A transformer family defect identification method and device based on bath curve fitting

Abstract

The application discloses a kind of based on bathtub curve fitting transformer family defect identification method and device, comprising: obtaining the sample set of different voltage grade transformer average failure probability;Based on sample set, the distribution fitting of different voltage grade transformer average failure probability in sample set is carried out, and the bathtub curve of transformer failure probability is constructed;Using CATS algorithm, the failure information of transformer that may exist family defect is matched with the similarity of transformer failure probability bathtub curve, and the similarity measure coefficient is obtained;Similarity measure coefficient is standardized, and then according to the family defect risk classification framework, the family defect of transformer is discriminated.The application uses Weibull-CATS method to establish the identification model of transformer family defect, using the data of power grid data platform, just can offline to high-risk transformer family defect fast identification, without another electrical test, without affecting transformer operation, with higher precision, robustness.

Description

Technical Field

[0001] This invention relates to the field of transformer fault detection technology, and in particular to a method and apparatus for identifying family-related defects in transformers based on bathtub curve fitting. Background Technology

[0002] The safety of the power system is the last line of defense for ensuring people's basic livelihoods and a vital lifeline for national economic development. The normal operation of transformers is crucial for preventing power grid safety accidents. Transformers have complex structures, making it difficult to pinpoint the causes of their failures. Some failures are due to normal aging leading to insulation damage, some to accidental operational errors, and some may originate during the transformer manufacturer's production process. Identifying family-related defects in power transformers plays a vital role in industrial applications. Timely identification of these defects and the discovery of design and material problems during equipment manufacturing will greatly contribute to the safe operation of the power system.

[0003] Currently, there are few methods for identifying family-related defects in transformers, relying mainly on expert knowledge and manual screening. This is not only costly in terms of manpower and resources, but also fails to meet the requirements of practical applications. With the gradual development of digital power grid construction, big data analytics technology is also playing a significant role in the electrical field. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and device for identifying transformer family defects based on bathtub curve fitting, which can quickly, efficiently and accurately identify transformer family defects using big data technology.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0006] In a first aspect, the present invention provides a method for identifying family-related defects in transformers based on bathtub curve fitting, the method comprising the following steps:

[0007] S1, obtain a sample set of average fault probabilities for transformers of different voltage levels;

[0008] S2, Based on the sample set, the distribution of the average failure probability of transformers at different voltage levels in the sample set is fitted to construct a bathtub curve of transformer failure probability;

[0009] S3. Using the CATS algorithm, the transformer fault information that may have family defects is matched with the transformer fault probability bathtub curve to obtain the similarity measurement coefficient. The similarity measurement coefficient is then standardized to obtain the standardized similarity.

[0010] S4.甄别 transformer familial defects according to the familial defect risk grading framework based on the similarity after standardized processing.

[0011] Further, in the S1, obtaining a sample set of the average failure probabilities of transformers with different voltage levels includes:

[0012] Collecting the in-service quantity, failure quantity, and operation years information of transformers with different models, different manufacturers, and different voltage levels that have been in operation for more than m years as failure samples, preprocessing the m-year failure samples, and classifying and summarizing according to the operation time and voltage level to obtain a sample set of the average failure probabilities of transformers with different voltage levels.

[0013] Further, in the S1, preprocessing the m-year failure samples includes:

[0014] Calculating the failure probability of the transformer , and the calculation formula is as follows:

[0015]

[0016] where m is the operation years of the device.

[0017] Further, in the step S2, based on the sample set, performing distribution fitting on the average failure probabilities of transformers with different voltage levels in the sample set to construct a bathtub curve of the transformer failure probability, including:

[0018] S201. The expression of the transformer failure probability h0(t) based on the Weibull distribution is:

[0019]

[0020] where k represents the Weibull distribution shape parameter, represents the Weibull distribution shape parameter scale parameter, t represents the operation years of the transformer; when k < 1, the failure rate shows a downward trend; when k = 1, the failure rate is a constant; when k > 1, the failure rate shows an upward trend;

[0021] S202. Obtaining the parameters k and η in each stage, and their calculation formulas are:

[0022]

[0023]

[0024]

[0025]

[0026] where x iFor the service life of the transformer, y i denoted as the failure rate, n as the length of the sample data points, a as the intermediate value of the Weibull distribution parameter calculation, and b as the fitted value of the Weibull distribution shape parameter.

[0027] S203. Using the calculated k and η parameters, bathtub curves of transformer fault probability for different voltage levels are plotted.

[0028] Further, in step S3, the CATS algorithm is used to perform similarity matching between transformer fault information that may have familial defects and the transformer fault probability bathtub curve to obtain a similarity metric coefficient. This similarity metric coefficient is then standardized to obtain a standardized similarity, including:

[0029] S301, using transformer-related information data with the probability of suspected familial defects, an analogy analysis is performed with the transformer failure probability bathtub curve to obtain a measure of similarity between the transformer failure probability bathtub curves, CATS.

[0030] S302, The similarity measure CATS between the transformer fault probability bathtub curves is standardized.

[0031] S303, CATS based on standardized similarity St Based on the familial defect risk classification framework, transformer familial defects are identified.

[0032] Further, in step S301, the similarity between transformer fault information that may have familial defects and the transformer fault probability bathtub curve is determined to obtain a similarity metric CATS between the transformer fault probability bathtub curves, including:

[0033] Define the transformer failure probability bathtub curve and the spatial attenuation function of transformer manufacturer failure information distance:

[0034]

[0035] Among them, P i,l Let P be a point on the bathtub curve of transformer failure probability. j,k This is a fault message from the transformer manufacturer. `dist` is a distance measurement formula, using Euclidean distance. It is the set threshold of the transformer failure probability bathtub curve space. For Euclidean distance values ​​greater than the set threshold, the attenuation function is judged to be 0.

[0036] Define the Score function:

[0037]

[0038] in, Indicates the transformer's service life T j Corresponding transformer manufacturer fault information, t i,l It is P i,l The corresponding service life of the transformer, It is the boundary of the service life of the transformer failure probability bathtub curve, that is, for a point P i,l On the transformer failure probability bathtub curve, among all points that satisfy the service life boundary, the point with the largest spatial decay function is the alignment point of that point.

[0039] Based on the alignment method of point-to-transformer fault probability bathtub curves, a similarity measure CATS is proposed between transformer fault probability bathtub curves:

[0040]

[0041] Among them, T i T represents the service life of the transformer corresponding to the i-th point on the bathtub curve. j It represents the service life of the transformer corresponding to point j on the bathtub curve.

[0042] Further, step S302, which standardizes the CATS (Similarity Measurement System) metric for the similarity between the transformer fault probability bathtub curves, includes:

[0043] The data standardization formula is as follows:

[0044]

[0045] Among them, CATS St For the standardized similarity, The average of all similarity CATS values. denoted as the standard deviation of the similarity CATS.

[0046] Further, step S4, based on the standardized similarity, identifies transformer family defects according to the family defect hazard classification framework, including:

[0047] Similarity measures between transformer fault probability bathtub curves: CATS-standardized similarity St To identify family-related defects in transformers, when CATS St A value less than 0.5 indicates a high probability of the transformer having a family-related defect. (CATS) St When the value is between 0.5 and 0.75, further identification of transformer family defects is conducted using professional knowledge. (CATS) St A value greater than 0.75 indicates that the transformer does not have a family-related defect.

[0048] Secondly, the present invention provides a transformer family defect identification device based on bathtub curve fitting, comprising:

[0049] Input module: Used to obtain a sample set of average fault probabilities for transformers of different voltage levels;

[0050] Fitting module: used to fit the distribution of the average failure probability of transformers at different voltage levels in the sample set based on the sample set, and construct a bathtub curve of transformer failure probability;

[0051] Similarity module: Used to match the transformer fault information that may have family defects with the transformer fault probability bathtub curve using the CATS algorithm, obtain the similarity measurement coefficient, and standardize the similarity measurement coefficient to obtain the standardized similarity.

[0052] Screening module: Used to screen transformer family defects based on the similarity after standardization and according to the family defect risk classification framework.

[0053] Thirdly, the present invention provides a transformer family defect identification device based on bathtub curve fitting, including a processor and a storage medium;

[0054] The storage medium is used to store instructions;

[0055] The processor is configured to operate according to the instructions to perform the steps of the method described in the first aspect.

[0056] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0057] This invention provides a method for identifying family-related defects in transformers based on bathtub curve fitting. It employs the Weibull-CATS algorithm to establish an identification model for these defects. Utilizing data from a power grid data platform, it enables rapid offline identification of high-risk transformer family-related defects without requiring additional electrical testing, thus avoiding any impact on transformer operation. Furthermore, it exhibits high accuracy and robustness. The CART algorithm used in this invention can mitigate errors caused by delays in the bathtub curve, such as inconsistencies in transformer operating environments, load conditions, and aging levels. It can also reduce errors caused by discontinuous failure rate variations in the bathtub curve, including issues related to proper transformer maintenance, timely repairs, sudden failures, and natural disasters. Finally, this invention can comprehensively evaluate the quality of transformers from different manufacturers, providing recommendations for transformer selection and purchase, and contributing to improved equipment management. Attached Figure Description

[0058] Figure 1This is a flowchart of the transformer family defect identification method based on bathtub curve fitting of the present invention;

[0059] Figure 2 This is a sample set of average fault probability for 35kV voltage level transformers according to embodiments of the present invention.

[0060] Figure 3 This is the bathtub curve of the failure probability of a 35kV voltage level transformer after processing by the method described in this invention.

[0061] Figure 4 This is a comparison chart showing the effect of analogy analysis between transformer manufacturer fault information and transformer fault probability bathtub curve in an embodiment of the present invention. Detailed Implementation

[0062] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0063] Example 1:

[0064] This embodiment provides a method and apparatus for identifying transformer family defects based on bathtub curve fitting, which can quickly, efficiently and accurately identify transformer family defects using big data technology.

[0065] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] like Figure 1 As shown, the transformer family defect identification method based on bathtub curve fitting provided by the present invention includes the following steps:

[0067] S1. Obtain a sample set of average failure probabilities of transformers at different voltage levels. Specifically, collect information on the number of transformers in service, the number of failures, and the years of operation of transformers of different models, manufacturers, and voltage levels that have been in operation for more than m years as failure samples. Preprocess the failure samples of the m years and classify and summarize them according to the operating time and voltage level to obtain a sample set of average failure probabilities of transformers at different voltage levels.

[0068] Preprocessing of fault samples from year m includes:

[0069] The formula for calculating the transformer failure probability is as follows:

[0070]

[0071] Where m represents the equipment's operating years. m is a natural number.

[0072] S2. Based on the sample set, use the Weibull method to perform distribution fitting on the average failure probabilities of transformers with different voltage levels in the sample set, and construct a bathtub curve of transformer failure probability, including:

[0073] S201. The failure probability h0(t) of a transformer based on the Weibull distribution is expressed as:

[0074]

[0075] where k represents the Weibull distribution shape parameter, represents the Weibull distribution scale parameter of the shape parameter, t represents the operating years of the transformer; when k < 1, the failure rate shows a downward trend; when k = 1, the failure rate is a constant; when k > 1, the failure rate shows an upward trend;

[0076] S202. Obtain the parameters k and η for each stage, and their calculation formulas are:

[0077]

[0078]

[0079]

[0080]

[0081] where x i is the in-service years of the transformer, y i is the failure rate, n is the length of the sample data points, a is the intermediate value of the Weibull distribution parameter calculation, and b represents the fitted value of the Weibull distribution shape parameter;

[0082] S203. Use the obtained parameters k and η to draw the bathtub curves of transformer failure probabilities for different voltage levels;

[0083] S3. Use the CATS algorithm to perform similarity matching between the failure information of transformers that may have family defects and the bathtub curves of transformer failure probabilities, and obtain a similarity measurement coefficient, including:

[0084] S301. Use the data of transformer-related information with a suspected probability of family defects to conduct analogical analysis with the bathtub curves of transformer failure probabilities, and obtain the measurement CATS of the similarity between the bathtub curves of transformer failure probabilities;

[0085] S302. Perform standardization processing on the measurement CATS of the similarity between the bathtub curves of transformer failure probabilities;

[0086] S303, based on the standardized similarity CATS, identifies transformer family defects according to the family defect hazard classification framework.

[0087] S4 standardizes the similarity measurement coefficients and then identifies transformer family defects based on the family defect risk classification framework.

[0088] For example, step S301 involves determining the similarity between transformer fault information that may have familial defects and the transformer fault probability bathtub curve, and obtaining a metric for the similarity between the transformer fault probability bathtub curves, including:

[0089] Define the transformer failure probability bathtub curve and the spatial attenuation function of transformer manufacturer failure information distance:

[0090]

[0091] Among them, P i,l Let P be a point on the bathtub curve of transformer failure probability. j,k This is a fault message from the transformer manufacturer. `dist` is a distance measurement formula; the Euclidean distance can be used. It is the set threshold of the transformer failure probability bathtub curve space. For Euclidean distance values ​​greater than the set threshold, the attenuation function is judged to be 0.

[0092] Define the Score function:

[0093]

[0094] in, Indicates the transformer's service life T j Corresponding transformer manufacturer fault information, t i,l It is P i,l The corresponding service life of the transformer, It is the boundary of the service life of the transformer failure probability bathtub curve, that is, for a point P i,l On the transformer failure probability bathtub curve, among all points that satisfy the service life boundary, the point with the largest spatial decay function is the alignment point of that point.

[0095] Based on the alignment method of point-to-transformer fault probability bathtub curves, a similarity measure CATS is proposed between transformer fault probability bathtub curves:

[0096]

[0097] T i T represents the service life of the transformer corresponding to the i-th point on the bathtub curve. j It represents the service life of the transformer corresponding to point j on the bathtub curve.

[0098] Step S302, which standardizes the similarity metric CATS between the transformer fault probability bathtub curves, includes:

[0099] The data standardization formula is as follows:

[0100]

[0101] Among them, CATS St For the standardized similarity, The average of all similarity CATS values. denoted as the standard deviation of the similarity CATS.

[0102] Finally, step S4 involves identifying transformer family defects based on a family defect hazard classification framework, including:

[0103] Similarity measures between transformer fault probability bathtub curves: CATS-standardized similarity St To identify family-related defects in transformers, when CATS St A value less than 0.5 indicates a high probability of the transformer having a family-related defect. (CATS) St When the value is between 0.5 and 0.75, further identification of transformer family defects is conducted using professional knowledge. (CATS) St A value greater than 0.75 indicates that the transformer does not have a family-related defect.

[0104] The entire process is a complete familial defect identification model. The previous steps are used to obtain similarity metrics, and the final conclusion is given based on the familial defect risk classification framework.

[0105] Taking a 35kV voltage level transformer as an example, the transformer family defect identification method based on bathtub curve fitting provided by this invention will be explained:

[0106] 1. Obtaining the sample set of average fault probability for 35kV voltage level transformers

[0107] Information on the number of transformers in service, the number of faults, and the years of operation of transformers of different models, manufacturers, and voltage levels over 27 years was collected. After preprocessing the fault samples from the 27 years, a sample set of average fault probabilities for 35kV voltage level transformers was obtained by classifying and summarizing them according to operating time and voltage level. Figure 2 As shown.

[0108] 2. Based on the average fault probability sample set of 35kV voltage level transformers, the Weibull method is used to fit the distribution of the average fault rate of the same voltage level in the sample set and construct the transformer fault probability bathtub curve.

[0109] 2.1 Calculate the transformer fault probability h0(t) based on the Weibull distribution;

[0110] 2.2 By analyzing historical statistical data of the same type of electrical equipment, the failure rate curve based on the Weibull distribution can be piecewise fitted to obtain the parameters k and η for each stage;

[0111] 2.3 Using the calculated k and η parameters, a bathtub curve of the failure probability of a 35kV voltage level transformer is plotted, as shown below. Figure 3 As shown.

[0112] 3. After obtaining the transformer failure probability bathtub curve, the transformer failure information that may have familial defects is matched with the transformer failure probability bathtub curve to obtain a similarity measurement coefficient, such as... Figure 4 As shown in Table 1, the pseudocode for the CATS algorithm is as follows:

[0113] Table 1

[0114]

[0115] Pick =1.5, =0.001 A set of transformer manufacturer fault information data of interest was selected and compared with the standard transformer fault probability bathtub curve of the same voltage level. During the CATS standardization process of the similarity between the transformer fault probability bathtub curves, the formula for data standardization is:

[0116]

[0117] Get CATS St The value is 0.952541131738429, because CATS St A value greater than 0.75 indicates that the transformer does not have a family-related defect.

[0118] In summary, the transformer family defect identification method based on bathtub curve fitting provided by this invention uses the Weibull-CATS method to establish an identification model for transformer family defects. Utilizing data from the power grid data platform, it can quickly identify high-risk transformer family defects offline without requiring additional electrical testing, thus not affecting transformer operation. It also exhibits high accuracy and robustness. Furthermore, this invention can fully assess the quality of transformers from different manufacturers, providing suggestions for transformer selection and purchase, and contributing to improved equipment management.

[0119] Example 2:

[0120] This embodiment provides a transformer family defect identification device based on bathtub curve fitting, including:

[0121] Input module: Used to obtain a sample set of average fault probabilities for transformers of different voltage levels;

[0122] Fitting module: used to fit the distribution of the average failure probability of transformers at different voltage levels in the sample set based on the sample set, and construct a bathtub curve of transformer failure probability;

[0123] Similarity module: Used to match the transformer fault information that may have family defects with the transformer fault probability bathtub curve using the CATS algorithm, obtain the similarity measurement coefficient, and standardize the similarity measurement coefficient to obtain the standardized similarity.

[0124] Screening module: Used to screen transformer family defects based on the similarity after standardization and according to the family defect risk classification framework.

[0125] The apparatus in this embodiment can be used to implement the method described in Embodiment 1.

[0126] Example 3:

[0127] This embodiment provides a transformer family defect identification device based on bathtub curve fitting, including a processor and a storage medium;

[0128] The storage medium is used to store instructions;

[0129] The processor is configured to operate according to the instructions to execute the steps of the method described in Embodiment 1.

[0130] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0132] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0133] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0134] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A transformer family defect identification method based on bath curve fitting, characterized in that, The method comprises the following steps: obtaining a sample set of average failure probabilities of transformers of different voltage grades; based on the sample set, using the Weibull method to perform distribution fitting on the average failure probabilities of transformers of different voltage grades in the sample set, and constructing a transformer failure probability bathtub curve; using the CATS algorithm, performing similarity matching between transformer failure information that may have a family defect and the transformer failure probability bathtub curve, and obtaining a similarity measurement coefficient; standardizing the similarity measurement coefficient, and according to the similarity after standardization, discriminating the transformer family defect according to a family defect risk grading framework; using the CATS algorithm, performing similarity matching between transformer failure information that may have a family defect and the transformer failure probability bathtub curve, and obtaining a similarity measurement coefficient, comprising: determining the similarity between the transformer failure probability bathtub curves by similarity determination, and obtaining a measurement CATS of the similarity between the transformer failure probability bathtub curves; standardizing the measurement CATS of the similarity between the transformer failure probability bathtub curves; determining the similarity between the transformer failure probability bathtub curves by similarity determination, and obtaining a measurement CATS of the similarity between the transformer failure probability bathtub curves, comprising: defining a decay function in the distance space of the transformer failure probability bathtub curve and the transformer manufacturer failure information: ； wherein, is a point on the transformer failure probability bathtub curve, is the transformer manufacturer failure information, dist is the distance measurement formula, and the Euclidean distance is adopted, is a set threshold value of the transformer failure probability bathtub curve space. For the Euclidean distance value greater than the set threshold value, the attenuation function is judged as 0. defining a Score function: ； wherein, represents the transformer operating age T j corresponding transformer manufacturer fault information, t i,l is P i,l corresponding transformer operating age, is the boundary of the transformer failure probability bathtub curve in-service age, that is, for a point P i,l On the transformer failure probability bathtub curve, among all points that satisfy the in-service age boundary, the point with the maximum spatial decay function is the alignment point of the point. according to a point alignment method to the transformer failure probability bathtub curve, proposing a measurement CATS of the similarity between the transformer failure probability bathtub curves: ； wherein T i is the transformer operating age corresponding to the i-th point on the bathtub curve, T j is the transformer operating age corresponding to the j-th point on the bathtub curve.

2. The bathtub curve fitting based transformer family defect identification method of claim 1, wherein, obtaining a sample set of average failure probabilities of transformers of different voltage grades, comprising: collecting the in-service quantity, failure quantity, and operation time information of transformers of different types, different manufacturers, and different voltage grades that have been in operation for more than m years as failure samples, pre-processing the m-year failure samples, and classifying and summarizing according to the operation time and voltage grade to obtain a sample set of average failure probabilities of transformers of different voltage grades.

3. The bathtub curve fitting based transformer family defect identification method of claim 2, wherein, pre-processing the m-year failure samples, comprising: The transformer failure probability is calculated by the following formula: ； wherein m is the operation time of the equipment.

4. The bathtub curve fitting based transformer family defect identification method of claim 1, wherein, based on the sample set, using the Weibull method to perform distribution fitting on the average failure probabilities of transformers of different voltage grades in the sample set, and constructing a transformer failure probability bathtub curve, comprising: the transformer failure probability h0(t) based on the Weibull distribution is expressed as: ； wherein k represents a Weibull distribution shape parameter, wherein k represents a Weibull distribution shape parameter, λ represents a Weibull distribution scale parameter, and t represents the running age of the transformer; when k < 1, the failure rate shows a downward trend; when k = 1, the failure rate is constant; and when k > 1, the failure rate shows an upward trend. obtaining the parameters k and η of each stage, and the calculation formula is: ； ； ； ； Wherein, x i is the service life of the transformer, y i is the failure rate, n is the length of the sample data points, a is the Weibull distribution parameter, and b represents the fitting value of the Weibull distribution shape parameter. using the obtained k and η parameters to draw the transformer failure probability bathtub curve of different voltage grades.

5. The bathtub curve fitting based transformer family defect identification method of claim 4, wherein, standardizing the measurement CATS of the similarity between the transformer failure probability bathtub curves, comprising: using the data standardization processing formula as follows: ； where CATS St is the standardized similarity, is the average of all similarities CATS, is the standard deviation of the similarities CATS.

6. The bathtub curve fitting based transformer family defect identification method of claim 5, wherein, according to the similarity after standardization, discriminating the transformer family defect according to a family defect risk grading framework, comprising: CATS similarity based on the similarity between the transformer failure probability bathtub curve CATS similarity after standardization St family defect risk identification of the transformer; When CATS St Less than 0.5 indicates a high probability that the transformer has a familial defect. CATS St When the value is between 0.5 and 0.75, further transformer family defect identification is required. CATS St A value greater than 0.75 indicates that the transformer does not have a familial defect.

7. A device for performing the bathtub curve fitting based transformer family defect identification method of claim 1, wherein, comprising: an input module: for obtaining a sample set of average failure probabilities of transformers of different voltage grades; The fitting module is configured to perform distribution fitting on the average failure probability of the transformers in different voltage grades in the sample set based on the sample set, and construct a transformer failure probability bathtub curve. The similarity module is configured to perform similarity matching on the transformer failure information that may exist family defect and the transformer failure probability bathtub curve by using a CATS algorithm, obtain a similarity measurement coefficient, and perform standardization processing on the similarity measurement coefficient to obtain a standardized similarity. The screening module is configured to screen the transformer family defect according to the standardized similarity and a family defect risk grading framework.

8. A device for identifying a family defect of a transformer based on a bathtub curve fitting, characterized by, The system comprises a processor and a storage medium. The storage medium is configured to store instructions. The processor is configured to operate according to the instructions to perform the steps of the method of any one of claims 1-6.

Images