Power Quality Disturbance Identification Method and System Based on Chaotic Tianying Optimizer

Abstract

The present invention relates to the technical field of power quality disturbance signals, and provides a power quality disturbance identification method and system based on a chaotic Tianying optimizer, including: acquiring power quality disturbance data; after extracting feature vectors from the power quality disturbance data, obtaining a classification result by using a neural network; wherein, during the training process of the neural network, the number of neurons in the middle layer is optimized by using a chaotic Tianying optimizer, and the exploitation coefficient of the chaotic Tianying optimizer is obtained through tent mapping and Sine mapping. The performance of the neural network for power quality disturbance identification is greatly improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power quality disturbance signals, and particularly relates to a power quality disturbance identification method and system based on a chaotic eagle optimizer. Background Technique

[0002] The statements in this part only provide background technical information related to the present invention, and do not necessarily constitute prior art.

[0003] Power quality encompasses a wide range of electromagnetic phenomena related to the interaction between power system networks and end-user equipment. With the increasing grid connection rate of new energy sources such as wind energy and solar energy conversion systems and complex power distribution, power quality has deteriorated further. Therefore, due to the increasing use of sensitive solid-state converters in non-linear energy-intensive equipment, the demand for clean energy has also increased. The voltage waveform pollution in the distribution system depends on the different patterns of the current waveforms generated by these non-linear loads. These distorted load currents will cause voltage distortion at the point of common coupling of the distribution system. Continuous processes, multi-stage batch operations, and data processing are mainly affected by poor power quality. Any power quality disturbance will interrupt the production process in industries such as papermaking and semiconductors, resulting in production suspension and huge production losses. In addition, restarting a large number of workstations and resuming suspended transactions is a time-consuming process. Summary of the Invention

[0004] In order to solve the technical problems existing in the above background technique, the present invention provides a power quality disturbance identification method and system based on a chaotic eagle optimizer, which uses the chaotic eagle optimizer to obtain the optimal number of intermediate layer nodes of the neural network, and by replacing the exploitation coefficients α and β in the chaotic eagle optimizer with the tent map and the Sine map respectively, greatly improves the performance of the neural network in identifying power quality disturbances.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] The first aspect of the present invention provides a power quality disturbance identification method based on a chaotic eagle optimizer, which includes:

[0007] Obtain power quality disturbance data;

[0008] After extracting the feature vectors from the power quality disturbance data, use the neural network to obtain the classification result;

[0009] Wherein, during the training process of the neural network, the chaotic eagle optimizer is used to optimize the number of intermediate layer neurons, and the exploitation coefficients of the chaotic eagle optimizer are obtained through the tent map and the Sine map.

[0010] Further, the specific steps of using the chaotic eagle optimizer to optimize the number of intermediate layer neurons are:

[0011] Initialize the population of the chaotic eagle optimizer, where each individual represents the number of neurons.

[0012] Train the neural network with the number of neurons.

[0013] Use the trained neural network to perform classification multiple times, calculate the average value of the classification result accuracy, and record the average value of the accuracy as the fitness value of the individual.

[0014] Use the chaotic eagle optimizer for iterative optimization, and determine the number of neurons in the middle layer of the neural network as the individual with the highest fitness value.

[0015] Further, the specific steps of using the chaotic eagle optimizer for iterative optimization are as follows:

[0016] When the number of iterations is in the first interval, the eagle optimizer randomly selects one of the expanding exploration scheme and the shrinking exploration scheme to execute.

[0017] When the number of iterations is in the second interval, the eagle optimizer randomly selects one of the expanding exploitation scheme and the shrinking exploitation scheme to execute.

[0018] Further, the convergence condition during the training process of the neural network includes: the average cosine similarity is not less than the threshold.

[0019] Further, use the S-transform to extract feature vectors from the power quality disturbance data.

[0020] Further, the feature vectors include: half of the average value at 50 Hz, the maximum peak value at 50 Hz, the minimum peak value at 50 Hz, the average value at 150 Hz, the average value at 250 Hz, the average value at 350 Hz, the average value at 450 Hz, the average value at 550 Hz, and the average value at 700 - 2500 Hz.

[0021] Further, the power quality disturbance data is divided into five categories, namely: normal data, voltage sag, harmonics, voltage flicker, and voltage swell data.

[0022] The second aspect of the present invention provides a power quality disturbance identification system based on a chaotic eagle optimizer, which includes:

[0023] A data acquisition module, which is configured to: acquire power quality disturbance data;

[0024] An identification module, which is configured to: after extracting feature vectors from the power quality disturbance data, obtain classification results using a neural network;

[0025] Among them, during the training process of the neural network, the chaotic eagle optimizer is used to optimize the number of neurons in the middle layer, and the exploitation coefficients of the chaotic eagle optimizer are obtained through tent mapping and Sine mapping.

[0026] The third aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the steps in the power quality disturbance recognition method based on the chaotic eagle optimizer as described above.

[0027] The fourth aspect of the present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the steps in the power quality disturbance recognition method based on the chaotic eagle optimizer as described above.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] The present invention provides a power quality disturbance recognition method based on the chaotic eagle optimizer. On the premise of obtaining a number of data with power quality disturbance labels, a classifier based on the BP neural network model is established, and this classifier can accurately identify different power quality disturbances.

[0030] The present invention provides a power quality disturbance recognition method based on the chaotic eagle optimizer. Based on the BP neural network model, the average cosine similarity is applied as an additional convergence condition, making the trained improved BP neural network model more accurate.

[0031] The present invention provides a power quality disturbance recognition method based on the chaotic eagle optimizer. Based on the eagle optimizer, by replacing the exploitation coefficients α and β in the chaotic eagle optimizer with tent mapping and Sine mapping respectively, the performance of the algorithm is greatly improved, and the chaotic eagle optimizer is used to obtain the optimal number of middle layer nodes of the improved BP neural network. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings forming a part of this invention are used to provide a further understanding of the invention. The schematic embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0033] Figure 1 is a flowchart of the power quality disturbance recognition method based on the chaotic eagle optimizer in Embodiment 1 of the present invention;

[0034] Figure 2 is a pie chart of the average classification accuracy of 20 times in Embodiment 1 of the present invention;

[0035] Figure 3It is the bar chart of the accuracy rate of different power quality disturbance signals in the first embodiment of the present invention;

[0036] Figure 4 It is the waveform schematic diagram of five power quality interference signals in the first embodiment of the present invention;

[0037] Figure 5 It is the algorithm flow chart of the chaotic eagle optimizer in the first embodiment of the present invention. Specific implementation manner

[0038] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0039] It should be noted that the following detailed description is illustrative and is intended to provide further description of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0040] Embodiment 1

[0041] This embodiment provides a power quality disturbance identification method based on a chaotic eagle optimizer, as Figure 1 shown, which specifically includes the following steps:

[0042] Step 1, obtain the data of power quality disturbances that have been classified.

[0043] Specifically, set the sampling interval to 0.0001 s, the signal step to 1000, the amplitude to 1 V, and the high - order harmonic index for power quality signal interference detection to 250. Use a mathematical model to obtain five synthesized power quality disturbance data, that is, the power quality disturbance data is divided into five categories, namely: normal data, voltage sag, harmonics, voltage flicker, and voltage swell data. The power quality disturbance waveforms are shown in Figure 4 , and the mathematical model formulas for the five - category power quality disturbance data are as follows:

[0044] Normal data:

[0045] V(t)=Asin(ωt) (1)

[0046] A = 1(pu) (2)

[0047] ω = 2*π*50rad / s (3)

[0048] Voltage sag:

[0049] V(t)=(1 - α(u(t - t 1 )) - u(t - t 2 )))sinωt (4)

[0050] 0.1≤α≤0.9 (5)

[0051] T ≤ t 2 -t 1 ≤ 9T (6)

[0052] Voltage harmonics:

[0053] V(t) = α 3 sin(3ωt) + α 5 sin(5ωt) + α 7 sin(7ωt) (7)

[0054] 0.05 ≤ α 3 , α 5 , α 7 ≤ 0.15 (8)

[0055]

[0056] Voltage fluctuation:

[0057] V(t) = (1 + α f sin(βωt))sinωt (10)

[0058] 0.1 ≤ α f ≤ 0.2 (11)

[0059] 5 < β < 20Hz (12)

[0060] Voltage swell:

[0061] V(t) = (1 + α(u(t - t 1 ) - u(t - t 2 )))sinωt (13)

[0062] 0.1 ≤ α ≤ 0.8 (14)

[0063] T ≤ t 2 -t 1 ≤ 9T (15)

[0064] Wherein, A is the amplitude of the standard signal; u(t) is the step function; ω is the angular velocity; α is the amplitude distortion degree; t 1 and t 2 are the start and end times; β is the oscillation index; t is the current time; α 3 , α 5 , α 7 are respectively the amplitude distortion degrees of the third harmonic, fifth harmonic and seventh harmonic; α f is the amplitude distortion degree during voltage fluctuation, and pu represents the per-unit value.

[0065] Step 2: Extract the feature vectors of the training samples by using the S transform.

[0066] Specifically, nine feature vectors of the training samples are extracted by using the S-transform, which are respectively: half of the average value at 50 Hz frequency, the maximum peak value at 50 Hz frequency, the minimum peak value at 50 Hz frequency, the average value at 150 Hz frequency, the average value at 250 Hz frequency, the average value at 350 Hz frequency, the average value at 450 Hz frequency, the average value at 550 Hz frequency, and the average value at 700 - 2500 Hz frequency.

[0067] The definition of the S-transform is specifically as follows:

[0068]

[0069]

[0070]

[0071] In the formula, y(t) is the signal; f is the frequency; τ is the time shift factor; ω G is the Gaussian window function; t is the current time; σ is the scale factor, which can adjust the height and width of the Gaussian window and is inversely proportional to the frequency f.

[0072] The S-transform is an extension of the short-time Fourier transform and the wavelet transform, and it has excellent time-frequency localization ability and feature extraction ability. The features extracted by the S-transform are intuitive and clear, and it is widely used especially in the extraction of power quality disturbance features.

[0073] Step 3: Use the extracted feature vectors as the input vectors of the improved BP neural network.

[0074] Specifically, the feature vectors of each data, that is, half of the average value at 50 Hz frequency, the maximum peak value at 50 Hz frequency, the minimum peak value at 50 Hz frequency, the average value at 150 Hz frequency, the average value at 250 Hz frequency, the average value at 350 Hz frequency, the average value at 450 Hz frequency, the average value at 550 Hz frequency, and the average value at 700 - 2500 Hz frequency, are used as the input vectors of the improved BP neural network.

[0075] Step 4: Divide the input vectors into a training set and a test set, and use the training set to train the improved BP neural network.

[0076] Specifically, the nine feature vectors are divided into a training set and a test set according to the ratio of 0.7:0.3, and the training set is used to train the improved BP neural network. The improvement scheme of the BP neural network is as follows:

[0077] The improved BP neural network determines the optimal number of neurons in the middle layer by using the chaotic Tianying optimizer on the basis of the original single-hidden-layer BP neural network. The specific steps are as follows:

[0078] (1) The range of the number of neurons in the middle layer of the improved BP neural network is set to an integer in [5, 100];

[0079] (2) Initialize the population of the chaotic Tianying optimizer, where an individual represents the number of neurons in the middle layer of a single-hidden-layer BP neural network;

[0080] (3) Train the improved BP neural network with this number of neurons;

[0081] (4) Use the trained improved BP neural network in step (3) to perform classification multiple times (it can be three times), and calculate the average value of the classification result accuracy;

[0082] (5) Denote the average value of the accuracy as the fitness value of the individual;

[0083] (6) Use the chaotic Tianying optimizer for iterative optimization;

[0084] (7) Determine the optimal candidate solution as the number of neurons in the middle layer of the improved BP neural network, that is, determine the individual with the highest fitness value as the number of neurons in the middle layer of the BP neural network.

[0085] Specifically, the process of the chaotic Tianying optimizer is as follows:

[0086] ① Initialize the algorithm parameters, including the population size, the maximum number of iterations T, the current number of iterations t, the exploitation coefficients α and β.

[0087] ② Initialize the population position X, calculate the initial population fitness and find the best individual.

[0088] ③ When the number of iterations is in the first interval, that is, when t < 0.7T, the Tianying optimizer randomly selects one of the two search schemes of "expanding exploration" and "shrinking exploration" to execute:

[0089] "Expanding exploration" search scheme: The Tianying identifies the prey area and selects the hunting area by flying high with a vertical bend. In this scheme, the Tianying determines the area of the search space at high altitude, that is, the position of the search prey, and the update formula is:

[0090]

[0091]

[0092] where: X 1 (t + 1) is the (t + 1)-th iteration solution generated by the "expanding exploration" search scheme; X best (t) is the optimal solution of the t-th iteration, which reflects the approximate position of the prey; (1 - t / T) represents controlling the exploration by the number of iterations; XM (t) represents the average value of the current solution at the t-th iteration; X i (t) represents the fitness value of the i-th candidate solution in the t-th generation; rand is a random number between [0, 1]; t and T represent the current iteration number and the maximum iteration number respectively; N is the number of candidate solutions.

[0093] For the "shrinking exploration" search scheme, when the eagle finds the prey area, it will fly in a spiral above the target prey and get ready to attack. This method is called contour flight of short glide attack. At this time, the eagle explores the selected area of the prey narrowly, and the update formula is:

[0094] X 2 (t + 1) = X best (t) × levy(D) + X R (t) + (y - x) × rand(21)

[0095] Where: X 2 (t + 1) is the solution of the (t + 1)-th iteration generated by the "shrinking exploration" search scheme; X R (t) is a candidate solution randomly selected within the value range of the population; D represents the dimensional space, and levy(D) is the Levy flight distribution function, and its calculation formula is:

[0096]

[0097]

[0098] Where: s is a constant value fixed at 0.01; u and v are random numbers between 0 and 1; Γ represents the gamma function; z is a constant fixed at 1.5.

[0099] Among them, y and x show a spiral form in the search, and their calculation formulas are as follows:

[0100] y = r × cos(θ) (24)

[0101] x = r × sin(θ) (25)

[0102] r = r 1 + U × D 1 (26)

[0103] θ = -ω × D 1 + θ 1 (27)

[0104] θ 1 = (3 × π) / 2 (28)

[0105] Where: r 1is a fixed - period exponent between 1 and 20; and U is a constant fixed at 0.00565; D 1 is an integer from 1 to the length of the search space; ω is a constant fixed at 0.005.

[0106] ④When the number of iterations is in the second interval, that is, when 0.7T < t < T, the Tianying optimizer starts to randomly select one of the two search schemes of "expanding exploitation" and "shrinking exploitation" to execute:

[0107] For the "expanding exploitation" search scheme, the Tianying locks the hunting area, gets ready to land and attack, then descends vertically and makes a preliminary attack to test the prey's reaction. The update formula is:

[0108] X 3 (t + 1)=(X best (t)-X M (t))×α - rand+((UB - LB)×rand+LB)×β (29)

[0109] where X 3 (t + 1) is the (t + 1)-th iteration solution generated by the "expanding exploitation" search scheme; the exploitation coefficients α and β, whose values are small, are in the range of (0, 1); LB and UB represent the lower and upper bounds of the given problem respectively.

[0110] For the "shrinking exploitation" search scheme, when the Tianying approaches the prey, it attacks the prey according to the random movement of the prey. The update formula is:

[0111] X 4 (t + 1)=QF×X best (t)-(G 1 ×X(t)×rand)-G 2 ×levy(D) (30)

[0112]

[0113] G 1 =2×rand - 1 (32)

[0114]

[0115] where X 4 (t + 1) is the (t + 1)-th iteration solution generated by the "shrinking exploitation" search scheme; X(t) represents the position of an individual at the t - th iteration; QF(t) represents the quality function used to balance the search strategy; G 1 represents different methods adopted by the Tianying during the prey's escape process; G 2is a decreasing value from 2 to 0, representing the flight slope of the eagle from the first position to the last position when tracking prey.

[0116] ⑤ Calculate and update the fitness of the population to obtain the current best individual position and fitness.

[0117] ⑥ Compare the fitness of the current best individual with the fitness of the best individual found up to the t-th generation, and retain the better individual position.

[0118] ⑦ When the iteration reaches the maximum number of iterations T, output the optimal solution and the best fitness.

[0119] Specifically, the chaotic eagle optimizer in this embodiment replaces the exploitation coefficients α and β in the original eagle optimizer with the tent map and the Sine map respectively.

[0120] The tent map expression is:

[0121]

[0122] where α(t + 1) is the (t + 1)-th term in the Sine map sequence, α(t) is the t-th term in the tent map sequence, the initial value α(1) in the tent map sequence is 0.6, and t is the current iteration number of the chaotic eagle optimizer.

[0123] The Sine map expression is:

[0124] β(t + 1) = sin(β(t)) (35)

[0125] where β(t + 1) is the (t + 1)-th term in the Sine map sequence, β(t) is the t-th term in the Sine map sequence, the initial value β(1) in the Sine map sequence is 0.7, and t is the current iteration number of the chaotic eagle optimizer.

[0126] Step 5: Use the test set to test the improved BP neural network trained in Step 4. If the convergence condition is met, use this improved BP neural network as the classifier for power quality disturbances; if the convergence condition is not met, return to Step 4 to retrain the improved BP neural network until the convergence condition is met.

[0127] Specifically, on the basis of the original minimum root mean square error, the convergence condition adds the requirement that the average cosine similarity is not less than the threshold of 0.98. The expression for the average cosine similarity is:

[0128]

[0129] In the formula, M is the number of results classified by the improved BP neural network; cos icos(θ) represents the cosine similarity between the i-th predicted classification result vector and the i-th true classification result vector. i The expression of cos(θ) is:

[0130]

[0131] In the formula, x (i,j) is the value of the j-th dimension in the i-th predicted classification result vector; y (i,j) is the value of the j-th dimension in the i-th true classification result vector; Dim represents the total dimension of the result vector.

[0132] Step 6: Extract the feature vectors of the unclassified power quality disturbance data by using the S transform.

[0133] Specifically, use the S transform to extract the unclassified power quality disturbance data samples, with 50 samples for each of the normal data, voltage sag, harmonics, voltage flicker, and voltage swell. A total of 250 data are respectively extracted with 9 kinds of feature vectors, which are: half of the average value at 50 Hz, the maximum peak value at 50 Hz, the minimum peak value at 50 Hz, the average value at 150 Hz, the average value at 250 Hz, the average value at 350 Hz, the average value at 450 Hz, the average value at 550 Hz, and the average value at 700 - 2500 Hz.

[0134] Step 7: Use the feature vectors processed by the S transform in Step 6 as the input vectors, and use the improved BP neural network trained in Step 5 to output the classification results.

[0135] Specifically, divide the 9 kinds of feature vectors into a training set and a test set according to the ratio of 0.7:0.3, and use the trained improved BP neural network to output the classification results.

[0136] The classification results are as Figure 3 shown. The optimal results of 20 classifications are shown in Table 1, and the worst results of 20 classifications are shown in Table 2.

[0137] Table 1. Optimal results of 20 classifications

[0138] Normal signal Voltage sag Harmonic Voltage flicker Voltage swell Normal signal 50 Voltage sag 50 Harmonic 50 Voltage flicker 50 Sag + Harmonic 50

[0139] In Table 1, the diagonal elements indicate that 50 normal signals are recognized as normal signals, and the same applies to voltage sag, harmonics, voltage flicker, and voltage swell. In the optimal classification, the classification accuracy of normal signals is 100%; the classification accuracy of voltage sag signals is 100%; the classification accuracy of harmonics is 100%; the classification accuracy of voltage flicker is 100%; the classification accuracy of voltage swell is 100%; and the final overall classification result accuracy is 100%.

[0140] Table 2, Worst Results of 20 Classifications

[0141] Normal signal Voltage sag Harmonic Voltage flicker Voltage swell Normal signal 50 Voltage sag 1 49 Harmonic 50 Voltage flicker 50 Sag + Harmonic 50

[0142] In Table 2, the data in the third row indicates that 1 out of 50 voltage sag signals was misclassified as a normal signal. In the worst classification, the classification accuracy of normal signals was 100%; the classification accuracy of voltage sag signals was 100%; the classification accuracy of harmonics was 100%; the classification accuracy of voltage flicker was 100%; the classification accuracy of voltage swell was 100%; and the overall classification accuracy was 99.60%.

[0143] The accuracy statistics of 20 classification results are shown in Table 3. The average accuracy of 20 classifications is as Figure 2 shown.

[0144] Table 3, Accuracy of 20 Classification Results

[0145] Number of times Accuracy rate Number of times Accuracy rate Number of times Accuracy rate Number of times Accuracy rate 1 100％ 6 100％ 11 100％ 16 100％ 2 100％ 7 99.60％ 12 100％ 17 100％ 3 100％ 8 100％ 13 100％ 18 100％ 4 99.60％ 9 100％ 14 100％ 19 100％ 5 99.60％ 10 100％ 15 100％ 20 100％

[0146] From Table 3, it can be obtained that the average classification accuracy is 99.94%. In 20 classifications, there were only three classification errors, and only one data had a classification error each time, which further shows that the present invention can stably solve the problem of power quality disturbance identification.

[0147] The results of Table 1, Table 2, and Table 3 fully illustrate that the present invention effectively solves the problem of power quality disturbance identification.

[0148] The present invention proposes a power quality disturbance identification method based on a chaotic Tianying optimizer. In order to effectively identify power quality problems, power quality is classified into the following five categories according to indicators: normal data, voltage sag, harmonics, voltage flicker, and voltage swell. Adding cosine similarity as the convergence condition for improving the BP neural network can increase the classification accuracy of the BP neural network; using the chaotic Tianying optimizer to optimize the number of neurons in the middle layer of the BP neural network can improve the classification accuracy of the BP neural network.

[0149] Embodiment 2

[0150] This embodiment provides a power quality disturbance identification system based on a chaotic Tianying optimizer, which specifically includes the following modules:

[0151] A data acquisition module, which is configured to: acquire power quality disturbance data;

[0152] An identification module, which is configured to: after extracting feature vectors from the power quality disturbance data, obtain classification results using a neural network;

[0153] Among them, during the training process of the neural network, the chaotic Tianying optimizer is used to optimize the number of neurons in the middle layer, and the exploitation coefficient of the chaotic Tianying optimizer is obtained through the tent map and the Sine map.

[0154] It should be noted here that each module in this embodiment corresponds one by one to each step in Embodiment 1, and the specific implementation process is the same, so it will not be repeated here.

[0155] Embodiment 3

[0156] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the steps in the power quality disturbance recognition method based on the chaotic Tianying optimizer as described in Embodiment 1 above.

[0157] Embodiment 4

[0158] This embodiment provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the steps in the power quality disturbance recognition method based on the chaotic Tianying optimizer as described in Embodiment 1 above.

[0159] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories and optical memories, etc.) containing computer-usable program codes.

[0160] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0161] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device that implements the functions specified in one or more processes and / or blocks Figure 1 in one or more processes and / or blocks Figure 1 specified in the function.

[0162] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are performed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes and / or blocks Figure 1 in one or more processes and / or blocks Figure 1 specified in the function.

[0163] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0164] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for identifying power quality disturbances based on a chaotic eagle optimizer, characterized in that, it includes: Obtain power quality disturbance data; After extracting the feature vectors from the power quality disturbance data, use a neural network to obtain the classification result; Among them, during the training process of the neural network, the chaotic eagle optimizer is used to optimize the number of neurons in the middle layer, and the exploitation coefficient of the chaotic eagle optimizer is obtained through tent mapping and Sine mapping; The specific steps for using the chaotic eagle optimizer to optimize the number of neurons in the middle layer are: Initialize the population of the chaotic eagle optimizer, and the individual represents the number of neurons; Under the number of neurons, train the neural network; Use the trained neural network to perform classification multiple times, calculate the average value of the classification result accuracy, and record the average value of the accuracy as the fitness value of the individual; Use the chaotic eagle optimizer for iterative optimization, and determine the number of neurons in the middle layer of the neural network as the individual with the highest fitness value.

2. The method for identifying power quality disturbances based on a chaotic eagle optimizer according to claim 1, characterized in that, The specific steps for using the chaotic eagle optimizer for iterative optimization are: When the number of iterations is in the first interval, the eagle optimizer randomly selects one of the expanding exploration scheme and the shrinking exploration scheme to execute; When the number of iterations is in the second interval, the eagle optimizer randomly selects one of the expanding exploitation scheme and the shrinking exploitation scheme to execute.

3. The method for identifying power quality disturbances based on a chaotic eagle optimizer according to claim 1, characterized in that, The convergence condition during the training process of the neural network includes: the average cosine similarity is not less than the threshold.

4. The method for identifying power quality disturbances based on a chaotic eagle optimizer according to claim 1, characterized in that, Use the S transform to extract feature vectors from the power quality disturbance data.

5. The method for identifying power quality disturbances based on a chaotic eagle optimizer according to claim 1, characterized in that, The feature vectors include: half of the average value at 50Hz, the maximum peak value at 50Hz, the minimum peak value at 50Hz, the average value at 150Hz, the average value at 250Hz, the average value at 350Hz, the average value at 450Hz, the average value at 550Hz, and the average value at 700 - 2500Hz.

6. The method for identifying power quality disturbances based on a chaotic eagle optimizer according to claim 1, characterized in that, The data of the power quality disturbances are divided into five categories, namely: normal data, voltage sag, harmonics, voltage flicker, and voltage swell data.

7. A power quality disturbance identification system based on a chaotic eagle optimizer, characterized in that, it includes: A data acquisition module, which is configured to: obtain power quality disturbance data; An identification module, which is configured to: after extracting the feature vectors from the power quality disturbance data, use a neural network to obtain the classification result; Among them, during the training process of the neural network, the chaotic eagle optimizer is used to optimize the number of neurons in the middle layer, and the exploitation coefficient of the chaotic eagle optimizer is obtained through tent mapping and Sine mapping; The specific steps for optimizing the number of intermediate layer neurons using the chaotic eagle optimizer are as follows: Initialize the population of the chaotic eagle optimizer, where each individual represents the number of neurons; Under the number of neurons, train the neural network; Use the trained neural network to perform classification multiple times, calculate the average value of the classification result accuracy, and record the average value of the accuracy as the fitness value of the individual; Use the chaotic eagle optimizer for iterative optimization, and determine the individual with the highest fitness value as the number of intermediate layer neurons of the neural network.

8. A computer-readable storage medium, on which a computer program is stored, characterized in that, when the program is executed by a processor, it implements the steps in the power quality disturbance recognition method based on the chaotic eagle optimizer described in any one of claims 1-6.

9. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, when the processor executes the program, it implements the steps in the power quality disturbance recognition method based on the chaotic eagle optimizer described in any one of claims 1-6.

Images