Harmonic source ensemble learning positioning method based on dominant scale feature transfer matching

Abstract

The invention provides a harmonic source integrated learning positioning method based on dominant scale feature transfer matching, and the method comprises the steps: carrying out the multi-scale decomposition of each monitoring quantity through employing a wavelet packet decomposition method, and extracting the unique fluctuation features of each harmonic source; the method comprises the following steps: randomly combining voltage or current monitoring quantities to generate a plurality of base learners, and estimating a regression coefficient corresponding to a dominant scale d sequence of each base learner on the basis of screening out an optimal window width and a data segment when only a dominant harmonic source acts by using an improved dynamic windowing algorithm; and integrating voting results of the base learners by using an improved weighted voting method, and outputting a dominant harmonic source position.

Description

Technical Field

[0001] The present invention relates to the technical field of harmonic source positioning in power systems, and in particular to a harmonic source integrated learning positioning method based on dominant scale feature transfer matching. Background Art

[0002] With the development of the economic level, a large number of distributed clean energy and nonlinear loads are connected to the power grid, which injects a lot of harmonic pollution into the power grid, and also makes the harmonic conditions in the power grid more complex and difficult to estimate. Harmonic source positioning plays an important role in clarifying the source of harmonic pollution, and is the premise for the division of harmonic responsibilities and the resolution of economic disputes. At present, the installation density of harmonic monitoring devices in the distribution network is low, which cannot meet the observability requirements, resulting in the inapplicability of existing harmonic source positioning methods. Therefore, how to use the existing statistical harmonic monitoring data to locate the harmonic source under the condition of limited measurement points is an urgent problem to be solved. In addition, massive, multi-dimensional, and long-time scale harmonic monitoring data contain rich time-varying information of harmonic sources. How to start from the data-driven idea and combine intelligent algorithms to mine harmonic source location information from statistical harmonic monitoring data deserves in-depth study. Summary of the invention

[0003] In view of the defects and shortcomings of the prior art, the purpose of the present invention is to provide a harmonic source integrated learning positioning method based on dominant scale feature transfer matching, which fully explores the fluctuation characteristics of different harmonic sources according to the data and distribution characteristics of existing harmonic monitoring devices, and realizes harmonic source positioning in combination with their transfer characteristics, thereby avoiding the deployment of a large number of dedicated synchronous measurement equipment.

[0004] The implementation of the scheme first uses the wavelet packet decomposition method to perform multi-scale decomposition on each monitored quantity and extract the unique fluctuation characteristics of each harmonic source; secondly, taking into account the non-synchronization of the monitoring data and the lack of phase angle information, the transfer equation of the fluctuation characteristics of each scale is established according to the system path information, and the improved dynamic windowing algorithm is used to screen out the optimal window width and the data segment when only the dominant harmonic source acts, and the transfer equation parameters are identified at the same time, and then the nodes where the dominant harmonic source exists are matched based on the transfer equation parameters; finally, any combination of voltage or current monitoring quantities is used to generate several basis learners, and the voting results of each basis learner are integrated using the improved weighted voting method to output the position of the dominant harmonic source.

[0005] The specific technical solutions are as follows:

[0006] A harmonic source integrated learning positioning method based on dominant scale feature transfer matching uses wavelet packet decomposition method to perform multi-scale decomposition on each monitoring quantity and extract the unique fluctuation characteristics of each harmonic source; arbitrarily combine voltage or current monitoring quantities to generate several basis learners, and use improved dynamic windowing algorithm to screen out the optimal window width and the data segment when only the dominant harmonic source acts, and then estimate the regression coefficient corresponding to the dominant scale d sequence of each basis learner. Then, the improved weighted voting method is used to integrate the voting results of each basis learner to output the dominant harmonic source position.

[0007] Furthermore, the wavelet packet decomposition method is used to perform multi-scale decomposition on each monitoring variable to extract the unique fluctuation characteristics of each harmonic source:

[0008] Obtain the statistical harmonic monitoring sequence saved by each harmonic monitoring device in the network; select the wavelet basis and the number of decomposition layers to decompose all monitoring data sequences into wavelet coefficient sequences of subspaces of different scales; select the decomposition results of the harmonic voltage average sequence, and calculate the energy of the wavelet coefficients of each subspace of each scale. The calculation formula is as follows:

[0009]

[0010] In the formula, q represents the number of decomposition levels, F represents the set of measurement points, j represents the subspace scale, and c f (q,j) represents the wavelet coefficient sequence of the qth layer j-scale subspace, ||.|| 2 represents the two-norm;

[0011] Calculate the sum of the energy of the wavelet coefficients of all scales except the 0-scale subspace at all measurement points. The calculation formula is as follows:

[0012]

[0013] Calculate the energy proportion of the wavelet coefficients in each scale subspace. The calculation formula is as follows:

[0014]

[0015] The scales with more than 5% of the subspace energy are defined as the dominant scales;

[0016] The wavelet coefficient sequence belonging to the dominant scale in the wavelet coefficient sequence is reconstructed.

[0017] Furthermore, the arbitrary combination of voltage or current monitoring quantities to generate a number of base learners is specifically:

[0018] By combining the harmonic voltage or branch harmonic current data of different statistical value types in pairs, K basis learners are formed. The calculation formula of K is as follows:

[0019]

[0020] In the formula, C u and C i They respectively represent the number of state quantities of harmonic voltage and current that can be obtained by the harmonic monitoring device in the network.

[0021] Furthermore, the improved dynamic windowing algorithm is used to filter out the data segment when only the dominant harmonic source acts:

[0022] The wavelet coefficients of the dominant scale d of the harmonic monitoring quantity associated with each learner are used to reconstruct the sequence and construct a linear model in the following form:

[0023] C ad =β d1 C bd +β d0 +ε

[0024] In the formula, C ad and C bd They represent the sequence reconstructed by the wavelet coefficients of the dominant scale d of the harmonic monitoring quantity associated with each base learner, β d1 and β d2 are the regression coefficients of the linear model respectively;

[0025] Set the initial window width and maximum window width H max ;

[0026] Divide the input sequence of the linear model according to the window width H;

[0027] Estimate the regression coefficient β and F statistic of the corresponding sequence in each window. The calculation formula is as follows;

[0028]

[0029] In the formula, x, y represent the input sequence of the kth linear model, and g represents the window number;

[0030] like Then discard the corresponding window, and assume that there are G windows in total. The final regression coefficient corresponding to the window width H is calculated by the following formula: and the value of the F statistic; if G<1, reduce the threshold of the F statistic:

[0031]

[0032] If H <H max , let H = H + 1, re-divide the input sequence of the linear model according to the window width H and estimate the regression coefficient of the corresponding sequence in each window and calculation

[0033] If H = H max, then Sort by window width, form a group of 5, and calculate the average value of each group. and select The largest group is the optimal window width range.

[0034] Furthermore, the estimated regression coefficient corresponding to the dominant scale d sequence of each base learner is specifically:

[0035] The average of the regression estimation results within the optimal window width range is taken to obtain the estimated result of the regression coefficient. The calculation formula is as follows:

[0036]

[0037] Wherein, H and H+5 represent the range of the optimal window width.

[0038] Furthermore, the process of integrating the voting results of each base learner using the improved weighted voting method and outputting the position of the dominant harmonic source is specifically as follows:

[0039] Calculate the matching coefficient M of the dominant scale d of each base learner ds , the calculation formula is as follows;

[0040]

[0041] Where s represents the node number of the power grid, a and b represent the node numbers associated with the harmonic voltage base learner input, q, w, j, k represent the node numbers associated with the harmonic current base learner input; Z is impedance, and Y is admittance;

[0042] Normalize the matching coefficient;

[0043] Calculate the difference of the matching coefficient, the calculation formula is as follows:

[0044]

[0045] In the formula, max(·) means taking the maximum value;

[0046] Calculate the weight of the base learner, the calculation formula is as follows:

[0047]

[0048] Calculate the voting value of the base learner. The calculation formula is as follows:

[0049]

[0050] Combining the voting results of each base learner, the probability that the harmonic source is located at the s node is calculated. The calculation formula is as follows

[0051]

[0052] According to the probability calculation results, the node with the highest probability of the harmonic source appearing is output;

[0053] The harmonic source location is completed by outputting the node with the highest probability of occurrence of the harmonic source obtained in each dominant scale.

[0054] And, an electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the harmonic source integrated learning positioning method based on dominant scale feature transfer matching as described above are implemented.

[0055] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the harmonic source integrated learning positioning method based on dominant scale feature transfer matching as described above.

[0056] Compared with the prior art, the present invention and its preferred solution use wavelet packet decomposition to divide the time series according to frequency, and screen the dominant scale through the wavelet coefficient energy, which can effectively extract the different fluctuation characteristics of each harmonic source and improve the efficiency of subsequent analysis. The proposed regression parameter estimation method of improved dynamic window width can eliminate the problem of large differences in results of similar window widths, screen out the time period of the dominant harmonic source, and accurately identify the regression coefficients of each linear model. The proposed dominant harmonic source positioning model based on ensemble learning makes full use of the data provided by the harmonic monitoring device, effectively integrates the path information matching results of multiple basis learners through weight coefficients, and can realize harmonic source positioning under the condition that the harmonic monitoring device does not meet the observability, and has certain robustness and generalization ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0058] Figure 1 The figure is a schematic diagram of a process flow for implementing the method of an embodiment of the present invention. DETAILED DESCRIPTION

[0059] Hereinafter, the specific embodiments of the present application will be described in detail with reference to the accompanying drawings. According to these detailed descriptions, those skilled in the art can clearly understand the present application and can implement the present application. Without violating the principles of the present application, the features in the various embodiments can be combined to obtain new implementations, or certain features in certain embodiments can be replaced to obtain other preferred implementations.

[0060] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0061] In order to make the features and advantages of this patent more obvious and easy to understand, the following embodiments are specifically described in detail as follows:

[0062] like Figure 1 As shown, the embodiment of the present invention provides a harmonic source integrated learning positioning method based on dominant scale feature transfer matching, and its implementation process includes the following steps:

[0063] Step S1, obtaining a statistical harmonic monitoring sequence stored in each harmonic monitoring device in the network;

[0064] Step S2, selecting a wavelet basis and a decomposition layer number to decompose all monitoring data sequences into wavelet coefficient sequences of subspaces of different scales;

[0065] Step S3: Select the decomposition result of the harmonic voltage average value sequence of the wavelet coefficient sequence obtained in step S2, and calculate the energy of the wavelet coefficients of each scale subspace. The calculation formula is as follows:

[0066]

[0067] In the formula, q represents the number of decomposition levels, F represents the set of measurement points, j represents the subspace scale, and c f (q,j) represents the wavelet coefficient sequence of the qth layer j-scale subspace, ||.|| 2 represents the two-norm;

[0068] Step S4, calculate the energy sum of the wavelet coefficients of all scales of all measurement points except the 0-scale subspace, and the calculation formula is as follows:

[0069]

[0070] Step S5, calculate the energy proportion of the wavelet coefficients of each scale subspace, and the calculation formula is as follows:

[0071]

[0072] Step S6, defining the scale whose subspace energy accounts for more than 5% as the dominant scale;

[0073] Step S7, reconstructing the wavelet coefficient sequence obtained in step S2 that belongs to the wavelet coefficient sequence of the dominant scale;

[0074] Step S8: K basis learners are formed by combining harmonic voltage or branch harmonic current data of different statistical value types in pairs. The calculation formula of K is as follows:

[0075]

[0076] In the formula, C u and C i They represent the number of harmonic voltage and current state quantities that can be obtained by the harmonic monitoring device in the network. Since statistical data includes statistical values ​​such as maximum value, minimum value, average value and 95% probability maximum value, the number of base learners should be multiplied by 4;

[0077] Step S8: Reconstruct the sequence using the wavelet coefficients of the dominant scale d of the harmonic monitoring quantity associated with the kth basis learner to construct a linear model in the following form:

[0078]

[0079] In the formula, and They represent the reconstructed sequence of wavelet coefficients of the dominant scale d of the harmonic monitoring quantity associated with the kth basis learner, and are the regression coefficients of the linear model respectively;

[0080] Step S9: Set the initial window width and the maximum window width H max ;

[0081] Step S10, dividing the input sequence of the linear model according to the window width H;

[0082] Step S11, estimate the regression coefficient and F statistic of the corresponding sequence in each window, and the calculation formula is as follows;

[0083]

[0084] Where x, y represent the input sequence of the kth linear model, and g represents the window number.

[0085] Step S12: If Then discard the window, assuming that there are G windows in total, and calculate the final regression coefficient corresponding to the window width H by the following formula: and the value of the F statistic. If G<1, reduce the threshold of the F statistic:

[0086]

[0087] Step S13: If H=H max , then proceed to the next step. Otherwise, set H = H + 1 and return to step S10;

[0088] Step S14: Sort by window width, form a group of 5, and calculate the average value of each group. and select The largest group is the optimal window width range;

[0089] Step S15, taking the average of the regression estimation results within the optimal window width range to obtain the estimation result of the final regression coefficient, the calculation formula is as follows:

[0090]

[0091] Wherein, H and H+5 represent the range of the optimal window width.

[0092] Step S16, through the loop calculation from step S9 to step S14, until the regression coefficient corresponding to the dominant scale d sequence of each base learner is estimated;

[0093] Step S17: Calculate the matching coefficient M of the dominant scale d of each base learner ds , the calculation formula is as follows;

[0094]

[0095] Where s represents the grid node number, a and b represent the node numbers associated with the harmonic voltage base learner input, and q, w, j, and k represent the node numbers associated with the harmonic current base learner input.

[0096] Step S18, normalizing the matching coefficient;

[0097] Step S19: Calculate the difference of the matching coefficient. The calculation formula is as follows:

[0098]

[0099] In the formula, max(·) means taking the maximum value;

[0100] Step S20: Calculate the weight of the base learner. The calculation formula is as follows:

[0101]

[0102] Step S21, calculate the voting value of the base learner, the calculation formula is as follows:

[0103]

[0104] Step S22: Based on the voting results of each base learner, the probability that the harmonic source is located at the s node is calculated. The calculation formula is as follows:

[0105]

[0106] Step S23: output the node with the highest probability of occurrence of harmonic source according to the probability calculation result.

[0107] By outputting the node with the highest probability of occurrence of the harmonic source obtained at each dominant scale, the harmonic source positioning is completed.

[0108] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0109] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0110] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0111] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0112] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0113] This patent is not limited to the above-mentioned optimal implementation mode. Anyone can derive other forms of harmonic source integrated learning positioning methods based on dominant scale feature transfer matching under the inspiration of this patent. All equal changes and modifications made according to the scope of the patent application of the present invention should be covered by this patent.

Claims

1. A harmonic source ensemble learning positioning method based on dominant scale feature transfer matching, characterized by: The wavelet packet decomposition method is used to perform multi-scale decomposition of each monitoring variable to extract the unique fluctuation characteristics of each harmonic source; Several base learners are generated by any combination of voltage or current monitoring quantities. On the basis of using the improved dynamic windowing algorithm to screen out the optimal window width and the data segment when only the dominant harmonic source acts, the regression coefficient corresponding to the dominant scale d sequence of each base learner is estimated, and then the voting results of each base learner are integrated using the improved weighted voting method to output the position of the dominant harmonic source.

2. The harmonic source integrated learning positioning method based on dominant scale feature transfer matching according to claim 1 is characterized in that: The wavelet packet decomposition method is used to perform multi-scale decomposition on each monitoring variable to extract the unique fluctuation characteristics of each harmonic source: Obtain the statistical harmonic monitoring sequence saved by each harmonic monitoring device in the network; select the wavelet basis and the number of decomposition layers to decompose all monitoring data sequences into wavelet coefficient sequences of subspaces of different scales; select the decomposition results of the harmonic voltage average sequence, and calculate the energy of the wavelet coefficients of each subspace of each scale. The calculation formula is as follows: In the formula, q represents the number of decomposition levels, F represents the set of measurement points, j represents the subspace scale, and c f (q,j) represents the wavelet coefficient sequence of the qth layer j-scale subspace, ||.|| 2 represents the two-norm; Calculate the sum of the energy of the wavelet coefficients of all scales except the 0-scale subspace at all measurement points. The calculation formula is as follows: Calculate the energy proportion of the wavelet coefficients in each scale subspace. The calculation formula is as follows: The scales with more than 5% of the subspace energy are defined as the dominant scales; The wavelet coefficient sequence belonging to the dominant scale in the wavelet coefficient sequence is reconstructed.

3. The harmonic source integrated learning positioning method based on dominant scale feature transfer matching according to claim 2 is characterized in that: The arbitrary combination of voltage or current monitoring quantities to generate a number of base learners is specifically: By combining the harmonic voltage or branch harmonic current data of different statistical value types in pairs, K basis learners are formed. The calculation formula of K is as follows: In the formula, C u and C i They respectively represent the number of state quantities of harmonic voltage and current that can be obtained by the harmonic monitoring device in the network.

4. The harmonic source integrated learning positioning method based on dominant scale feature transfer matching according to claim 3 is characterized in that: The data segment when the optimal window width and only the dominant harmonic source act by using the improved dynamic windowing algorithm is specifically: The wavelet coefficients of the dominant scale d of the harmonic monitoring quantity associated with each base learner are used to reconstruct the sequence and construct a linear model in the following form: C ad =b d1 C bd +b d0 +e In the formula, C ad and C bd They represent the sequence reconstructed by the wavelet coefficients of the dominant scale d of the harmonic monitoring quantity associated with each base learner, β d1 and β d2 are the regression coefficients of the linear model respectively; Set the initial window width and maximum window width H max ; Divide the input sequence of the linear model according to the window width H; Estimate the regression coefficient β and F statistic of the corresponding sequence in each window. The calculation formula is as follows; In the formula, x, y represent the input sequence of the kth linear model, and g represents the window number; like Then discard the corresponding window, and assume that there are G windows in total. Calculate the regression coefficient and F statistic value corresponding to the window width H by the following formula; if G<1, reduce the threshold of F statistic: If H <H max , let H = H + 1, re-divide the input sequence of the linear model according to the window width H and estimate the regression coefficient of the corresponding sequence in each window and calculation If H = H max , then Sort by window width, form a group of 5, and calculate the average value of each group. and select The largest group is the optimal window width range.

5. The harmonic source integrated learning positioning method based on dominant scale feature transfer matching according to claim 4 is characterized in that: The estimated regression coefficient corresponding to the dominant scale d sequence of each base learner is specifically: The average of the regression estimation results within the optimal window width range is taken to obtain the estimated result of the regression coefficient. The calculation formula is as follows: Wherein, H and H+5 represent the range of the optimal window width.

6. The harmonic source integrated learning positioning method based on dominant scale feature transfer matching according to claim 5 is characterized in that: The process of integrating the voting results of each base learner by using the improved weighted voting method and outputting the position of the dominant harmonic source is specifically as follows: Calculate the matching coefficient M of the dominant scale d of each base learner ds , the calculation formula is as follows; or Where s represents the node number of the power grid, a and b represent the node numbers associated with the harmonic voltage base learner input, q, w, j, k represent the node numbers associated with the harmonic current base learner input; Z is impedance, and Y is admittance; Normalize the matching coefficient; Calculate the difference of the matching coefficient, the calculation formula is as follows: In the formula, max(·) means taking the maximum value; Calculate the weight of the base learner, the calculation formula is as follows: Calculate the voting value of the base learner. The calculation formula is as follows: Combining the voting results of each base learner, the probability that the harmonic source is located at the s node is calculated. The calculation formula is as follows According to the probability calculation results, the node with the highest probability of the harmonic source appearing is output; The harmonic source location is completed by outputting the node with the highest probability of occurrence of the harmonic source obtained in each dominant scale.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the harmonic source integrated learning positioning method based on dominant scale feature transfer matching as described in any one of claims 1 to 6 are implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the harmonic source integrated learning positioning method based on dominant scale feature transfer matching as described in any one of claims 1 to 6 are implemented.

Images