Method, device and storage medium for determining parameters of a three-phase three-level circuit

Abstract

The application relates to a three-phase three-level circuit parameter determination method, device, equipment and storage medium. The method comprises the following steps: acquiring and processing three-phase three-level circuit data, establishing a two-stage circuit optimization model, including a first-stage prediction model, which is used for predicting output alternating-current side filter inductance values and direct-current side energy storage capacitance values, and outputting confidence values corresponding to the inductance and the capacitance to control the learning ability of a second-stage optimization model, and the second stage outputs a switch time angle sequence which minimizes the harmonics of the three-phase three-level circuit. The method can simultaneously solve the midpoint imbalance problem and the harmonic suppression problem of the three-phase three-level circuit, and has good effects.

Description

Technical Field

[0001] This invention relates to the field of circuit optimization technology, and in particular to a method, apparatus, device, and storage medium for determining parameters of a three-phase three-level circuit. Background Technology

[0002] The energy characteristics of AC / DC microgrids require equipment to be regulated by output power to connect distributed energy sources to the low-voltage grid. Distributed energy sources connected to the DC bus require AC / DC or DC / DC conversion, while those connected to the AC bus require AC / AC or DC / AC conversion. In large power grids (conventional power systems), due to the constantly changing power consumption and generation conditions, fully controllable voltage / frequency permanent magnet synchronous generators cannot be controlled in microgrids. Therefore, how to implement AC / DC and DC / AC control is a key issue in microgrid research.

[0003] The most significant characteristic of a voltage-source AC / DC bidirectional power converter (VSC) topology is that the DC-side circuit is filtered by capacitors, resulting in the low-impedance characteristic of the VSC voltage source. For a three-phase symmetrical AC system, based on the mathematical model of a three-phase VSC, attention should also be paid to the design of the AC-side filter inductor parameters and the DC-side energy storage capacitor parameters to ensure the effective functioning of these components. However, this approach faces challenges such as voltage midpoint imbalance and the difficulty in eliminating harmonics.

[0004] Currently, the problems of midpoint potential imbalance and harmonic suppression are often solved separately. For example, the midpoint potential balance problem often uses the space vector control method, while the harmonic suppression problem often uses the solution of nonlinear equations of the switching time angle sequence. Such methods cannot solve the circuit's midpoint balance problem and harmonic suppression problem at the same time. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, device, and storage medium for determining the parameters of a three-phase three-level circuit that can simultaneously address the midpoint potential balance problem and harmonic suppression problem, in order to solve the above-mentioned technical problems.

[0006] In a first aspect, embodiments of this application provide a method for determining parameters of a three-phase three-level circuit, characterized in that the method includes:

[0007] Acquire data samples of a three-phase three-level circuit, including: three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, average rectified voltage Ud0, switching time angle sequence, load line current I11, effective value of line voltage V11, AC side filter inductor value, and DC side energy storage capacitor value;

[0008] The data samples are preprocessed, which includes dividing the data samples into input data samples and output data samples, performing correlation analysis on the input data samples and output data samples and adding weights to obtain the final training input data samples.

[0009] The preprocessed data samples are input into the preset three-phase three-level circuit parameter determination model for training to obtain the optimal switching angle sequence and the corresponding AC side filter inductance value and DC side energy storage capacitor value.

[0010] The parameter determination model of the three-phase three-level circuit includes: a first-stage prediction model of the AC side filter inductor and DC side energy storage capacitor of the three-phase three-level circuit and a second-stage optimization model of the harmonics of the three-phase three-level circuit, wherein the first-stage prediction model and the second-stage optimization model are connected in series.

[0011] The first-stage prediction model is used to predict and output the AC-side filter inductor value, the DC-side energy storage capacitor value, the confidence value T1 corresponding to the AC-side filter inductor value, and the confidence value T2 corresponding to the energy storage capacitor value. Based on the confidence values, the learning factor value of the second-stage optimization model is determined. The learning factor value is used to determine the learning ability of the second-stage optimization model; the larger the learning factor value, the stronger the learning ability of the second-stage optimization model. The second-stage optimization model is used to optimize the three-phase three-level switching time angle sequence to find the switching time angle sequence that minimizes the output voltage harmonics. In conjunction with the first aspect of this application, in another optional embodiment, the method further includes: data preprocessing, wherein the input data samples include:

[0012] Three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, rectified voltage average value Ud0, switching time angle sequence, load line current I11, and line voltage effective value V11.

[0013] The harmonic voltage value is obtained by performing a Fourier series expansion on the DC side voltage Ud;

[0014] The output data samples include: AC side filter inductance value, DC side energy storage capacitor value, and harmonic voltage value.

[0015] Correlation analysis is performed on the input and output data samples, and the input data samples are assigned positive correlation weights based on the degree of correlation. The stronger the correlation, the larger the weight, thus obtaining the final training input samples.

[0016] In conjunction with the first aspect of this application, in another alternative embodiment, the first-stage prediction model includes:

[0017] Training phase and prediction phase;

[0018] The training phase is used to establish a mapping relationship between the AC-side filter inductance value and the DC-side energy storage capacitor value in the training input data sample and the output data sample.

[0019] The prediction stage is used to predict the AC-side filter inductance value, DC-side energy storage capacitor value, confidence value T1 corresponding to the AC-side filter inductance value, and confidence value T2 corresponding to the energy storage capacitor value, which are one or more newly set input sample values.

[0020] If either the confidence value T1 or the confidence value T2 is less than 0.8, then the first-stage prediction model will be re-performed.

[0021] In conjunction with the first aspect of this application, in another optional embodiment, determining the learning factor value of the second-stage optimization model using the confidence score includes:

[0022] If the confidence value is greater than or equal to 0.8 and less than 0.9, then the learning factor value = the confidence value - 0.4;

[0023] If the confidence value is greater than or equal to 0.9 and less than 1, then the learning factor value = the confidence value - 0.7.

[0024] The learning factor value includes a learning factor value C1 and a learning factor value C2, wherein the learning factor value C1 is the larger learning factor value calculated from the confidence value T1 corresponding to the AC side filter inductor value or the confidence value T2 corresponding to the energy storage capacitor value, and the learning factor value C2 is the smaller learning factor value calculated from the confidence value T1 corresponding to the AC side filter inductor value or the confidence value T2 corresponding to the energy storage capacitor value.

[0025] In conjunction with the first aspect of this application, in another alternative embodiment, the second-stage optimization model includes:

[0026] The model parameter initialization phase includes acquiring harmonic voltage values, learning factor values, iteration counts, and inertia weights from training input data samples and output data samples.

[0027] The model optimization stage includes constructing a fitness function, performing iterative optimization, and outputting the switching time angle sequence that minimizes the output voltage harmonics.

[0028] Secondly, embodiments of this application provide a three-phase three-level circuit parameter determination device, the device comprising:

[0029] The data acquisition module is used to acquire data samples of a three-phase three-level circuit. The data samples include: three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, average rectified voltage Ud0, switching time angle sequence, load line current I11, effective value of line voltage V11, AC side filter inductor value, and DC side energy storage capacitor value.

[0030] The data processing module is used to preprocess the data samples. The preprocessing includes dividing the data samples into input data samples and output data samples, performing correlation analysis on the input data samples and output data samples and adding weights to obtain the final training input data samples.

[0031] The training module is used to input the preprocessed data samples into the preset three-phase three-level circuit parameter determination model for training, so as to obtain the optimal switching angle sequence and the corresponding AC side filter inductance value and DC side energy storage capacitor value.

[0032] The parameter determination model of the three-phase three-level circuit includes: a first-stage prediction model of the AC side filter inductor and DC side energy storage capacitor of the three-phase three-level circuit and a second-stage optimization model of the harmonics of the three-phase three-level circuit, wherein the first-stage prediction model and the second-stage optimization model are connected in series.

[0033] The first-stage prediction model is used to predict and output the AC-side filter inductance value, the DC-side energy storage capacitor value, the confidence value T1 corresponding to the AC-side filter inductance value, and the confidence value T2 corresponding to the energy storage capacitor value. Based on the confidence values, the learning factor value of the second-stage optimization model is determined. The learning factor value is used to determine the learning ability of the second-stage optimization model. The larger the learning factor value, the stronger the learning ability of the second-stage optimization model. The second-stage optimization model is used to optimize the three-phase three-level switching time angle sequence to find the switching time angle sequence that minimizes the output voltage harmonics.

[0034] Thirdly, one embodiment of this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method for determining the parameters of a three-phase three-level circuit provided in any one of the first aspects.

[0035] Fourthly, one embodiment of this application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for determining the parameters of a three-phase three-level circuit provided in any one of the first aspects.

[0036] The above-mentioned method, device, equipment, and storage medium for determining the parameters of a three-phase three-level circuit, by establishing a two-stage optimization model, simultaneously solves the problems of neutral point potential imbalance and harmonic suppression in three-phase three-level circuits. The first-stage prediction model predicts the filter inductor and energy storage capacitor, and judges the confidence value of the predicted value to determine whether the current prediction value is reliable. The current reliable confidence value determines the learning ability of the second-stage optimization model, which can effectively reduce the computational cost of the model. At the same time, under the optimization of the second-stage optimization model, the optimal switching time angle sequence is found. Attached Figure Description

[0037] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0038] Figure 1 This is a flowchart illustrating the method for determining the parameters of a three-phase three-level circuit according to an embodiment of this application.

[0039] Figure 2 This is a schematic diagram of the particle swarm optimization algorithm in the method for determining parameters of a three-phase three-level circuit according to an embodiment of this application.

[0040] Figure 3 This is a schematic diagram of the structure of the three-phase three-level circuit parameter determination device according to an embodiment of this application. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0042] This application provides a method for determining the parameters of a three-phase three-level circuit. Specifically, this method can be applied to computer equipment, such as a terminal or server.

[0043] The method for determining the parameters of a three-phase three-level circuit is as follows: Figure 1 As shown, the method for determining the parameters of the three-phase three-level circuit includes:

[0044] Step S1: Obtain data samples of the three-phase three-level circuit. The data samples include: three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, average rectified voltage Ud0, switching time angle sequence, load line current I11, effective value of line voltage V11, AC side filter inductor value, and DC side energy storage capacitor value.

[0045] Step S2: Preprocess the data samples. The preprocessing includes dividing the data samples into input data samples and output data samples, performing correlation analysis on the input data samples and output data samples and adding weights to obtain the final training input data samples.

[0046] Step S3: Input the preprocessed data samples into the preset three-phase three-level circuit parameter determination model for training to obtain the optimal switching angle sequence and the corresponding AC side filter inductance value and DC side energy storage capacitor value;

[0047] The three-phase three-level optimization model includes: a first-stage prediction model of the AC side filter inductor and DC side energy storage capacitor of the three-phase three-level circuit, and a second-stage optimization model of the harmonics of the three-phase three-level circuit, wherein the first-stage prediction model and the second-stage optimization model are connected in series.

[0048] Step S31: Using the first-stage prediction model, predict and output the AC-side filter inductor value, the DC-side energy storage capacitor value, the confidence value T1 corresponding to the AC-side filter inductor value, and the confidence value T2 corresponding to the energy storage capacitor value.

[0049] Step S32: Based on the confidence value, determine the learning factor value of the second-stage optimization model. The learning factor value is used to determine the learning ability of the second-stage optimization model. The larger the learning factor value, the stronger the learning ability of the second-stage optimization model.

[0050] Step S33: Using the second-stage optimization model, train the three-phase three-level switching time angle sequence to find the switching time angle sequence that minimizes the output voltage harmonics.

[0051] In this embodiment, data sample acquisition, data sample processing, the first-stage prediction model, and the second-stage optimization model jointly realize the method for determining the parameters of a three-phase three-level circuit.

[0052] First, the acquired three-phase three-level circuit data is preprocessed, and the DC-side voltage Ud is expanded using Fourier series to obtain the harmonic voltage value. Thus, the obtained output sample data includes the AC-side filter inductor value, the DC-side energy storage capacitor value, and the harmonic voltage value. The input data samples include: three-phase three-level AC-side electromotive force vector E, AC-side phase voltage vector V, AC-side filter inductor voltage vector Vl, AC-side phase current vector I, maximum phase voltage utilization rate M, DC-side voltage Ud, DC-side rated power Pe, average rectified voltage Ud0, switching time angle sequence, load line current I11, and effective value of line voltage V11.

[0053] Correlation analysis is performed on the input and output data samples, and the input data samples are assigned positive correlation weights based on the degree of correlation. The stronger the correlation, the larger the weight, thus obtaining the final training input samples.

[0054] The correlation analysis employs multi-input variable and multi-output variable analysis to determine the degree of correlation between multiple input variables of the input samples and multiple output variables of the output samples of the three-phase three-level circuit. At the same time, it distinguishes between input data samples as independent variables and output sample data as dependent variables. After analyzing the degree of correlation, positive correlation weights are assigned to the input sample data, thereby reducing the impact of independent variables with low correlation on the model and improving the model's predictive optimization ability.

[0055] Secondly, the first-stage prediction model includes a training stage and a prediction stage: the training stage is used to establish the mapping relationship between the AC side filter inductance value and the DC side energy storage capacitor value in the training input data sample and the output data sample;

[0056] The prediction phase is used to predict the AC-side filter inductance value, DC-side energy storage capacitor value, confidence value T1 corresponding to the AC-side filter inductance value, and confidence value T2 corresponding to the energy storage capacitor value for one or more newly set input sample values. The confidence value ranges from 0 to 1, with a higher confidence value indicating greater reliability. If either the confidence value T1 or the confidence value T2 is less than 0.8, the prediction in the first phase of the prediction model is repeated. By judging the magnitude of the confidence value, the accuracy of the model prediction is measured. This reduces the difficulty of obtaining test data and effectively reduces the computational cost of the first-phase prediction model.

[0057] The first-stage prediction model is connected in series with the second-stage optimization model. The confidence value output by the first-stage prediction model is used to determine the learning factor value of the second-stage optimization model. If the confidence value is greater than or equal to 0.8 and less than 0.9, then the learning factor value = the confidence value - 0.4.

[0058] If the confidence value is greater than or equal to 0.9 and less than 1, then the learning factor value = the confidence value - 0.7.

[0059] When the confidence level is in the range of 0.9 to 1, the second-stage optimization model considers the prediction of the first-stage prediction model to be very reliable. Therefore, its learning cost will be reduced, and the learning factor value will be in the range of 0.2 to 0.3, thereby improving the optimization speed of the second-stage optimization model and achieving faster optimization.

[0060] When the confidence level is in the range of 0.8 to 0.9, the second-stage optimization model considers the prediction of the first-stage prediction model to be highly reliable. Therefore, its learning cost will be reduced, and the learning factor value will be in the range of 0.4 to 0.5, thereby improving the learning ability of the second-stage optimization model and improving the accuracy of optimization.

[0061] The second stage of model optimization includes: a model parameter initialization stage, which includes acquiring harmonic voltage values, learning factor values, iteration counts, and inertia weights from training input data samples and output data samples. The learning factor values ​​include learning factor value C1 and learning factor value C2, wherein learning factor value C1 is the larger learning factor value calculated from the confidence value T1 corresponding to the AC side filter inductor value or the confidence value T2 corresponding to the energy storage capacitor value, and learning factor value C2 is the smaller learning factor value calculated from the confidence value T1 corresponding to the AC side filter inductor value or the confidence value T2 corresponding to the energy storage capacitor value.

[0062] The model optimization stage includes constructing a fitness function, performing iterative optimization, and outputting the switching time angle sequence that minimizes the output voltage harmonics.

[0063] Furthermore, as a preferred embodiment of this application, the specific steps in step S31 include:

[0064] Establish the first-stage prediction model, which can adopt a Gaussian process model, as follows:

[0065] Gaussian process models are a natural generalization of multivariate random variables to an infinite (countable or continuous) index set, and can be divided into two parts: training and prediction.

[0066] (1) Training a Gaussian Process (GP)

[0067] The properties of GP are determined by the mean function and the covariance function, and can be expressed as:

[0068]

[0069] Where x, x' ∈ Rd Let be any random variable, d be the dimension of the input data, and m(x) and k(x,x') be the mean function and covariance function, respectively, which can be further expressed as follows:

[0070] f(x)~GP(m(x),k(x,x')) (2)

[0071] Suppose the training set D = {(x i ,y i There exist n observations in the subset |i=1,2,…,n}, X i =[x1,x2,…,x n ] is a d×n dimensional input training matrix composed of n d-dimensional input training vectors, y=[y1,y2,…,y n ] T It consists of n training output scalars y i The training output vector is composed of these components. If the observed value y is contaminated with additive noise ε, the regression model can be expressed as y = f(x) + ε. Here, ε is a normally distributed random variable with a mean of 0 and a variance of 0. It can be represented as,

[0072]

[0073] Then the prior distribution of the observed target value y is:

[0074] y~N(0,K+σ n 2 I) (4)

[0075] Where K = K(X,X) is an n×n symmetric positive definite covariance matrix, and I is the identity matrix. The output values ​​y of the n training samples and n... * The output value f of each test sample * The joint Gaussian prior distribution is formed as follows:

[0076]

[0077] Where K(X,X) * ) is n * The n×n pairs of test output samples and n training output samples * The order covariance matrix, K(X) * ,X * ) is n * n test output samples * ×n * The order covariance matrix.

[0078] Typically, the covariance matrix (Gaussian kernel) in a GP regression model uses the ARD (Automatic Relevance Determination) series of quadratic exponential kernels. However, considering that the actual optimization problem does not have an ideally smooth input, this paper uses the ARD Matern 5 / 2 kernel function.

[0079]

[0080] in θ = {θ0,...,θ d Let} be a set of hyperparameters of the Gaussian kernel, which determine the properties of the GP model. In actual modeling, the maximum likelihood function is typically used to determine the hyperparameters in the GP model: by establishing the log-likelihood function of the conditional probability of the training samples, taking the partial derivative with respect to the hyperparameters, and then using the conjugate gradient optimization method to find the optimal solution for the hyperparameters. The negative log-likelihood function has the form:

[0081]

[0082] After obtaining the optimal hyperparameters, the trained GP can be used for relevant predictions to obtain the predicted output energy storage capacitor value and filter inductance value, as well as the confidence values ​​corresponding to the energy storage capacitor value and filter inductance value.

[0083] Furthermore, as a preferred embodiment of this application, the specific steps in step S33 include:

[0084] A second-stage optimization model is established, which can employ a particle swarm optimization algorithm. Since the output voltage exhibits 1 / 4 periodic symmetry and satisfies Dirichlet's theorem, it is expanded into a Fourier series form:

[0085]

[0086] In the formula, A0 is the DC component of the voltage Ud, and A0 = 0 when the waveform is odd-symmetric at half its length.

[0087]

[0088]

[0089] Since the phase voltage waveform is odd-symmetric about half the period and axially symmetric about one-quarter of the period, it contains only odd-order sinusoidal harmonics. Therefore:

[0090] A n =0 (11)

[0091] When n is odd When n is even, Bn is 0;

[0092] In the formula: Usm(n) is the voltage value of the nth harmonic; Ud is the DC side voltage; n is the harmonic order; ω is the angular frequency; αk is the switching angle; N is the number of switching angles within 1 / 4 cycle. The switching angle should meet the following conditions:

[0093] 0 < α1 < α2 < ... < α N <π / 2 (12)

[0094] In a three-phase three-level system, harmonics that are multiples of 3 can be automatically eliminated, and zero-sequence harmonics can be eliminated based on the three-phase three-wire output. Therefore, it is only necessary to consider eliminating harmonics of the 5th, 7th, 11th, ..., 6n-1th and 6n+1th orders.

[0095] To eliminate the 5th and 7th harmonics, which have a relatively high content, simply set B1 = m and B5 and B7 to 0 in equation (5) to make the fundamental frequency the amplitude under any modulation, and the 5th and 7th harmonics will be eliminated.

[0096] Establish the fitness function equation:

[0097]

[0098] In the formula, USM(1), USM(5), and USM(7) are the voltage values ​​of the fundamental, 5th, and 7th harmonics, respectively; f1(α), f2(α), and f3(α) are functions established for the fundamental, 5th, and 7th harmonics with the switching angle α as the variable; α1, α2, and α3 are three values ​​of α; m is the fundamental modulation index, and we have:

[0099]

[0100] Solving the fitness function equation yields a series of switching angles for eliminating the 5th and 7th harmonics.

[0101] In the PSO algorithm, if n particles are chosen as the initial random population, then in a d-dimensional space, the position of each particle represents a solution to the problem to be optimized. The performance of each particle depends on the fitness value determined by the objective function of the problem. Particles iteratively adjust their direction and velocity until they find the global optimum. In the population, suppose the position of the i-th particle is Xi = [xi1, xi2, ..., xid], and its velocity (i.e., the rate of change of position) is Vi = [vi1, vi2, ..., vid]. During its movement, the individual optimum found by this particle is Pbesti = [pbesti1, pbesti2, ..., pbestid]. The current optimum found by the entire population, i.e., the global optimum, is Gbest = [gbest1, gbest2, ..., gbestd]. Then, during the (p+1)-th iteration, the particle updates its velocity and position information according to the method defined by the following formula.

[0102]

[0103] In the formula: i = 1, 2, ..., n; j = 1, 2, ..., d; p is the iteration number; r1 and r2 are arbitrary random numbers in the interval [0, 1]; c1 and c2 are learning factors, representing the particle's self-learning ability and global learning ability, respectively; w is the inertia weight, used to control the influence of the previous velocity on the current state, and this value is calculated according to the following formula.

[0104]

[0105] In the formula, w1 and w2 are weighting factors, and M axgen G generation These are the maximum number of iterations and the current number of iterations, respectively.

[0106] like Figure 2 As shown, the switching time angle sequence is optimized according to the flowchart.

[0107] Step 1: Three-phase three-level data input;

[0108] Step 2: Randomly generate the switching angle sequence X and initialize the particle's position x and velocity v;

[0109] Step 3: Initialize parameter settings such as w and number of iterations;

[0110] Step 4: Calculate pbest and gbest;

[0111] Step 5: Update x and v, generate a new particle X(t+1), and calculate the fitness value F(X(t+1)) of the new particle.

[0112] Step 6: Determine if F(Xi(t+1)) is greater than F(pbesti). If yes, proceed to Step 7; otherwise, proceed to Step 8.

[0113] Step 7, Pbesti = Xi(t+1);

[0114] Step 8: Determine if F(Xi(t+1)) is greater than F(gbesti). If yes, proceed to Step 9; otherwise, proceed to Step 10.

[0115] Step9, gbesti = Xi(t+1);

[0116] Step 10: Confirm that all particles have been calculated? If yes, proceed to Step 11; otherwise, proceed to Step 3.

[0117] Step 11: Determine if the maximum number of iterations has been reached. If yes, proceed to Step 12; otherwise, proceed to Step 13.

[0118] Step 12: Output the optimal solution.

[0119] This application also provides a device for determining the parameters of a three-phase three-level circuit, such as... Figure 3 As shown, the device includes:

[0120] The data acquisition module is used to acquire data samples of a three-phase three-level circuit. The data samples include: three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, average rectified voltage Ud0, switching time angle sequence, load line current I11, effective value of line voltage V11, AC side filter inductor value, and DC side energy storage capacitor value.

[0121] The data processing module is used to preprocess the data samples. The preprocessing includes dividing the data samples into input data samples and output data samples, performing correlation analysis on the input data samples and output data samples and adding weights to obtain the final training input data samples.

[0122] The training module is used to input the preprocessed data samples into the preset three-phase three-level circuit parameter determination model for training, so as to obtain the optimal switching angle sequence and the corresponding AC side filter inductance value and DC side energy storage capacitor value.

[0123] The parameter determination model of the three-phase three-level circuit includes: a first-stage prediction model of the AC side filter inductor and DC side energy storage capacitor of the three-phase three-level circuit and a second-stage optimization model of the harmonics of the three-phase three-level circuit, wherein the first-stage prediction model and the second-stage optimization model are connected in series.

[0124] The first-stage prediction model is used to predict and output the AC-side filter inductance value, the DC-side energy storage capacitor value, the confidence value T1 corresponding to the AC-side filter inductance value, and the confidence value T2 corresponding to the energy storage capacitor value. Based on the confidence values, the learning factor value of the second-stage optimization model is determined. The learning factor value is used to determine the learning ability of the second-stage optimization model. The larger the learning factor value, the stronger the learning ability of the second-stage optimization model. The second-stage optimization model is used to optimize the three-phase three-level switching time angle sequence to find the switching time angle sequence that minimizes the output voltage harmonics.

[0125] Specific limitations regarding the three-phase three-level circuit parameter determination device can be found in the limitations of the three-phase three-level circuit parameter determination method described above, and will not be repeated here. Each module in the aforementioned three-phase three-level circuit parameter determination device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in a computer device, or stored in software in the memory of a computer device, so that the processor can call and execute the corresponding operations of each module.

[0126] This application embodiment also provides a computer device, the computer device comprising:

[0127] processor;

[0128] Memory used to store executable instructions for a computer;

[0129] The processor is configured to execute the computer-executable instructions to implement the method for determining the parameters of a three-phase three-level circuit as described in any of the above embodiments.

[0130] A processor may be a central processing unit (CPU) or other form of processing unit with data processing and / or instruction execution capabilities, and may control other components in a computer device to perform desired functions.

[0131] The memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 1501 may execute the program instructions to implement the steps in the text recognition methods of the various embodiments of this application described above, and / or other desired functions.

[0132] In one example, the computer device may also include input devices and output devices, which are interconnected via a bus system and / or other forms of connection mechanism (not shown in the figure).

[0133] In addition, input devices may include, for example, a keyboard, a mouse, a microphone, etc. Output devices can output various information to the outside, and may include, for example, a monitor, speakers, a printer, and communication networks and their connected remote output devices, etc.

[0134] This application also provides a computer-readable storage medium storing instructions that, when executed by a processor of a computer device, enable the computer device to perform the three-phase three-level circuit parameter determination method described in any of the above embodiments.

[0135] Embodiments of this application may be systems, methods, and / or computer program products. A computer program product may include a computer-readable storage medium on which computer-readable program instructions are loaded to cause a processor to implement various aspects of this application. The computer program product may be written in any combination of one or more programming languages ​​to perform operations of embodiments of this application. Programming languages ​​include object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code may execute entirely on a user's computing device, partially on a user's device, as a standalone software package, partially on a user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computers, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuits, such as programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), are personalized by utilizing state information of computer-readable program instructions. These electronic circuits can execute computer-readable program instructions to implement various aspects of this application.

[0136] Computer-readable storage media can take the form of any combination of one or more readable media. A readable medium can be a readable signal medium or a readable storage medium. A computer-readable storage medium is a tangible device capable of holding and storing instructions for use by an instruction execution device. A readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combinations thereof. The computer-readable storage medium as used herein is not to be construed as a transient signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0137] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0138] Various aspects of this application are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0139] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0140] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0141] It should be noted that the embodiments of the method for determining parameters of a three-phase three-level circuit, the device for determining parameters of a three-phase three-level circuit, the computer device, and the computer-readable storage medium provided in this application belong to the same concept; the technical features in the technical solutions described in each embodiment can be arbitrarily combined without conflict.

[0142] It should be understood that the above embodiments are exemplary and are not intended to encompass all possible implementations included in the claims. Various modifications and changes can be made to the above embodiments without departing from the scope of this disclosure. Similarly, the various technical features of the above embodiments can be arbitrarily combined to form other embodiments of the present invention that may not be explicitly described. Therefore, the above embodiments only illustrate several implementations of the present invention and do not limit the scope of protection of this patent.

Claims

1. A method for determining parameters of a three-phase three-level circuit, characterized in that, The method includes: Acquire data samples of a three-phase three-level circuit, including: three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, average rectified voltage Ud0, switching time angle sequence, load line current I11, effective value of line voltage V11, AC side filter inductor value, and DC side energy storage capacitor value; The data samples are preprocessed, which includes dividing the data samples into input data samples and output data samples, performing correlation analysis on the input data samples and output data samples and adding weights to obtain the final training input data samples. The preprocessed data samples are input into the preset three-phase three-level circuit parameter determination model for training to obtain the optimal switching angle sequence and the corresponding AC side filter inductance value and DC side energy storage capacitor value. The parameter determination model of the three-phase three-level circuit includes: a first-stage prediction model of the AC side filter inductor and DC side energy storage capacitor of the three-phase three-level circuit and a second-stage optimization model of the harmonics of the three-phase three-level circuit, wherein the first-stage prediction model and the second-stage optimization model are connected in series. The first-stage prediction model is used to predict and output the AC-side filter inductance value, the DC-side energy storage capacitor value, the confidence value T1 corresponding to the AC-side filter inductance value, and the confidence value T2 corresponding to the energy storage capacitor value. Based on the confidence values, the learning factor value of the second-stage optimization model is determined. The learning factor value is used to determine the learning ability of the second-stage optimization model. The larger the learning factor value, the stronger the learning ability of the second-stage optimization model. The second-stage optimization model is used to optimize the three-phase three-level switching time angle sequence to find the switching time angle sequence that minimizes the output voltage harmonics.

2. The method for determining parameters of a three-phase three-level circuit according to claim 1, characterized in that, The data preprocessing includes: The input data samples include: Three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, rectified voltage average value Ud0, switching time angle sequence, load line current I11, and line voltage effective value V11. The harmonic voltage value is obtained by performing a Fourier series expansion on the DC side voltage Ud; The output data samples include: AC side filter inductance value, DC side energy storage capacitor value, and harmonic voltage value. Correlation analysis is performed on the input and output data samples, and the input data samples are assigned positive correlation weights based on the degree of correlation. The stronger the correlation, the larger the weight, thus obtaining the final training input samples.

3. The method for determining parameters of a three-phase three-level circuit according to claim 1, characterized in that, The first-stage prediction model includes: Training phase and prediction phase; The training phase is used to establish a mapping relationship between the AC-side filter inductance value and the DC-side energy storage capacitor value in the training input data sample and the output data sample. The prediction stage is used to predict the AC-side filter inductance value, DC-side energy storage capacitor value, confidence value T1 corresponding to the AC-side filter inductance value, and confidence value T2 corresponding to the energy storage capacitor value, which are one or more newly set input sample values.

4. The method for determining the parameters of a three-phase three-level circuit according to claim 3, characterized in that, The confidence level values ​​include: If either the confidence value T1 or the confidence value T2 is less than 0.8, then the first-stage prediction model will be re-performed.

5. The method for determining parameters of a three-phase three-level circuit according to claim 1, characterized in that, The confidence score determines the learning factor value of the second-stage optimization model, including: If the confidence value is greater than or equal to 0.8 and less than 0.9, then the learning factor value = the confidence value - 0.4; If the confidence value is greater than or equal to 0.9 and less than 1, then the learning factor value = the confidence value - 0.

7.

6. The method for determining parameters of a three-phase three-level circuit according to claim 5, characterized in that, The learning factor value includes a learning factor value C1 and a learning factor value C2, wherein the learning factor value C1 is the larger learning factor value calculated from the confidence value T1 corresponding to the AC side filter inductor value or the confidence value T2 corresponding to the energy storage capacitor value, and the learning factor value C2 is the smaller learning factor value calculated from the confidence value T1 corresponding to the AC side filter inductor value or the confidence value T2 corresponding to the energy storage capacitor value.

7. The method for determining parameters of a three-phase three-level circuit according to claim 1, characterized in that, The second-stage optimization model includes: The model parameter initialization phase includes acquiring harmonic voltage values, learning factor values, iteration counts, and inertia weights from training input data samples and output data samples. The model optimization stage includes constructing a fitness function, performing iterative optimization, and outputting the switching time angle sequence that minimizes the output voltage harmonics.

8. A device for determining parameters of a three-phase three-level circuit, characterized in that, The device includes: The data acquisition module is used to acquire data samples of a three-phase three-level circuit. The data samples include: three-phase three-level AC side electromotive force vector E, AC side phase voltage vector V, AC side filter inductor voltage vector Vl, AC side phase current vector I, phase voltage maximum utilization rate M, DC side voltage Ud, DC side rated power Pe, average rectified voltage Ud0, switching time angle sequence, load line current I11, effective value of line voltage V11, AC side filter inductor value, and DC side energy storage capacitor value. The data processing module is used to preprocess the data samples. The preprocessing includes dividing the data samples into input data samples and output data samples, performing correlation analysis on the input data samples and output data samples and adding weights to obtain the final training input data samples. The training module is used to input the preprocessed data samples into the preset three-phase three-level circuit parameter determination model for training, so as to obtain the optimal switching angle sequence and the corresponding AC side filter inductance value and DC side energy storage capacitor value. The parameter determination model of the three-phase three-level circuit includes: a first-stage prediction model of the AC side filter inductor and DC side energy storage capacitor of the three-phase three-level circuit and a second-stage optimization model of the harmonics of the three-phase three-level circuit, wherein the first-stage prediction model and the second-stage optimization model are connected in series. The first-stage prediction model is used to predict and output the AC-side filter inductance value, the DC-side energy storage capacitor value, the confidence value T1 corresponding to the AC-side filter inductance value, and the confidence value T2 corresponding to the energy storage capacitor value. Based on the confidence values, the learning factor value of the second-stage optimization model is determined. The learning factor value is used to determine the learning ability of the second-stage optimization model. The larger the learning factor value, the stronger the learning ability of the second-stage optimization model. The second-stage optimization model is used to optimize the three-phase three-level switching time angle sequence to find the switching time angle sequence that minimizes the output voltage harmonics.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images