A power quality disturbance feature screening method based on multi-index fusion evaluation

Abstract

The present application relates to a kind of based on multi-index fusion evaluation power quality disturbance feature screening method, comprising: step 1: input power quality disturbance signal, extract multiple power quality disturbance features, construct power quality disturbance feature universe;Step 2: based on the fusion evaluation index of intersection degree index, redundancy index, separation index, using Cuckoo search method is searched and each dimension power quality disturbance feature subset selected is obtained;Step 3: based on cost factor, the selected power quality disturbance feature subset is evaluated, and the optimal power quality disturbance feature subset is obtained, and based on the optimal power quality disturbance feature subset, power quality disturbance signal is classified.The method is conducive to improving power quality disturbance classification efficiency and accuracy.

Description

Technical Field

[0001] This invention belongs to the field of power quality disturbance classification, and specifically relates to a method for optimizing power quality disturbance characteristics based on multi-index fusion evaluation. Background Technology

[0002] With a high proportion of new energy sources being integrated into the power grid, the power quality problems faced by the power system are becoming increasingly complex. Power quality disturbances reduce the efficiency of power utilization and the lifespan of electrical equipment, leading to a decline in the reliability of power grid supply and even causing power accidents. Efficient and accurate classification of power quality disturbance signals helps to uncover the root causes of disturbances and implement targeted power quality management measures. Traditional power quality disturbance classification methods mainly extract features through signal analysis and directly use these features for classification. However, this process neglects the quality of the feature set, resulting in poor classification performance. Therefore, feature optimization is necessary before classifying power quality disturbances. Feature optimization methods mainly include encapsulated feature optimization, filtering feature optimization, and embedded feature optimization. The feature selection evaluation indicators involved in the optimization process are mainly divided into two categories. One is to directly use the classification accuracy and its derived indicators, and the other is to use indicators related to feature redundancy and separation to pre-select classification features before disturbance classification. However, the above indicators tend to evaluate individual power quality disturbance features and ignore the evaluation of power quality disturbance feature subsets, which will affect the accuracy of power quality disturbance classification. Summary of the Invention

[0003] The purpose of this invention is to provide a method for optimizing power quality disturbance characteristics based on multi-index fusion evaluation, which is beneficial to improving the efficiency and accuracy of power quality disturbance classification.

[0004] To achieve the above objectives, the technical solution adopted by this invention is: a method for optimizing power quality disturbance characteristics based on multi-index fusion evaluation, comprising:

[0005] Step 1: Input the power quality disturbance signal, extract various power quality disturbance features, and construct a complete set of power quality disturbance features;

[0006] Step 2: Based on the fusion evaluation index of intersection degree index, redundancy degree index, and separation degree index, the cuckoo search method is used to search for the candidate power quality disturbance feature subsets of each dimension.

[0007] Step 3: Evaluate the selected power quality disturbance feature subset based on the cost factor to obtain the optimal power quality disturbance feature subset, and classify the power quality disturbance signal based on the optimal power quality disturbance feature subset.

[0008] Furthermore, the extracted power quality disturbance features include: maximum voltage amplitude, minimum voltage amplitude, average voltage amplitude, standard deviation of voltage, disturbance duration, maximum instantaneous amplitude, minimum instantaneous amplitude, average instantaneous amplitude, standard deviation of instantaneous amplitude, marginal spectral energy, fundamental frequency marginal spectral energy, high frequency marginal spectral energy, first intrinsic mode energy, second intrinsic mode energy, third intrinsic mode energy, fourth intrinsic mode energy, fifth intrinsic mode energy, transient component factor, spectral energy corresponding to 150Hz, and spectral energy corresponding to 250Hz.

[0009] Furthermore, the specific implementation method of step 2 is as follows:

[0010] First, taking the complete set of power quality disturbance features as the object, a subset of power quality disturbance features is randomly searched from it using the cuckoo search method, and the intersection index of the power quality disturbance feature subsets is defined as follows:

[0011] Assuming two types of power quality disturbance signals S i S j Extracting features F k The power quality disturbance feature sets of the two types of signals were extracted as SF k i and SF k j, then the power quality disturbance characteristic values ​​are located in the interval [min(SF k i), max(SF) k i)] and [min(SF k j), max(SF k j)] in; where min() represents taking the minimum value, and max() represents taking the maximum value; the positional relationship of the characteristic value interval of power quality disturbance is divided into 4 categories:

[0012] I)max(SF k i)≤min(SF k j) and min(SF) k i)≥max(SF k j);

[0013] II) min(SF k i)<min(SF k j)<max(SF k i) < max(SF) k j);

[0014] III) min(SF) k j)<min(SF k i) < max(SF) k j)<max(SF k i);

[0015] IV) [min(SF] k i), max(SF k i)] ⊆ [min(SF k j), max(SF k j)];

[0016] Based on the positional relationship of the characteristic value intervals of power quality disturbance, the power quality disturbance S is defined. i With power quality disturbance S j Feature F k The degree of intersection xF of the power quality disturbance characteristic value interval k ij is as follows:

[0017]

[0018] When xF k When ij approaches 0, it indicates that feature F k Able to completely distinguish disturbance S i and S j Conversely, when xF k The larger the value of ij, the stronger the feature F k Distinguish disturbance S i and S j The worse the ability;

[0019] Therefore, the intersection matrix X of the power quality perturbation is adopted. Fk Describe all power quality disturbances in characteristic F k The degree of intersection is calculated using the following formula:

[0020]

[0021] When the intersection matrix X of the power quality disturbance Fk When the matrix is ​​zero, it indicates the power quality disturbance characteristic F. k It can distinguish all types of power quality disturbances; conversely, it indicates the power quality disturbance characteristic F. k It is impossible to distinguish all types of power quality disturbances, that is, the disturbance type corresponding to the non-zero position of the intersection degree matrix of power quality disturbances cannot be distinguished;

[0022] Based on the definition of the power quality disturbance intersection degree index, the total intersection degree matrix X is calculated. Assuming a power quality disturbance feature subset A = [F1, F2, F3, F4, …] and the number of power quality disturbance signal categories is M, the power quality disturbance intersection degree matrix corresponding to each power quality disturbance feature is calculated as X. F1 X F2 XF3 X F4 …and then the total intersection matrix of the power quality disturbance feature subset A is calculated as follows:

[0023]

[0024] In the formula, * represents the dot product symbol, where each element in the total intersection matrix of power quality disturbances is multiplied accordingly; the total intersection matrix is ​​an M×M matrix; the elements x in the total intersection matrix X... ij This indicates the degree of intersection between each pair of power quality disturbance signals when extracting the power quality disturbance feature subset A;

[0025] Then, the intersection index ξ of the power quality disturbance feature subset A is calculated. If the intersection index ξ approaches 0, it indicates that the power quality disturbance feature subset A has the ability to distinguish all power quality disturbances; otherwise, the power quality disturbance feature subset does not meet the requirements. The formula for calculating the intersection index ξ is as follows:

[0026]

[0027] After obtaining the subset of power quality disturbance features that meet the intersection degree requirement, this subset is input into the power quality disturbance feature optimization objective function G(S), and the optimization objective function value is calculated. In the optimization objective function G(S), mutual information and Fisher Score are used as redundancy and separation indices, respectively, so that the selected power quality disturbance features meet the feature redundancy index while having the maximum separation capability. The calculation formula of the power quality disturbance feature optimization objective function G(S) is as follows:

[0028]

[0029] Where A represents the power quality disturbance feature subset, V represents the dimension of the power quality disturbance feature subset; n is the number of features in the complete power quality disturbance feature set; ω1, ω2, ... , It is calculated by the following formula:

[0030]

[0031] Finally, for power quality disturbance feature subsets with the same dimension, their preferred objective function values ​​are compared, and the power quality disturbance feature subset with the largest preferred objective function value is selected as the candidate power quality disturbance feature subset for that dimension.

[0032] Furthermore, the specific implementation method of step 3 is as follows:

[0033] A cost factor is defined based on subset dimension and classification accuracy, and used as the basis for ranking power quality disturbance feature subsets. This ensures that the optimal power quality disturbance feature subset achieves high power quality disturbance classification accuracy while maintaining a small subset dimension. The cost factor R is defined as follows:

[0034]

[0035] In the formula, V represents the dimension of the power quality disturbance feature subset, V max V represents the maximum dimension of the subset of candidate power quality perturbation features. i <V max θ represents the classification accuracy, defined as follows:

[0036]

[0037] θ max N is the maximum classification accuracy of the subset of candidate power quality disturbance features during classification. true N represents the number of correctly classified signals. all The total number of signals;

[0038] The value of the cost factor R is jointly determined by the power quality disturbance classification accuracy and the dimension of the power quality disturbance feature subset. When the power quality disturbance classification accuracy θ is consistent, the smaller V is, the smaller the value of R is, and the more concise the power quality disturbance feature subset is.

[0039] Calculate the cost factor R of the candidate power quality disturbance feature subsets, sort the candidate power quality disturbance feature subsets according to the cost factor R value, and finally select the candidate power quality disturbance feature subset with the smallest R value as the optimal power quality disturbance feature subset.

[0040] Compared with the prior art, the present invention has the following beneficial effects: It provides a method for optimizing power quality disturbance features based on multi-index fusion evaluation. This method uses the intersection degree index to quantify the degree of intersection of sets of different combinations of power quality disturbance features when classifying power quality disturbances. Based on the fusion evaluation index of intersection degree index, redundancy index, and separation degree index, it optimizes the power quality disturbance feature subset. Then, it combines the cost factor R defined based on the dimension of the power quality disturbance feature subset and the power quality disturbance classification accuracy to establish a connection between the power quality disturbance feature subset and the power quality disturbance classifier, thereby improving the power quality disturbance classification accuracy. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating the method implementation of an embodiment of the present invention.

[0042] Figure 2 This describes the intersection of characteristic value intervals for different power quality disturbances in embodiments of the present invention. Detailed Implementation

[0043] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0044] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0045] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0046] like Figure 1 As shown, this embodiment provides a method for optimizing power quality disturbance characteristics based on multi-index fusion evaluation, including:

[0047] Step 1: Input the power quality disturbance signal and extract various power quality disturbance features, including maximum voltage amplitude, minimum voltage amplitude, average voltage amplitude, standard deviation of voltage, disturbance duration, maximum instantaneous amplitude, minimum instantaneous amplitude, average instantaneous amplitude, standard deviation of instantaneous amplitude, marginal spectral energy, fundamental frequency marginal spectral energy, high frequency marginal spectral energy, first intrinsic mode energy, second intrinsic mode energy, third intrinsic mode energy, fourth intrinsic mode energy, fifth intrinsic mode energy, transient component factor, spectral energy corresponding to 150Hz, spectral energy corresponding to 250Hz, etc., to construct a complete set of power quality disturbance features.

[0048] Step 2: Based on the fusion evaluation index of intersection degree index, redundancy degree index, and separation degree index, the cuckoo search method is used to search for the candidate power quality disturbance feature subsets of each dimension.

[0049] In this embodiment, the specific implementation method of step 2 is as follows:

[0050] First, taking the complete set of power quality disturbance features as the object, a subset of power quality disturbance features is randomly searched from it using the cuckoo search method, and the intersection index of the power quality disturbance feature subsets is defined as follows:

[0051] Assuming two types of power quality disturbance signals Si S j Extracting features F k The power quality disturbance feature sets of the two types of signals were extracted as SF k i and SF k j, then the power quality disturbance characteristic values ​​are located in the interval [min(SF k i), max(SF) k i)] and [min(SF k j), max(SF k In the context of j), min() represents taking the minimum value, and max() represents taking the maximum value. The positional relationships of the characteristic value intervals of power quality disturbances are divided into four categories, such as... Figure 2 As shown:

[0052] I)max(SF k i)≤min(SF k j) and min(SF) k i)≥max(SF k j);

[0053] II) min(SF k i)<min(SF k j)<max(SF k i) < max(SF) k j);

[0054] III) min(SF) k j)<min(SF k i) < max(SF) k j)<max(SF k i);

[0055] IV) [min(SF] k i), max(SF k i)] ⊆ [min(SF k j), max(SF k j)];

[0056] Based on the positional relationship of the characteristic value intervals of power quality disturbance, the power quality disturbance S is defined. i With power quality disturbance S j Feature F k The degree of intersection xF of the power quality disturbance characteristic value interval k ij is as follows:

[0057]

[0058] When xFk When ij approaches 0, it indicates that feature F k Able to completely distinguish disturbance S i and S j Conversely, when xF k The larger the value of ij, the stronger the feature F k Distinguish disturbance S i and S j The worse their ability is.

[0059] Therefore, the intersection matrix X of the power quality perturbation is adopted. Fk Describe all power quality disturbances in characteristic F k The degree of intersection is calculated using the following formula:

[0060]

[0061] Specifically, when the intersection matrix X of the power quality disturbance... Fk When the matrix is ​​zero, it indicates the power quality disturbance characteristic F. k It can distinguish all types of power quality disturbances; conversely, it indicates the power quality disturbance characteristic F. k It is impossible to distinguish all types of power quality disturbances, that is, the disturbance type corresponding to the non-zero position of the intersection degree matrix of power quality disturbances cannot be distinguished.

[0062] Based on the definition of the power quality disturbance intersection degree index, the total intersection degree matrix X is calculated. Assuming a power quality disturbance feature subset A = [F1, F2, F3, F4, …] and the number of power quality disturbance signal categories is M, the power quality disturbance intersection degree matrix corresponding to each power quality disturbance feature is calculated as X. F1 X F2 X F3 X F4 …and then the total intersection matrix of the power quality disturbance feature subset A is calculated as follows:

[0063]

[0064] In the formula, * represents the dot product symbol, where each element in the total intersection matrix of power quality disturbances is multiplied accordingly; the total intersection matrix is ​​an M×M matrix; the elements x in the total intersection matrix X... ij This indicates the degree of intersection between each pair of power quality disturbance signals when extracting the power quality disturbance feature subset A.

[0065] Then, the intersection index ξ of the power quality disturbance feature subset A is calculated. If the intersection index ξ approaches 0, it indicates that the power quality disturbance feature subset A has the ability to distinguish all power quality disturbances; otherwise, the power quality disturbance feature subset does not meet the requirements. The formula for calculating the intersection index ξ is as follows:

[0066]

[0067] After obtaining the power quality disturbance feature subset that meets the intersection index requirement, the power quality disturbance feature subset that meets the intersection index requirement is input into the power quality disturbance feature optimization objective function G(S), and the optimization objective function value is calculated. In the optimization objective function G(S), mutual information and Fisher Score are used as redundancy index and separation index, respectively, so that the selected power quality disturbance features meet the feature redundancy index and have the maximum separation capability.

[0068] The formula for calculating the objective function G(S) for optimal power quality disturbance characteristics is as follows:

[0069]

[0070] Where A represents the power quality disturbance feature subset, V represents the dimension of the power quality disturbance feature subset; n is the number of features in the complete power quality disturbance feature set; ω1, ω2, ... , It is calculated by the following formula:

[0071]

[0072] Finally, for power quality disturbance feature subsets with the same dimension, their preferred objective function values ​​are compared, and the power quality disturbance feature subset with the largest preferred objective function value is selected as the candidate power quality disturbance feature subset for that dimension.

[0073] Step 3: Evaluate the selected power quality disturbance feature subset based on the cost factor to obtain the optimal power quality disturbance feature subset, and classify the power quality disturbance signal based on the optimal power quality disturbance feature subset.

[0074] In this embodiment, the specific implementation method of step 3 is as follows:

[0075] A cost factor is defined based on subset dimension and classification accuracy, and used as the basis for ranking the power quality disturbance feature subsets. This ensures that the optimal power quality disturbance feature subset achieves high power quality disturbance classification accuracy while maintaining a small subset dimension. The cost factor R is defined as follows:

[0076]

[0077] In the formula, V represents the dimension of the power quality disturbance feature subset, V max V represents the maximum dimension of the subset of candidate power quality perturbation features. i <V max θ represents the classification accuracy, defined as follows:

[0078]

[0079] θ max N is the maximum classification accuracy of the subset of candidate power quality disturbance features during classification. true N represents the number of correctly classified signals. all This represents the total number of signals.

[0080] The value of the cost factor R is jointly determined by the power quality disturbance classification accuracy and the dimension of the power quality disturbance feature subset. When the power quality disturbance classification accuracy θ is consistent, the smaller V is, the smaller the value of R will be, and the more concise the power quality disturbance feature subset will be.

[0081] Calculate the cost factor R of the candidate power quality disturbance feature subsets, sort the candidate power quality disturbance feature subsets according to the cost factor R value, and finally select the candidate power quality disturbance feature subset with the smallest R value as the optimal power quality disturbance feature subset.

[0082] 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.

[0083] 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.

[0084] 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.

[0085] 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.

[0086] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for screening power quality disturbance characteristics based on multi-index fusion evaluation, characterized in that, include: Step 1: Input the power quality disturbance signal, extract various power quality disturbance features, and construct a complete set of power quality disturbance features; Step 2: Based on the fusion evaluation index of intersection degree index, redundancy degree index, and separation degree index, the cuckoo search method is used to search for the candidate power quality disturbance feature subsets of each dimension. Step 3: Evaluate the selected power quality disturbance feature subset based on the cost factor to obtain the optimal power quality disturbance feature subset, and classify the power quality disturbance signal based on the optimal power quality disturbance feature subset; The specific implementation method of step 3 is as follows: A cost factor is defined based on subset dimension and classification accuracy, and used as the basis for ranking power quality disturbance feature subsets. This ensures that the optimal power quality disturbance feature subset achieves high power quality disturbance classification accuracy while maintaining a small subset dimension. The cost factor R is defined as follows: In the formula, V represents the dimension of the power quality disturbance feature subset, V max V represents the maximum dimension of the subset of candidate power quality perturbation features. i <V max θ represents the classification accuracy, defined as follows: θ max N is the maximum classification accuracy of the subset of candidate power quality disturbance features during classification. true N represents the number of correctly classified signals. all The total number of signals; The value of the cost factor R is jointly determined by the power quality disturbance classification accuracy and the dimension of the power quality disturbance feature subset. When the power quality disturbance classification accuracy θ is consistent, the smaller V is, the smaller the value of R is, and the more concise the power quality disturbance feature subset is. Calculate the cost factor R of the candidate power quality disturbance feature subsets, sort the candidate power quality disturbance feature subsets according to the cost factor R value, and finally select the candidate power quality disturbance feature subset with the smallest R value as the optimal power quality disturbance feature subset.

2. The method for screening power quality disturbance characteristics based on multi-index fusion evaluation according to claim 1, characterized in that, The extracted power quality disturbance features include: maximum voltage amplitude, minimum voltage amplitude, average voltage amplitude, standard deviation of voltage, disturbance duration, maximum instantaneous amplitude, minimum instantaneous amplitude, average instantaneous amplitude, standard deviation of instantaneous amplitude, marginal spectral energy, fundamental frequency marginal spectral energy, high frequency marginal spectral energy, first intrinsic mode energy, second intrinsic mode energy, third intrinsic mode energy, fourth intrinsic mode energy, fifth intrinsic mode energy, transient component factor, spectral energy corresponding to 150Hz, and spectral energy corresponding to 250Hz.

3. The method for screening power quality disturbance characteristics based on multi-index fusion evaluation according to claim 1, characterized in that, The specific implementation method of step 2 is as follows: First, taking the complete set of power quality disturbance features as the object, a subset of power quality disturbance features is randomly searched from it using the cuckoo search method, and the intersection index of the power quality disturbance feature subsets is defined as follows: Assuming two types of power quality disturbance signals S i S j Extracting features F k The power quality disturbance feature sets of the two types of signals were extracted as SF k i and SF k j, then the power quality disturbance characteristic values ​​are located in the interval [min(SF k i), max(SF k i)] and [min(SF k j), max(SF k j)] in; where min() represents taking the minimum value, and max() represents taking the maximum value; the positional relationship of the characteristic value interval of power quality disturbance is divided into 4 categories: I) max(SF k i) ≤ min(SF k j) and min(SF k i) ≥ max(SF k j); II)min(SF k i)<min(SF k j)<max(SF k i)<max(SF k j); III)min(SF k j)<min(SF k i)<max(SF k j)<max(SF k i); IV)[min(SF k i), max(SF k i)] ⊆ [min(SF k j), max(SF k j)]; Based on the positional relationship of the characteristic value intervals of power quality disturbance, the power quality disturbance S is defined. i With power quality disturbance S j Feature F k The degree of intersection xF of the power quality disturbance characteristic value interval k ij is as follows: When xF k When ij approaches 0, it indicates that feature F k Able to completely distinguish disturbance S i and S j Conversely, when xF k The larger the value of ij, the stronger the feature F k Distinguish disturbance S i and S j The worse the ability; Therefore, the intersection matrix X of the power quality perturbation is adopted. Fk Describe all power quality disturbances in characteristic F k The degree of intersection is calculated using the following formula: When the intersection matrix X of the power quality disturbance Fk When the matrix is ​​zero, it indicates the power quality disturbance characteristic F. k It can distinguish all types of power quality disturbances; conversely, it indicates the power quality disturbance characteristic F. k It is impossible to distinguish all types of power quality disturbances, that is, the disturbance type corresponding to the non-zero position of the intersection degree matrix of power quality disturbances cannot be distinguished; Based on the definition of the power quality disturbance intersection degree index, the total intersection degree matrix X is calculated. Assuming a power quality disturbance feature subset A = [F1, F2, F3, F4, …] and the number of power quality disturbance signal categories is M, the power quality disturbance intersection degree matrix corresponding to each power quality disturbance feature is calculated as X. F1 X F2 X F3 X F4 …and then the total intersection matrix of the power quality disturbance feature subset A is calculated as follows: In the formula, * represents the dot product symbol, where each element in the total intersection matrix of power quality disturbances is multiplied accordingly; the total intersection matrix is ​​an M×M matrix; the elements x in the total intersection matrix X... ij This indicates the degree of intersection between each pair of power quality disturbance signals when extracting the power quality disturbance feature subset A; Then, the intersection index ξ of the power quality disturbance feature subset A is calculated. If the intersection index ξ approaches 0, it indicates that the power quality disturbance feature subset A has the ability to distinguish all power quality disturbances; otherwise, the power quality disturbance feature subset does not meet the requirements. The formula for calculating the intersection index ξ is as follows: After obtaining the subset of power quality disturbance features that meet the intersection degree requirement, this subset is input into the power quality disturbance feature optimization objective function G(S), and the optimization objective function value is calculated. In the optimization objective function G(S), mutual information and Fisher Score are used as redundancy and separation indices, respectively, so that the selected power quality disturbance features meet the feature redundancy index while having the maximum separation capability. The calculation formula of the power quality disturbance feature optimization objective function G(S) is as follows: Where A represents the power quality disturbance feature subset, V represents the dimension of the power quality disturbance feature subset; n is the number of features in the complete power quality disturbance feature set; ω1, ω2, ... , It is calculated by the following formula: Finally, for power quality disturbance feature subsets with the same dimension, their preferred objective function values ​​are compared, and the power quality disturbance feature subset with the largest preferred objective function value is selected as the candidate power quality disturbance feature subset for that dimension.

Images