Transformer area line loss rate anomaly detection method and system based on improved grey correlation analysis and isolated forest model

Through improved gray correlation analysis and isolated forest model, the line loss rate detection is simplified by weighted correlation degree and line loss rate feature number, solving the problem of complex and time-consuming detection in the existing technology, and achieving fast and accurate line loss rate abnormal judgment .

CN119066533APending Publication Date: 2024-12-03STATE GRID ANHUI ELECTRIC POWER CO LTD MARKETING SERVICE CENTER +1

Patent Information

Application Number
CN202411101811.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing linear loss rate abnormality detection process is complex and time-consuming, and requires a variety of electrical characteristic parameters to calculate and determine whether it is abnormal, and the overall detection efficiency is low.

Method used

The improved grey correlation analysis method is used to calculate the correlation between electrical characteristic parameters and line loss rate, and the line loss rate abnormality is judged by the isolated forest model, and the detection process is simplified by the weighted correlation degree and line loss rate feature number.

Benefits of technology

It realizes fast and accurate detection of abnormal line loss rate, reduces detection time, enhances the accuracy of judgment, and avoids data inaccuracy caused by the same weight of electrical characteristic parameters.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119066533A_ABST
    Figure CN119066533A_ABST
Patent Text Reader

Abstract

The invention discloses a transformer area line loss rate anomaly detection method and system based on improved grey correlation analysis and an isolated forest model. The method comprises the following steps of 1, collecting electrical characteristic parameters in a transformer area once per hour; 2, processing the electrical characteristic parameters according to the weighted correlation degree to obtain a line loss rate characteristic number; 3, inputting the obtained line loss rate characteristic number into an isolated forest model to obtain an abnormal score; and 4, judging whether the transformer area line loss rate is abnormal or not according to the abnormal score, if the transformer area line loss rate is normal, continuing to detect, and if the transformer area line loss rate is abnormal, giving an alarm. According to the invention, the line loss rate abnormity is detected through the line loss rate characteristic number, the accuracy of line loss rate detection is greatly improved, and the detection process and time are saved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of power grids, and in particular to a method and system for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and an isolation forest model. Background Art

[0002] Line loss is the sum of the energy losses incurred in the transmission, transformation, distribution and consumption stages during the process of transmitting electricity from the power grid to the user. The line loss rate is the ratio of energy loss to total power supply in a certain period of time. It is an important indicator for measuring the technical and economic feasibility of power grid enterprises. However, the existing line loss rate anomaly detection is basically carried out after the line loss rate is obtained through manual or computer calculation. During the detection, it is necessary to roughly calculate the line loss rate through a variety of electrical characteristic parameters, and then determine whether the line loss rate is abnormal. The overall detection process is complicated and time-consuming.

[0003] Existing patents, such as the power distribution line line loss rate calculation system and method (publication number CN101958544A), calculate the line loss rate based on the collected electrical characteristic parameters by the on-site management terminal. It requires many collection devices and the calculation of the line loss rate is too complicated. Summary of the invention

[0004] In order to at least partially solve the problem that the line loss rate detection process is time-consuming and complicated, the present invention provides a method and system for detecting line loss rate anomalies in a substation based on improved grey correlation analysis and isolation forest model. The correlation between multiple electrical characteristic parameters and line loss rate is obtained through improved grey correlation analysis method, and then the parameters are processed according to the correlation. The processed data is input into the constructed isolation forest model to determine whether the line loss rate is abnormal.

[0005] In order to achieve the above object, the technical solution of the present invention is:

[0006] In a first aspect, the present invention proposes a method for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model, comprising the following steps:

[0007] Step 1: Collect electrical characteristic parameters in the substation area once an hour to detect abnormal line loss rate;

[0008] Step 2: Process the electrical characteristic parameters according to the weighted correlation to obtain the line loss rate characteristic number, which is used to establish a correlation between the electrical characteristic parameters and the line loss rate, so as to facilitate the judgment of the abnormality of the line loss rate according to the line loss rate characteristic number;

[0009] Step 3: Input the obtained line loss rate feature number into the isolation forest model to obtain the abnormal score, so as to judge the line loss rate situation according to the abnormal score;

[0010] Step 4: Determine whether the line loss rate of the substation area is abnormal based on the abnormal score. If the line loss rate of the substation area is normal, continue to detect. If the line loss rate of the substation area is abnormal, issue a warning to prompt the staff to conduct timely maintenance.

[0011] In the step 2, the weighted correlation between the electrical characteristic parameter and the line loss rate is calculated according to the following formula:

[0012]

[0013] Among them, δ j (k) is the correlation coefficient of the jth electrical characteristic parameter in the kth hour, X j ′(k) is the jth electrical characteristic parameter at the kth hour, X0′(k) is the line loss rate corresponding to the kth hour, ρ is the resolution coefficient, w(k) is the correlation coefficient weight, n is the total number of the jth electrical characteristic parameters, γ j is the weighted correlation of the jth electrical characteristic parameter, 1≤j≤m, m is the total number of electrical characteristic parameter types, and the weighted correlation of electrical characteristic parameters can better represent the relationship between electrical characteristic parameters and line loss rate;

[0014] The correlation coefficient weight is expressed as follows:

[0015]

[0016] Among them, c(k) is the initial weight, is the average value of the jth electrical characteristic parameter at the kth hour, w(k) is the correlation coefficient weight, and n is the total number of the jth electrical characteristic parameters;

[0017] In the step 2, the line loss rate characteristic number is calculated according to the following formula:

[0018]

[0019] Among them, γ j is the weighted correlation of the jth electrical characteristic parameter, X j ′(k) is the jth electrical characteristic parameter at the kth hour, 1≤j≤m, m is the total number of electrical characteristic parameter types, and T(k) is the line loss rate characteristic number corresponding to the kth hour. It is convenient to judge the line loss rate anomaly by the line loss rate characteristic number, without calculating the line loss rate, making the line loss rate anomaly detection simpler.

[0020] Furthermore, in step 3, an isolation forest model is constructed by the following method:

[0021] Collect historical electrical characteristic parameters and line loss rates as training data;

[0022] Perform grey correlation analysis on historical electrical characteristic parameters to obtain the weighted correlation between electrical characteristic parameters and line loss rate, and obtain the line loss rate characteristic number based on the weighted correlation, which is convenient for subsequent line loss rate detection;

[0023] An initial isolation forest model is constructed and trained using the line loss rate feature number to obtain a trained isolation forest model, thereby ensuring the detection accuracy of the isolation forest model.

[0024] Furthermore, in step 3, the obtained line loss rate feature number is input into the isolation forest model to obtain anomaly scores, which specifically includes:

[0025] The line loss rate feature numbers collected within 24 hours and the normal line loss rate feature numbers of the substation area in the previous three days are combined to construct a feature number set to avoid the impact caused by too many abnormal situations within 24 hours;

[0026] The feature set is placed in the root node of the binary search tree, and a split attribute value is randomly selected between the maximum and minimum values ​​of the feature set. The data is classified according to the split attribute value to obtain the path length of the line loss rate feature number. The anomaly score is calculated based on the path length to determine the abnormality of the line loss rate.

[0027] The second aspect of the present invention proposes a system for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model, comprising:

[0028] The collection module is used to collect electrical characteristic parameters in the substation area once an hour to detect abnormal line loss rate;

[0029] A data processing module is used to process the electrical characteristic parameters according to the weighted correlation degree to obtain a line loss rate characteristic number, which is used to establish a correlation between the electrical characteristic parameters and the line loss rate, so as to facilitate the judgment of abnormal line loss rate according to the line loss rate characteristic number;

[0030] An evaluation module is used to input the obtained line loss rate feature number into the isolation forest model to obtain anomaly scores, so as to judge the line loss rate situation according to the anomaly scores;

[0031] The alarm module is used to determine whether the line loss rate of the substation area is abnormal according to the abnormal points. If the line loss rate of the substation area is normal, the detection will continue. If the line loss rate of the substation area is abnormal, an alarm will be issued to prompt the staff to conduct timely maintenance;

[0032] In the data processing module, the weighted correlation between the electrical characteristic parameter and the line loss rate is calculated according to the following formula:

[0033]

[0034] Among them, δ j (k) is the correlation coefficient of the jth electrical characteristic parameter in the kth hour, Xj ′(k) is the jth electrical characteristic parameter at the kth hour, X0′(k) is the line loss rate corresponding to the kth hour, ρ is the resolution coefficient, w(k) is the correlation coefficient weight, n is the total number of the jth electrical characteristic parameters, γ j is the weighted correlation of the jth electrical characteristic parameter, 1≤j≤m, m is the total number of electrical characteristic parameter types, and the weighted correlation of electrical characteristic parameters can better represent the relationship between electrical characteristic parameters and line loss rate;

[0035] The correlation coefficient weight is expressed as follows:

[0036]

[0037] Among them, c(k) is the initial weight, is the average value of the jth electrical characteristic parameter at the kth hour, w(k) is the correlation coefficient weight, and n is the total number of the jth electrical characteristic parameters;

[0038] In the data processing module, the line loss rate characteristic number is calculated according to the following formula:

[0039]

[0040] Among them, γ j is the weighted correlation of the jth electrical characteristic parameter, X j ′(k) is the jth electrical characteristic parameter at the kth hour, 1≤j≤m, m is the total number of electrical characteristic parameter types, and T(k) is the line loss rate characteristic number corresponding to the kth hour. It is convenient to judge the line loss rate anomaly by the line loss rate characteristic number, without calculating the line loss rate, making the line loss rate anomaly detection simpler.

[0041] Furthermore, in the evaluation module, an isolation forest model is constructed by the following method:

[0042] Collect historical electrical characteristic parameters and line loss rates as training data;

[0043] Perform grey correlation analysis on historical electrical characteristic parameters to obtain the weighted correlation between electrical characteristic parameters and line loss rate, and obtain the line loss rate characteristic number based on the weighted correlation, which is convenient for subsequent line loss rate detection;

[0044] An initial isolation forest model is constructed and trained using the line loss rate feature number to obtain a trained isolation forest model, thereby ensuring the detection accuracy of the isolation forest model.

[0045] Furthermore, in the evaluation module, the obtained line loss rate feature number is input into the isolation forest model to obtain anomaly scores, specifically including:

[0046] The line loss rate feature numbers collected within 24 hours and the normal line loss rate feature numbers of the substation area in the previous three days are combined to construct a feature number set to avoid the impact caused by too many abnormal situations within 24 hours;

[0047] The feature set is placed in the root node of the binary search tree, and a split attribute value is randomly selected between the maximum and minimum values ​​of the feature set. The data is classified according to the split attribute value to obtain the average path length of the line loss rate feature number. The anomaly score is calculated based on the average path length to determine the abnormality of the line loss rate.

[0048] Beneficial effects of the present invention:

[0049] (1) The present invention constructs an isolation forest model to detect line loss rate anomalies, calculates the path length of the line loss rate feature number, and then obtains the anomaly score based on the path length. The line loss rate anomaly is quickly and accurately identified based on the size of the anomaly score, thereby realizing rapid detection of line loss rate anomalies.

[0050] (2) The present invention establishes the relationship between the electrical characteristic parameters and the line loss rate through an improved grey correlation analysis method, and weights the electrical characteristic parameters when solving the correlation between the electrical characteristic parameters and the line loss rate, thereby avoiding inaccurate data caused by the same weights of the electrical characteristic parameters when subsequently calculating the line loss rate characteristic numbers based on the electrical characteristic parameters, thereby enhancing the accuracy of abnormal judgment of the line loss rate.

[0051] (3) The present invention performs anomaly detection by constructing a line loss rate characteristic number. The line loss rate characteristic number is obtained by multiplying the electrical characteristic parameter by the correlation degree of each type of electrical characteristic parameter and adding them together. The line loss rate characteristic number is used instead of the electrical characteristic parameter to input into the constructed isolation forest model. In this setting, only one line loss rate characteristic number needs to be input to obtain the line loss rate abnormality, which is convenient for rapid detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A flowchart of a method for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model is provided in an embodiment of the present invention.

[0053] Figure 2 A flowchart of a method for constructing an isolation forest model provided in an embodiment of the present invention.

[0054] Figure 3 An architectural diagram of a system for detecting abnormal line loss rate in a substation based on improved grey correlation analysis and isolation forest model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all 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.

[0056] Example 1

[0057] like Figure 1 As shown, a method for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model includes the following steps:

[0058] S101: Collect electrical characteristic parameters in the substation once every hour.

[0059] S102: Processing the electrical characteristic parameters according to the weighted correlation to obtain a line loss rate characteristic number.

[0060] S103: Input the obtained line loss rate feature number into the isolation forest model to obtain an abnormal score.

[0061] S104: Determine whether the line loss rate of the substation area is abnormal based on the abnormal score. If the line loss rate of the substation area is normal, continue to detect. If the line loss rate of the substation area is abnormal, issue a warning.

[0062] Example 2

[0063] Based on the above embodiments, Figure 2 As shown, the embodiment of the present invention provides a method for constructing a grey relational analysis and an isolation forest model, which specifically includes:

[0064] S201: Collect historical electrical characteristic parameters and line loss rate as training data. Specifically, the electrical characteristic parameters include power consumption in the substation, active power in the substation, power supply in the substation, reactive power in the substation, and line length in the substation.

[0065] S202: Perform grey correlation analysis on the collected historical electrical characteristic parameters to obtain the weighted correlation between the electrical characteristic parameters and the line loss rate. Specifically, the weighted correlation between the electrical characteristic parameters and the line loss rate is calculated according to the following formula:

[0066]

[0067] Among them, δ j (k) is the correlation coefficient of the jth electrical characteristic parameter in the kth hour, X j′(k) is the jth electrical characteristic parameter at the kth hour, X0′(k) is the line loss rate corresponding to the kth hour, ρ is the resolution coefficient, w(k) is the correlation coefficient weight, n is the total number of the jth electrical characteristic parameters, γ j is the weighted correlation degree of the jth electrical characteristic parameter, 1≤j≤m, and m is the total number of electrical characteristic parameter types.

[0068] After weighting the electrical characteristic parameters, the correlation between them and the line loss rate can be solved to more accurately obtain the correlation between each electrical characteristic parameter and the line loss rate, which is convenient for subsequent line loss rate abnormality detection. The correlation coefficient weight can distinguish the degree to which different types of electrical characteristic parameters affect the line loss rate, avoiding treating all electrical characteristic parameters as having the same impact on the line loss rate, thereby causing large errors. The correlation coefficient weight is expressed by the following formula:

[0069]

[0070] Among them, c(k) is the initial weight, is the average value of the j-th electrical characteristic parameter at the k-th hour, w(k) is the correlation coefficient weight, and n is the total number of the j-th electrical characteristic parameter.

[0071] The electrical characteristic parameters are processed according to the weighted correlation. Specifically, the line loss rate characteristic number is calculated according to the following formula:

[0072]

[0073] Among them, γ j is the weighted correlation of the jth electrical characteristic parameter, X j (k) is the jth electrical characteristic parameter at the kth hour, 1≤j≤m, m is the total number of electrical characteristic parameter types, and T(k) is the line loss rate characteristic number corresponding to the kth hour.

[0074] S203: Train the initial isolation forest model using the line loss rate feature number to obtain a trained isolation forest model.

[0075] Example 3

[0076] On the basis of the above embodiments, the embodiment of the present invention provides a method for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model, which specifically includes:

[0077] S301: Collect electrical characteristic parameters in the substation once every hour. The electrical characteristic parameters include the power consumption, active power, power supply, reactive power and line length in the substation. Pre-process the electrical characteristic parameters and calibrate and fill in the electrical characteristic parameters with abnormalities or missing data.

[0078] S302: Process the electrical characteristic parameters according to the weighted correlation to obtain the line loss rate characteristic number. Process according to the following formula:

[0079]

[0080] Among them, c(k) is the initial weight, is the average value of the j-th electrical characteristic parameter at the k-th hour, w(k) is the correlation coefficient weight, and n is the total number of the j-th electrical characteristic parameter.

[0081]

[0082] Among them, δ j (k) is the correlation coefficient of the jth electrical characteristic parameter in the kth hour, X j ′(k) is the jth electrical characteristic parameter at the kth hour, X0′(k) is the line loss rate corresponding to the kth hour, ρ is the resolution coefficient, w(k) is the correlation coefficient weight, n is the total number of the jth electrical characteristic parameters, γ j is the weighted correlation degree of the jth electrical characteristic parameter, 1≤j≤m, and m is the total number of electrical characteristic parameter types.

[0083]

[0084] Among them, γ j is the weighted correlation of the jth electrical characteristic parameter, X j (k) is the jth electrical characteristic parameter at the kth hour, 1≤j≤m, m is the total number of electrical characteristic parameter types, and T(k) is the line loss rate characteristic number corresponding to the kth hour.

[0085] S303: Input the obtained line loss rate feature number into the isolation forest model to obtain anomaly scores. The specific process is as follows:

[0086] The electrical characteristic parameters are collected once every 24 hours and the line loss rate characteristic numbers are calculated based on the electrical characteristic parameters. The 24 line loss rate characteristic numbers are combined with the normal line loss rate characteristic numbers of the substation in the previous three days to construct a characteristic number set, and the characteristic number set is input into the isolation forest model.

[0087] Put the feature set into the root node of the binary search tree, randomly select a split attribute value p between the maximum and minimum values ​​of the feature set, classify the data according to the split attribute value p, and put the sample data points whose feature values ​​are less than the split attribute value p in the left branch of the binary search tree, and the sample data points whose feature values ​​are greater than or equal to the split attribute value p in the right branch of the binary search tree. Recursively divide the left branch and the right branch until the following conditions are met:

[0088] The feature set cannot be further divided, or the height of the binary search tree reaches the preset limit.

[0089] Then, according to the divided results, the anomaly score is calculated according to the following formula:

[0090]

[0091] Where c(n) is the average path length of a binary search tree constructed with n samples, n is the number of samples, h(x) is the length of the scalable vector path of the sample on the tree, E(h(x)) is the mean length of the scalable vector path of the sample on t trees, and s(x, n) is the anomaly score.

[0092] S304: Determine whether the line loss rate of the substation area is abnormal based on the abnormality score. If the line loss rate of the substation area is normal, continue to detect. If the line loss rate of the substation area is abnormal, issue a warning. Specifically, if the abnormality score is close to 1, then it must be an abnormal point, and the line loss rate at this time is abnormal. If the abnormality score is less than 0.5, then it is definitely not an abnormal point, and the line loss rate is normal. If the abnormality scores are all around 0.5, then there is no abnormal point in the sample, and the line loss rate is normal.

[0093] Example 4

[0094] Corresponding to the above method, such as Figure 3 As shown, an embodiment of the present invention provides a system for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model, comprising:

[0095] The collection module is used to collect the electrical characteristic parameters in the substation once an hour.

[0096] The data processing module is used to process the electrical characteristic parameters according to the weighted correlation degree to obtain the line loss rate characteristic number.

[0097] The evaluation module is used to input the obtained line loss rate feature number into the isolation forest model to obtain the anomaly score.

[0098] The alarm module is used to determine whether the line loss rate of the substation area is abnormal based on the abnormal points. If the line loss rate of the substation area is normal, the detection will continue; if the line loss rate of the substation area is abnormal, an alarm will be issued.

[0099] It should be noted that the embodiment of the present invention provides a substation line loss rate anomaly detection system based on improved grey correlation analysis and isolation forest model, which is intended to realize the above-mentioned substation line loss rate anomaly detection method based on improved grey correlation analysis and isolation forest model. Its specific functions can be referred to the above-mentioned method embodiments, which will not be repeated here.

[0100] In summary, the present invention performs abnormal detection of line loss rate by constructing an isolated forest model, calculates the path length of the line loss rate characteristic number, obtains the abnormal score according to the path, and quickly and accurately identifies the abnormal situation of the line loss rate according to the size of the abnormal score, thereby realizing the rapid detection of abnormal line loss rate. The present invention establishes the relationship between the electrical characteristic parameters and the line loss rate through the gray correlation analysis method, and weights the electrical characteristic parameters when solving the correlation between the electrical characteristic parameters and the line loss rate, thereby avoiding the inaccurate data caused by the same weight of the electrical characteristic parameters when the line loss rate characteristic number is calculated according to the electrical characteristic parameters, thereby enhancing the accuracy of the abnormal judgment of the line loss rate. The present invention performs abnormal detection by constructing a line loss rate characteristic number, obtains the line loss rate characteristic number by multiplying the correlation degree of each electrical characteristic parameter with the electrical characteristic parameter and adding them together, and uses the line loss rate characteristic number instead of the electrical characteristic parameter to input into the constructed isolated forest model, so that only one line loss rate characteristic number needs to be input to obtain the abnormal situation of the line loss rate, which is convenient for rapid detection.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model, characterized in that: The following steps are involved: Step 1: Collect electrical characteristic parameters in the substation once every hour; Step 2: Process the electrical characteristic parameters according to the weighted correlation to obtain the line loss rate characteristic number; Step 3: Input the obtained line loss rate feature number into the isolation forest model to obtain the anomaly score; Step 4: Determine whether the line loss rate of the substation area is abnormal based on the abnormal score. If the line loss rate of the substation area is normal, continue to detect. If the line loss rate of the substation area is abnormal, issue a warning. In the step 2, the weighted correlation between the electrical characteristic parameter and the line loss rate is calculated according to the following formula: Among them, δ j (k) is the correlation coefficient of the jth electrical characteristic parameter at the kth hour, X′ j (k) is the jth electrical characteristic parameter at the kth hour, X′0(k) is the line loss rate corresponding to the kth hour, ρ is the resolution coefficient, w(k) is the correlation coefficient weight, n is the total number of the jth electrical characteristic parameters, γ j is the weighted correlation degree of the jth electrical characteristic parameter, 1≤j≤m, m is the total number of electrical characteristic parameter types; The correlation coefficient weight is expressed as follows: Among them, c(k) is the initial weight, is the average value of the jth electrical characteristic parameter at the kth hour, w(k) is the correlation coefficient weight, and n is the total number of the jth electrical characteristic parameters; In the step 2, the line loss rate characteristic number is calculated according to the following formula: Among them, γ j is the weighted correlation of the jth electrical characteristic parameter, X′ j (k) is the jth electrical characteristic parameter at the kth hour, 1≤j≤m, m is the total number of electrical characteristic parameter types, and T(k) is the line loss rate characteristic number corresponding to the kth hour.

2. According to claim 1, a method for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model, characterized in that: In step 3, the isolation forest model is constructed by the following method: Collect historical electrical characteristic parameters and line loss rates as training data; Perform grey correlation analysis on historical electrical characteristic parameters to obtain the weighted correlation between the electrical characteristic parameters and the line loss rate, and obtain the line loss rate characteristic number based on the weighted correlation; An initial isolation forest model is constructed and trained using the line loss rate feature number to obtain a trained isolation forest model.

3. According to claim 1, a method for detecting abnormal line loss rate in a transformer area based on improved grey correlation analysis and isolation forest model is characterized in that: In the step 3, the obtained line loss rate feature number is input into the isolation forest model to obtain anomaly scores, which specifically includes: The line loss rate feature numbers collected within 24 hours and the normal line loss rate feature numbers of the substation area in the previous three days are combined to form a feature number set; The feature set is placed in the root node of the binary search tree, and a split attribute value is randomly selected between the maximum and minimum values ​​of the feature set. The data is classified according to the split attribute value to obtain the path length of the line loss rate feature number, and the anomaly score is calculated based on the path length.

4. A line loss rate anomaly detection system for a substation area based on improved grey correlation analysis and isolation forest model, characterized in that: include: A collection module is used to collect electrical characteristic parameters in the substation once an hour; A data processing module is used to process the electrical characteristic parameters according to the weighted correlation degree to obtain a line loss rate characteristic number; An evaluation module is used to input the obtained line loss rate feature number into the isolation forest model to obtain anomaly scores; The alarm module is used to determine whether the line loss rate of the substation area is abnormal according to the abnormal points. If the line loss rate of the substation area is normal, the detection will continue. If the line loss rate of the substation area is abnormal, an alarm will be issued; In the data processing module, the weighted correlation between the electrical characteristic parameter and the line loss rate is calculated according to the following formula: Among them, δ j (k) is the correlation coefficient of the jth electrical characteristic parameter at the kth hour, X′ j (k) is the jth electrical characteristic parameter at the kth hour, X′0(k) is the line loss rate corresponding to the kth hour, ρ is the resolution coefficient, w(k) is the correlation coefficient weight, n is the total number of the jth electrical characteristic parameters, γ j is the weighted correlation degree of the jth electrical characteristic parameter, 1≤j≤m, m is the total number of electrical characteristic parameter types; The correlation coefficient weight is expressed as follows: Among them, c(k) is the initial weight, is the average value of the jth electrical characteristic parameter at the kth hour, w(k) is the correlation coefficient weight, and n is the total number of the jth electrical characteristic parameters; In the data processing module, the line loss rate characteristic number is calculated according to the following formula: Among them, γ j is the weighted correlation of the jth electrical characteristic parameter, X′ j (k) is the jth electrical characteristic parameter at the kth hour, 1≤j≤m, m is the total number of electrical characteristic parameter types, and T(k) is the line loss rate characteristic number corresponding to the kth hour.

5. According to claim 4, a substation line loss rate anomaly detection system based on improved grey relational analysis and isolation forest model is characterized in that: In the evaluation module, the isolation forest model is constructed by the following method: Collect historical electrical characteristic parameters and line loss rates as training data; Perform grey correlation analysis on historical electrical characteristic parameters to obtain the weighted correlation between the electrical characteristic parameters and the line loss rate, and obtain the line loss rate characteristic number based on the weighted correlation; An initial isolation forest model is constructed and trained using the line loss rate feature number to obtain a trained isolation forest model.

6. According to claim 4, a substation line loss rate anomaly detection system based on improved grey correlation analysis and isolation forest model is characterized in that: In the evaluation module, the obtained line loss rate feature number is input into the isolation forest model to obtain anomaly scores, specifically including: The line loss rate feature numbers collected within 24 hours and the normal line loss rate feature numbers of the substation area in the previous three days are combined to form a feature number set; The feature set is placed in the root node of the binary search tree, and a split attribute value is randomly selected between the maximum and minimum values ​​of the feature set. The data is classified according to the split attribute value to obtain the path length of the line loss rate feature number, and the anomaly score is calculated based on the path length.

Citation Information

Patent Citations

  • System and method for calculating line loss rate of electric power distribution line

    CN101958544A

Cited By

  • District line loss rate abnormity diagnosis method based on data driving

    CN120850156A