A method for modeling harmonic sources based on synchronized phasor measurement data

Abstract

The application provides a harmonic source modeling method based on synchronous phasor measurement data, extracts current characteristics of a harmonic source load by VMD decomposition, and clusters working states of the load by a DBSCAN method; utilizes voltage and current data measured by a system to reflect nonlinear characteristics of the harmonic source by load equivalent impedance; and thus establishes a model for the harmonic source. The working states of the load can be classified, the application range is expanded, and the practicability is higher, and the harmonic source modeling method can effectively support harmonic treatment, trace analysis and other work.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power systems, harmonic governance and the like, and in particular to a harmonic source modeling method based on synchronous phasor measurement data. BACKGROUND

[0002] In the process of continuous development of power systems, the types and quantities of harmonic sources are also increasing. The access of various power electronic devices and new energy equipment makes the voltage and current waveform distortion of power systems more serious. In order to find a method to govern harmonic problems, the existing technology improves the power quality level from many aspects, and the demand for modeling of harmonic source load is more and more obvious. As an important process of analyzing harmonics, an accurate harmonic source model will play an important role in understanding and analyzing harmonic phenomena, predicting and estimating the influence of harmonics on the entire power grid system, etc.

[0003] However, the current prior art scheme still has the following obvious defects and deficiencies:

[0004] 1) The equivalent impedance of nonlinear load can essentially reflect the influence of harmonic source load in the power system, which is an important link between voltage waveform and current waveform, and contains the cause of harmonic generation. The process of obtaining the equivalent impedance of nonlinear load is to extract the characteristics of harmonic source load. In theory, the instantaneous voltage and current values measured by the system are used to solve the equivalent impedance of the load, which will not be affected by the system side harmonic source, but the current method of calculating the equivalent impedance of the load will cause large errors due to the large system side harmonic distortion rate and the ill-conditioned measurement matrix.

[0005] 2) In order to meet the different needs of users, the existing load often has multiple working states. The harmonic source load often has different load characteristics in different working states, and different load characteristics will also have unique effects on the power system. The current research method focuses on the establishment of single-state model of load, and there are few identification methods for unlabeled multi-state load or state-continuous load.

[0006] 3) The current harmonic source modeling method can be divided into mechanism modeling method and data-driven modeling method according to the idea. The former accurately describes and analyzes the behavior characteristics of harmonic sources through physical principles, but it is difficult to accurately describe the mechanism of complex harmonic sources generating harmonics. The latter uses a large amount of existing power system data to establish a model by extracting the potential relationship between data and harmonics, but lacks physical interpretability. SUMMARY

[0007] With the accelerated construction of new power systems, the types and quantities of harmonic sources have increased rapidly, and the access of various power electronic devices and new energy equipment has made the voltage and current waveform distortion of the power system more serious. In order to find a method to solve the harmonic problem and improve the power quality of the power grid from multiple aspects, the demand for modeling of harmonic source load is more and more prominent. The present application combines two ideas of harmonic source modeling, uses the voltage and current data measured by the PMU device, reflects the nonlinear characteristics of the harmonic source through the equivalent impedance of the load, proposes a modeling method for distinguishing the working state of the harmonic source based on the equivalent impedance of the load and VMD-DBSCAN, compared with other modeling methods, the present application can classify the working state of the load, expand the scope of application while having stronger practicality, and can effectively support harmonic control, source analysis and other work.

[0008] Specifically, the following technical solutions are adopted:

[0009] A harmonic source modeling method based on synchronous phasor measurement data, which uses VMD decomposition to extract the current characteristics of the harmonic source load, and clusters the working state of the load through the DBSCAN method;

[0010] The voltage and current data measured by the system are used to reflect the nonlinear characteristics of the harmonic source through the equivalent impedance of the load, thereby establishing a model of the harmonic source.

[0011] Further, the following steps are included:

[0012] Step S1: using a synchronous phasor measurement device PMU to collect voltage and current sampling data of the load port;

[0013] Step S2: VMD decomposition is performed on the sampling current data of the harmonic source load, and the energy and energy entropy of the harmonic source current are calculated;

[0014] Step S3: using DBSCAN to identify the working state of the harmonic source load at each time;

[0015] Step S4: combining the working state of the harmonic source load at each time obtained in step S3, solving the equivalent impedance parameters of the harmonic source load under different working conditions;

[0016] Step S5: using the voltage and the equivalent impedance parameters of the harmonic source load under each working condition to estimate the harmonic source current, thereby establishing a harmonic source load model.

[0017] Further, in step S2, a VMD constrained variational model is constructed, and it is assumed that the harmonic source current is decomposed into several intrinsic modal components, and the expression is:

[0018]

[0019] In the formula, i k(t) represents the harmonic source current I. IMFk (t) The amplitude of the envelope; It is the instantaneous phase;

[0020] Performing a Hilbert transform on each decomposed modal component yields the estimated bandwidth expression for each mode:

[0021]

[0022] In the formula, I IMFk (t) represents the intrinsic mode components decomposed from the harmonic source current; ω k (t) represents each I IMFk The center frequency of (t);

[0023] Equation (2) should satisfy the constraint that each decomposed mode has the minimum estimated bandwidth and that the superposition of each mode component is the original harmonic source current:

[0024]

[0025] In the above formula, I(t) is the original harmonic source load current;

[0026] Introducing the augmented Lagrange function solution equation (3), the solution function is as follows:

[0027]

[0028] In the formula, α is the penalty particle and λ is the Lagrange multiplier;

[0029] To obtain the optimal solution of equation (4), the alternating direction multiplier method is adopted, and the iterative parameters are continuously updated. The optimal value is found through an iterative formula as follows:

[0030]

[0031]

[0032]

[0033]

[0034] In the formula, n is the number of iterations, and ω is the frequency. These are the eigenmode components decomposed from the harmonic source current in the nth iteration. For each The center frequency, For the Lagrange multipliers in the nth iteration, for The Fourier transform result of the residual error, τ represents the step length of iteration, τ > 0, ε is a residual error threshold, and the output result is output when the residual error is less than ε, and ε > 0;

[0035] Energy and energy entropy are introduced as the characteristics of the harmonic source load; the expressions of energy and energy entropy are as follows:

[0036] E i =∫I IMFk (t)dt (9)

[0037]

[0038] In the formula, E i is the energy of the i-th modal component, H EN is the energy entropy of the harmonic source load current.

[0039] Further, in step S3, two parameters are introduced in the process of identifying and solving, which are Eps and MinPts, wherein Eps is a distance threshold between sample points, and MinPts is a minimum value of the number of clustering samples; the data is divided into three types, which are core data, boundary data and noise data, and the principle is that the number of other data points in the Eps field of the data point; when the number of points in the field is more than MinPts, it is core data; when the number of points in the field is less than MinPts but more than 0, it is boundary data; when there is no other data in the field, it is noise data; the cluster formed by the core data is regarded as a working state;

[0040] In the process of load state identification step:

[0041] By arbitrarily selecting a feature data of an un-identified state, the number of other data in the Eps field of the data is calculated to determine the data type;

[0042] Until the states of all feature data are identified, the working states of the current data of each segment are output.

[0043] Further, in step S4:

[0044] The voltage and current measurement values at the load port are represented by u(t) and i(t); the time-varying equivalent impedance parameters of the load are represented by R(t) and L(t);

[0045] t1 and t2 are two consecutive sampling points, and it is assumed that R and L of the equivalent impedance parameters of the load are stable and unchanged in the process from t1 to t2;

[0046] u(t1) is the voltage sampling data of the load port at t1, i(t1) is the current sampling data of the load port at t1, R(t1) is the equivalent resistance parameter at t1, L(t1) is the equivalent reactance parameter at t1, u(t2) is the voltage sampling data of the load port at t2, i(t2) is the current sampling data of the load port at t2, R(t2) is the equivalent resistance parameter at t2, and L(t2) is the equivalent reactance parameter at t2;

[0047] For each sampling interval, the solution of the load equivalent impedance parameter is realized by the following formula:

[0048]

[0049] The harmonic source current is solved by combining the load equivalent impedance parameter and the voltage data by the following formula:

[0050]

[0051] In the formula, i(t K ) is the predicted current value at t K , u(t K ) is the actual voltage value at t K , R(t K ) and L(t K ) are the equivalent resistance and reactance parameters at t K , i(t K+1 ) is the predicted current value at t K+1 , and u(t K+1 ) is the actual voltage value at t K .

[0052] Further, the harmonic source load model comprises a data acquisition module, a state identification module and an impedance module.

[0053] The data acquisition module acquires voltage and current sampling data at the load port through a synchronous phasor measurement device; the state identification module performs VMD decomposition on the load current state, extracts the energy and energy entropy of the load at each time, and takes the clustering results as the feature data of DBSCAN clustering to represent the working state of the load; and the impedance module saves the equivalent impedance parameters of the harmonic source load under different working states for harmonic source load current prediction.

[0054] Further, the specific steps for establishing the harmonic source load model are as follows:

[0055] First, the voltage and current data at the harmonic source load port are collected by using a synchronous phasor measurement device, and the sampling frequency is recorded;

[0056] Then, the working state of the harmonic source load is identified;

[0057] Finally, select the voltage and current data of several cycles in each working state, calculate the equivalent impedance parameters of the harmonic source load in each working state, and store them in the impedance module, together with the data acquisition module and the state recognition module to form a harmonic source load model.

[0058] And a harmonic source load model based on synchronous phasor measurement data is obtained by the above method:

[0059] It comprises a data acquisition module, a state recognition module and an impedance module.

[0060] The data acquisition module obtains voltage and current sampling data at the load port through a synchronous phasor measurement device; the state recognition module performs VMD decomposition on the load current state, extracts the energy and energy entropy of the load at each time, and uses them as feature data for DBSCAN clustering, and uses the clustering results to represent the working state of the load; and the impedance module stores the equivalent impedance parameters of the harmonic source load in different working states for harmonic source load current prediction.

[0061] In view of the defects and deficiencies of the prior art and the new requirements of the power system, the present application proposes a harmonic source modeling method based on synchronous phasor measurement data, which mainly includes the following design points:

[0062] 1) Based on the differences in harmonic characteristics of different types of harmonic sources and different working states of the same harmonic source, the current characteristics of the harmonic source load are extracted by VMD decomposition, and the working state of the load is clustered by DBSCAN method, which reduces the complexity of system analysis and calculation compared with directly using current clustering, and improves the calculation efficiency.

[0063] 2) Combining two ideas of harmonic source modeling, using the voltage and current data measured by the system, the nonlinear characteristics of the harmonic source are reflected by the equivalent impedance of the load, and a modeling method based on the equivalent impedance of the load and VMD-DBSCAN is proposed to distinguish the working state of the harmonic source, which can classify the working state of the load compared with other modeling methods, expand the application range, and has stronger practicality.

[0064] Compared with the prior art, the present application and its preferred scheme have the following outstanding effects:

[0065] 1) VMD decomposition is used, and energy entropy is introduced as a characteristic quantity of the working state of the harmonic source, and DBSCAN method is used for clustering, which improves the recognition accuracy of the multi-state load, and the present application can be applied not only to the state recognition of single-phase residential load, but also to three-phase circuits.

[0066] 2) The load equivalent impedance is the link between the voltage waveform and the current waveform, and the cause of harmonic generation can be essentially reflected by calculating the equivalent impedance of the nonlinear load. By combining the two ways of modeling the harmonic source, a model of the harmonic source is established based on the measured harmonic data of the system, which has strong physical interpretation and can be extended to three-phase user modeling to reflect power quality problems. BRIEF DESCRIPTION OF DRAWINGS

[0067] The application will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0068] Figure 1 The harmonic source modeling method flowchart of the embodiment of the application is based on load equivalent impedance parameters and VMD-DBSCAN;

[0069] Figure 2 The data type classification diagram of the embodiment of the application;

[0070] Figure 3 The wave source load state recognition flowchart of the embodiment of the application;

[0071] Figure 4 The time-domain load equivalent impedance parameter schematic diagram of the embodiment of the application;

[0072] Figure 5 The harmonic source load model diagram of the embodiment of the application. DETAILED DESCRIPTION

[0073] In the following, specific embodiments of the present application will be described in detail with reference to the accompanying drawings, and those skilled in the art can clearly understand the present application and implement the present application according to these detailed descriptions. The features in each different embodiment can be combined to obtain new embodiments or replace some features in some embodiments to obtain other preferred embodiments without departing from the principles of the present application.

[0074] It should be noted that the terms used herein are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and it should also be understood that when the terms "comprise" and / or "include" are used in the specification, there is a feature, step, operation, device, component and / or combination thereof.

[0075] In order to make the features and advantages of the patent more obvious and easy to understand, the following specific examples are described in detail as follows:

[0076] 1 Harmonic source modeling flow based on load equivalent impedance parameters and VMD-DBSCAN

[0077] The load equivalent impedance, as an equivalent time-varying impedance parameter based on the Thevenin theorem, contains the essential reason for the harmonic generated by the harmonic source load connected to the power system. Considering the differences in harmonic characteristics of different types of harmonic sources and different working states of the same harmonic source, the application introduces VMD-DBSCAN to extract and identify the characteristics of the harmonic source, obtains the working state of the multi-state harmonic source, and uses the classified voltage and current data to model the different working states of the harmonic source load in combination with the load equivalent impedance parameter. The principle flowchart of the harmonic source modeling based on the load equivalent impedance parameter and VMD-DBSCAN is shown in Figure 1 The specific steps are as follows:

[0078] 1) Use the synchronous phasor measurement device (PMU, phasor measurement unit) to collect the voltage and current sampling data of the load port;

[0079] 2) Perform VMD decomposition on the sampling current data of the harmonic source load to obtain the energy and energy entropy of the harmonic source current;

[0080] 3) Use DBSCAN to identify the working state of the harmonic source load at each time;

[0081] 4) Combine the working state of the harmonic source load at each time obtained in (3) to solve the load equivalent impedance parameter of the harmonic source load under different working states;

[0082] 5) Use the voltage and the load equivalent impedance parameter of the harmonic source load under each working state to estimate the harmonic source current;

[0083] 6) Establish a harmonic source load model.

[0084] 2 Load operating state identification based on VMD-DBSCAN

[0085] Existing nonlinear loads often have multiple operating states, such as electric arc furnaces, refrigerators, microwave ovens, etc. The harmonic currents generated by different types of harmonic sources and different working states of the same harmonic source are different, and the modal components obtained by VMD decomposition are significantly different. Therefore, the application takes the energy and energy entropy of each intrinsic modal component obtained by load decomposition as the characteristics of the harmonic source state, and identifies the working state of the harmonic source load through DBSCAN clustering.

[0086] 2.1 Harmonic source feature extraction based on VMD

[0087] VMD is a signal decomposition method, which has the nature of completely non-recursive, can divide the frequency domain of the signal, and can divide the multi-component signal into multiple single-component amplitude modulation frequency modulation signals at one time, and obtain the effective components of the signal. Compared with directly using the current as the characteristic data of the harmonic source load, the energy and energy entropy of the modal component obtained by VMD decomposition contain more harmonic source load information, so that the accuracy of load state identification can be improved.

[0088] The VMD constrained variational model is constructed, and the harmonic source current is assumed to be decomposed into several eigenmodal components, and the expression is as follows:

[0089]

[0090] In the formula, i k (t) is the amplitude of the envelope of the harmonic source current I IMFk (t); is the instantaneous phase.

[0091] The decomposed modal components are subjected to Hilbert transform (HT), so that they have stable center frequency and limited bandwidth, and the estimated bandwidth expression of each modal is as follows:

[0092]

[0093] In the formula, I IMFk (t) is the eigenmodal component decomposed from the harmonic source current; ω k (t) is the center frequency of each I IMFk (t).

[0094] In addition, the above formula should satisfy the constraint condition that the estimated bandwidth of each modal is minimum, and each modal component is superimposed to obtain the original harmonic source current.

[0095]

[0096] In the formula, I(t) is the original harmonic source load current.

[0097] The present application utilizes the advantages of the quadratic penalty term and the Lagrange multiplier method, introduces an augmented Lagrange function to solve formula (3), and the solving function is as follows:

[0098]

[0099] In the formula, alpha is the penalty particle, and lambda is the Lagrange multiplier.

[0100] In order to obtain the optimal solution of formula (4), the present application selects to use the alternating direction multiplier method, and the value of iteration parameter is updated constantly to optimize, and the iteration formula is as follows:

[0101]

[0102]

[0103]

[0104]

[0105] where n is the iteration number, ω is the frequency, is the eigenmode component decomposed from the harmonic source current in the n th iteration, is the center frequency of each is the Lagrange multiplier of the n th iteration, is the Fourier transform result of , τ represents the step size of iteration, τ > 0, and ε is a residual threshold, and the output result is less than ε, ε > 0;

[0106] In order to better extract the information contained in each modal component, energy and energy entropy are introduced as the characteristics of the harmonic source load. The expressions of energy and energy entropy are as follows:

[0107] E i = ∫I IMFk (t)dt (9)

[0108]

[0109] where E i is the energy of the i th modal component, H EN is the energy entropy of the harmonic source load current.

[0110] 2.2 Wave source load state recognition based on DBSCAN

[0111] DBSCAN is a clustering algorithm that classifies data based on density. Considering that the energy and energy entropy of the eigenmode component of the harmonic source load are different under different working states, and a large amount of measured data can be obtained through multiple measurements or long-time measurements, the DBSCAN method is selected to recognize the load state. Compared with the commonly used K-means clustering algorithm, DBSCAN does not need to extract the input clustering number, and has better clustering effect for non-convex data sets. Compared with other unsupervised clustering algorithms such as AP clustering, DBSCAN reduces the complexity of calculation, reduces the requirement for storage space of the solving system, and improves the processing capacity for discrete data and noise data.

[0112] ​In order to better describe the closeness of the data sample, the solving process introduces two parameters, Eps and MinPts, wherein Eps is the distance threshold between sample points, and MinPts is the minimum value of the number of clustering samples. Figure 2 As shown in the figure, the application divides data into three types, namely core data, boundary data and noise data, and the principle is that the number of other data points in the Eps field of the data point. When the number of points in the field is more than MinPts, it is core data; when the number of points in the field is less than MinPts but more than 0, it is boundary data; when there is no other data in the field, it is noise data. The cluster formed by the core data is regarded as a working state, which is used for the establishment of the model hereinafter.

[0113] The load state identification steps of the application are summarized as follows:

[0114] 1) Using a synchronous phasor measurement device (PMU, phasor measurement unit) to collect current sampling data of a load port and performing segmented processing;

[0115] 2) Initializing modal components, solving the optimal modal components of each segment of current data through iteration, and extracting the energy and energy entropy of each modal component;

[0116] 3) Arbitrarily selecting a feature data of an un-identified state, calculating the number of other data in the Eps field of the feature data, and judging the data type of the feature data;

[0117] 4) Until the states of all feature data are identified, the working states of each segment of current data are output.

[0118] The identification process is shown in the form of a flow chart as shown in Figure 3 .

[0119] 3 Load equivalent impedance calculation

[0120] 3.1 Time-domain load equivalent impedance model

[0121] Based on the characteristic that most of the harmonic source loads are inductive at present, the application selects the RL series equivalent impedance expression of the load in the time domain, which is widely used in various fields and has been studied a lot. In Figure 4 , the voltage and current measurement values at the load port are represented by u(t) and i(t); and the time-varying equivalent impedance parameters of the load are represented by R(t) and L(t).

[0122] 3.2 Solving load equivalent impedance parameters

[0123] From Figure 4 , the following relationship between the current and voltage of the measurement point can be obtained:

[0124]

[0125] Assuming t1 and t2 are two continuous sampling points, for formula (11), the following formula is satisfied at t1 and t2:

[0126]

[0127] In the formula, u(t1) is the voltage sampling data of the load port at t1, i(t1) is the current sampling data of the load port at t1, R(t1) is the equivalent resistance parameter at t1, L(t1) is the equivalent reactance parameter at t1, u(t2) is the voltage sampling data of the load port at t2, i(t2) is the current sampling data of the load port at t2, R(t2) is the equivalent resistance parameter at t2, and L(t2) is the equivalent reactance parameter at t2.

[0128] When the sampling frequency and the variation frequency of the load parameter are very different, that is, a large number of sampling data can be obtained within one waveform of the load parameter variation, it is generally considered that the parameter is stable within the sampling interval time. That is, during t1 to t2, the R and L of the load equivalent impedance parameter are stable and unchanged, and the above equation can be linearized.

[0129] Therefore, formula (12) can be modified as:

[0130]

[0131] Transform formula (13) into a matrix form:

[0132]

[0133] Observe formula (14) to obtain the expressions of R and L, as follows:

[0134]

[0135] Generalize formula (15) to each sampling interval of the entire state, and the solution of the load equivalent impedance parameter can be realized.

[0136] 4 Load current prediction

[0137] Change the matrix processed by formula (14) to obtain the following formula:

[0138]

[0139] In the formula, i(t K ) is the predicted current value at t K , u(t K ) is the actual voltage value at t K , R(t K ) and L(tK ) are the instantaneous equivalent resistance, reactance parameters, i(t K K+1 ) are the instantaneous predicted current values, u(t K+1 K+1 ) are the instantaneous actual voltage values. K

[0140] Based on equation (16), the present application can obtain the harmonic source current by combining the equivalent impedance parameters and voltage data of the load.

[0141] 5 Building the harmonic source load model

[0142] The harmonic source load model contains three modules: data acquisition module, state identification module, and impedance module. The data acquisition module obtains the voltage and current sampling data at the load port through the synchronized phasor measurement device. The state identification module can perform VMD decomposition on the load current state to extract the energy and energy entropy of the load at each time, which are used as the feature data for DBSCAN clustering. The clustering results represent the working state of the load. The details are shown in sections 2.1 and 2.2 above. The impedance module saves the equivalent impedance parameters of the harmonic source load under different working states for harmonic source load current prediction. The harmonic source load model is shown in FIG. 2. Figure 5

[0143] The specific steps for building the harmonic source load model can be summarized as follows:

[0144] 1) Use the synchronized phasor measurement device to collect the voltage and current data at the harmonic source load port, and record the sampling frequency.

[0145] 2) Identify the working state of the harmonic source load according to the method provided in section 2.

[0146] 3) Select 10 cycles of voltage and current data in each working state, calculate the equivalent impedance parameters of the harmonic source load under each working state, and store them in the impedance module. The data acquisition module, state identification module, and impedance module form the harmonic source load model.

[0147] Those skilled in the art will appreciate that embodiments of the present application can be provided as methods, systems, or computer program products. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage media, etc.) having computer-usable program code embodied in the medium.

[0148] ​​​​The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 one or more flow or blocks

[0149] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 one or more flow or blocks

[0150] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 one or more flow or blocks

[0151] It should be noted that, unless otherwise defined, technical terms or scientific terms used in the present application shall have the common meaning understood by one of ordinary skill in the art to which the present application pertains. The terms "first", "second", and similar terms used in the present application do not denote any order, quantity, or importance, but are used to distinguish different components. The terms "include", "contain", and similar terms mean that the elements or objects before the terms encompass the elements or objects listed after the terms and their equivalents, and do not exclude other elements or objects. The terms "connected" or "coupled" and similar terms do not limit to physical or mechanical connections or couplings, but can include electrical connections or couplings, whether direct or indirect. The terms "upper", "lower", "left", "right", and the like are used only to indicate relative positional relationships when the absolute positions of the described objects are changed, and the relative positional relationships can also be changed accordingly.

[0152] It should be noted that the technical terms or scientific terms used in the present application should be the general meanings understood by those skilled in the art unless otherwise defined, unless otherwise defined. The "first", "second" and similar words used in the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Include" or "contain" and similar words mean that the elements or objects before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connected" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to represent the relative positional relationship, when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0153] The present patent is not limited to the above best mode, and anyone can derive other various forms of a harmonic source modeling method based on synchronous phasor measurement data under the inspiration of the present patent. Any equivalent changes and modifications made within the scope of the patent application of the present application shall be within the scope of the present patent.

Claims

1. A harmonic source modeling method based on synchronous phasor measurement data, characterized in that: Includes the following steps: Step S1: Use a synchronous phasor measurement unit (PMU) to acquire voltage and current sampling data at the load port; Step S2: Perform VMD decomposition on the sampled current data of the harmonic source load to obtain the energy and energy entropy of the harmonic source current; In step S2, a VMD-constrained variational model is constructed, assuming that the harmonic source current is decomposed into several intrinsic mode components, expressed as: In the formula, i k (t) represents the harmonic source current I. IMFk (t) The amplitude of the envelope; It is the instantaneous phase; Performing a Hilbert transform on each decomposed modal component yields the estimated bandwidth expression for each mode: In the formula, I IMFk (t) represents the intrinsic mode components decomposed from the harmonic source current; ω k (t) represents each I IMFk The center frequency of (t); Equation (2) should satisfy the constraint that each decomposed mode has the minimum estimated bandwidth and that the superposition of each mode component is the original harmonic source current: In the above formula, I(t) is the original harmonic source load current; Introducing the augmented Lagrange function solution equation (3), the solution function is as follows: In the formula, α is the penalty particle and λ is the Lagrange multiplier; To obtain the optimal solution of equation (4), the alternating direction multiplier method is adopted, and the iterative parameters are continuously updated. The optimal value is found through an iterative process as follows: In the formula, n is the number of iterations, and ω is the frequency. These are the eigenmode components decomposed from the harmonic source current in the nth iteration. For each The center frequency, For the Lagrange multipliers in the nth iteration, for The Fourier transform result, where τ represents the iteration step size (τ > 0), and ε is the residual threshold. The output result is given when the residual is less than ε (ε > 0). Energy and energy entropy are introduced as characteristics of harmonic source loads; the expressions for energy and energy entropy are as follows: E i =∫I IMFk (t)dt (9) In the formula, E i H is the energy of the i-th modal component. EN This represents the energy entropy of the load current of the harmonic source. Step S3: Use DBSCAN to identify the operating status of the harmonic source load at each time. In step S3, two parameters are introduced during the identification and solution process: Eps and MinPts. Eps is the distance threshold between sample points, and MinPts is the minimum number of clustered samples. The data is divided into three types: core data, boundary data, and noisy data. The principle is that the number of other data points in the neighborhood of a data point is equal to the number of other data points in the neighborhood of Eps. When the number of points in the neighborhood is greater than MinPts, it is considered core data. When the number of points in the neighborhood is less than MinPts but more than 0, it is considered boundary data; when there is no other data in the neighborhood, it is considered noisy data; the cluster formed by the core data is regarded as a working state. During the load condition identification process: By arbitrarily selecting feature data of an unidentified state, the number of other data in its EPS neighborhood is calculated to determine its data type; The working status of each segment of current data is output once the status of all feature data has been identified. Step S4: Combine the operating status of the harmonic source load at each moment obtained in Step S3, and solve for the equivalent impedance parameters of the harmonic source load under different operating conditions. In step S4: The voltage and current measurements at the load port are represented by u(t) and i(t); the time-varying equivalent impedance parameters of the load are represented by R(t) and L(t). t1 and t2 are two consecutive sampling points. It is assumed that the load equivalent impedance parameters R and L remain stable during the process from t1 to t2. u (t1) represents the voltage sampling data of the load port at instant t1, i(t1) represents the current sampling data of the load port at instant t1, R(t1) represents the equivalent resistance parameter at instant t1, L(t1) represents the equivalent reactance parameter at instant t1, u(t2) represents the voltage sampling data of the load port at instant t2, i(t2) represents the current sampling data of the load port at instant t2, R(t2) represents the equivalent resistance parameter at instant t2, and L(t2) represents the equivalent reactance parameter at instant t2. For each sampling interval, the equivalent impedance parameter of the load is solved using the following formula: The harmonic source current can be obtained by using the following formula, combined with the load equivalent impedance parameter and voltage data: In the formula, i(t) K ) for t K The instantaneous predicted current value, u(t) K ) for t K The instantaneous actual voltage value, R(t) K ) and L(t K ) are respectively t K Instantaneous equivalent resistance and reactance parameters, i(t) K+1 ) for t K+1 The instantaneous predicted current value, u(t) K+1 ) for t K The instantaneous actual voltage value; Step S5: Estimate the harmonic source current using the equivalent impedance parameters of the load under various operating conditions of the voltage and harmonic source load; thereby establishing the harmonic source load model.

2. The harmonic source modeling method based on synchronous phasor measurement data according to claim 1, characterized in that: The harmonic source load model includes: a data acquisition module, a state identification module, and an impedance module; The data acquisition module acquires voltage and current sampling data at the load port through a synchronous phasor measurement device; the state identification module performs VMD decomposition on the load current state, extracts the energy and energy entropy of the load at each moment, and uses them as feature data for DBSCAN clustering, with the clustering results representing the working state of the load; the impedance module saves the equivalent impedance parameters of the harmonic source load under different working states for harmonic source load current prediction.

3. The harmonic source modeling method based on synchronous phasor measurement data according to claim 2, characterized in that: The specific steps for establishing a harmonic source load model are as follows: First, the voltage and current data at the load port of the harmonic source are collected using a synchronous phasor measurement device, and the sampling frequency is recorded. Then identify the operating status of the harmonic source load; Finally, select voltage and current data for several cycles in each operating state, calculate the equivalent impedance parameters of the harmonic source load in each operating state, store them in the impedance module, and form a harmonic source load model together with the data acquisition module and the state identification module.

Images