A method for evaluating the operating status of transformer low-voltage busbar insulation sheath

By establishing a transformer low-voltage busbar insulating sheath status evaluation system for information fusion, using the extreme learning machine and D-S evidence theory combined with quantitative and qualitative indicators, the problem of insulating sheath fault detection error in the existing technology is solved, and the transformer low-voltage bus is realized The fast and intelligent online inspection of the insulating sheath ensures the safe and stable operation of the equipment.

CN115034606BActive Publication Date: 2024-11-26MAANSHAN POWER SUPPLY COMPANY STATE GRID ANHUI ELECTRIC POWER
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210643896.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-08
Publication Date
2024-11-26
Estimated Expiration
2042-06-08

AI Technical Summary

Technical Problem

In the prior art, the fault detection method for low-voltage busbar insulation sheath of transformers is single and has high errors. It is impossible to detect short-circuit failures caused by insulation deterioration and foreign objects climbing in time, affecting the safe operation of the transformer.

Method used

Establish a status evaluation system for low-voltage busbar insulation sheath based on information fusion. Through the data processing module and state evaluation module, use the limit learning machine and D-S evidence theory, and combine quantitative and qualitative indicators to achieve the operation of busbar insulation sheath Accurate evaluation of status.

Benefits of technology

It improves the accuracy of the status evaluation of the insulation sheath, avoids errors and misjudgments of manual detection, and realizes rapid and intelligent online detection of the insulating sheath faults of the transformer's low-voltage busbar, ensuring the safe and stable operation of the equipment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115034606B_ABST
    Figure CN115034606B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for evaluating the operating status of a transformer low-voltage busbar insulation sheath, and relates to the technical field of online monitoring of power equipment. The method for evaluating the operating status of a transformer low-voltage busbar insulation sheath first extracts characteristic data related to the operating status of the transformer low-voltage busbar insulation sheath as an evaluation index, and achieves the comprehensiveness of fault evaluation through the diversification of evaluation indicators, thereby avoiding errors and misjudgments caused by the small amount of information in manual detection. Secondly, an evaluation model for the operating status of the transformer low-voltage busbar insulation sheath is established, and the evaluation indicators are learned through an extreme learning machine, and the D-S evidence theory is used to fuse the evaluation results of the operating status at different times to ensure the accuracy of the detection. Finally, a mature operating status evaluation model makes a final decision on the operating status of the transformer low-voltage busbar insulation sheath based on a comprehensive analysis of the diagnostic evaluation results.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of online monitoring of electric power equipment, and in particular to a method for evaluating the operating status of an insulating sheath of a low-voltage busbar of a transformer. Background Art

[0002] The busbar is a conductor used to collect and distribute current, so it is also called a busbar. In general high-voltage power distribution devices, bare conductors are often used for busbars. Insulating sheaths are widely used in the outer covering of substation busbars, cable joints, transformer terminals, etc. due to their excellent electrical and mechanical properties. They mainly play the role of corrosion prevention and insulation protection. In recent years, in the actual operation of substations, short-circuit accidents between phases of the wrapped insulating sheathed busbars caused by foreign matter have occurred frequently, which has become one of the main faults threatening the safe operation of substations.

[0003] The transformer is the core component of the safe operation of the power grid. The medium and low voltage of the main transformer has the characteristics of large short-circuit current and current electrodynamics. If the busbar is in operation due to insulation degradation (aging, insulation degradation of metal wear noise at the fixed position of the busbar), the existence of foreign objects (small animals) climbing, and the insulation sheath itself will also have insulation defects such as bubbles in the production process, if these conditions are not discovered in time, they may eventually lead to the occurrence of transformer near-area short-circuit faults, seriously affecting the safe operation of the main transformer. At present, the detection of the busbar insulation condition is generally carried out by measuring the relative leakage current and the insulation resistance test between the busbar and the porcelain bottle hardware, and manually judging the insulation condition of the busbar sheath, which may cause misjudgment and missed judgment. Therefore, it is necessary to study the detection technology of the operating status of the medium and low voltage busbar insulation sheath of the main transformer, and obtain effective rapid detection technology of the operating status of the busbar sheath through research, and obtain the fault information of the medium and low voltage busbar insulation sheath of the main transformer before the short-circuit fault occurs, so as to provide technical support for the safe and stable operation of the main transformer.

[0004] Therefore, the applicant, based on many years of rich experience in design, development and actual production in the relevant industry, has conducted research and improvements on the existing structures and deficiencies, and provided a method for evaluating the operating status of the insulating sheath of a transformer low-voltage busbar, in order to achieve a more practical purpose. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In view of the shortcomings of the prior art, the present invention provides a method for evaluating the operating status of a transformer low-voltage busbar insulating sheath, which solves the problem that the existing transformer low-voltage busbar insulating sheath fault detection method is single and the manual detection technology has high errors.

[0007] (II) Technical solution

[0008] In order to achieve the above objectives and improve the accuracy of transformer low-voltage busbar insulation sheath operating status assessment, a transformer low-voltage busbar insulation sheath status assessment system based on information fusion was established. The entire assessment process is divided into a data processing module and a status assessment module. The specific steps are as follows:

[0009] Data processing module: Based on historical inspection records and real-time detection data, select effective and easy-to-obtain feature information as the evaluation object; analyze and organize the collected data to build an evaluation index system. To facilitate the establishment of the evaluation model, the feature data that actually reflects the operating status of the equipment is divided into a training set and a test set, and a sample library of the operating status of the insulating sheath is established;

[0010] Status evaluation module: Based on the test set sample data, establish a transformer low-voltage busbar insulation sheath operation status evaluation model based on information fusion; input the training set feature data, evaluate and calculate the training set data according to the above evaluation model, and obtain the corresponding status evaluation results. If the operating status is within the allowable range, continue to test while ensuring the safe operation of the equipment. If the evaluation result exceeds the limit, the equipment will be shut down for maintenance.

[0011] Preferably, the data processing module uses five characteristic information including quantitative indicators: leakage current indication (F1), insulation resistance value between busbar and porcelain bottle fittings (F2), and qualitative indicators: infrared temperature measurement record (F3), on-site noise and vibration conditions (F4) and busbar insulation sheath surface degradation conditions (F5) as evaluation indicators for the operating status of the transformer low-voltage busbar insulation sheath.

[0012] Furthermore, the quantitative indicators have different dimensions and quantification standards. In order to facilitate subsequent evaluation, normalization is required. i To characterize the relative severity, its range is [0,1]. Since the indicators F1 and F2 are both smaller, the better the data, when x i The closer it is to 0, the better the parameter state is, so the ascending semi-trapezoidal model is used for normalization processing. a and b are the thresholds of the indicator respectively. The formula is as follows:

[0013]

[0014] Furthermore, the qualitative indicators are scored and classified and converted into intuitive status data. Scoring is performed based on the severity of the impact of the insulation sheath failure on the equipment operation status in the historical inspection records. The more severe the impact, the higher the score, and vice versa. The scoring rules are shown in Table 1:

[0015] Table 1. Scoring rules for qualitative indicators of transformer low-voltage busbar insulation sheath

[0016]

[0017] Preferably, the state assessment module uses an extreme learning machine (ELM) combined with DS evidence theory to establish a transformer low-voltage busbar insulation sheath operation state assessment model, and the specific steps are as follows:

[0018] Step 1: Construct an evaluation set and determine the evaluation factors. Referring to the actual operating conditions of the transformer low-voltage busbar insulation sheath, in order to avoid unnecessary power outage maintenance operations, the present invention divides the insulation sheath operating status into three levels: normal, abnormal and dangerous.

[0019] Assume that the evaluation set is Y = {y1,y2,...,y m}, indicating the three operating states included in the evaluation results; the evaluation factor is X = {x1, x2, ..., x n}, representing the five evaluation indicators involved in the evaluation.

[0020] Step 2: Establish an evaluation structure based on extreme learning machine. Extreme learning machine is a new neural network, which is characterized by not needing to set the connection weights between the input layer and the hidden layer and the bias of the hidden layer nodes in advance. The ELM algorithm essentially uses the least squares regression formula to calculate the output weight matrix. When using it, it only needs to set the excitation function and the number of hidden layer neurons. The present invention assumes that the ELM network has n input nodes, 1 hidden layer nodes, m output layer nodes, the weight between the input layer and the hidden layer is w, the weight between the hidden layer and the output layer is β, the hidden layer activation function g(x) is a Sigmoid function, and the bias is b. The expected output of ELM is mathematically expressed as:

[0021]

[0022] In the formula, w i is the weight vector between the input node and the i-th hidden node, i.e., the input weight; b i is the bias of the i-th hidden node; β i is the weight vector between the ith hidden node and the output node, i.e., the output weight. In the training phase, in order to minimize the output error, i.e., the difference between the expected and actual output of the ELM network is close to zero, the objective function can be expressed as:

[0023]

[0024] That is, there exists w i , β i and b i , so that:

[0025]

[0026] The matrix form is: Hβ=T. Among them, H is the hidden layer output, β is the hidden layer weight, and T is the target output of the network. The model training of the extreme learning machine uses the randomly generated input weights w and bias values ​​b to solve the least squares solution of the minimum norm of the linear equation to determine the only set of output weights β=H that meet the model error requirements. + Y, obtain a set of model parameters (w, b, β), where H + is the MP generalized inverse matrix of the hidden layer output matrix H, H + =(H T H) -1 H T .

[0027] Preferably, if the output result of the traditional ELM is the evaluation level divided as above, further information fusion cannot be performed, so the output result of the extreme learning machine is converted into a probability form, and the conversion process is:

[0028]

[0029] The probability of the output of the extreme learning machine corresponding to each evaluation level is p(u i ), and normalize it:

[0030]

[0031] For each output node, the output probability matrix is:

[0032]

[0033] where p ik (y) is the output probability of the k-th group of input data corresponding to the i-th category label, which is determined by the maximum value method.

[0034] Step 3: Fusion of information based on improved DS evidence theory. The probability form of the output of the extreme learning machine is used as the basic probability distribution of evidence theory fusion, and the improved DS evidence theory is used for fusion to output the corresponding three evaluation levels, so that equipment maintenance can be performed according to the recommendations. For two pieces of evidence m1 and m2, their respective basic probability distributions are expressed as: i (L j )=b j ,i=1,2;j=1,2,...,n. Calculate the compatibility coefficient R and credibility W of evidence:

[0035]

[0036]

[0037] R ij (L k) indicates evidence m i 、m j When one of them is 0 (indicating complete negation) and the other is not 0 (indicating a certain degree of support), the two pieces of evidence are incompatible, so R ij (L k ) takes the value 0; when the two pieces of evidence are equal, they are completely consistent, then R ij (L k ) takes the value 1. i (L k ) represents the i-th piece of evidence about event L k The absolute compatibility of D is , where N is the number of evidences. When all N pieces of evidence are consistent, D i (L k ) takes the value N-1, then W i (L k ) takes the value 1 to indicate absolute credibility; when all evidence is conflicting, D i (L k ) takes the value 0, then W i (L k ) takes the value 0 to indicate that it is completely untrustworthy.

[0038] Resynthesize the mass function of the evaluation object to obtain the final probability distribution of the event. The calculation formula is as follows:

[0039] m i (L k )=W i (L k )b k ,i=1,2;k=1,2,...,n

[0040]

[0041]

[0042] Based on the fused probability distribution, according to the rule: m(L max )-(L max )>ε makes a decision. m(L max ) and m(L min ) are the maximum and second largest values ​​of the probability distribution respectively, and ε is a set constant, that is, a decision is made only when a certain result is larger than other results by a certain degree.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The state assessment method of the present invention first extracts characteristic data related to the operating state of the transformer low-voltage busbar insulation sheath as an evaluation index, and realizes the comprehensiveness of fault assessment through the diversification of evaluation indicators, thereby avoiding errors and misjudgments caused by the small amount of information in manual detection. Secondly, an evaluation model for the operating state of the transformer low-voltage busbar insulation sheath is established, and the evaluation index is learned by an extreme learning machine, and the DS evidence theory is used to fuse the evaluation results of the operating state at different times to ensure the accuracy of the detection. Finally, the mature operating state assessment model makes a final decision on the operating state of the transformer low-voltage busbar insulation sheath based on a comprehensive analysis of the diagnostic evaluation results. In short, the state assessment method of the present invention completes the online detection and evaluation of the transformer low-voltage busbar insulation sheath through the digitization of fault characteristics, thereby realizing the rapid and intelligent evaluation of the busbar sheath operating state.

[0045] (III) Beneficial effects

[0046] The present invention provides a method for evaluating the operating status of a transformer low-voltage busbar insulation sheath. It has the following beneficial effects:

[0047] The transformer low-voltage busbar insulation sheath operation status assessment method first extracts characteristic data related to the transformer low-voltage busbar insulation sheath operation status as an assessment index, and realizes the comprehensiveness of fault assessment through the diversification of the assessment index, thereby avoiding the errors and misjudgments caused by the small amount of information in manual detection. Secondly, a transformer low-voltage busbar insulation sheath operation status assessment model is established, the assessment index is learned through an extreme learning machine, and the DS evidence theory is used to fuse the assessment results of the operation status at different times to ensure the accuracy of the detection. Finally, the mature operation status assessment model makes a final decision on the operation status of the transformer low-voltage busbar insulation sheath based on the comprehensive analysis of the diagnostic assessment results. In short, the state assessment method of the present invention completes the online detection and evaluation of the transformer low-voltage busbar insulation sheath through the digitization of fault characteristics, thereby realizing the rapid and intelligent evaluation of the busbar sheath operation status. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a structural diagram of the extreme learning machine of the present invention;

[0049] Figure 2 This is an evaluation flow chart of the present invention based on extreme learning machine and improved DS evidence theory. DETAILED DESCRIPTION

[0050] 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.

[0051] See also Figure 2 The present invention provides an embodiment of a method for evaluating the operating status of a transformer low-voltage busbar insulation sheath. The specific implementation method and principle include the following steps:

[0052] Step 1: Evaluation data preprocessing: In order to test the rationality and correctness of the algorithm, the present invention uses the historical inspection records and real-time detection data of the transformer low-voltage busbar insulation sheath to train the extreme learning machine model, among which there are 100 groups of historical inspection records and 20 groups of real-time detection data, a total of 120 groups of sample sets. The training set selects 70% of the total number of samples, and the rest are used as test sets. For each group of data, quantitative indicators are selected: leakage current indication (F1), insulation resistance value between busbar and porcelain bottle hardware (F2), and qualitative indicators: infrared temperature measurement records (F3), on-site noise and vibration conditions (F4) and busbar insulation sheath surface degradation conditions (F5), a total of 5 evaluation indicators. The quantitative indicators are normalized, and their range is [0,1]. The indicators F1 and F2 are both smaller and better data. When x i The closer it is to 0, the better the parameter state is, so the ascending semi-trapezoidal model is used for normalization processing. a and b are the thresholds of the indicator, which are initially set to 10mA and 50mA. The formula is as follows:

[0053]

[0054] Qualitative indicators are scored and classified and converted into intuitive status data. Scoring is performed based on the severity of the impact of the insulation sheath failure on the equipment operation status in the historical inspection records. The more severe the impact, the higher the score, and vice versa. The scoring rules are shown in Table 1:

[0055] Table 1 Scoring rules for qualitative indicators of transformer low-voltage busbar insulation sheath

[0056]

[0057] Step 2: Extreme learning machine model training: The present invention divides the operating status of the insulating sheath into three levels: normal, abnormal and dangerous. Figure 1The structure diagram of the extreme learning machine shown in the figure sets the number of output nodes of the extreme learning machine to 3, the activation function is Sigmoid, and the number of output nodes is determined according to the number of evaluation indicators. The number of hidden layer nodes is set from 1 to 30, and the parameters are set 100 times. In the iterative test, as the number of hidden layer nodes increases, the test accuracy also increases. When the number of hidden layer nodes is 20, the highest test accuracy occurs the most times, so the number of hidden layer nodes is set to 20.

[0058] Step 3: Extreme learning machine model test analysis: The parameters of the trained extreme learning machine model are fixed, and the evaluation index data of the test set is used as input. The output result is the evaluation level of the operating status of the insulating sheath. In order to facilitate information fusion, the output result is converted into a probability form. The conversion process is:

[0059]

[0060] The probability of the output of the extreme learning machine corresponding to each evaluation level is p(u i ), and normalize it:

[0061]

[0062] Among them, p i (y) is the output probability of the input data corresponding to the i-th category label.

[0063] According to the above extreme learning machine model testing process, the comprehensive evaluation results of 5 evaluation indicators are obtained. Among them, the evaluation result of the quantitative indicator is B1=[0.07330.07490.8518], which can be preliminarily considered that the operating state of the insulating sheath is a warning state; the evaluation result of the qualitative indicator is B2=[0.06190.77620.1619], which means that the operating state of the insulating sheath is an abnormal state.

[0064] Step 4: Information fusion evaluation: From the above test results, it can be seen that when quantitative indicators and qualitative indicators are used for diagnostic evaluation respectively, the evaluation results conflict with each other. Taking any one result will discard the other two evaluation information, which is likely to cause misjudgment. According to the improved DS evidence fusion method, the two evaluation results are used as the basic probability distribution of evidence theory fusion, and the compatibility coefficient R and evidence credibility W of the three operating states are calculated, where R1, R2, and R3 represent the three operating states of "normal", "abnormal" and "serious", respectively.

[0065]

[0066]

[0067]

[0068]

[0069]

[0070] The credibility of the evidence of the two indicators about the three evaluation results is calculated: W1 = [0.90590.96060.9121]; W2 = [0.82300.87750.9250]. The mass function of the evaluation object is resynthesized to obtain the final probability distribution of the event. The calculation formula is as follows:

[0071] m i (L k )=W i (L k )b k ,i=1,2;k=1,2,...,n

[0072]

[0073]

[0074] According to the above formula, the basic probability distribution is redistributed, C1 = [0.1777 0.3916 0.3358 0.0949]; C2 = [0.2232 0.3423 0.3607 0.0738]. The newly allocated probability dimension increases by one. The first three data are the probabilities of the three evaluation results, and the fourth represents the uncertain probability. The newly allocated probabilities are reintegrated to obtain the final probability distribution of the three evaluation results, which is C = [0.2017 0.3459 0.4523 0.0200]. According to the rule m(L max )-(L max )>ε decision, when ε is 0.25, the evaluation result is “dangerous state”.

[0075] The present invention diagnoses and evaluates the operating status of the transformer busbar insulation sheath through an extreme learning machine and an improved DS evidence theory, thereby realizing the digitization, rapidity and diversification of online detection and evaluation, and can solve the problem in the prior art that the operating status of the transformer busbar insulation sheath cannot be effectively evaluated in time, so that maintenance personnel can timely and efficiently perform transformer busbar insulation sheath maintenance, avoiding unnecessary economic losses.

[0076] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting from any point of view, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0077] It should be noted that, in the description of the invention, the directions or positional relationships indicated by the terms "up", "down", "left", "right", "front", "back", etc. are descriptions of the structure of the present invention based on the accompanying drawings, and are only for the convenience of describing the present invention, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0078] The "first" and "second" in this technical solution are only used to distinguish the names of the same or similar structures, or corresponding structures with similar functions, and are not an arrangement of the importance of these structures, nor do they have a ranking, comparison of size, or other meanings.

[0079] In addition, unless otherwise clearly specified and limited, the terms "installation" and "connection" should be understood in a broad sense. For example, the connection can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be a connection between the two structures. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood based on the overall idea of ​​the present invention and the specific context of the present solution.

Claims

1. A method for evaluating the operating status of a transformer low-voltage busbar insulation sheath, characterized in that: The evaluation method includes a data processing module and a status evaluation module; The data processing module selects effective and easy-to-obtain feature information as the evaluation object based on historical inspection records and real-time detection data; analyzes and organizes the collected data to build an evaluation index system; to facilitate the establishment of an evaluation model, the feature data that actually represents the operating status of the equipment is divided into a training set and a test set, and a sample library of the operating status of the insulating sheath is established; The state assessment module establishes an information fusion-based transformer low-voltage busbar insulation sheath operation state assessment model based on the test set sample data; inputs the training set feature data, and performs assessment calculation on the training set data according to the assessment model to obtain the corresponding state assessment result; If the operating status is within the allowable range, continue testing while ensuring safe operation of the equipment; If the evaluation result exceeds the limit, the equipment will be shut down for maintenance; The data processing module uses five characteristic information, including quantitative indicators: leakage current indication (F1), insulation resistance value between busbar and porcelain bottle fittings (F2), and qualitative indicators: infrared temperature measurement record (F3), on-site noise and vibration conditions (F4) and busbar insulation sheath surface degradation conditions (F5), as evaluation indicators for the operating status of transformer low-voltage busbar insulation sheath; The state assessment module uses extreme learning machine (ELM) combined with DS evidence theory to establish a transformer low-voltage busbar insulation sheath operation state assessment model. The specific steps are as follows: Step 1: Construct an evaluation set and determine the evaluation factors; refer to the actual operating conditions of the transformer low-voltage busbar insulation sheath, and divide the insulation sheath operating status into three levels: normal, abnormal, and dangerous to avoid unnecessary power outage maintenance operations; Assume that the evaluation set is Y = {y1,y2,...,y m }, indicating the three operating states included in the evaluation results; the evaluation factor is X = {x1, x2, ..., x n }, representing the five evaluation indicators involved in the evaluation; Step 2: Establish an evaluation structure based on extreme learning machine. Assume that the ELM network has n input nodes, l hidden layer nodes, m output layer nodes, the weight between the input layer and the hidden layer is w, the weight between the hidden layer and the output layer is β, the hidden layer activation function g(x) is the Sigmoid function, and the bias is b; then the expected output of ELM is mathematically expressed as: In the formula, wi is the weight vector between the input node and the i-th hidden node, that is, the input weight; bi is the bias of the i-th hidden node; βi is the weight vector between the i-th hidden node and the output node, that is, the output weight; in the training stage, in order to minimize the output error, that is, the difference between the expected and actual output of the ELM network is close to zero, the objective function can be expressed as: That is, there exist wi, βi and bi such that: The matrix form is: Hβ = T; Among them, H is the hidden layer output, β is the hidden layer weight, and T is the target output of the network; the model training of the extreme learning machine uses the randomly generated input weights w and bias values ​​b to determine the only set of output weights β=H that meet the model error requirements by solving the least squares solution of the minimum norm of the linear equation. + Y, obtain a set of model parameters (w, b, β), where H + is the MP generalized inverse matrix of the hidden layer output matrix H, H + =(H T H) -1 H T ; Step 3: Based on the improved DS evidence theory, the probability form of the extreme learning machine output is used as the basic probability distribution of evidence theory fusion, and the improved DS evidence theory is used for fusion to output the corresponding three evaluation levels, so as to perform equipment maintenance according to the suggestions; for two pieces of evidence m1 and m2; their respective basic probability distributions are expressed as: m i (L j )=b j ,i=1,2;j=1,2,...,n;Calculate the compatibility coefficient R and credibility W of evidence: R ij (L k ) indicates evidence m i 、m j degree of consistency; When one of them is 0, which means complete denial, and the other is not 0, which means a certain degree of support, the two pieces of evidence are incompatible, so R ij (L k ) takes the value 0; When two pieces of evidence are equal, they are completely consistent, then R ij (L k ) takes the value 1; and D i (L k ) represents the i-th piece of evidence about event L k The absolute compatibility of , N represents the number of evidences; When all N pieces of evidence are consistent, D i (L k ) takes the value N-1, then W i (L k ) takes the value 1 to indicate absolute credibility; when all evidence is conflicting, D i (L k ) takes the value 0, then W i (L k ) takes the value 0 to indicate that it is completely untrustworthy; Resynthesize the mass function of the evaluation object to obtain the final probability distribution of the event. The calculation formula is as follows: m i (L k )=W i (L k )b k ,i=1,2;k=1,2,...,n Based on the fused probability distribution, according to the rule: m(L max )-(L max )>ε makes a decision; m(L max ) and m(L min ) are the maximum and second largest values ​​of the probability distribution respectively, and ε is a set constant, that is, a decision is made only when a certain result is larger than other results by a certain degree.

2. A method for evaluating the operating status of a transformer low-voltage busbar insulation sheath according to claim 1, characterized in that: The quantitative indicators mentioned above have different dimensions and quantification standards. In order to facilitate subsequent evaluation, they need to be normalized; Each indicator is characterized by its relative severity through xi, which ranges from [0,1]; Since the indicators F1 and F2 are both smaller, the better the data, when xi is closer to 0, it means that the parameter state is better, so the ascending semi-trapezoidal model is used for normalization processing, a and b are the thresholds of the indicator respectively, and the formula is as follows:

3. A method for evaluating the operating status of a transformer low-voltage busbar insulation sheath according to claim 1, characterized in that: The qualitative indicators are scored and classified and converted into intuitive status data; A score is given based on the severity of the impact of the insulation sheath failure on the equipment operation status in the historical inspection records. The more severe the impact, the higher the score, and vice versa.

4. A method for evaluating the operating status of a transformer low-voltage busbar insulation sheath according to claim 1, characterized in that: If the output of the traditional ELM is the evaluation level divided above, further information fusion cannot be performed. Therefore, the output of the extreme learning machine is converted into a probability form. The conversion process is: The probability of the output of the extreme learning machine corresponding to each evaluation level is p(u i ), and normalize it: For each output node, the output probability matrix is: where p ik (y) is the output probability of the k-th group of input data corresponding to the i-th category label, which is determined by the maximum value method.