) where: V i，t represents the current speed; V i，t-1 represents the initial speed; γ i,t represents a random number in the range [0, 1]; β best represents the position of the lead hiker; α i，t represents that the sweep factor SF of the hiker is in the range [1, 3]; The position of the walker in the Tobler walker function is represented as: β i,t+1 ＝β i,t +V i,t where: β i,t+1 represents the current position; Initialize the position of each walker, each walker corresponds to a result, the result contains the input layer weight and the hidden layer bias; According to the Tobler walker function, the walking speed of each walker is obtained, and the real-time position of each walker is updated, and the current fitness and the terrain are updated to adjust the step length; The walking speed and real-time position of each walker are continuously iterated and optimized to obtain the optimal fitness threshold and step length.

6. A transformer fault diagnostic device characterized by comprising: The method comprises the following steps: The data acquisition module is used for collecting the terahertz time domain spectrum signal data of the transformer insulating oil and forming time domain sequence data; The data processing module is used for decomposing the time domain sequence data by using the complementary ensemble empirical mode decomposition (CEEMD) to obtain intrinsic mode component (IMF) sub-sequences with different center frequencies; the fuzzy entropy (FE) method is used to obtain the fuzzy entropy value of each intrinsic mode component (IMF) sub-sequence, and the fuzzy entropy values of each intrinsic mode component (IMF) sub-sequence are combined to form multiple feature column vectors; The model module is used for constructing a transformer fault diagnosis model based on a regularized extreme learning machine (RELM); the multiple feature column vectors are used to train the transformer fault diagnosis model, and the hill climbing optimization algorithm (HOA) is used to simulate the constraint of the transformer fault diagnosis model during the training process to update the model parameters of the transformer fault diagnosis model, thereby obtaining the trained transformer fault diagnosis model; The fault recognition module is used for inputting the feature column vector of the transformer to be tested into the trained transformer fault diagnosis model to obtain the fault type of the diagnosed transformer.

7. An electronic device, comprising: The method comprises the following steps: A memory and a processor are provided. The memory is configured to store a computer program. The processor is configured to execute the computer program stored in the memory to implement the steps of the transformer fault diagnosis method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer program is stored in the memory and is executed by the processor to implement the steps of the transformer fault diagnosis method according to any one of claims 1-5.

Images