Abnormal distribution transformer identification method and device considering sample imbalance processing
By acquiring and processing the daily report data and population data of the distribution transformer, combining clustering and oversampling technology, an abnormal prediction model is constructed, which solves the accuracy and sample imbalance of the prediction of heavy overload during major holidays, and improves the prediction accuracy rate .
Patent Information
- Application Number
- CN202411454454.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-17
- Publication Date
- 2025-02-14
AI Technical Summary
The prior art cannot accurately predict heavy overload abnormalities in distribution transformers during special periods such as major holidays, and in the prediction of distribution and heavy overload, due to the unbalanced data set samples, the model's prediction effect on heavy overload situation is poor.
By obtaining the daily report data and population data of the distribution transformer in the target area, the indicator characteristics used to determine whether it is overloaded are selected, and the change rate of the number of people returning home is calculated, and data classification and normalization are performed. Then, the K-Means clustering algorithm is used for clustering and stratification undersampling, the SMOTE oversampling algorithm is used for oversampling, and finally anomaly prediction model is built through the SVM algorithm optimized by the particle swarm intelligence algorithm.
The accuracy of prediction of heavy and overload of the distribution model during the return of people during major holidays such as the Spring Festival has been improved, and the problem of low accuracy of the distribution model of heavy and overload of the distribution model caused by sample imbalance is solved.
Smart Images
Figure CN119441866A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of power transmission technology control, and in particular to a method, system, readable storage medium and computer equipment for identifying abnormal distribution transformers considering sample imbalance processing. Background Art
[0002] Distribution transformers are important nodes for power transmission to users. If a distribution transformer is in a heavy or overloaded state for a long time, it will reduce the life of the equipment. On the other hand, it may also cause line failures and low voltage at the end of the low-voltage line. The operating status of the distribution transformer largely determines the quality and safety of power supply in the area. Therefore, accurate prediction of heavy overload anomalies of distribution transformers is crucial.
[0003] However, there are the following defects in the prior art:
[0004] 1) It is impossible to accurately predict the abnormal overload of distribution transformers during special periods such as the Spring Festival and other major holidays when people return home;
[0005] 2) In the prediction of heavy overload anomalies of distribution transformers, the data set often has sample imbalance, that is, the number of samples of heavy overload events is relatively small, which may lead to the model's better prediction effect for normal situations, but poor prediction effect for heavy overload situations. Summary of the invention
[0006] The present invention aims to solve at least one of the technical problems existing in the prior art or related art.
[0007] To this end, the purpose of the present invention is to provide a method and device for identifying abnormal distribution transformers taking into account sample imbalance processing, which can solve the problem of low accuracy of the distribution transformer heavy overload prediction model caused by actual sample imbalance such as too many samples of normally operating distribution transformers and too few samples of heavily overloaded distribution transformers.
[0008] To achieve the above-mentioned purpose, the technical solution of the first aspect of the present invention provides a method for identifying abnormal distribution transformers taking into account sample unbalance processing, comprising:
[0009] Obtaining daily data of distribution transformers in a target area and population data of the target area;
[0010] Screening out indicator features for judging whether there is a heavy overload in the daily data;
[0011] The rate of change in the number of people returning home calculated based on the population data;
[0012] Classify the daily data into non-heavy overload data, heavy load data and overload data according to the indicator characteristics and the change rate of the number of people returning home, and obtain the daily classification data corresponding to the distribution transformer;
[0013] Normalizing the daily classified data to obtain daily classified normalized data, and dividing the daily classified normalized data into a training set and a test set;
[0014] Performing clustering and stratified undersampling on the non-severely overloaded data in the training set and the test set according to the K-Means clustering algorithm to obtain non-severely overloaded distribution transformer samples in the training set and the test set;
[0015] Oversampling the heavy load data and the overload data according to the SMOTE oversampling algorithm to obtain heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set;
[0016] Model training is performed on non-severely overloaded distribution transformer samples, heavily loaded distribution transformer samples and overloaded distribution transformer samples in the training set and the test set according to the SVM algorithm optimized by the particle swarm intelligence algorithm to obtain a distribution transformer abnormality prediction model;
[0017] According to the distribution transformer abnormality prediction model, it is predicted whether the distribution transformer will have abnormal behavior of heavy load or overload at the target time.
[0018] In the above technical solution, preferably, the daily data includes:
[0019]
[0020] The indicator characteristics include:
[0021]
[0022]
[0023] In the above technical solution, preferably, the expression of the rate of change of the number of returning people is:
[0024] Among them, γ represents the change rate of the number of people returning home, α represents the population data during holidays, and β represents the population data before holidays.
[0025] In the above technical solution, preferably, the expression of the normalization process is:
[0026] Normalization Among them, X i Represented as the i-th indicator feature, X min It is represented as the minimum value of all the daily classification data corresponding to the i-th indicator feature, X max It is represented as the maximum value of all the daily classification data corresponding to the i-th indicator feature.
[0027] In the above technical solution, preferably, clustering and stratifying undersampling the non-heavy overload data in the training set and the test set according to the K-Means clustering algorithm includes:
[0028] Use K-Means clustering algorithm to cluster the daily classified data and divide the data into k clusters, where the optimal k value is calculated by the following formula:
[0029]
[0030] E is the error sum of squares, which is the sum of the squares of the distances between the centroid of the cluster and its internal members, representing the quality of the clustering effect. i is the i-th cluster, p is C i The sample points in m i It is C i The center of mass, C i The mean of all samples in ;
[0031] Select representative samples: Select representative samples from each cluster by randomly selecting multiple samples in the cluster without replacement.
[0032] In the above technical solution, preferably, oversampling the heavy load data and the overload data according to the SMOTE oversampling algorithm includes:
[0033] For each sample x in the heavy data and the overloaded data, respectively, the distance between all samples in the heavy data and the overloaded data is calculated based on the Euclidean distance to obtain its nearest neighbors;
[0034] Determine the sampling rate U, and for each of the overload data and the overload data y, randomly select a number of samples from its neighbors, and the selected neighbors are recorded as y n .
[0035] For each randomly selected neighbor y n , respectively, construct new reload data and overload data with the original reload data and overload data according to the following formula:
[0036] y new =y+rand(0,1)×yy n ; rand(0,1) means generating a random number between 0 and 1.
[0037] The technical solution of the second aspect of the present invention provides an abnormal distribution transformer identification system considering sample unbalance processing, comprising:
[0038] An acquisition module is configured to acquire daily data of distribution transformers in a target area and population data of the target area;
[0039] A screening module, configured to screen out indicator features for determining whether the daily data is overloaded;
[0040] A return home change rate calculation module is configured to calculate the return home number change rate based on the population data;
[0041] A classification module is configured to classify the daily data into non-heavy overload data, heavy load data and overload data according to the indicator characteristics and the rate of change of the number of returnees, so as to obtain daily classification data corresponding to the distribution transformer;
[0042] A normalization calculation module is configured to perform normalization processing on the daily classified data to obtain daily classified normalized data, and divide the daily classified normalized data into a training set and a test set;
[0043] A clustering stratified undersampling module is configured to perform clustering stratified undersampling on the non-heavy overload data in the training set and the test set according to a K-Means clustering algorithm to obtain non-heavy overload distribution transformer samples in the training set and the test set;
[0044] An oversampling module is configured to oversample the heavy load data and the overload data according to a SMOTE oversampling algorithm to obtain heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set;
[0045] an abnormality prediction model building module, configured to perform model training on non-heavy overload distribution transformer samples, heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set according to the SVM algorithm optimized by the particle swarm intelligence algorithm, so as to obtain an abnormality prediction model for the distribution transformer;
[0046] The heavy overload abnormality judgment module is configured to predict whether the distribution transformer will have heavy load or overload abnormal behavior at a target time according to the distribution transformer abnormality prediction model.
[0047] The technical solution of the third aspect of the present invention provides a readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the abnormal distribution transformer identification method considering sample unbalance processing provided by the technical solution of the first aspect are implemented.
[0048] The technical solution of the fourth aspect of the present invention provides a computer device, including a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to implement the steps of the abnormal distribution transformer identification method considering sample imbalance processing provided by the technical solution of the first aspect above.
[0049] Compared with the prior art, the advantages of the abnormal distribution transformer identification method and device taking into account sample imbalance processing provided by the present invention are: 1. Improving the accuracy of distribution transformer heavy overload prediction during major holidays such as the Spring Festival when people return home; 2. Solving the problem of low accuracy of the distribution transformer heavy overload prediction model caused by actual sample imbalance, such as too many samples of normally operating distribution transformers and too few samples of heavily overloaded distribution transformers. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:
[0051] Figure 1 A flowchart of a method for identifying an abnormal distribution transformer according to an embodiment of the present invention is shown;
[0052] Figure 2 The structure block diagram of the abnormal distribution transformer identification system involved in the embodiment of the present invention is shown;
[0053] Figure 3 The figure shows the calculation results of the change rate of the number of returning home residents in a certain area in the past three years involved in the embodiment of the present invention. DETAILED DESCRIPTION
[0054] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.
[0056] like Figure 1 As shown, according to an embodiment of the present invention, a method for identifying abnormal distribution transformers taking into account sample unbalance processing includes:
[0057] S1, obtaining daily data of distribution transformers in a target area and population data of the target area;
[0058] S2, screening out indicator features for determining whether there is a severe overload in the daily data;
[0059] S3, calculating the change rate of the number of people returning home based on the population data;
[0060] S4, classifying the daily data into non-heavy overload data, heavy load data and overload data according to the indicator characteristics and the change rate of the number of returning home people, and obtaining daily classification data corresponding to the distribution transformer;
[0061] In this step, the classified label information is as follows:
[0062] category Label Non-heavy overload 0 Overload 1 Overload 2 .
[0063] S5, normalizing the daily classified data to obtain daily classified normalized data, and dividing the daily classified normalized data into a training set and a test set;
[0064] In this step, the data set of a certain embodiment is divided into data set information as shown in the following table:
[0065]
[0066] S6, performing clustering and stratified undersampling on the non-severely overloaded data in the training set and the test set according to a K-Means clustering algorithm to obtain non-severely overloaded distribution transformer samples in the training set and the test set;
[0067] S7, oversampling the heavy load data and the overload data according to the SMOTE oversampling algorithm to obtain heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set;
[0068] In this step, the data set after imbalance processing in the above embodiment is shown in the following table:
[0069]
[0070] It can be seen from the above table that after the unbalanced sample processing, the number of training set samples of non-heavy overload distribution transformers decreased from 180,503 to 3,000, the number of training set samples of heavy load distribution transformers increased from 2,547 to 2,978, and the number of training set samples of overload distribution transformers increased from 612 to 2,978. After the unbalanced sample processing, the number of samples of non-heavy overload and heavy load and overload distribution transformers is equivalent.
[0071] S8, performing model training on the non-severely overloaded distribution transformer samples, heavily loaded distribution transformer samples and overloaded distribution transformer samples in the training set and the test set according to the SVM algorithm optimized by the particle swarm intelligence algorithm, to obtain a distribution transformer abnormality prediction model;
[0072] In this step, the model is selected and built, and the support vector machine (SVM) is selected as the prediction model, and the two hyperparameters of SVM (regularization coefficient and kernel parameter) are optimized and determined using the particle swarm intelligence algorithm PSO.
[0073] S9, predicting whether the distribution transformer will have abnormal behavior such as heavy load or overload at the target time according to the distribution transformer abnormality prediction model.
[0074] In this embodiment, the non-heavy overload distribution transformer samples, heavy load distribution transformer samples and overload distribution transformer samples in the training set are used as the input of the PSO-SVM model for training. After completion, the generalization ability of the PSO-SVM model is verified with the test set. The performance evaluation indicators are correctness, precision and F1 score.
[0075] The performance of the algorithm on the test set when the sample imbalance is not handled
[0076]
[0077] The performance of the algorithm on the test set after sample imbalance processing
[0078]
[0079] Gradient Boosting Machine (GBM), Classification And Regression Tree (CART), and Random Forest (RF) are used as comparison models.
[0080] It can be seen from the above two tables that after the undersampling and oversampling processing of sample imbalance, the prediction accuracy, precision and F1 score of the PSO-SVM model in the two types of abnormal distribution transformers, namely, heavy load and overload, are significantly improved.
[0081] In the above embodiment, preferably, the daily report data includes:
[0082]
[0083] The indicator characteristics include:
[0084]
[0085]
[0086] In the above embodiment, preferably, the expression of the rate of change of the number of returning home people is:
[0087] Among them, γ represents the change rate of the number of people returning home, α represents the population data during holidays, and β represents the population data before holidays.
[0088] Take the change in the number of people in a certain area before and during the Spring Festival in the past three years as an example:
[0089]
[0090] The calculation results are as follows: Figure 3 shown.
[0091] In the above embodiment, preferably, the expression of the normalization process is:
[0092] Normalization Among them, X i Represented as the i-th indicator feature, X min It is represented as the minimum value of all the daily classification data corresponding to the i-th indicator feature, X max It is represented as the maximum value of all the daily classification data corresponding to the i-th indicator feature.
[0093] In the above embodiment, preferably, clustering and stratifying undersampling the non-heavy overload data in the training set and the test set according to the K-Means clustering algorithm includes:
[0094] Use K-Means clustering algorithm to cluster the daily classified data and divide the data into k clusters, where the optimal k value is calculated by the following formula:
[0095]
[0096] E is the error sum of squares, which is the sum of the squares of the distances between the centroid of the cluster and its internal members, representing the quality of the clustering effect. i is the i-th cluster, p is C i The sample points in m i It is C i The center of mass, C i The mean of all samples in ;
[0097] Select representative samples: Select representative samples from each cluster by randomly selecting multiple samples in the cluster without replacement.
[0098] In the above embodiment, preferably, oversampling the heavy load data and the overload data according to the SMOTE oversampling algorithm includes:
[0099] For each sample x in the heavy data and the overloaded data, respectively, the distance between all samples in the heavy data and the overloaded data is calculated based on the Euclidean distance to obtain its nearest neighbors;
[0100] Determine the sampling rate, and for each of the heavy load data and the overload data y, randomly select a number of samples from its neighbors, and the selected neighbors are recorded as y n .
[0101] For each randomly selected neighbor y n , respectively, construct new reload data and overload data with the original reload data and overload data according to the following formula:
[0102] y new =y+rand(0,1)×yy n ; rand(0,1) means generating a random number between 0 and 1.
[0103] like Figure 2 As shown, according to another embodiment of the present invention, an abnormal distribution transformer identification system 100 considering sample unbalance processing includes:
[0104] An acquisition module 1 is configured to acquire daily data of distribution transformers in a target area and population data of the target area;
[0105] A screening module 2 is configured to screen out indicator features for determining whether the daily data is overloaded;
[0106] A return home change rate calculation module 3 is configured to calculate the return home number change rate according to the population data;
[0107] The classification module 4 is configured to classify the daily data into non-heavy overload data, heavy load data and overload data according to the indicator characteristics and the change rate of the number of returnees, and obtain the daily classification data corresponding to the distribution transformer;
[0108] A normalization calculation module 5 is configured to perform normalization processing on the daily classified data to obtain daily classified normalized data, and divide the daily classified normalized data into a training set and a test set;
[0109] The clustering stratified undersampling module 6 is configured to perform clustering stratified undersampling on the non-heavy overload data in the training set and the test set according to the K-Means clustering algorithm to obtain non-heavy overload distribution transformer samples in the training set and the test set;
[0110] An oversampling module 7 is configured to oversample the heavy load data and the overload data according to a SMOTE oversampling algorithm to obtain heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set;
[0111] The abnormality prediction model building module 8 is configured to perform model training on the non-heavy overload distribution transformer samples, heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set according to the SVM algorithm and the particle swarm intelligence algorithm to obtain the distribution transformer abnormality prediction model;
[0112] The heavy overload abnormality judgment module 9 is configured to predict whether the distribution transformer has heavy load or overload abnormal behavior at the target time according to the distribution transformer abnormality prediction model.
[0113] Based on the above Figure 1 The method shown, accordingly, an embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the abnormal distribution transformer identification method considering sample imbalance processing in any of the above embodiments are implemented.
[0114] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present application.
[0115] Based on the above Figure 1 The method shown, and Figure 2 The virtual device embodiment shown, in order to achieve the above-mentioned purpose, the embodiment of the present application also provides a computer device, including a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to implement the steps of the abnormal distribution transformer identification method considering sample imbalance processing in any of the above-mentioned embodiments.
[0116] Optionally, the computer device may further include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a Bluetooth interface, a WI-FI interface), etc.
[0117] Those skilled in the art will appreciate that the computer device structure provided in this embodiment does not limit the computer device, and may include more or fewer components, or a combination of certain components, or different component arrangements.
[0118] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages and saves the hardware and software resources of the computer device, and supports the operation of information processing programs and other software and / or programs. The network communication module is used to realize communication between the components inside the storage medium, and communication with other hardware and software in the physical device.
[0119] In the present invention, the terms "first", "second", and "third" are used for descriptive purposes only and should not be understood as indicating or implying relative importance; the term "plurality" refers to two or more, unless otherwise clearly defined. The terms "installed", "connected", "connected", "fixed", etc. should be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; "connected" can be a direct connection or an indirect connection through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0120] In the description of the present invention, it is necessary to understand that the directions or positional relationships indicated by terms such as “upper”, “lower”, “left”, “right”, “front” and “back” are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or unit referred to must have a specific direction, be constructed and operated in a specific orientation, and therefore, cannot be understood as a limitation on the present invention.
[0121] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "specific embodiments", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0122] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for identifying abnormal distribution transformers considering sample unbalance processing, characterized in that: include: Obtaining daily data of distribution transformers in a target area and population data of the target area; Screening out indicator features for judging whether there is a heavy overload in the daily data; The rate of change in the number of people returning home calculated based on the population data; Classify the daily data into non-heavy overload data, heavy load data and overload data according to the indicator characteristics and the change rate of the number of people returning home, and obtain the daily classification data corresponding to the distribution transformer; Normalizing the daily classified data to obtain daily classified normalized data, and dividing the daily classified normalized data into a training set and a test set; Performing clustering and stratified undersampling on the non-severely overloaded data in the training set and the test set according to the K-Means clustering algorithm to obtain non-severely overloaded distribution transformer samples in the training set and the test set; Oversampling the heavy load data and the overload data according to the SMOTE oversampling algorithm to obtain heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set; Model training is performed on non-severely overloaded distribution transformer samples, heavily loaded distribution transformer samples and overloaded distribution transformer samples in the training set and the test set according to the SVM algorithm optimized by the particle swarm intelligence algorithm to obtain a distribution transformer abnormality prediction model; According to the distribution transformer abnormality prediction model, it is predicted whether the distribution transformer will have abnormal behavior of heavy load or overload at the target time.
2. The abnormal distribution transformer identification method considering sample unbalance processing according to claim 1 is characterized in that: The daily data includes: The indicator characteristics include:
3. The abnormal distribution transformer identification method considering sample unbalance processing according to claim 1 is characterized in that: The expression of the change rate of the number of returning home is: Among them, γ represents the change rate of the number of people returning home, α represents the population data during holidays, and β represents the population data before holidays.
4. The abnormal distribution transformer identification method considering sample unbalance processing according to claim 1 is characterized in that: The expression of the normalization process is: Normalization Among them, X i Represented as the i-th indicator feature, X min It is represented as the minimum value of all the daily classification data corresponding to the i-th indicator feature, X max It is represented as the maximum value of all the daily classification data corresponding to the i-th indicator feature.
5. The abnormal distribution transformer identification method considering sample unbalance processing according to claim 1 is characterized in that: Clustering and stratifying undersampling the non-heavy overload data in the training set and the test set according to the K-Means clustering algorithm includes: Use K-Means clustering algorithm to cluster the daily classified data and divide the data into k clusters, where the optimal k value is calculated by the following formula: E is the error sum of squares, which is the sum of the squares of the distances between the centroid of the cluster and its internal members, representing the quality of the clustering effect. i is the i-th cluster, p is C i The sample points in m i It is C i The center of mass, C i The mean of all samples in ; Select representative samples: Select representative samples from each cluster by randomly selecting multiple samples in the cluster without replacement.
6. The abnormal distribution transformer identification method considering sample unbalance processing according to claim 5 is characterized in that: Oversampling the heavy load data and the overload data according to the SMOTE oversampling algorithm includes: For each sample x in the heavy data and the overloaded data, respectively, the distance between all samples in the heavy data and the overloaded data is calculated based on the Euclidean distance to obtain its nearest neighbors; Determine the sampling rate, and for each of the heavy load data and the overload data y, randomly select a number of samples from its neighbors, and the selected neighbors are recorded as y n . For each randomly selected neighbor y n , respectively, construct new reload data and overload data with the original reload data and overload data according to the following formula: y new =y+rand(0,1)×|yy n |; rand(0,1) means generating a random number between 0 and 1.
7. An abnormal distribution transformer identification system considering sample unbalance processing, characterized in that: include: An acquisition module is configured to acquire daily data of distribution transformers in a target area and population data of the target area; A screening module, configured to screen out indicator features for determining whether the daily data is overloaded; A return home change rate calculation module is configured to calculate the return home number change rate based on the population data; A classification module is configured to classify the daily data into non-heavy overload data, heavy load data and overload data according to the indicator characteristics and the change rate of the number of returnees, so as to obtain daily classification data corresponding to the distribution transformer; A normalization calculation module is configured to perform normalization processing on the daily classified data to obtain daily classified normalized data, and divide the daily classified normalized data into a training set and a test set; A clustering stratified undersampling module is configured to perform clustering stratified undersampling on the non-heavy overload data in the training set and the test set according to a K-Means clustering algorithm to obtain non-heavy overload distribution transformer samples in the training set and the test set; An oversampling module is configured to oversample the heavy load data and the overload data according to a SMOTE oversampling algorithm to obtain heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set; an abnormality prediction model building module, configured to perform model training on non-heavy overload distribution transformer samples, heavy load distribution transformer samples and overload distribution transformer samples in the training set and the test set according to the SVM algorithm optimized by the particle swarm intelligence algorithm, so as to obtain an abnormality prediction model for the distribution transformer; The heavy overload abnormality judgment module is configured to predict whether the distribution transformer will have heavy load or overload abnormal behavior at a target time according to the distribution transformer abnormality prediction model.
8. A readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the steps of the abnormal distribution transformer identification method considering sample unbalance processing as described in any one of claims 1 to 6 are implemented.
9. A computer device, characterized in that: The method comprises a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to implement the steps of the abnormal distribution transformer identification method considering sample unbalance processing as described in any one of claims 1 to 6.