A system and method for selecting harmonic current suppression technology in power grid

Abstract

The present invention belongs to the technical field of harmonic current suppression, and discloses a system and method for selecting a harmonic current suppression technology in a power grid; current signals of n nodes are collected; current signals are processed and analyzed to obtain harmonic current characteristic data and the number of harmonic currents in the current signals; Q groups of historical power grid state data are collected, the historical power grid state data include power grid characteristic data and suppression technologies corresponding to the power grid characteristic data, the power grid characteristic data include power grid characteristic data and harmonic current characteristic data; based on the historical power grid state data, a method selection model for predicting suppression technology labels is trained; whether to mark a suppression area is determined according to the number of harmonic currents and the total harmonic current number; harmonic currents are effectively suppressed to reduce the impact of harmonic currents on the power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of harmonic current suppression, and more specifically, to a system and method for selecting a harmonic current suppression technology in a power grid. Background Art

[0002] Harmonic current refers to the additional current component in the power grid whose frequency is an integer multiple of the fundamental frequency. Nonlinear loads introduce nonlinear current during operation, causing the current waveform of the power grid to deform, thereby generating harmonic current. Harmonic current has a significant impact on the power grid. For example, harmonic current will bring additional load to power equipment, causing equipment overheating, damage or even failure. Harmonic current will cause voltage fluctuations, causing instability in the power grid, which may lead to drastic voltage fluctuations and frequency deviations. Harmonic current will cause additional loss of electric energy during transmission, thereby reducing the energy efficiency of the power grid. Therefore, it is necessary to effectively suppress harmonic current to reduce the impact of harmonic current on the power grid.

[0003] In order to effectively suppress harmonic current, the selection of suppression technology is particularly important. In the existing technology, the selection of suppression technology depends on workers to detect harmonics in the current of the power grid, obtain relevant information of harmonic current, and then select the corresponding suppression technology based on the experience of workers in combination with the characteristics of the power grid. In addition, workers are required to use the selected suppression technology to suppress harmonic current in the power grid and detect the suppression effect. If the suppression method is not properly selected, the suppression effect is poor and the suppression technology needs to be reselected. This is not only time-consuming and labor-intensive, but also increases costs, and workers need to have a high level of professionalism.

[0004] Of course, there are also intelligent suppression technology options. The patent with publication number CN116846226A discloses a single-phase DC transformer and its DC current harmonic suppression method and device, including collecting the DC current of the single-phase DC transformer, extracting the current through a filter to obtain the current harmonic component, and performing coordinate transformation on the current harmonic component and the current harmonic component after a certain time delay, and then converting the current component after the coordinate transformation into a current component of equal magnitude and opposite direction, and adjusting and controlling each reverse current component through controllers on several branches respectively, and the controller on each branch selects a PI controller or a PR controller according to the suppression demand; for another example, the Chinese patent with authorization announcement number CN115642599B discloses a harmonic current suppression method, control device and power generation system;

[0005] However, the above technology only selects the appropriate suppression technology through the parameters of harmonic current. There are various harmonic currents in the power grid, and different harmonic currents correspond to different suppression technologies. If all corresponding suppression technologies are selected and used, the suppression cost and suppression efficiency will be increased, the energy consumption of the power grid will be increased, and the space requirement will be large.

[0006] In view of this, the present invention proposes a harmonic current suppression technology selection system and method in a power grid to solve the above-mentioned problem. Summary of the invention

[0007] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for selecting harmonic current suppression technology in a power grid, comprising:

[0008] Collect current signals of n nodes;

[0009] Process and analyze the current signal to obtain harmonic current characteristic data and the amount of harmonic current in the current signal;

[0010] Collecting Q groups of historical grid status data, the historical grid status data including grid characteristic data and suppression technology corresponding to the grid characteristic data, the grid characteristic data including grid characteristic data and harmonic current characteristic data;

[0011] Based on historical grid status data, a method selection model is trained to predict the suppression technology labels;

[0012] The power grid is divided into v areas, where v is an integer greater than 1 and v <n;

[0013] According to the number of harmonic currents in each of the v areas and the total number of harmonic currents in the v areas, determine whether to mark it as a suppression area;

[0014] The real-time power grid characteristic data of the suppression area is input into the trained method selection model, and the corresponding real-time suppression technology label is output. The number of real-time suppression technology labels of each suppression area is counted, and the mean and standard deviation of the number of suppression technology labels of each suppression area are calculated. It is determined whether discrete selection instructions or centralized selection instructions are generated, and the real-time suppression technology of the suppression area is selected according to the generated instructions.

[0015] Furthermore, the power grid topology is obtained through the power grid planning drawings, and the nodes are set according to the power grid topology.

[0016] Furthermore, the harmonic current characteristic data includes harmonic order, amplitude and phase, the amplitude is the harmonic current amplitude, and the phase is the harmonic current phase;

[0017] Methods for obtaining harmonic current characteristic data and harmonic current quantity include:

[0018] By performing fast Fourier transform on the current signal, the current signal in the time domain can be converted into an energy spectrum in the frequency domain. The fast Fourier transform is Where X(m) represents the mth frequency component in the frequency domain, X(n) represents the nth current signal in the time domain, N represents the total number of current signals, e is the natural logarithmic constant, j is the imaginary unit, and π is the circumference of a circle;

[0019] The fundamental frequency is obtained according to the energy spectrum; the frequency component includes harmonic frequency, amplitude and phase; the harmonic order c is the harmonic frequency f c The ratio of the fundamental frequency f1 is

[0020] Furthermore, the grid characteristic data includes the voltage level, current capacity and grid structure of the grid; the training process of the method selection model includes:

[0021] Label the suppression techniques with numerical labels;

[0022] Historical power grid status data is used as a sample set, and the sample set is divided into a training set and a test set. A classifier is constructed, and the power grid feature data in the training set is used as input data, and the suppression technology label in the training set is used as output data. The classifier is trained to obtain an initial classifier, and the initial classifier is tested using the test set. The initial classifier that meets the prediction error is output as a method selection model, and the classifier is one of a logistic regression model, a naive Bayes model, or a support vector machine model.

[0023] Furthermore, the method for determining whether to mark as an inhibition area is as follows:

[0024] n nodes are evenly distributed in v areas, and the number of harmonic currents and the total number of harmonic currents in each of the v areas are collected; the number of harmonic currents in each area is divided by the total number of harmonic currents to obtain the proportion of harmonic currents in each area;

[0025] If the proportion of harmonic current is greater than or equal to Then the area is marked as an inhibition area;

[0026] If the proportion of harmonic current is less than The area is not marked as an inhibition area.

[0027] Furthermore, the mean number of suppression technology labels is calculated as follows:

[0028]

[0029] In the formula, is the mean number of suppression technology labels in the h-th suppression region, YZ hi is the number of labels of the ith suppression technique in the hth suppression region, h∈[1,v], R h is the number of suppression technique label types in the h-th suppression region, i∈[1,R h ];

[0030] The standard deviation is calculated as follows:

[0031]

[0032] Where, σY h is the standard deviation of the h-th inhibition zone.

[0033] Further, the method for determining whether to generate a discrete selection instruction or a centralized selection instruction includes:

[0034] A standard deviation threshold is preset, and if the standard deviation of the suppression area is greater than the standard deviation threshold, a discrete selection instruction is generated; according to the discrete selection instruction, a suppression technology label whose number is greater than the average number of suppression technology labels is selected, and the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area;

[0035] If the standard deviation of the suppression area is less than or equal to the standard deviation threshold, a centralized selection instruction is generated; according to the centralized selection instruction, suppression technology labels whose number is greater than the selection threshold are selected; and the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area.

[0036] Further, the number corresponding to each suppression technology label before suppression of the suppression area is marked as the number before suppression, the number corresponding to each suppression technology label after suppression of the suppression area is marked as the number after suppression, the number before suppression and the suppression technology corresponding to the number before suppression are used as suppression feature data, and the suppression feature data and the number after suppression corresponding to the suppression feature data are used as suppression state data; based on the suppression state data, an effect detection model for predicting the number after suppression is trained;

[0037] The specific training process of the effect detection model includes:

[0038] In an experimental environment, the corresponding suppression technology is used to suppress multiple harmonic currents on power equipment with the same suppression characteristic data in turn, and the suppressed quantity of the power equipment after suppression is recorded in turn, and the average of the suppressed quantity after multiple harmonic current suppression is used as the suppressed quantity corresponding to the suppression characteristic data; the corresponding suppressed quantity is obtained in turn for power equipment with different suppression characteristic data according to the method;

[0039] The inhibition state data is used as the input of the effect detection model, the effect detection model takes the predicted inhibition quantity corresponding to each group of inhibition feature data as the output, takes the actual inhibition quantity corresponding to the group of inhibition feature data as the prediction target, and takes minimizing the sum of prediction errors of all inhibition feature data as the training target; the effect detection model is trained until the sum of prediction errors reaches convergence, and the training is stopped; the effect detection model is any one of a deep neural network model or a deep belief network model.

[0040] Further, the influence coefficient of the suppression technology corresponding to the largest number of suppression technology labels on the remaining harmonic currents is calculated to determine whether to generate a reselection instruction;

[0041] The calculation method of the influence coefficient is as follows:

[0042]

[0043] In the formula, YX i is the influence coefficient of the i-th harmonic current, C i is the harmonic order of the i-th harmonic current, FZ i is the amplitude of the i-th harmonic current, XW i is the phase of the i-th harmonic current, JS is the suppression technology value, ω1, ω2, ω3, ω4 are the preset weights;

[0044] The method for determining whether to generate a reselection instruction includes:

[0045] Compare the impact coefficient to the impact threshold;

[0046] If the influence coefficients are all less than the influence threshold, no reselection instruction is generated;

[0047] If there is an influence coefficient greater than or equal to the influence threshold, a reselection instruction is generated;

[0048] If a reselection instruction is generated, the real-time suppression feature data is input into the effect detection model, the suppression quantity is output, and the subsequent suppression technology is reselected.

[0049] A system for selecting a harmonic current suppression technology in a power grid is implemented according to the method for selecting a harmonic current suppression technology in a power grid, and comprises:

[0050] A first data acquisition module collects current signals of n nodes;

[0051] The data analysis module processes and analyzes the current signal to obtain the harmonic current characteristic data and the harmonic current quantity of the harmonic current in the current signal;

[0052] The second data acquisition module collects Q groups of historical grid status data, the historical grid status data includes grid characteristic data and suppression technology corresponding to the grid characteristic data, the grid characteristic data includes grid characteristic data and harmonic current characteristic data;

[0053] The first model training module trains a method selection model for predicting suppression technology labels based on historical power grid status data;

[0054] The regional division module divides the power grid into v regions, where v is an integer greater than 1 and v <n;

[0055] The area screening module determines whether to mark the area as a suppression area based on the number of harmonic currents in each of the v areas and the total number of harmonic currents in the v areas;

[0056] The suppression technology selection module inputs the real-time power grid characteristic data of the suppression area into the trained method selection model, outputs the corresponding real-time suppression technology label, counts the number of real-time suppression technology labels of each suppression area, calculates the mean and standard deviation of the number of suppression technology labels of each suppression area, determines whether to generate discrete selection instructions or centralized selection instructions, and selects the real-time suppression technology of the suppression area according to the generated instructions.

[0057] The technical effects and advantages of the system and method for selecting harmonic current suppression technology in a power grid of the present invention are as follows:

[0058] 1. For the harmonic currents in different areas of the power grid, analyze the number of harmonic currents and the number of suppression technology tags, and select the best harmonic current suppression technology, which can effectively suppress the harmonic current, reduce the impact of harmonic current on the power grid, save suppression costs, and reduce power grid energy consumption; at the same time, it saves the worker test link and improves work efficiency; and it can monitor the suppression effect in real time and adjust the suppression strategy in time.

[0059] 2. By predicting the suppression effect, select multiple appropriate suppression technologies for the same area to avoid excessive suppression of harmonic currents, reduce suppression costs, and prevent power loss and harmonic resonance. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of a harmonic current suppression technology selection system in a power grid according to Embodiment 1 of the present invention;

[0061] Figure 2 This is a schematic diagram of area division in Example 1 of the present invention;

[0062] Figure 3 A schematic diagram of a harmonic current suppression technology selection system in a power grid according to Embodiment 2 of the present invention;

[0063] Figure 4 A schematic diagram of a method for selecting a harmonic current suppression technology in a power grid according to Embodiment 3 of the present invention;

[0064] Figure 5 This is a schematic diagram of the structure of an electronic device according to Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of 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.

[0066] Example 1

[0067] See also Figure 1 As shown, a harmonic current suppression technology selection system in a power grid described in this embodiment includes a first data acquisition module, a data analysis module, a second data acquisition module, a first model training module, a region division module, a region screening module and a suppression technology selection module, and each module is connected by wire and / or wirelessly;

[0068] The first data acquisition module collects current signals of n nodes, where n is an integer greater than 1; the current signals are acquired by current transformers installed on the n nodes, and one current transformer is installed on each node; the node acquisition method is: the grid topology is acquired through the grid planning drawing, and the nodes are set according to the grid topology. The topology is the connection mode and layout form between various components in the grid, and the n components and branch points in the grid are used as nodes, such as substations, transformers, etc.;

[0069] The data analysis module processes and analyzes the current signal to obtain the harmonic current characteristic data and the harmonic current quantity of the harmonic current in the current signal;

[0070] In the power grid, a current signal can contain multiple frequency components, among which the fundamental current is the most important component, representing the standard operating frequency of the power grid; however, due to various factors, such as nonlinear loads, power electronic equipment, etc., there may be other frequency components in the current signal, which are harmonic currents; harmonic current characteristic data include type, amplitude and phase; type, amplitude and phase are harmonic current type, harmonic current amplitude and harmonic current phase; harmonic current characteristic data is obtained by calculating and analyzing the current signal;

[0071] The method for obtaining harmonic current characteristic data and harmonic current quantity is as follows:

[0072] By performing fast Fourier transform on the current signal, the current signal in the time domain can be converted into an energy spectrum in the frequency domain. The fast Fourier transform is Where X(m) represents the mth frequency component in the frequency domain, X(n) represents the nth current signal in the time domain, N represents the total number of current signals, e is the natural logarithmic constant, j is the imaginary unit, and π is the circumference of a circle;

[0073] In the frequency domain, harmonic currents appear as frequency components different from the fundamental frequency f1. These frequency components correspond to the harmonic frequencies f c , the frequency component includes harmonic frequency, amplitude and phase. Different harmonic current amplitudes and phases are caused by different nonlinear loads and have different effects on the power grid; the harmonic order c is the ratio of the harmonic frequency to the fundamental frequency, that is, The harmonic order determines the type of harmonic current. Different types of harmonic current have different effects on the power grid and use different suppression technologies. For example:

[0074] Odd harmonics refer to harmonic components whose frequencies are odd multiples of the fundamental frequency. They can be suppressed by symmetrical three-phase filters or by using the harmonic suppression function of transformers. Odd harmonics include the 3rd, 5th, and 7th harmonics.

[0075] Even harmonics refer to harmonic components whose frequencies are even multiples of the fundamental frequency. They can be filtered out by harmonic filters or transformers with low-pass filters. Even harmonics include 2nd, 4th, 6th, etc.

[0076] That is, different suppression technologies are required for harmonic currents with different harmonic current characteristic data;

[0077] A second data acquisition module collects Q groups of historical grid status data, where Q is an integer greater than 1, and the historical grid status data includes grid characteristic data and suppression technology corresponding to the grid characteristic data, and the grid characteristic data includes grid characteristic data and harmonic current characteristic data;

[0078] The grid characteristic data includes the voltage level, current capacity and grid structure of the grid. The grid characteristic data is obtained through the grid planning drawings and input into the grid control system by workers. In a grid with high voltage and large current capacity, there are more power equipment and loads, which will introduce more harmonic currents. Different grid structures will affect the propagation and distribution of harmonic currents. Therefore, the grid characteristic data will affect the selection of suppression technology.

[0079] In an experimental environment, current signals with different harmonic current characteristic data are subjected to harmonic current suppression in accordance with a plurality of preset suppression technologies in a grid with different grid characteristic data, and the current signal conditions are recorded in sequence. The current signal is subjected to fast Fourier transform decomposition to obtain harmonic current characteristic data of different harmonic currents, and the suppression technology corresponding to the minimum harmonic current characteristic data is used as the suppression technology corresponding to the harmonic current of the grid characteristic data. The suppression technologies corresponding to multiple groups of grid characteristic data may be the same.

[0080] The first model training module trains a method selection model for predicting suppression technology labels based on historical power grid status data. The training process of the method selection model includes:

[0081] Label the suppression technology with a tag, and the tag is a numerical tag;

[0082] Use the historical power grid status data as a sample set, divide the sample set into a training set and a test set, construct a classifier, use the power grid feature data in the training set as input data, use the suppression technology tags in the training set as output data, train the classifier to obtain an initial classifier, and use the test set to test the initial classifier. Output the initial classifier that meets the prediction error as the method selection model, where the calculation formula for the prediction error is z k =(α k -μ k ) 2 , where z k is the prediction accuracy, k is the number of the power grid feature data, α k is the predicted suppression technology tag corresponding to the kth group of power grid feature data, and μ k is the actual suppression technology tag corresponding to the kth group of power grid feature data;

[0083] The above classifier is specifically one of a logistic regression model, a naive Bayes model or a support vector machine model;

[0084] The area division module divides the power grid into v areas, where v is an integer greater than 1 and v < n, and n nodes are evenly distributed in v areas. Please refer to Figure 2 as shown;

[0085] The area screening module determines whether to mark it as a suppression area according to the harmonic current quantity in each of the v areas and the total harmonic current quantity in the v areas;

[0086] The method for determining whether to mark it as a suppression area includes:

[0087] n nodes are evenly distributed in v areas. Collect the harmonic current quantity in each of the v areas and the total harmonic current quantity; divide the harmonic current quantity in each area by the total harmonic current quantity to obtain the proportion of the harmonic current quantity in each area;

[0088] If the proportion of the harmonic current quantity is greater than or equal to , then mark this area as a suppression area, indicating that the harmonic current quantity in this area is greater than the average value of the harmonic current quantity, and the harmonic current quantity in this area is relatively large and needs to be suppressed in time;

[0089] If the proportion of the harmonic current quantity is less than , then do not mark this area as a suppression area, indicating that the harmonic current quantity in this area is relatively small;

[0090] The suppression technology selection module inputs the real-time power grid characteristic data of the suppression area into the trained method selection model, outputs the corresponding real-time suppression technology label, counts the number of real-time suppression technology labels of each suppression area, calculates the mean and standard deviation of the number of suppression technology labels of each suppression area, determines whether to generate discrete selection instructions or centralized selection instructions, and selects the real-time suppression technology of the suppression area according to the generated instructions;

[0091] The reason for selecting suppression technology is that suppression technology is divided into active suppression technology and passive suppression technology. Active suppression technology is a technology that actively suppresses harmonic currents by adding external suppression equipment, such as active filters, harmonic filters, and harmonic suppression transformers. Passive suppression technology is a technology that passively suppresses harmonic currents by introducing appropriate circuit elements, such as inductors and capacitors, to change the circuit impedance. In the existing technology, compared with passive suppression technology, active suppression technology can more effectively suppress harmonic currents and improve the power quality of power grids and load equipment, with better suppression effects. However, a current signal in the power grid may have multiple harmonic currents, which requires the use of a combination of multiple suppression technologies, that is, adding multiple suppression devices, which makes the control system more complex and the suppression cost higher. Therefore, it is necessary to select a suitable suppression technology.

[0092] The number of inhibition technology labels in an inhibition area is taken as an analysis set, and the mean and standard deviation of the number of inhibition technology labels in each analysis set are calculated in turn;

[0093] The mean number of suppression technology labels is calculated as follows:

[0094]

[0095] In the formula, is the mean number of suppression technology labels in the h-th suppression region, YZ hi is the number of labels of the ith suppression technique in the hth suppression region, h∈[1,v], R h is the number of suppression technique label types in the h-th suppression region, i∈[1,R h ];

[0096] The standard deviation is calculated as follows:

[0097]

[0098] Where, σY h is the standard deviation of the h-th inhibition area;

[0099] The method for determining whether to generate a discrete selection instruction or a centralized selection instruction includes:

[0100] A standard deviation threshold is preset. If the standard deviation of the suppression area is greater than the standard deviation threshold, a discrete selection instruction is generated. According to the discrete selection instruction, a suppression technology label whose number is greater than the mean number of suppression technology labels is selected, and the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area. This indicates that the number of suppression technology labels in the suppression area is not concentrated near the mean number of suppression technology labels, and the number of each suppression technology label varies greatly. It is necessary to select a suppression technology label with a large number of suppression technology labels, thereby reducing the suppression cost and improving the suppression efficiency.

[0101] If the standard deviation of the suppression area is less than or equal to the standard deviation threshold, a concentrated selection instruction is generated; according to the concentrated selection instruction, suppression technology labels whose number is greater than the selection threshold are selected, and the selection threshold is the mean number of suppression technology labels minus the standard deviation; the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area; this indicates that the number of suppression technology labels in the area is generally concentrated near the mean number of suppression technology labels, and the number of each suppression technology label is not much different, and all need to be selected;

[0102] The suppression method selected for each suppression area is sent to the power grid control system and displayed on the central control screen, so that workers can obtain the specific suppression method corresponding to each suppression area;

[0103] It should be noted that the standard deviation threshold is the worker's collection of multiple suppression technology labels in the experimental environment, which are concentrated in the suppression area near the mean number of suppression technology labels, and the corresponding standard deviation is calculated, and the mean of multiple standard deviations is used as the standard deviation threshold;

[0104] This embodiment analyzes the harmonic currents in different areas of the power grid according to the number of harmonic currents and the number of suppression technology tags, and selects the best harmonic current suppression technology, which can effectively suppress harmonic currents, reduce the impact of harmonic currents on the power grid, save suppression costs, and reduce power grid energy consumption; at the same time, it eliminates the worker test link and improves work efficiency; and it can monitor the suppression effect in real time and adjust the suppression strategy in time.

[0105] Example 2

[0106] See also Figure 3As shown, this embodiment further improves the design on the basis of embodiment 1. In embodiment 1, at least one harmonic current suppression technology is selected for each suppression area. In the actual harmonic current suppression process, after using a suppression technology, other harmonic currents in the same area will also be affected, resulting in a change in the number of suppression technology labels used to select the harmonic current suppression technology in embodiment 1. At this time, if the suppression technology selected by the suppression technology selection module is used for re-suppression, excessive harmonic current suppression will occur, thereby increasing the suppression cost, causing power loss, and generating harmonic resonance. Therefore, this embodiment provides a harmonic current suppression technology selection system in a power grid, which also includes a second model training module, an impact analysis module, and a suppression effect prediction module. It predicts the degree of influence on the remaining harmonic currents after using a suppression technology. If the degree of influence is large, the change in the corresponding number of each suppression technology label in the same suppression area is predicted to determine whether the corresponding suppression technology needs to be changed.

[0107] The second model training module marks the quantity corresponding to each inhibition technology label before inhibition of the inhibition area as the quantity before inhibition, marks the quantity corresponding to each inhibition technology label after inhibition of the inhibition area as the quantity after inhibition, uses the quantity before inhibition and the inhibition technology corresponding to the quantity before inhibition as inhibition feature data, and uses the inhibition feature data and the quantity after inhibition corresponding to the inhibition feature data as inhibition state data; based on the inhibition state data, trains an effect detection model for predicting the quantity after inhibition;

[0108] The specific training process of the effect detection model includes:

[0109] In an experimental environment, the corresponding suppression technology is used to suppress multiple harmonic currents for power equipment with the same suppression characteristic data in turn, and the suppressed quantity of the power equipment after suppression is recorded in turn, and the average of the suppressed quantity after multiple harmonic current suppression is used as the suppressed quantity corresponding to the suppression characteristic data; the corresponding suppressed quantity is obtained in turn for power equipment with different suppression characteristic data according to the same logic;

[0110] The inhibition state data is used as the input of the effect detection model, and the effect detection model uses the predicted inhibition quantity corresponding to each group of inhibition feature data as the output, the actual inhibition quantity corresponding to the group of inhibition feature data as the prediction target, and the minimization of the sum of the prediction errors of all inhibition feature data as the training target; wherein the calculation formula of the prediction error is z p =(α p -μ p ) 2 , where z p is the prediction error, p is the number of suppressed feature data, α p is the predicted suppression quantity corresponding to the p-th group of suppression feature data, μ pis the actual number of suppressed features corresponding to the p-th group of suppressed feature data; the effect detection model is trained until the sum of the prediction errors reaches convergence and the training is stopped;

[0111] The effect detection model is any one of a deep neural network model or a deep belief network model;

[0112] The impact analysis module calculates the impact coefficient of the suppression technology corresponding to the largest number of suppression technology labels on the remaining harmonic currents and determines whether to generate a reselection instruction;

[0113] The calculation method of the influence coefficient is as follows:

[0114]

[0115] In the formula, YX i is the influence coefficient of the i-th harmonic current, C i is the harmonic order of the i-th harmonic current, FZ i is the amplitude of the i-th harmonic current, XW i is the phase of the i-th harmonic current, JS is the suppression technology value, ω1, ω2, ω3, ω4 are the preset weights;

[0116] The specific value of the weight coefficient in the formula can be set according to the actual situation. The weight coefficient reflects the importance of each harmonic current characteristic data and the suppression technology value. The technicians in this field can preset the corresponding proportional coefficient according to the importance of each harmonic current characteristic data and the suppression technology value, so as to accurately evaluate the influence coefficient;

[0117] The suppression technology value is assigned by the staff to the suppression technology. When the staff records the suppressed number of the suppressed power equipment in sequence, the staff assigns a smaller value to the suppression technology corresponding to the smaller suppressed number, that is, the smaller the suppression technology value, the greater the impact on the remaining harmonic currents;

[0118] The method for determining whether to generate a reselection instruction includes:

[0119] Compare the impact coefficient to the impact threshold;

[0120] If the influence coefficients are all less than the influence threshold, no reselection instruction is generated; this indicates that the use of the suppression technology at this time has little impact on the remaining harmonic currents, and there is no need to regenerate the remaining suppression technologies;

[0121] If there is an influence coefficient greater than or equal to the influence threshold, a reselection instruction is generated; this indicates that the use of the suppression technology at this time has a greater impact on the remaining harmonic currents, and the remaining suppression technologies need to be regenerated;

[0122] The impact threshold is that the staff uses the suppression technology corresponding to the largest number of suppression technology labels on the harmonic current multiple times under the experimental environment, and calculates the impact coefficients in turn, taking the average of the impact coefficients as the impact threshold;

[0123] The suppression effect prediction module, if a reselection instruction is generated, inputs the real-time suppression feature data into the effect detection model, outputs the suppression quantity, sends the suppression quantity to the suppression technology selection module, and reselects the subsequent suppression technology. The suppression technology in the real-time suppression feature data is the suppression technology corresponding to the actual suppression technology label with the largest number in the same suppression area;

[0124] This embodiment predicts the suppression effect and selects multiple appropriate suppression technologies for the same area, thereby avoiding excessive suppression of harmonic currents, reducing suppression costs, and preventing power loss and harmonic resonance.

[0125] Example 3

[0126] See also Figure 4 As shown, the part not described in detail in this embodiment is described in the description of embodiments 1 and 2, and a method for selecting a harmonic current suppression technology in a power grid is provided, including:

[0127] Collect current signals of n nodes;

[0128] Process and analyze the current signal to obtain harmonic current characteristic data and the amount of harmonic current in the current signal;

[0129] Collecting Q groups of historical grid status data, the historical grid status data including grid characteristic data and suppression technology corresponding to the grid characteristic data, the grid characteristic data including grid characteristic data and harmonic current characteristic data;

[0130] Based on historical grid status data, a method selection model is trained to predict the suppression technology labels;

[0131] The power grid is divided into v areas, where v is an integer greater than 1 and v <n;

[0132] According to the number of harmonic currents in each of the v areas and the total number of harmonic currents in the v areas, determine whether to mark it as a suppression area;

[0133] The real-time power grid characteristic data of the suppression area is input into the trained method selection model, and the corresponding real-time suppression technology label is output. The number of real-time suppression technology labels of each suppression area is counted, and the mean and standard deviation of the number of suppression technology labels of each suppression area are calculated. It is determined whether discrete selection instructions or centralized selection instructions are generated, and the real-time suppression technology of the suppression area is selected according to the generated instructions.

[0134] Furthermore, the power grid topology is obtained through the power grid planning drawings, and the nodes are set according to the power grid topology.

[0135] Furthermore, the harmonic current characteristic data includes harmonic order, amplitude and phase, the amplitude is the harmonic current amplitude, and the phase is the harmonic current phase;

[0136] Methods for obtaining harmonic current characteristic data and harmonic current quantity include:

[0137] By performing fast Fourier transform on the current signal, the current signal in the time domain can be converted into an energy spectrum in the frequency domain. The fast Fourier transform is Where X(m) represents the mth frequency component in the frequency domain, X(n) represents the nth current signal in the time domain, N represents the total number of current signals, e is the natural logarithmic constant, j is the imaginary unit, and π is the circumference of a circle;

[0138] The fundamental frequency is obtained according to the energy spectrum; the frequency component includes harmonic frequency, amplitude and phase; the harmonic order c is the harmonic frequency f c The ratio of the fundamental frequency f1 is

[0139] Furthermore, the grid characteristic data includes the voltage level, current capacity and grid structure of the grid; the training process of the method selection model includes:

[0140] Label the suppression techniques with numerical labels;

[0141] Historical power grid status data is used as a sample set, and the sample set is divided into a training set and a test set. A classifier is constructed, and the power grid feature data in the training set is used as input data, and the suppression technology label in the training set is used as output data. The classifier is trained to obtain an initial classifier, and the initial classifier is tested using the test set. The initial classifier that meets the prediction error is output as a method selection model, and the classifier is one of a logistic regression model, a naive Bayes model, or a support vector machine model.

[0142] Furthermore, the method for determining whether to mark as an inhibition area is as follows:

[0143] n nodes are evenly distributed in v areas, and the number of harmonic currents and the total number of harmonic currents in each of the v areas are collected; the number of harmonic currents in each area is divided by the total number of harmonic currents to obtain the proportion of harmonic currents in each area;

[0144] If the proportion of harmonic current is greater than or equal to Then the area is marked as an inhibition area;

[0145] If the proportion of harmonic current is less than The area is not marked as an inhibition area.

[0146] Furthermore, the mean number of suppression technology labels is calculated as follows:

[0147]

[0148] In the formula, is the mean number of suppression technology labels in the h-th suppression region, YZ hi is the number of labels of the ith suppression technique in the hth suppression region, h∈[1,v], R h is the number of suppression technique label types in the h-th suppression region, i∈[1,R h ];

[0149] The standard deviation is calculated as follows:

[0150]

[0151] Where, σY h is the standard deviation of the h-th inhibition zone.

[0152] Further, the method for determining whether to generate a discrete selection instruction or a centralized selection instruction includes:

[0153] A standard deviation threshold is preset, and if the standard deviation of the suppression area is greater than the standard deviation threshold, a discrete selection instruction is generated; according to the discrete selection instruction, a suppression technology label whose number is greater than the average number of suppression technology labels is selected, and the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area;

[0154] If the standard deviation of the suppression area is less than or equal to the standard deviation threshold, a centralized selection instruction is generated; according to the centralized selection instruction, suppression technology labels whose number is greater than the selection threshold are selected; and the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area.

[0155] Further, the number corresponding to each suppression technology label before suppression of the suppression area is marked as the number before suppression, the number corresponding to each suppression technology label after suppression of the suppression area is marked as the number after suppression, the number before suppression and the suppression technology corresponding to the number before suppression are used as suppression feature data, and the suppression feature data and the number after suppression corresponding to the suppression feature data are used as suppression state data; based on the suppression state data, an effect detection model for predicting the number after suppression is trained;

[0156] The specific training process of the effect detection model includes:

[0157] In an experimental environment, the corresponding suppression technology is used to suppress multiple harmonic currents on power equipment with the same suppression characteristic data in turn, and the suppressed quantity of the power equipment after suppression is recorded in turn, and the average of the suppressed quantity after multiple harmonic current suppression is used as the suppressed quantity corresponding to the suppression characteristic data; the corresponding suppressed quantity is obtained in turn for power equipment with different suppression characteristic data according to the method;

[0158] The inhibition state data is used as the input of the effect detection model, the effect detection model takes the predicted inhibition quantity corresponding to each group of inhibition feature data as the output, takes the actual inhibition quantity corresponding to the group of inhibition feature data as the prediction target, and takes minimizing the sum of prediction errors of all inhibition feature data as the training target; the effect detection model is trained until the sum of prediction errors reaches convergence, and the training is stopped; the effect detection model is any one of a deep neural network model or a deep belief network model.

[0159] Further, the influence coefficient of the suppression technology corresponding to the largest number of suppression technology labels on the remaining harmonic currents is calculated to determine whether to generate a reselection instruction;

[0160] The calculation method of the influence coefficient is as follows:

[0161]

[0162] In the formula, YX i is the influence coefficient of the i-th harmonic current, C i is the harmonic order of the i-th harmonic current, FZ i is the amplitude of the i-th harmonic current, XW i is the phase of the i-th harmonic current, JS is the suppression technology value, ω1, ω2, ω3, ω4 are the preset weights;

[0163] The method for determining whether to generate a reselection instruction includes:

[0164] Compare the impact coefficient to the impact threshold;

[0165] If the influence coefficients are all less than the influence threshold, no reselection instruction is generated;

[0166] If there is an influence coefficient greater than or equal to the influence threshold, a reselection instruction is generated;

[0167] If a reselection instruction is generated, the real-time suppression feature data is input into the effect detection model, the suppression quantity is output, and the subsequent suppression technology is reselected.

[0168] Example 4

[0169] See also Figure 5As shown, this embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a method for selecting a harmonic current suppression technology in a power grid provided by the above methods is implemented.

[0170] Since the electronic device introduced in this embodiment is an electronic device used to implement a method for selecting a harmonic current suppression technology in a power grid in the embodiment of the present application, based on the method for selecting a harmonic current suppression technology in a power grid introduced in the embodiment of the present application, a person skilled in the art can understand the specific implementation of the electronic device of the present embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present application is not described in detail here. As long as a person skilled in the art implements the electronic device used in the method for selecting a harmonic current suppression technology in a power grid in the embodiment of the present application, it belongs to the scope of protection of the present application.

[0171] Example 5

[0172] This embodiment discloses a computer-readable storage medium, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a method for selecting a harmonic current suppression technology in a power grid provided by the above methods is implemented.

[0173] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0174] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for selecting harmonic current suppression technology in a power grid, characterized in that: include: Collect current signals of n nodes; Process and analyze the current signal to obtain harmonic current characteristic data and the amount of harmonic current in the current signal; Collecting Q groups of historical grid status data, the historical grid status data including grid characteristic data and suppression technology corresponding to the grid characteristic data, the grid characteristic data including grid characteristic data and harmonic current characteristic data; Based on historical grid status data, a method selection model is trained to predict the suppression technology labels; The power grid is divided into v areas, where v is an integer greater than 1 and v <n; According to the number of harmonic currents in each of the v areas and the total number of harmonic currents in the v areas, determine whether to mark it as a suppression area; Methods for determining whether to mark an inhibition area include: n nodes are evenly distributed in v areas, and the number of harmonic currents in each of the v areas and the total number of harmonic currents in the v areas are collected; the number of harmonic currents in each area is divided by the total number of harmonic currents to obtain the proportion of harmonic currents in each area; If the proportion of harmonic current is greater than or equal to Then the area is marked as an inhibition area; If the proportion of harmonic current is less than Then the area is not marked as an inhibition area; Input the real-time power grid characteristic data of the suppression area into the trained method selection model, output the corresponding real-time suppression technology label, count the number of real-time suppression technology labels of each suppression area, calculate the mean and standard deviation of the number of suppression technology labels of each suppression area, determine whether to generate discrete selection instructions or centralized selection instructions, and select the real-time suppression technology of the suppression area according to the generated instructions; The mean number of suppression technology labels is calculated as follows: In the formula, is the mean number of suppression technology labels in the h-th suppression region, YZ hi is the number of labels of the ith suppression technique in the hth suppression region, h∈[1,v], R h is the number of suppression technique label types in the h-th suppression region, i∈[1,R h ]; The standard deviation is calculated as follows: Where, σY h is the standard deviation of the h-th inhibition area; The method for determining whether to generate a discrete selection instruction or a centralized selection instruction includes: A standard deviation threshold is preset, and if the standard deviation of the suppression area is greater than the standard deviation threshold, a discrete selection instruction is generated; according to the discrete selection instruction, a suppression technology label whose number is greater than the average number of suppression technology labels is selected, and the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area; If the standard deviation of the suppression area is less than or equal to the standard deviation threshold, a centralized selection instruction is generated; according to the centralized selection instruction, suppression technology labels whose number is greater than the selection threshold are selected; and the suppression technology corresponding to the suppression technology label is used as the suppression technology used in the area; Mark the quantity corresponding to each suppression technology label before suppression of the suppression area as the quantity before suppression, mark the quantity corresponding to each suppression technology label after suppression of the suppression area as the quantity after suppression, use the quantity before suppression and the suppression technology corresponding to the quantity before suppression as suppression feature data, use the suppression feature data and the quantity after suppression corresponding to the suppression feature data as suppression state data; based on the suppression state data, train an effect detection model for predicting the quantity after suppression; The effect detection model training process includes: Preliminarily, the corresponding suppression technology is used for the power equipment with the same suppression characteristic data in sequence to suppress multiple harmonic currents, and the suppressed quantity of the power equipment after suppression is recorded in sequence, and the average of the suppressed quantity after the multiple harmonic current suppression is used as the suppressed quantity corresponding to the suppression characteristic data; and the corresponding suppressed quantity is obtained in sequence for the power equipment with different suppression characteristic data according to the same logic; The inhibition state data is used as input of an effect detection model, the effect detection model uses the predicted inhibition quantity corresponding to each group of inhibition feature data as output, uses the actual inhibition quantity corresponding to the group of inhibition feature data as a prediction target, and uses minimizing the sum of prediction errors of all inhibition feature data as a training target; the effect detection model is trained until the sum of prediction errors reaches convergence and the training is stopped; the effect detection model is any one of a deep neural network model or a deep belief network model; Calculate the influence coefficient of the suppression technology corresponding to the largest number of suppression technology labels on the remaining harmonic currents, and determine whether to generate a reselection instruction; The calculation method of the influence coefficient is as follows: In the formula, YX i is the influence coefficient of the i-th harmonic current, C i is the harmonic order of the i-th harmonic current, FZ i is the amplitude of the i-th harmonic current, XW i is the phase of the i-th harmonic current, JS is the suppression technology value, ω1, ω2, ω3, ω4 are the preset weights; The method for determining whether to generate a reselection instruction includes: Compare the impact coefficient to the impact threshold; If the influence coefficients are all less than the influence threshold, no reselection instruction is generated; If there is an influence coefficient greater than or equal to the influence threshold, a reselection instruction is generated; If a reselection instruction is generated, the real-time suppression feature data is input into the effect detection model, the suppression quantity is output, and the subsequent suppression technology is reselected.

2. A method for selecting harmonic current suppression technology in a power grid according to claim 1, characterized in that: Obtain the grid topology through the grid planning drawings and set nodes according to the grid topology.

3. The method for selecting harmonic current suppression technology in a power grid according to claim 2, characterized in that: The harmonic current characteristic data includes harmonic order, amplitude and phase, the amplitude is the harmonic current amplitude, and the phase is the harmonic current phase; Methods for obtaining harmonic current characteristic data and harmonic current quantity include: By performing fast Fourier transform on the current signal, the current signal in the time domain is converted into an energy spectrum in the frequency domain. The fast Fourier transform is Where X(m) represents the mth frequency component in the frequency domain, X(n) represents the nth current signal in the time domain, N represents the total number of current signals, e is the natural logarithmic constant, j is the imaginary unit, and π is the circumference of a circle; The fundamental frequency is obtained according to the energy spectrum; the frequency component includes harmonic frequency, amplitude and phase; the harmonic order c is the harmonic frequency f c The ratio of the fundamental frequency f1 is 4. The method for selecting harmonic current suppression technology in a power grid according to claim 3, characterized in that: The grid characteristic data includes the voltage level, current capacity and grid structure of the grid; the training process of the method selection model includes: Label the suppression techniques with numerical labels; Historical power grid status data is used as a sample set, and the sample set is divided into a training set and a test set. A classifier is constructed, and the power grid feature data in the training set is used as input data, and the suppression technology label in the training set is used as output data. The classifier is trained to obtain an initial classifier, and the initial classifier is tested using the test set. The initial classifier that meets the prediction error is output as a method selection model, and the classifier is one of a logistic regression model, a naive Bayes model, or a support vector machine model.

5. A system for selecting harmonic current suppression technology in a power grid, implemented according to a method for selecting harmonic current suppression technology in a power grid according to any one of claims 1 to 4, characterized in that: include: A first data acquisition module collects current signals of n nodes; The data analysis module processes and analyzes the current signal to obtain the harmonic current characteristic data and the harmonic current quantity of the harmonic current in the current signal; The second data acquisition module collects Q groups of historical grid status data, the historical grid status data includes grid characteristic data and suppression technology corresponding to the grid characteristic data, the grid characteristic data includes grid characteristic data and harmonic current characteristic data; The first model training module trains a method selection model for predicting suppression technology labels based on historical power grid status data; The regional division module divides the power grid into v regions, where v is an integer greater than 1 and v <n; The area screening module determines whether to mark the area as a suppression area based on the number of harmonic currents in each of the v areas and the total number of harmonic currents in the v areas; The suppression technology selection module inputs the real-time power grid characteristic data of the suppression area into the trained method selection model, outputs the corresponding real-time suppression technology label, counts the number of real-time suppression technology labels of each suppression area, calculates the mean and standard deviation of the number of suppression technology labels of each suppression area, determines whether to generate discrete selection instructions or centralized selection instructions, and selects the real-time suppression technology of the suppression area according to the generated instructions.

Images