Voltage transformer reliability analysis method and storage medium

Abstract

The invention discloses a voltage transformer reliability analysis method, which comprises the following steps of: acquiring a plurality of preset reliability index data of a voltage transformer, and performing standardization processing to eliminate dimensional difference and generate a standardized data matrix; on the basis of the standardized data matrix, objective weights of the indexes are calculated through information entropy, and a comprehensive weight is generated through fusion in combination with subjective structure weights determined by expert experience; constructing a membership function for each reliability index, and describing the degree of the index data belonging to different grades in a preset reliability grade set; according to the membership function and the comprehensive weight, calculating the comprehensive membership degree of the voltage transformer to each reliability level, determining the current reliability level according to the comprehensive membership degree, and triggering a corresponding maintenance decision; according to the method, the problem of boundary fuzziness in a traditional classification method is solved, reliability level description better conforms to practical application, and the accuracy and reliability of evaluation are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power monitoring, and in particular to a voltage transformer reliability analysis method and a storage medium. Background Art

[0002] With the continuous improvement of industrial automation, online voltage transformer monitoring devices are becoming increasingly widely used in industrial systems, becoming a key component in ensuring the stable operation of power systems. However, the reliability of voltage transformers can be affected by various factors over the long term, such as battery life, hardware wear, communication error rates, and environmental adaptability. Therefore, accurately assessing the reliability of voltage transformers is crucial for ensuring the safety and stability of power systems.

[0003] Traditional reliability assessment methods often rely on fixed thresholds or simple statistical models, which fail to fully account for the multiple uncertainties in complex systems. This limits the accuracy and practicality of assessment results. This is especially true for indicators with fuzzy boundaries and dynamic characteristics, making it difficult for traditional methods to accurately classify and assess them. Summary of the Invention

[0004] The present invention addresses the problems of the prior art by providing a voltage transformer reliability analysis method and storage medium. This method addresses the fuzzy boundary issues in traditional classification methods, making the description of reliability levels more consistent with the fuzzy characteristics of practical applications, thereby significantly improving the accuracy and reliability of the assessment.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A voltage transformer reliability analysis method comprises the following steps:

[0007] S1. Collecting multiple preset reliability index data of a voltage transformer, and performing standardization processing on the data to eliminate dimensional differences and generate a standardized data matrix;

[0008] S2. Based on the standardized data matrix, the objective weight of each indicator is calculated by information entropy, and the subjective structural weight determined by expert experience is combined to generate the comprehensive weight of each indicator;

[0009] S3. Constructing a membership function for each reliability indicator to describe the degree to which the indicator data belongs to different levels in the preset reliability level set;

[0010] S4. Calculating the comprehensive membership of the voltage transformer to each reliability level according to the membership function and the comprehensive weight;

[0011] S5. Determine the current reliability level of the voltage transformer based on the comprehensive membership, and trigger a corresponding maintenance decision according to the level result.

[0012] Step S1 includes:

[0013] Collect m reliability index data of voltage transformer, the number of samples is n, and the original data matrix is:

[0014] X=[x ij ] n×m ;

[0015] Among them, x ij represents the original value of the jth indicator of the i-th sample, n represents the number of samples, and m represents the number of indicators;

[0016] According to the nature of the indicators, standardization is performed separately;

[0017] The standardized processing formula for positive indicators is:

[0018]

[0019] Among them, min(x j ) represents the minimum value of the jth indicator in all samples, max(x j ) represents the maximum value of the jth indicator in all samples;

[0020] The standardized processing formula for negative indicators is:

[0021]

[0022] Process the original data matrix X column by column, and finally get the standardized matrix Y, the formula is:

[0023] Y=[y ij ] n×m ;

[0024] y ij It represents the normalized value of the jth indicator of the i-th sample, and its range is limited to [0,100].

[0025] Step 2 includes:

[0026] Based on the standardized data matrix, the information entropy is calculated as follows:

[0027]

[0028] Among them, E j represents the information entropy of the jth indicator, ranging from [0,1];

[0029] Among them, p ijIt represents the proportion of the i-th sample in the j-th indicator. The formula is:

[0030]

[0031] If p ij =0, then p ij lnp ij =0;

[0032] The objective weight of each indicator is calculated by information entropy, and the formula is:

[0033]

[0034] Among them, w j represents the entropy weight of the jth indicator; (1-Ej) represents the difference coefficient of the indicator. The larger the value, the higher the degree of data dispersion.

[0035] Step S2 includes:

[0036] Experts score the importance of m indicators based on prior knowledge and obtain the subjective structural weight vector:

[0037] θ=[θ j ] 1×m ;

[0038] And must meet

[0039]

[0040] The objective weight and subjective structural weight are integrated to generate the comprehensive weight of each indicator. The formula is:

[0041] W j =αw j +(1-α)θ j , 0≤α≤1;

[0042] Among them, W j represents the comprehensive weight of the jth indicator, and α is the preset weight coefficient, which is used to adjust the balance between subjective and objective weights.

[0043] Step S3 includes:

[0044] The preset reliability level sets are:

[0045] V = {υ1,υ2,υ3,υ4} = {very reliable, reliable, unreliable, very unreliable};

[0046] Construct the membership function, and define each level function as follows:

[0047]

[0048] Step S4 includes:

[0049] For each indicator j, according to its standardized value y j , calculate its membership A to the four levels v1, v2, v3, v4 through the membership function j1 ,A j2 ,A j3 ,A j4 ;

[0050] For each level v k , the membership degree A of all indicators jk Multiply by its weight W j , sum to get the comprehensive membership B k , the formula is:

[0051]

[0052] B k Indicates the device is level v k The comprehensive membership degree of W is greater, and the greater the value, the more likely it is to belong to the level; j represents the comprehensive weight of the jth indicator; A jk Indicates the jth index for level v k The degree of membership;

[0053] The final comprehensive membership vector is expressed as:

[0054] B=[B1,B2,B3,B4].

[0055] Step S5 includes:

[0056] From the comprehensive membership vector B = [B1, B2, B3, B4], find the k value corresponding to the maximum value;

[0057] According to the k value, it is mapped to the corresponding reliability level v k .

[0058] When the level is v3 unreliable or v4 very unreliable, an early warning signal is automatically triggered.

[0059] Step S6: Obtain historical moments t1, t2, ..., t N The comprehensive membership vector B k (t) , where t is the timestamp;

[0060] The historical comprehensive membership B k (t) As a time series, input into the LSTM network;

[0061] LSTM network output predicts B in the future T period k (t+1) ,...,B k(t+T) ;

[0062] Define warning threshold B thk , when the predicted value satisfies B k (t+T) >B thk triggering early warnings.

[0063] In step S6, the structure of the LSTM network includes an input layer, an LSTM layer, and a fully connected layer;

[0064] The input layer is used to receive historical time series data as the input of the model; the input dimension is L×1, representing L consecutive time steps;

[0065] The first LSTM layer is configured with 64 units and returns the output of all time steps;

[0066] The second LSTM layer is configured with 32 units and returns the output of the last time step;

[0067] The dropout layer dropout rate is set to 0.2;

[0068] The number of units in the fully connected layer is configured as T, corresponding to the number of future time steps to be predicted, and the activation function is configured as Sigmoid; the output is the predicted value for the next T time period.

[0069] A computer-readable storage medium stores a computer program, wherein the computer program implements the above method steps when executed by a processor.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] 1. By combining the objective weights calculated by information entropy and the subjective structural weights determined by expert experience, the comprehensive weights of each indicator are generated. A membership function is constructed for each reliability indicator, converting the continuous indicator value into a discrete reliability grade membership. This solves the boundary fuzziness problem in traditional classification methods and makes the description of the reliability grade more consistent with the fuzzy characteristics in actual applications, thereby significantly improving the accuracy and reliability of the assessment.

[0072] 2. By collecting data from multiple preset reliability indicators, the reliability status of the voltage transformer is fully reflected, ensuring the comprehensiveness of the evaluation. Standardization processing eliminates the dimensional differences between different indicators and generates a unified standardized data matrix, providing a reliable data basis for subsequent analysis and enhancing the scientific nature and comparability of the evaluation results.

[0073] 3. The LSTM network is used to perform time series modeling on historical comprehensive membership vectors to predict comprehensive membership values ​​for future time periods, improving the accuracy and reliability of the prediction results. By defining warning thresholds and triggering early warnings based on the prediction results, timely measures can be taken before the equipment reliability condition deteriorates, providing forward-looking support for operation and maintenance decisions, enhancing the system's preventive maintenance capabilities, and thus improving the safety and stability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0075] Figure 1 This is a schematic diagram of the overall process of a voltage transformer reliability analysis method in an embodiment of the present application. DETAILED DESCRIPTION

[0076] The following will clearly and completely describe the technical solution of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0077] The size of the serial numbers of each step in the description of this application does not 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 embodiments of this application.

[0078] In the description of this application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish descriptions and should not be understood as indicating or implying relative importance. It should also be understood that although the terms "first", "second", etc. are used in the text to describe various elements in some embodiments of the present application, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first table can be named a second table, and similarly, a second table can be named a first table without departing from the scope of the various described embodiments. Both the first table and the second table are tables, but they are not the same table.

[0079] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0080] With the continuous improvement of industrial automation, online voltage transformer monitoring devices are becoming increasingly widely used in industrial systems, becoming a key component in ensuring the stable operation of power systems. However, the reliability of voltage transformers can be affected by various factors over the long term, such as battery life, hardware wear, communication error rates, and environmental adaptability. Therefore, accurately assessing the reliability of voltage transformers is crucial for ensuring the safety and stability of power systems.

[0081] Traditional reliability assessment methods often rely on fixed thresholds or simple statistical models, which fail to fully account for the multiple uncertainties in complex systems. This limits the accuracy and practicality of assessment results. This is especially true for indicators with fuzzy boundaries and dynamic characteristics, making it difficult for traditional methods to accurately classify and assess them.

[0082] In view of the above technical problems, the embodiment of the present application provides a voltage transformer reliability analysis method, including the following steps S1-S6. Figure 1 shown.

[0083] Step S1: collecting a plurality of preset reliability index data of a voltage transformer, and performing standardization processing on the data to eliminate dimensional differences and generate a standardized data matrix.

[0084] In some embodiments, step S1 includes the following steps S1.1-S1.3.

[0085] Step S1.1: Collect m reliability index data of voltage transformer, the number of samples is n, and the original data matrix is:

[0086] X=[x ij ] n×m ;

[0087] Among them, x ij represents the original value of the jth indicator of the i-th sample, n represents the number of samples, and m represents the number of indicators.

[0088] For example, reliability-related indicators include the device's battery life, hardware wear, communication bit error rate, environmental adaptability, firmware update frequency, data storage capacity, and data transmission rate.

[0089] Step S1.2: Standardize the indicators according to their properties;

[0090] The standardized processing formula for positive indicators is:

[0091]

[0092] Among them, min(x j ) represents the minimum value of the jth indicator in all samples, max(x j ) represents the maximum value of the jth indicator in all samples;

[0093] The standardized processing formula for negative indicators is:

[0094]

[0095] Step S1.3: Process the original data matrix X column by column to obtain the standardized matrix Y. The formula is:

[0096] Y=[y ij ] n×m ;

[0097] y ij It represents the normalized value of the jth indicator of the i-th sample, and its range is limited to [0,100].

[0098] The above steps S1.1-S1.3 can comprehensively reflect the reliability status of the voltage transformer by collecting multiple reliability index data; the standardization process eliminates the dimensional differences between different indicators and provides a unified data basis for subsequent analysis.

[0099] Step S2: Based on the standardized data matrix, the objective weight of each indicator is calculated by information entropy, and the subjective structural weight determined by expert experience is combined to generate the comprehensive weight of each indicator.

[0100] In some embodiments, step 2 includes the following steps S2.1-S2.4.

[0101] Step S2.1: Based on the standardized data matrix, calculate the information entropy using the formula:

[0102]

[0103] Among them, E j represents the information entropy of the jth indicator, ranging from [0,1];

[0104] Among them, pij It represents the proportion of the i-th sample in the j-th indicator. The formula is:

[0105]

[0106] If p ij =0, then p ij lnp ij =0;

[0107] Step S2.2: Calculate the objective weight of each indicator using information entropy. The formula is:

[0108]

[0109] Among them, w j represents the entropy weight of the jth indicator; (1-Ej) represents the difference coefficient of the indicator. The larger the value, the higher the degree of data dispersion.

[0110] Step S2.3: Experts score the importance of m indicators based on prior knowledge and obtain the subjective structural weight vector:

[0111] θ=[θ j ] 1×m ;

[0112] And must meet

[0113]

[0114] Step S2.4: Fusion of objective weights and subjective structural weights to generate comprehensive weights for each indicator. The formula is:

[0115] W j =αw j +(1-α)θ j , 0≤α≤1;

[0116] Among them, W j represents the comprehensive weight of the jth indicator, and α is the preset weight coefficient, which is used to adjust the balance between subjective and objective weights.

[0117] In steps S2.1-S2.4 above, the information entropy of each indicator is calculated using the information entropy formula based on the standardized data matrix. Information entropy reflects the degree of dispersion of the indicator data; smaller entropy values ​​indicate greater information content. The objective weight of each indicator is calculated using information entropy. The entropy weight represents the coefficient of variation of the indicator; larger values ​​indicate a more significant impact on overall reliability. Experts score the importance of each indicator based on prior knowledge, generating a subjective structural weight vector to supplement the practical experience factors that may be overlooked by the objective weight. Using preset weight coefficients, the objective weights are combined with the subjective structural weights to generate a comprehensive weight for each indicator. This implementation step considers both the objective characteristics of the data itself and the subjective experience of the experts, ensuring the scientific and rationality of the weight calculation.

[0118] Step S3: constructing a membership function for each reliability indicator to describe the degree to which the indicator data belongs to different levels in the preset reliability level set.

[0119] In some embodiments, step S3 includes the following steps S3.1-S3.2.

[0120] Step S3.1: The preset reliability level set is:

[0121] V = {υ1,υ2,υ3,υ4} = {very reliable, reliable, unreliable, very unreliable}.

[0122] Step S3.2: Construct membership functions. The functions of each level are defined as follows:

[0123]

[0124]

[0125] In the above steps S3.1-S3.2, by constructing a membership function, the continuous index value is converted into a discrete reliability grade membership, which solves the problem of inaccurate classification caused by fuzzy boundaries in traditional methods; the introduction of the membership function makes the description of the reliability grade more consistent with the fuzzy characteristics in actual applications, and can more realistically reflect the reliability status of the voltage transformer; the preset reliability grade set and the standardized membership function construction method provide a unified standard framework for reliability analysis in different application scenarios, enhancing the versatility of the method.

[0126] Step S4: Calculating the comprehensive membership of the voltage transformer to each reliability level according to the membership function and the comprehensive weight.

[0127] In some embodiments, step S4 includes the following steps S4.1-S4.2.

[0128] Step S4.1: For each indicator j, according to its standardized value yj , calculate its membership A to the four levels v1, v2, v3, v4 through the membership function j1 ,A j2 ,A j3 ,A j4 .

[0129] Step S4.2: For each level v k , the membership degree A of all indicators jk Multiply by its weight W j , sum to get the comprehensive membership B k , the formula is:

[0130]

[0131] B k Indicates the device is level v k The comprehensive membership degree of W is greater, and the greater the value, the more likely it is to belong to the level; j represents the comprehensive weight of the jth indicator; A jk Indicates the jth index for level v k The degree of membership.

[0132] Step S4.3: The final comprehensive membership vector is expressed as:

[0133] B=[B1,B2,B3,B4].

[0134] In the above steps S4.1-S4.3, the values ​​of each indicator are converted into membership to the reliability level through the membership function, which can effectively deal with the ambiguity and uncertainty in the data and improve the accuracy and reliability of the evaluation. The membership of each indicator is weighted and summed in combination with the comprehensive weight, ensuring that the contribution of different indicators to the final evaluation result matches their importance. The representation of the comprehensive membership vector intuitively reflects the degree of membership of the voltage transformer to each reliability level, providing a reliable data basis for subsequent level judgment and decision-making.

[0135] Step S5: determining the current reliability level of the voltage transformer based on the comprehensive membership, and triggering a corresponding maintenance decision according to the level result.

[0136] In some embodiments, step S5 includes the following steps S5.1-S5.3.

[0137] Step S5.1: Find the k value corresponding to the maximum value from the comprehensive membership vector B = [B1, B2, B3, B4].

[0138] Step S5.2: Map the k value to the corresponding reliability level v k .

[0139] Step S5.3: When the level is v3 unreliable or v4 very unreliable, an early warning signal is automatically triggered.

[0140] The above steps S5.1-S5.3, based on the maximum value determination method of the comprehensive membership, can quickly determine the current reliability level of the voltage transformer, avoiding the interference of complex calculations and subjective judgments; through the mechanism of automatically triggering the early warning signal, it can promptly remind the operation and maintenance personnel when the reliability condition of the voltage transformer deteriorates, prevent serious consequences caused by equipment failure, and enhance the safety of the system; by combining the reliability level determination with maintenance decision-making, closed-loop management from analysis to execution is realized, providing strong technical support for the intelligent operation and maintenance of the voltage transformer.

[0141] Step S6: By obtaining the historical comprehensive membership vector and inputting it into the LSTM network as time series data, the comprehensive membership value of the future time period is predicted, and an early warning is triggered when the predicted value exceeds the preset warning threshold, reminding the operation and maintenance personnel to take maintenance measures in advance.

[0142] In some embodiments, step S6 specifically includes steps S6.1-S6.4.

[0143] Step S6.1: Get historical moments t1, t2, ..., t N The comprehensive membership vector B k (t) , where t is the timestamp.

[0144] Step S6.2: The historical comprehensive membership B k (t) As a time series, input into the LSTM network.

[0145] Step S6.3: LSTM network output predicts B in the future T period k (t+1) ,...,B k (t+T) .

[0146] Step S6.4: Define the warning threshold B thk , when the predicted value satisfies B k (t+T) >B thk triggering early warnings.

[0147] In the above steps S6.1-S6.4, the modeling capability of the LSTM network for time series data is utilized to accurately capture the long-term trend and periodic characteristics of the voltage transformer reliability changes, thereby improving the accuracy and reliability of the prediction results. By predicting the comprehensive membership value for future time periods, defining the warning threshold, and combining the prediction results to trigger early warnings, timely measures can be taken before the equipment reliability condition deteriorates, providing forward-looking support for operation and maintenance decisions, enhancing the system's preventive maintenance capabilities, and thus improving the safety and stability of the power system.

[0148] Specifically, in step S6, the structure of the LSTM network includes an input layer, an LSTM layer, and a fully connected layer; wherein the input layer is used to receive historical time series data as the input of the model; the input dimension is L×1, representing L consecutive time steps; the first layer LSTM is configured with 64 units and returns the output of all time steps; the second layer LSTM is configured with 32 units and returns the output of the last time step; the dropout rate of the Dropout layer is configured to 0.2; the number of units in the fully connected layer is configured to T, corresponding to the number of future time steps to be predicted, and the activation function is configured to Sigmoid; the output is the predicted value for the future T time period.

[0149] In summary, this implementation method comprehensively reflects the reliability status of voltage transformers by collecting data from multiple preset reliability indicators. Through standardization, dimensional differences between different indicators are eliminated, generating a unified, standardized data matrix. This provides a reliable data foundation for subsequent analysis and ensures the scientific nature and comparability of the evaluation results.

[0150] The comprehensive weights of each indicator are generated by combining the objective weights calculated from information entropy and the subjective structural weights determined by expert experience. Information entropy reflects the degree of dispersion of the indicator data, while the entropy weight represents the coefficient of variation of the indicator. A larger value indicates a more significant impact on the overall reliability of the indicator. Expert scoring, on the other hand, supplements the practical experience factors that may be overlooked by the objective weights. By using preset weight coefficients to adjust the balance between subjective and objective weights, the scientific and rational weight calculation is ensured, making the evaluation results more consistent with actual conditions.

[0151] A membership function was constructed for each reliability indicator, converting continuous indicator values ​​into discrete reliability grade memberships. This addresses the inaccurate classification issues inherent in traditional methods due to fuzzy boundaries. The introduction of membership functions makes the description of reliability grades more consistent with the fuzzy characteristics of practical applications, more realistically reflecting the reliability status of voltage transformers and significantly improving the accuracy and reliability of the assessment.

[0152] An LSTM network is used to perform time series modeling of historical comprehensive membership vectors, predicting comprehensive membership values ​​for future time periods and triggering early warnings when the predicted value exceeds a preset warning threshold. The LSTM network accurately captures the long-term trends and cyclical characteristics of voltage transformer reliability changes, improving the accuracy and reliability of prediction results. This early warning mechanism enables timely action before equipment reliability deteriorates, providing forward-looking support for operation and maintenance decisions, enhancing the system's preventive maintenance capabilities, and ultimately improving the safety and stability of the power system.

[0153] In another aspect of the embodiments of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program implements the above-mentioned method steps when executed by a processor.

[0154] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A voltage transformer reliability analysis method, characterized in that: The following steps are involved: S1. Collecting multiple preset reliability index data of a voltage transformer, and performing standardization processing on the data to generate a standardized data matrix; S2. Based on the standardized data matrix, the objective weight of each indicator is calculated by information entropy, and the subjective structural weight determined by expert experience is combined to generate the comprehensive weight of each indicator; S3. Constructing a membership function for each reliability indicator to describe the degree to which the indicator data belongs to different levels in the preset reliability level set; S4. Calculating the comprehensive membership of the voltage transformer to each reliability level according to the membership function and the comprehensive weight; S5. Determine the current reliability level of the voltage transformer based on the comprehensive membership, and trigger a corresponding maintenance decision according to the level result.

2. A voltage transformer reliability analysis method according to claim 1, characterized in that: Step S1 includes: Collect m reliability index data of voltage transformer, the number of samples is n, and the original data matrix is: X=[x ij ] n×m ; Among them, x ij represents the original value of the jth indicator of the i-th sample, n represents the number of samples, and m represents the number of indicators; According to the nature of the indicators, standardization is performed separately; The standardized processing formula for positive indicators is: Among them, min(x j ) represents the minimum value of the jth indicator in all samples, max(x j ) represents the maximum value of the jth indicator in all samples; The standardized processing formula for negative indicators is: Process the original data matrix X column by column, and finally get the standardized matrix Y, the formula is: And=[and ij ] n×m ; Among them, y ij It represents the normalized value of the jth indicator of the i-th sample, and its range is limited to [0,100].

3. A voltage transformer reliability analysis method according to claim 1, characterized in that: Step 2 includes: Based on the standardized data matrix, the information entropy is calculated as follows: Among them, E j represents the information entropy of the jth indicator, ranging from [0,1]; Among them, p ij It represents the proportion of the i-th sample in the j-th indicator. The formula is: If p ij =0, then p ij lnp ij =0; The objective weight of each indicator is calculated by information entropy, and the formula is: Among them, w j represents the entropy weight of the jth indicator; (1-Ej) represents the difference coefficient of the indicator. The larger the value, the higher the degree of data dispersion.

4. A voltage transformer reliability analysis method according to claim 3, characterized in that: Step S2 includes: Experts score the importance of m indicators based on prior knowledge and obtain the subjective structural weight vector: θ=[θ j ] 1×m ; And must meet The objective weight and subjective structural weight are integrated to generate the comprehensive weight of each indicator. The formula is: W j =αw j +(1-α)θ j ,0≤α≤1; Among them, W j represents the comprehensive weight of the jth indicator, and α is the preset weight coefficient, which is used to adjust the balance between subjective and objective weights.

5. A voltage transformer reliability analysis method according to claim 1, characterized in that: Step S3 includes: The preset reliability level sets are: V = {v1, v2, v3, v4} = {very reliable, reliable, unreliable, very unreliable}; Construct the membership function, and define each level function as follows:

6. A voltage transformer reliability analysis method according to claim 1, characterized in that: Step S4 includes: For each indicator j, according to its standardized value y j , calculate its membership A to the four levels v1, v2, v3, v4 through the membership function j1 ,A j2 ,A j3 ,A j4 ; For each level v k , the membership degree A of all indicators jk Multiply by its weight W j , sum to get the comprehensive membership B k , the formula is: B k Indicates the device is level v k The comprehensive membership degree of W is greater, and the greater the value, the more likely it is to belong to the level; j represents the comprehensive weight of the jth indicator; A jk Indicates the jth index for level v k The degree of membership; The final comprehensive membership vector is expressed as: B=[B1,B2,B3,B4].

7. A voltage transformer reliability analysis method according to claim 1, characterized in that: Step S5 includes: From the comprehensive membership vector B = [B1, B2, B3, B4], find the k value corresponding to the maximum value; According to the k value, it is mapped to the corresponding reliability level v k . When the level is v3 unreliable or v4 very unreliable, an early warning signal is automatically triggered.

8. A voltage transformer reliability analysis method according to claim 1, characterized in that: The following steps are also included: S6. Get historical moments t1, t2, ..., t N The comprehensive membership vector B k (t) , where t is the timestamp; The historical comprehensive membership B k (t) As a time series, input into the LSTM network; LSTM network output predicts B in the future T period k (t+1) ,...,B k (t+T) ; Define warning threshold B thk , when the predicted value satisfies B k (t+T) >B thk triggering early warnings.

9. A voltage transformer reliability analysis method according to claim 8, characterized in that: In step S6, the structure of the LSTM network includes an input layer, an LSTM layer, and a fully connected layer; The input layer is used to receive historical time series data as the input of the model; the input dimension is L×1, representing L consecutive time steps; The first LSTM layer is configured with 64 units and returns the output of all time steps; The second LSTM layer is configured with 32 units and returns the output of the last time step; The dropout layer dropout rate is set to 0.2; The number of units in the fully connected layer is configured as T, corresponding to the number of future time steps to be predicted, and the activation function is configured as Sigmoid; the output is the predicted value for the next T time period.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein the computer program implements the method steps according to any one of claims 1 to 9 when executed by a processor.

Images