GOA-VMD-LSSVM-based DC electric energy quality disturbance identification method

Abstract

The invention discloses a GOA-VMD-LSSVM-based direct-current electric energy quality disturbance identification method, belongs to the field of direct-current electric energy quality analysis and treatment, and aims to solve the problem of direct-current electric energy disturbance identification in a direct-current distribution network. The method comprises the following steps: step 1, collecting power data under different DC electric energy quality disturbances; step 2, completing direct current power signal decomposition based on a GOA-VMD algorithm; 3, performing feature extraction on the decomposed signal by using permutation entropy; and step 4, optimizing LSSVM parameters by using GOA, and completing direct current electric energy quality disturbance identification. The method is suitable for the field of direct-current electric energy quality analysis and treatment. According to the method, power data under different DC electric energy quality disturbances are collected, DC power signal decomposition is completed based on a GOA-VMD algorithm, feature extraction is performed on the decomposed signals by using permutation entropy, LSSVM parameters are optimized by using GOA, DC electric energy quality disturbance identification is completed, and the method has the characteristics of high precision and high speed and has relatively high robustness.

Description

Technical Field

[0001] The present invention relates to the field of direct current power quality analysis and management, and in particular to a direct current power quality disturbance identification method based on GOA-VMD-LSSVM. Background Art

[0002] In the process of development, the new power system has shown the characteristics of "double high", namely, a high proportion of new energy access and a high proportion of power electronic equipment. At the same time, the application of DC technology is becoming more and more extensive in new energy scenarios such as distributed photovoltaics and charging piles. With the continuous development of power electronics technology and new energy technology, DC power grid has attracted the attention of many researchers with its economic reliability and unique technical advantages. However, the randomness of new energy generation such as wind power and solar power often causes voltage fluctuations in the DC bus, thereby affecting the power quality of the DC power grid, resulting in different types of DC power quality disturbances that may exist in the DC distribution network. Classifying and identifying different power quality disturbances that may exist in the DC power grid is of great significance for proposing corresponding governance strategies and ensuring the power supply stability of the DC distribution network.

[0003] There are many methods for identifying power quality disturbances in AC power grids, such as support vector machines, Fourier transform, wavelet transform, Hilbert transform, Prony algorithm, etc. However, there are few studies on disturbance identification in DC power grids, and a systematic theoretical system has not yet been formed. Although the DC distribution network does not need to consider reactive power and phase issues, which reduces the computational complexity of power quality disturbance feature extraction and identification, the nonlinear problem of DC distribution network is more prominent, which brings new challenges to the integrity and accuracy of DC disturbance signal feature extraction.

[0004] In order to solve the above problems, the present invention proposes a DC power quality disturbance identification method based on GOA-VMD-LSSVM.

[0005] Therefore, it is urgent to design a DC disturbance feature extraction and identification method with high accuracy, fast speed and high robustness. Summary of the invention

[0006] The purpose of the present invention is to overcome the problems of low accuracy, slow speed and low robustness of DC disturbance feature extraction and identification in the prior art, to meet the actual needs, and to provide a DC power quality disturbance identification method based on GOA-VMD-LSSVM to solve the above technical problems.

[0007] The technical solution adopted by the present invention is as follows:

[0008] A method for identifying direct current power quality disturbance based on GOA-VMD-LSSVM comprises the following steps:

[0009] Step 1: Collect power data under different DC power quality disturbances;

[0010] Furthermore, a simulation model was built using MATLAB / Simulink to collect voltage data for six types of DC power quality disturbances, including voltage fluctuation, voltage swell, voltage ripple, voltage sag caused by large-scale load input, and voltage sag caused by single-pole grounding and inter-pole short circuit in the DC distribution network.

[0011] Step 2: Decompose the DC power signal based on the GOA-VMD algorithm:

[0012] Further, in step 2.1, a DC voltage signal is input, the value range of the variational mode decomposition parameters to be optimized (the optimal decomposition level K and the penalty factor δ) is set, and parameters such as the maximum number of iterations F of the Grasshopper Optimization Algorithm (GOA) are initialized;

[0013] Step 2.2: Decompose the DC signal using variational mode decomposition (VMD), and select envelope entropy as the evaluation function to assist in determining the optimal value of the VWD parameter. The envelope entropy is shown in the following formula:

[0014]

[0015] Step 2.3, determine whether the maximum number of iterations is met, if so, terminate the iteration, if not, continue the iteration;

[0016] Step 2.4: Obtain and save the optimal parameters of VMD, decompose the DC power signal, and extract each modal component.

[0017] Step 3: Use permutation entropy to extract features from the decomposed signal:

[0018] Furthermore, assuming that a certain signal component after decomposition is Y = {u(1), u(2), L, u(N)}, its length is N, and the permutation entropy is selected to extract the features of the decomposed DC voltage signal. This function uses the phase space reconstruction delayed coordinate method to map the time series to the m-dimensional space reconstruction vector. Each subsequence is represented by X(i):

[0019] X(i)={u(i),u(i+L),L,u(i+(m-1)L)}

[0020] Then X(i) is arranged incrementally to obtain the arrangement index π. For an m-dimensional sequence, there are m possible arrangement indexes. The probability p of the jth arrangement is defined as j for:

[0021]

[0022] in represents the frequency of occurrence of arrangement j, then the arrangement entropy can be calculated by the following formula:

[0023]

[0024] In the formula, k represents the number of possible permutation categories.

[0025] Step 4: Use GOA to optimize LSSVM parameters and complete DC power quality disturbance identification:

[0026] Furthermore, in step 4.1, the inequality constraints in the support vector machine (SVM) are replaced by the linear least squares criterion to obtain the least squares support vector machine (LSSVM), and its hyperparameters are optimized:

[0027] Assume that a sample is Z = {(x1, y1), (x2, y2), L, (x l ,y l )}, xl and yl are the input vector and the corresponding output category respectively, and the mathematical expression is:

[0028]

[0029] In the formula, w is the weight, b is the bias term, The main purpose of the nonlinear function is to map the input space to the feature space, so LSSVM can be described as an optimization problem:

[0030]

[0031] In the formula, γ is the regularization parameter, e i is the regression error. The LSSVM model can be further expressed as:

[0032]

[0033] Where α is the Lagrange multiplier, K(x,x i ) is the kernel function. This paper adopts radial basis function (RBF) as the kernel function, and its expression is:

[0034]

[0035] In the formula, σ is the bandwidth of the kernel function. Since the hyperparameters γ and σ have a great influence on the performance of the LSSVM model, refer to step 2 to optimize the parameters of the LSSVM model;

[0036] Step 4.2: Complete DC power quality disturbance identification based on GOA-VMD-LSSVM:

[0037] Furthermore, the DC signal is first decomposed using VMD optimized by GOA algorithm, and then the signal features are extracted using permutation entropy;

[0038] Then the LSSVM model is trained, and the GOA algorithm is used to optimize the parameters of LSSVM. Finally, the test data and LSSVM are used to realize the identification of DC power quality disturbance.

[0039] The present invention has beneficial effects: the present invention provides a method for identifying DC power quality disturbances based on GOA-VMD-LSSVM, which collects power data under different DC power quality disturbances, decomposes DC power signals based on the GOA-VMD algorithm, extracts features of the decomposed signals using permutation entropy, and optimizes LSSVM parameters using GOA to identify DC power quality disturbances. The method has the characteristics of high precision and high speed, and has high robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0041] Figure 1 This is the flow chart of the DC power quality disturbance identification algorithm based on GOA-VMD-LSSVM. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the embodiments described are only part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0043] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0044] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0045] A DC signal ripple detection scheme based on GOA-VMD-LSSVM includes the following steps:

[0046] Step 1: Collect power data under different DC power quality disturbances:

[0047] A simulation model was built using Matlab / Simulink, and voltage data were collected for six types of DC power quality disturbances: voltage fluctuation, voltage swell, voltage ripple, voltage sag caused by large-scale load input, and voltage sag caused by single-pole grounding and inter-pole short circuit in the DC distribution network.

[0048] Step 2: Decompose the DC power signal based on the GOA-VMD algorithm:

[0049] Step 2.1, input DC voltage signal. Set the value range of variational mode decomposition parameters to be optimized (optimal decomposition level K and penalty factor δ), initialize the maximum number of iterations F of the grasshopper optimization algorithm (GOA) and other parameters;

[0050] Step 2.2: Decompose the DC signal using variational mode decomposition (VMD), and select envelope entropy as the evaluation function to assist in determining the optimal value of the VWD parameter. The envelope entropy is shown in the following formula:

[0051]

[0052] Step 2.3, determine whether the maximum number of iterations is met, if so, terminate the iteration, if not, continue the iteration;

[0053] Step 2.4: Obtain and save the optimal parameters of VMD, decompose the DC power signal, and extract each modal component.

[0054] Step 3: Use permutation entropy to extract features from the decomposed signal:

[0055] Assume that a certain signal component after decomposition is Y = {u(1), u(2), L, u(N)}, and its length is N. Select the permutation entropy to extract the features of the decomposed DC voltage signal. This function uses the phase space reconstruction delayed coordinate method to map the time series to the m-dimensional space reconstruction vector. Each subsequence is represented by X(i):

[0056] X(i)={u(i),u(i+L),L,u(i+(m-1)L)}

[0057] Then X(i) is arranged incrementally to obtain the arrangement index π. For an m-dimensional sequence, there are m possible arrangement indexes. The probability p of the jth arrangement is defined as j for:

[0058]

[0059] in represents the frequency of occurrence of arrangement j, then the arrangement entropy can be calculated by the following formula:

[0060]

[0061] In the formula, k represents the number of possible permutation categories.

[0062] Step 4: Use GOA to optimize LSSVM parameters and complete DC power quality disturbance identification;

[0063] Step 4.1: Use the linear least squares criterion to replace the inequality constraints in the support vector machine (SVM) to obtain the least squares support vector machine (LSSVM), and optimize its hyperparameters.

[0064] Assume that a sample is Z = {(x1, y1), (x2, y2), L, (x l ,y l )}, xl and yl are the input vector and the corresponding output category respectively, and the mathematical expression is:

[0065]

[0066] In the formula, w is the weight, b is the bias term, The main purpose of the nonlinear function is to map the input space to the feature space, so LSSVM can be described as an optimization problem:

[0067]

[0068] In the formula, γ is the regularization parameter, e i is the regression error. The LSSVM model can be further expressed as:

[0069]

[0070] Where α is the Lagrange multiplier, K(x,x i ) is the kernel function. This paper adopts radial basis function (RBF) as the kernel function, and its expression is:

[0071]

[0072] In the formula, σ is the bandwidth of the kernel function. Since the hyperparameters γ and σ have a great influence on the performance of the LSSVM model, refer to step 2 to optimize the parameters of the LSSVM model.

[0073] Step 4.2: Complete DC power quality disturbance identification based on GOA-VMD-LSSVM:

[0074] Firstly, the DC signal is decomposed by VMD optimized by GOA algorithm, and then the signal features are extracted by permutation entropy.

[0075] Then the LSSVM model is trained, and the GOA algorithm is used to optimize the parameters of LSSVM. Finally, the test data and LSSVM are used to realize the identification of DC power quality disturbance.

[0076] 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 and improvements 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 identifying DC power quality disturbances based on GOA-VMD-LSSVM, characterized in that: The steps include: Step 1: Collect power data under different DC power quality disturbances; Step 2: Decomposing the DC power signal based on the GOA-VMD algorithm; Step 3: Use permutation entropy to extract features from the decomposed signal; Step 4: Use GOA to optimize LSSVM parameters and complete DC power quality disturbance identification.

2. A method for identifying DC power quality disturbances based on GOA-VMD-LSSVM according to claim 1, characterized in that: Step 1 specifically includes: A simulation model was built using Matlab / Simulink to collect voltage data for six types of DC power quality disturbances, including voltage fluctuation, voltage swell, voltage ripple, voltage sag caused by large-scale load input, single-pole grounding of the DC distribution network, and voltage sag caused by inter-pole short circuit.

3. The method for identifying DC power quality disturbance based on GOA-VMD-LSSVM according to claim 1 is characterized in that: Step 2 specifically includes: Step 2.1: Input the DC voltage signal, set the value range of the variational mode decomposition parameters to be optimized (optimal decomposition level K and penalty factor δ), and initialize the grasshopper optimization algorithm (GOA) to optimize the parameters, including: (1) Modeling locust swarm behavior: X i =S i +G i +A i Where, X i is the position of the i-th locust, S i and G i are the individual interaction force and gravity of the i-th locust respectively; Among them, S i and G i It can be calculated by the following formula: Where, d ij =|x i -x j | is the distance between the i-th locust and the j-th locust, s(g) is the social status of the locust, is the unit vector from the i-th to the j-th locust, f and l are the attraction length and attraction length scale respectively. In the present invention, f and l are taken as 1.5 and 0.5 respectively, g is the gravitational constant, is a unit vector pointing to the center of the earth; u represents a constant drift, is the unit vector of wind direction, so we get X i for: Where N is the number of locusts, ignoring gravity and wind direction: ub d lb d They are s(r)=fe -r / l -e -r The upper and lower bounds in d dimension, Td is the optimal solution currently obtained, that is, the corresponding optimal number of decomposition layers and the optimal penalty factor, c represents the decreasing coefficient, which is obtained by the following formula: l and L represent the current iteration number and the total iteration number respectively, c_max and c_min represent the maximum and minimum values ​​of parameter c respectively; Step 2.2: Decompose the DC signal using variational mode decomposition (VMD), and select envelope entropy as the evaluation function to assist in determining the optimal value of the VWD parameter. The envelope entropy is shown in the following formula: Where, e(i) is the signal obtained after envelope analysis of the original DC signal, and γ(i) is the signal after normalization of e(i); Step 2.3, determine whether the maximum number of iterations is met, if so, terminate the iteration, if not, continue the iteration; Step 2.4: Obtain and save the optimal VMD parameters.

4. The method for identifying DC power quality disturbance based on GOA-VMD-LSSVM according to claim 1, characterized in that: Step 3 specifically includes: Assume that a certain signal component after decomposition is Y = {u(1), u(2), L, u(N)}, whose length is N, and select permutation entropy to extract features of the decomposed DC voltage signal. This function uses the phase space reconstruction delayed coordinate method to map the time series to the m-dimensional space reconstruction vector. Each subsequence is represented by X(i): X(i)={u(i),u(i+L),L,u(i+(m-1)L)} Then X(i) is arranged incrementally to obtain the arrangement index π. For an m-dimensional sequence, there are m! possible arrangement indexes. The probability p of the jth arrangement is defined as j for: in represents the frequency of occurrence of arrangement j, then the arrangement entropy can be calculated by the following formula: In the formula, k represents the number of possible permutation categories.

5. The method for identifying DC power quality disturbance based on GOA-VMD-LSSVM according to claim 1, characterized in that: Step 4 specifically includes: Step 4.1: Use the linear least squares criterion to replace the inequality constraints in the support vector machine (SVM) to obtain the least squares support vector machine (LSSVM), and optimize its hyperparameters: Assume that a sample is Z = {(x1, y1), (x2, y2), L, (x l ,y l )},x l ,y l They are the input vector and the corresponding output category, and the mathematical expression is: In the formula, w is the weight, b is the bias term, The main purpose of the nonlinear function is to map the input space to the feature space, so LSSVM can be described as an optimization problem: In the formula, γ is the regularization parameter, e i is the regression error, and the LSSVM model can be further expressed as: Where α is the Lagrange multiplier, K(x,x i ) is the kernel function. This paper adopts radial basis function (RBF) as the kernel function, and its expression is: In the formula, σ is the bandwidth of the kernel function. Since the hyperparameters γ and σ have a great influence on the performance of the LSSVM model, the parameters of the LSSVM model are optimized by referring to step 2. Step 4.2: Complete DC power quality disturbance identification based on GOA-VMD-LSSVM: Firstly, the DC signal is decomposed by VMD optimized by GOA algorithm, and then the signal features are extracted by permutation entropy. Then the LSSVM model is trained, and the GOA algorithm is used to optimize the parameters of LSSVM. Finally, the test data and LSSVM are used to realize the identification of DC power quality disturbance.

Images