Broadband oscillation suppression method and system, electronic equipment and storage medium

Abstract

The invention discloses a broadband oscillation suppression method and system, electronic equipment and a storage medium, and relates to the technical field of power systems. The method comprises the following steps: firstly, acquiring a three-phase current data set at an outlet branch of a grid-side converter of the following-grid-constructing hybrid fan, then performing parameter identification on the three-phase current data set by using a matrix beam-LM method, further determining a current component in a dominant oscillation mode, transmitting the current component to a virtual impedance control link to obtain a virtual voltage drop, and finally performing impedance control on the three-phase current data set. The equivalent damping of the grid-side converter is increased, the suppression of broadband oscillation is realized, and the stability of the system is improved at the same time.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular, to a broadband oscillation suppression method, system, electronic device, and storage medium. Background Art

[0002] In view of the fact that the hybrid operation of grid-connected converters and grid-forming converters in new power systems has gradually become normal, resulting in the increasingly serious problem of broadband oscillation. Existing research shows that the oscillation frequency range of broadband oscillation is relatively wide, and its generation mechanism is also relatively complex. Its existence seriously threatens the safety and stability of new power systems. Therefore, it is necessary to provide a broadband oscillation suppression method for grid-connected and grid-forming hybrid power systems. Summary of the Invention

[0003] The purpose of the present invention is to provide a broadband oscillation suppression method, system, electronic device, and storage medium, aiming to solve or improve at least one of the above technical problems.

[0004] To achieve the above purpose, the present invention provides the following solutions:

[0005] A broadband oscillation suppression method includes:

[0006] Obtaining a three-phase current data set at the outlet branch of the grid-side converter of a grid-connected and grid-forming hybrid wind turbine;

[0007] Using the matrix pencil-LM method to perform parameter identification on the three-phase current data set to determine the current components in the dominant oscillation mode; the matrix pencil-LM method includes the matrix pencil algorithm and the LM algorithm;

[0008] Sending the current components in the dominant oscillation mode to a virtual impedance control link to obtain a virtual voltage drop, increasing the equivalent damping of the grid-side converter, and realizing the suppression of broadband oscillation.

[0009] Optionally, the processing steps of the matrix pencil algorithm include:

[0010] Selecting the pencil parameter L and making the pencil parameter L satisfy the condition: n ≤ L ≤ M - n; where M is the number of sampling values and n is the number of expected eigenvalues;

[0011] Constructing a Hankel matrix [Y] and performing singular value decomposition to obtain: [Y] = [U][S][V] T ; where [U] and [V] both represent unitary matrices. The unitary matrix [U] contains the eigenvectors of the matrix [Y][Y] T The unitary matrix [V] contains the eigenvectors of the matrix [Y] T [Y], and [S] represents a diagonal matrix of the same order as [Y];

[0012] Construct the singular value diagonal matrices [V1] and [V2] to obtain:

[0013]

[0014] where v i is the i-th right singular eigenvector of matrix [V];

[0015] Construct the Hankel matrices [Y1] and [Y2] to obtain:

[0016]

[0017] Find the eigenvalues of the matrix {[Y1]; [Y2]} to obtain the poles z i ;

[0018] Based on the poles z i Solve for the system eigenvalues λ i , expressed as:

[0019]

[0020] where Δt represents the sampling time interval;

[0021] Solve for the residual vector B j , and let all j satisfy: Expressed as:

[0022]

[0023] Perform simplification to obtain the formula ZB = Y. Use the least squares method to solve the formula ZB = Y, and solve for the waveform according to the formula to complete parameter identification.

[0024] Optionally, the processing steps of the LM algorithm include:

[0025] For the matrix pencil algorithm, give the initial value x0 and the initial optimization radius μ during iteration;

[0026] When performing the j-th iteration, superimpose the trust region on the basis of the Gauss-Newton method and solve the expression of the nonlinear least squares problem :

[0027] where the matrix J is expressed as the Jacobian matrix;

[0028] Calculate the quantization index ρ to quantify the approximation degree. The expression of the quantization index ρ is:

[0029]

[0030] If the ρ value is greater than 0.75, it is regarded as approximately feasible. Let μ = 2μ, and let x j+1 = x j + Δx j ; if the ρ value is less than or equal to 0.75, then let μ = 0.5μ;

[0031] Judge whether it converges according to the set convergence condition. If it converges, end the algorithm process; if it does not converge, return for the next iteration operation.

[0032] The present invention also provides a broadband oscillation suppression system, including:

[0033] A data acquisition unit, configured to acquire a three-phase current data set at the outlet branch of the grid-side converter of the grid-connected and network-forming hybrid wind turbine;

[0034] A parameter identification unit, configured to perform parameter identification on the three-phase current data set by using the matrix pencil-LM method to determine the current component in the dominant oscillation mode; the matrix pencil-LM method includes a matrix pencil algorithm and an LM algorithm;

[0035] A broadband oscillation suppression unit, configured to deliver the current component in the dominant oscillation mode to a virtual impedance control link to obtain a virtual voltage drop, increase the equivalent damping of the grid-side converter, and achieve the suppression of broadband oscillation.

[0036] The present invention also provides an electronic device, including a memory and a processor. The memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the broadband oscillation suppression method according to the above.

[0037] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the broadband oscillation suppression method as described above is implemented.

[0038] According to the specific embodiments provided by the present invention, the following technical effects are disclosed by the present invention:

[0039] The present invention discloses a broadband oscillation suppression method, system, electronic device and storage medium. The method includes first acquiring a three-phase current data set at the outlet branch of the grid-side converter of the grid-connected and network-forming hybrid wind turbine, then performing parameter identification on it by using the matrix pencil-LM method, and further determining the current component in the dominant oscillation mode, delivering it to the virtual impedance control link to obtain a virtual voltage drop, increasing the equivalent damping of the grid-side converter, achieving the suppression of broadband oscillation, and improving the stability of the system at the same time. Description of the Drawings

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0041] Figure 1 It is the oscillating current amplitude diagram in each mode of this embodiment;

[0042] Figure 2 It is the mode energy amplitude diagram before and after adding virtual impedance in this embodiment;

[0043] Figure 3 It is the comparison diagram of line active power before and after adding virtual impedance in this embodiment;

[0044] Figure 4 It is the output voltage spectrum diagram of the converter in this embodiment; among them, part (a) is a schematic diagram without adding virtual impedance; part (b) is a schematic diagram with adding virtual impedance;

[0045] Figure 5 It is the flow schematic diagram of the broadband oscillation suppression method in this embodiment. Specific implementation manners

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0047] The purpose of the present invention is to provide a broadband oscillation suppression method, system, electronic device and storage medium, aiming to solve or improve at least one of the above technical problems.

[0048] To make the above objects, features and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the drawings and specific implementation manners.

[0049] Such as Figures 1 - 5As shown in the figure, the present invention provides a wide - frequency oscillation suppression method, including: First, based on the mode energy idea and virtual impedance theory, construct a wide - frequency oscillation suppression model based on dominant oscillation energy and virtual impedance; Second, use the constructed wide - frequency oscillation suppression model to propose a wide - frequency oscillation suppression method based on dominant oscillation energy and virtual impedance theory: By obtaining the three - phase current data set at the outlet branch of the grid - connected and grid - forming hybrid wind turbine grid - side converter, use the matrix pencil - LM method to identify its parameters, determine the current components under the dominant oscillation mode, and send them to the virtual impedance control link to obtain the corresponding virtual voltage drop, improve the equivalent damping level of the grid - side converter, and theoretically achieve the suppression of wide - frequency oscillation; Third, based on the constructed model and the proposed method, verify in the electromagnetic transient simulation example of a grid - connected and grid - forming hybrid wind turbine connected to an infinite - bus system, and finally prove the effectiveness of the proposed method.

[0050] As a specific implementation manner, the following specific processing steps are provided as shown below.

[0051] 1. Construction of a wide - frequency oscillation suppression model based on dominant oscillation energy and virtual impedance theory

[0052] (1) Virtual impedance theory

[0053] With the normalization trend of the "dual - high" power system, the dynamic characteristics of the power system have changed significantly. The emergence of distributed power sources has gradually replaced the original synchronous machines, but this trend has also brought new problems to the new - type power system, and wide - frequency oscillation is one of the typical problems.

[0054] In the new - type power system, virtual impedance technology has been widely applied to the control of VSG. The principle of this technology lies in adding a virtual impedance link to the control system to equivalently simulate the impedance characteristics of the synchronous machine and improve the dynamic characteristics of the original system. In existing research, virtual impedance is usually used to adjust the output impedance characteristics of the inverter and is applied before the voltage or current control loop. Obviously, reasonable input of the virtual impedance parameters can effectively suppress the wide - frequency oscillation phenomenon.

[0055] If this control method of virtual impedance is to be realized, it can be carried out in two ways: series and parallel. The principle of the former lies in adjusting the output impedance of the inverter to effectively improve the damping characteristics of the system; the principle of the latter lies in adjusting the equivalent impedance of the system to effectively improve the voltage stability of the system. Therefore, the design of this link needs to comprehensively consider the actual situation of the system, starting from different operating conditions and control objectives, to ensure that wide - frequency oscillation can be effectively suppressed under different conditions. At present, in VSG control, virtual impedance technology mainly suppresses wide - frequency oscillation in three ways: enhancing damping characteristics, optimizing impedance matching, and adjusting dynamic characteristics.

[0056] (2) Construction of broadband oscillation suppression model

[0057] Existing research has shown that in order to quickly and effectively suppress oscillation problems in power systems, a damping control link is usually added, and virtual impedance is generally considered to be one of the typical means to increase the damping control link. Currently, it has been gradually applied to suppress broadband oscillation phenomena. The principle of this technology is to equivalent its specific impedance characteristics at the output end of power electronic devices through a control algorithm, changing the dynamic response characteristics of the system from the mechanism to achieve the purpose of suppressing oscillation. Compared with traditional damping methods, the advantage of virtual impedance technology lies in its higher flexibility and adaptability. It can adjust the impedance parameters in real time as the system operating state changes, and is very suitable for complex and changeable broadband oscillation scenarios. In addition, to construct a broadband oscillation suppression model based on virtual impedance, it can be combined with the existing control strategy of the system to suppress broadband oscillation, thereby improving the stability of the system. Taking the grid-side converter as an example, the virtual impedance control link is applied before its voltage-current double closed-loop control link.

[0058] 2. Broadband oscillation suppression method based on virtual impedance theory and dominant oscillation energy

[0059] This part intends to propose a broadband oscillation suppression strategy based on virtual impedance theory and dominant oscillation energy by combining the virtual impedance theory in the previous text with the mode energy idea. This part first obtains the three-phase current data set at the outlet branch of the grid-side converter of the grid-connected and structure-connected hybrid wind turbine, and then uses the matrix pencil-LM method to identify its parameters, and then determines the current components under the dominant oscillation mode, and transports them to the virtual impedance control link to obtain a virtual voltage drop, increasing the equivalent damping of the grid-side converter, realizing the suppression of broadband oscillation, and improving the stability of the system at the same time.

[0060] In this method, the tuning of virtual impedance parameters is the key step in the entire virtual impedance control strategy, and its existence directly determines the dynamic characteristics and stability trend of the system.

[0061] (1) Matrix pencil algorithm

[0062] The matrix pencil algorithm is a modal parameter estimation method proposed in the last century and has been relatively maturely applied in industry. Existing research has shown that when there is noise in the signal to be identified, the matrix pencil algorithm can be used to identify its parameters. The principle is to construct a certain function expression and use this function expression to fit the function expression of the given signal. The fitting function expression is shown in the following formula:

[0063]

[0064] Set N sampling values and perform equidistant sampling with a sampling time interval of Δt. The finally fitted signal is set as y(t), and the expression of y(t) is shown as follows:

[0065] y(t k ) = y(k), k = 0, 1, …, N - 1 (2)

[0066] Express Equation (1) in the form of complex exponential, as shown below:

[0067]

[0068] Further simplify Equation (3) to obtain the following equation:

[0069]

[0070] After that, to solve for z i , a matrix pencil can be constructed, and the matrix pencil expression is shown as follows:

[0071] [Y2] - λ[Y1] = [Z1][B]{[Z0] - λ[I]}[Z2] (5)

[0072] In Equation (5), the expressions of each component element are shown as follows:

[0073]

[0074] In Equation (6), [B] represents the residual matrix, [I] represents the n-order identity matrix, n represents the number of expected eigenvalues, L represents the pencil parameter, and the relationship between L, the number of sampling values M, and the number of expected eigenvalues n is shown as Equation (7) below:

[0075] n ≤ L ≤ M - n (7)

[0076] The implementation steps of the matrix pencil algorithm are as follows:

[0077] First, select an appropriate pencil parameter L and make it satisfy Equation (7);

[0078] Second, construct the Hankel matrix [Y] and perform singular value decomposition on it to obtain the following equation:

[0079] [Y] = [U][S][V] T (8)

[0080] In Equation (8), both [U] and [V] represent unitary matrices. The former contains the eigenvectors of the matrix [Y][Y] T , and the latter contains the eigenvectors of the matrix [Y] T [Y]. [S] represents a diagonal matrix of the same order as [Y], and is denoted by σ iDenote the i-th singular value.

[0081] Third, the expressions of the finally constructed singular value diagonal matrices [V1] and [V2] are shown as follows:

[0082]

[0083] In Equation (9), v i is the i-th right singular eigenvector of matrix [V], where the right singular eigenvector is the right eigenvector of the singular value diagonal matrix.

[0084] Fourth, construct Hankel matrices [Y1] and [Y2], and their expressions are shown as follows:

[0085]

[0086] Fifth, obtain the eigenvalues of the matrix {[Y1]; [Y2]}, and then the poles z i ;

[0087] The system eigenvalues λ i are shown as follows:

[0088]

[0089] When the solution of the poles z i is completed, the system eigenvalues λ i can be obtained;

[0090] Sixth, solve for B j , and ensure that for all j subscripts, the following equation holds:

[0091]

[0092] From Equation (4), the following equation can be obtained:

[0093]

[0094] Further simplify Equation (13), and the following equation can be obtained:

[0095] ZB = Y(14)

[0096] Seventh, use the least squares method to solve Equation (14), and solve for the estimated waveform according to Equation (3), then the entire identification process can be completed.

[0097] (2) LM algorithm

[0098] The LM algorithm belongs to one of the optimization algorithms. This algorithm has been relatively maturely applied in many fields. The

[0099] The basic principle is to transform the optimization problem from an unconstrained least - squares problem into a constrained least - squares problem.

[0100] Generally, the expression of the non - linear least - squares problem is as follows:

[0101] minf(x0+Δx)=||f(x0+Δx)|| 2 (15)

[0102] In formula (15), y i represents the measured value of the system output at time t i ; x represents the vector composed of identification parameters such as amplitude, phase, angular frequency, and damping coefficient in formula (1), which depends on the eigenvalues of the system state matrix.

[0103] If we want to solve formula (15), we can expand the function f(x) near x0 using the Taylor series, and we can get the following formula:

[0104]

[0105] In formula (16), the matrix J represents the Jacobian matrix. Then, we take the derivative of formula (16) and set its first - order derivative to 0, and we can get its extreme value as follows:

[0106] JJ T Δx=-Jf(x0)(17)

[0107] To ensure that JJ T is invertible, we introduce the identity matrix:

[0108] H≈J T J+μI(18)

[0109] By updating the iteration formula, we can get the following formula:

[0110] x j+1 =x j -(J T J+μI) -1 g j (19)

[0111] In formula (19), g j represents the negative gradient direction.

[0112] The whole algorithm process is as follows:

[0113] First, given the initial value x0 and the initial optimization radius μ. Since the initial value x0 directly determines whether the algorithm converges, it is necessary to choose it carefully;

[0114] Second, when performing the j-th iteration, a trust region is superimposed on the basis of the Gauss-Newton method and solved:

[0115]

[0116] Third, calculate the quantization index ρ to quantify the approximation degree. The expression of the quantization index ρ is shown as follows:

[0117]

[0118] If the value of ρ is greater than 0.75, then let μ = 2μ; otherwise, let μ = 0.5μ.

[0119] Fourth, if ρ is greater than a certain threshold, it can be regarded as approximately feasible, and the following formula can be set:

[0120] x j+1 = x j + Δx j (22)

[0121] Fifth, judge whether it converges. If it converges, end the algorithm process; otherwise, return to the second step.

[0122] As a specific embodiment, the content is as follows.

[0123] To verify the effectiveness of the method proposed in the present invention, in this part, a grid-connected and network-forming hybrid wind turbine is connected to an infinite bus system, and this electromagnetic transient simulation example is used for verification. To simulate long-distance power transmission to the greatest extent, when the time t = 3 s, the length of the transmission line is increased to excite broadband oscillations, and parameter identification is performed on the amplitude and frequency of the measured current. The final identification results are shown in Table 1.

[0124] Table 1 Oscillating current identification results

[0125]

[0126] Among them, the amplitude of the current is as Figure 1 shown. It can be easily known from Figure 1 that there are a total of 7 oscillation modes, which are consistent with the 7 oscillation modes in Table 1. Obviously, mode 5 is the dominant oscillation mode. Then, a virtual resistance of 0.04 p.u. and a virtual reactance of 0.2 p.u. are taken, and the current component corresponding to mode 5 is sent to the virtual impedance control link. The system is set with the same disturbance again to excite broadband oscillations again, and the mode energy amplitude under this condition is compared with the mode energy before the virtual impedance control is added. The comparison effect is as Figure 2 shown.

[0127] From Figure 2It is easy to know that after introducing the virtual impedance control link, the modal energies of the 7 oscillation modes all decrease significantly. Among them, the decrease amplitude of mode 5 is the most significant. To ensure the rigor of the proposed method, the active power of the line before and after introducing the virtual impedance control link is compared, and the comparison effect is as Figure 3 shown.

[0128] It can be easily known from Figure 3 that after introducing the virtual impedance control link, the waveform of the active power curve of the line changes from the original equal-amplitude oscillation to damped oscillation. Obviously, the existence of the virtual impedance control link effectively suppresses the broadband oscillation of the system. Then, the Fourier analysis is carried out on the output voltage of the grid-side converter before and after introducing the virtual impedance control link, and the obtained spectrogram is as Figure 4 shown. It can be easily known from Figure 4 that after introducing the virtual impedance control link, only the fundamental frequency component exists, and even the oscillation component does not exist. Obviously, the original broadband oscillation in the system has been effectively suppressed at this time.

[0129] In summary, the method proposed in the present invention can effectively suppress broadband oscillation.

[0130] In this specification, each embodiment is described in a progressive manner. The key points of each embodiment are the differences from other embodiments. The same and similar parts among the embodiments can be referred to each other.

[0131] In this article, specific examples are used to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the core idea of the present invention; at the same time, for those of ordinary skill in the art, based on the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A broadband oscillation suppression method, characterized in that Including: Obtain a three-phase current data set at the outlet branch of the grid-connected and network-forming hybrid fan's grid-side converter; Use the matrix pencil-LM method to perform parameter identification on the three-phase current data set to determine the current components in the dominant oscillation mode; the matrix pencil-LM method includes the matrix pencil algorithm and the LM algorithm; Send the current components in the dominant oscillation mode to the virtual impedance control link to obtain a virtual voltage drop, increase the equivalent damping of the grid-side converter, and achieve the suppression of broadband oscillation.

2. The broadband oscillation suppression method according to claim 1, wherein The processing steps of the matrix pencil algorithm include: Select the pencil parameter L, and make the pencil parameter L satisfy the condition: n ≤ L ≤ M - n; where M is the number of sampling values and n is the number of expected eigenvalues; Construct the Hankel matrix [Y] and perform singular value decomposition to obtain: [Y] = [U][S][V] T ; where both [U] and [V] represent unitary matrices. The unitary matrix [U] contains the eigenvectors of the matrix [Y][Y] T , and the unitary matrix [V] contains the eigenvectors of the matrix [Y] T [Y], and [S] represents a diagonal matrix of the same order as [Y]; Construct the singular value diagonal matrices [V1] and [V2] to obtain: where, v i is the i-th right singular eigenvector of the matrix [V]; Construct the Hankel matrices [Y1] and [Y2] to obtain: Find the eigenvalues of the matrix {[Y1]; [Y2]} to obtain the poles z i ; Based on the pole z i Solve for the system eigenvalue λ i , expressed as: where Δt represents the sampling time interval; Solve the residual vector B j , and let all j satisfy:,, j = 0, 1,..., M - 1, expressed as: Simplify the processing to obtain the formula ZB = Y. Use the least squares method to solve the formula ZB = Y, and according to the formula Solve the waveform to complete the parameter identification.

3. The broadband oscillation suppression method according to claim 1, wherein, The processing steps of the LM algorithm include: For the matrix pencil algorithm, give the initial value x0 and the initial optimization radius μ during iteration; When the j-th iteration is performed, a trust region is superimposed on the basis of the Gauss-Newton method, and the expression of the nonlinear least squares problem is solved: where the matrix J is expressed as the Jacobian matrix; Calculate the quantization index ρ to quantify the approximation degree, and the expression of the quantization index ρ is: If the ρ value is greater than 0.75, it is regarded as approximately feasible. Let μ = 2μ, and let x j+1 = x j + Δx j ; if the ρ value is less than or equal to 0.75, then let μ = 0.5μ; Judge whether to converge according to the set convergence condition. If it converges, end the algorithm process; if it does not converge, return for the next iteration operation.

4. A wideband oscillation suppression system, characterized in that, Including: A data acquisition unit for obtaining a three-phase current data set at the outlet branch of the grid-connected and network-forming hybrid fan's grid-side converter; ​ ​ ​ 5. An electronic device, characterized in that, ​ 6. A computer-readable storage medium, characterized in that, ​

Images