CN116826735A - Broadband oscillation identification method and device for new energy station - Google Patents
Broadband oscillation identification method and device for new energy station Download PDFInfo
- Publication number
- CN116826735A CN116826735A CN202310825537.0A CN202310825537A CN116826735A CN 116826735 A CN116826735 A CN 116826735A CN 202310825537 A CN202310825537 A CN 202310825537A CN 116826735 A CN116826735 A CN 116826735A
- Authority
- CN
- China
- Prior art keywords
- broadband oscillation
- modal
- broadband
- denoising
- oscillation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000010355 oscillation Effects 0.000 title claims abstract description 138
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 51
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 30
- 238000005457 optimization Methods 0.000 claims abstract description 17
- 239000002245 particle Substances 0.000 claims description 11
- 230000009466 transformation Effects 0.000 claims description 6
- 238000003064 k means clustering Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 12
- 238000004364 calculation method Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 230000000694 effects Effects 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013016 damping Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/002—Flicker reduction, e.g. compensation of flicker introduced by non-linear load
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/24—Arrangements for preventing or reducing oscillations of power in networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/10—Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/28—The renewable source being wind energy
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Physics & Mathematics (AREA)
- Nonlinear Science (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a method and a device for identifying broadband oscillation of a new energy station, wherein the method for identifying broadband oscillation of the new energy station comprises the steps of obtaining broadband oscillation voltage signals of a wind turbine generator set side and a boosting low-voltage side; aiming at broadband oscillation voltage signals at two sides, respectively adopting an optimization algorithm to optimize parameters in variation modal decomposition to obtain the optimal values of the modal number and the penalty factors; according to the obtained optimal values of the modal number and the penalty factors, adopting variation modal decomposition to obtain a group of optimal time domain modal components, and carrying out denoising reconstruction based on a clustering method to obtain a denoising broadband oscillation voltage signal; extracting characteristic parameters of each mode of broadband oscillation from the denoising broadband oscillation voltage signal; and identifying the characteristic parameters of each mode of broadband oscillation extracted according to the two-side denoising broadband oscillation voltage signals to obtain an identification result. The invention can effectively identify whether broadband oscillation occurs.
Description
Technical Field
The invention relates to a broadband oscillation identification method and device for a new energy station, and belongs to the field of power system automation.
Background
With the mass access of the high-proportion renewable energy sources and the high-proportion power electronic equipment, the development trend of the double high is increasingly remarkable, so that the power signal has the characteristics of wide frequency band, quick change and high noise. In order to measure the frequency of the harmonic wave and the inter-harmonic wave of the distribution uncertainty in a refined way, various broadband measurement algorithms and devices are presented. The novel large-scale wind power centralized access system has the characteristics of complex influence factors and wide-range time variation, nonlinear continuous oscillation is finally shown, aliasing possibly occurring in each mode of multimode broadband electromagnetic oscillation causes inaccuracy in broadband measurement, the oscillation frequency of the source double-fed wind power plant rises by 2-4 Hz along with time, remarkable voltage fluctuation (about 4% of amplitude change in 0.5 s) occurs in the wind power plant in the oscillation process, and in addition, a large amount of field wave recording data also show that the amplitude and frequency of harmonic wave and inter-harmonic wave components have time-varying behaviors (3% -10% of the amplitude change occurs in seconds, and the frequency oscillates within 1Hz in a plurality of periods even seconds), and the novel challenge is provided for a broadband measurement algorithm because the dynamic change process of the broadband electric quantity of the new energy power system is faster.
Disclosure of Invention
The invention provides a method and a system for identifying broadband oscillation of a new energy station, which are used for realizing the identification of the broadband oscillation of the new energy station by combining variation modal decomposition and a Prony algorithm.
The technical scheme of the invention is as follows:
according to an aspect of the present invention, there is provided a broadband oscillation identification method for a new energy station, including: broadband oscillation voltage signals of a wind turbine generator side and a boosting low-voltage side are obtained; aiming at broadband oscillation voltage signals at two sides, respectively adopting an optimization algorithm to optimize parameters in variation modal decomposition to obtain the optimal values of the modal number and the penalty factors; according to the obtained optimal values of the modal number and the penalty factors, adopting variation modal decomposition to obtain a group of optimal time domain modal components, and carrying out denoising reconstruction based on a clustering method to obtain a denoising broadband oscillation voltage signal; extracting characteristic parameters of each mode of broadband oscillation from the denoising broadband oscillation voltage signal; and identifying the characteristic parameters of each mode of broadband oscillation extracted according to the two-side denoising broadband oscillation voltage signals to obtain an identification result.
The optimizing algorithm is adopted to optimize parameters in variation modal decomposition, and the optimal values of the modal number and the penalty factor are obtained, and the optimizing method comprises the following steps: setting an initial value of a parameter in a particle swarm optimization algorithm; setting a change interval of the number of modes and penalty factors; decomposing the broadband oscillation voltage signal through variation modal decomposition to obtain decomposed time domain modal components; calculating fitness function envelope entropy according to the time domain modal components; determining the overall optimal solution of the number of modes and the penalty factor according to the fitness function envelope entropy; repeating until the termination condition is met, and taking the global history optimal solution after iteration is finished as the optimal value of the modal number and the penalty factor of the particle swarm optimization algorithm.
Denoising reconstruction is carried out based on a clustering method to obtain a denoising broadband oscillation voltage signal, and the denoising method comprises the following steps: obtaining the amplitude-frequency characteristic of each optimal time domain modal component according to Fourier transformation, and extracting the effective frequency component of each optimal time domain modal component; for each optimal time-domain modal component, determining a frequency component with an amplitude greater than or equal to 10% of the maximum amplitude as an effective frequency component of each time-domain modal component; taking the amplitude value of the effective frequency component of each optimal time domain modal component as initial data, and carrying out K-means clustering on the amplitude value of the effective frequency component according to the Euclidean distance to obtain a clustering center of each optimal time domain modal component, wherein K clustering centers are taken as first clustering centers; clustering the K first clustering centers of each optimal time domain modal component again according to the Euclidean distance to obtain a second clustering center; comparing Euclidean distances after the weight coefficients of the obtained K first clustering centers and the obtained K second clustering centers are updated with a set threshold value: if the updated Euclidean distance is smaller than or equal to the threshold value, judging the Euclidean distance as an effective modal component, and if the updated Euclidean distance is larger than the threshold value, judging the Euclidean distance as an ineffective modal component; and reconstructing the effective modal component to obtain the denoising broadband oscillation voltage signal.
The Euclidean distance after the weight coefficient of the obtained K first clustering centers and the weight coefficient of the obtained second clustering centers are updated is expressed as follows:
D k ′＝D k /ρ k
wherein D is k ' is D k The updated value of the first cluster center and the second cluster center represents the Euclidean distance after the weight coefficient is updated; d (D) k For the kth first cluster center and the second cluster center x 0 A Euclidean distance; ρ k And the weight coefficient of the kth first cluster center.
The weight coefficient has the following expression:
wherein x is k The K is the clustering center of the K optimal time domain modal components, namely the K first clustering center, and K is the number of the optimal time domain modal components.
The characteristic parameters of each mode of broadband oscillation are extracted from the denoising broadband oscillation voltage signal, and specifically the characteristic parameters are as follows: and analyzing the denoising broadband oscillation voltage signal as an input signal of the Prony algorithm, and extracting characteristic parameters of each mode of broadband oscillation.
The characteristic parameters of each mode of broadband oscillation extracted according to the two-side denoising broadband oscillation voltage signals are identified, and an identification result is obtained, specifically:
if the conditions that the absolute value of the error between the sum of the frequencies of two oscillation modes smaller than 100Hz in the signals at two sides and the frequency of the vibration modes at two sides is smaller than 1Hz and the absolute value of the error of the attenuation factor at the corresponding modes of the signals at two sides is smaller than 0.5Hz are met, the amplitude of the low-voltage side of the boosted voltage is larger than that of the wind turbine generator set side are met, the problem of broadband oscillation of the new energy station is judged, and the two analyzed oscillation modes smaller than 100Hz are dominant oscillation modes of the broadband oscillation, otherwise, the problem is not solved.
According to another aspect of the present invention, there is provided a broadband oscillation identification system for a new energy station, including: the acquisition module is used for acquiring broadband oscillation voltage signals of the wind turbine generator system side and the boosting low-voltage side; the first obtaining module is used for optimizing parameters in variation modal decomposition by adopting an optimization algorithm aiming at broadband oscillation voltage signals at two sides to obtain the optimal values of the modal number and the penalty factor; the second obtaining module is used for obtaining a group of optimal time domain modal components by adopting variation modal decomposition according to the obtained modal number and the optimal value of the penalty factor, and carrying out denoising reconstruction based on a clustering method to obtain a denoising broadband oscillation voltage signal; the extraction module is used for extracting characteristic parameters of each mode of broadband oscillation for the denoising broadband oscillation voltage signal; and the third obtaining module is used for identifying the characteristic parameters of each mode of broadband oscillation extracted by the two-side denoising broadband oscillation voltage signal to obtain an identification result.
According to another aspect of the present invention, there is provided a processor for running a program, wherein the program executes any one of the above methods for identifying broadband oscillation of a new energy station.
According to another aspect of the present invention, there is provided a computer readable storage medium, the computer readable storage medium including a stored program, wherein when the program is executed, the device in which the computer readable storage medium is located is controlled to execute any one of the above method for identifying broadband oscillation of a new energy station.
The beneficial effects of the invention are as follows: on one hand, the invention considers that a large amount of noise exists in the actual data and the prony decomposition method is extremely sensitive to the noise, so that the original data is denoised by using the variational modal decomposition before the prony decomposition, thereby eliminating dimension disasters which are easy to occur in the traditional prony algorithm and improving the identification precision; on the other hand, two important parameters of the variant modal decomposition are considered: the mode number and the penalty factor are selected to have great influence on the final decomposition effect, so that the denoising effect is influenced, and therefore, the particle swarm optimization algorithm is adopted to optimize parameters of the variation mode decomposition, so that the variation mode decomposition algorithm has self-adaptability and better denoising effect; furthermore, a K-means clustering method is introduced to screen modal components obtained through variation modal decomposition, so that a better denoising effect is achieved; meanwhile, the invention can effectively identify whether broadband oscillation occurs or not through simulation verification.
Drawings
FIG. 1 is a schematic flow chart of the present invention;
FIG. 2 is a schematic diagram of a voltage transformer installation;
FIG. 3 is a schematic diagram of an original broadband oscillation voltage signal on the 690V side of a wind turbine;
FIG. 4 is a diagram of the original broadband oscillation voltage signal at the low-voltage side of the boost converter;
FIG. 5 is a schematic diagram of a time domain of a wind turbine generator 690V-side voltage signal decomposed into modal components by a variational modal decomposition;
FIG. 6 is a schematic diagram of a frequency domain of a wind turbine generator 690V-side voltage signal decomposed into modal components by a variational modal decomposition;
FIG. 7 is a schematic diagram of a time domain of the boost low-voltage side voltage signal decomposed into modal components by a variational modal decomposition;
fig. 8 is a schematic diagram showing the frequency domain of the boost low-voltage side voltage signal decomposed into modal components by the change mode decomposition.
Detailed Description
The invention will be further described with reference to the drawings and examples, but the invention is not limited to the scope.
Example 1: as shown in fig. 1 to 8, according to an aspect of the embodiment of the present invention, there is provided a broadband oscillation identification method for a new energy station, including: broadband oscillation voltage signals of a wind turbine generator side and a boosting low-voltage side are obtained; aiming at broadband oscillation voltage signals at two sides, respectively adopting an optimization algorithm to optimize parameters in variation modal decomposition to obtain the optimal values of the modal number and the penalty factors; according to the obtained optimal values of the modal number and the penalty factors, adopting variation modal decomposition to obtain a group of optimal time domain modal components, and carrying out denoising reconstruction based on a clustering method to obtain a denoising broadband oscillation voltage signal; extracting characteristic parameters of each mode of broadband oscillation from the denoising broadband oscillation voltage signal; and identifying the characteristic parameters of each mode of broadband oscillation extracted according to the two-side denoising broadband oscillation voltage signals to obtain an identification result.
Further, the invention is given in the following alternative embodiments:
step 1: voltage transformers are respectively arranged on the 690V side and the boosting low-voltage side of the wind turbine generator; acquiring broadband oscillation voltage signals at two sides according to the installed voltage transformers;
step 2: optimizing parameters in variation modal decomposition by using a particle swarm optimization algorithm; parameters in the optimized variational modal decomposition comprise the number of modes and penalty factors; the method comprises the following specific steps:
step 2-1: setting initial values of parameters in a particle swarm optimization algorithm, such as iteration times, population scale, learning factors and noise tolerance; setting a change interval of the number of modes and penalty factors;
step 2-2: decomposing the broadband oscillation voltage signal in the step 1 through variation modal decomposition to obtain decomposed modal components; calculating fitness function envelope entropy E from modal components p The expression is as follows:
wherein N is the data size of each modal component obtained by decomposition; a (i) is an envelope signal of a modal component decomposed by a variational modal decomposition VMD after Hilbert demodulation, and K represents the total number of the decomposed modal components;
step 2-3: envelope entropy E of fitness function obtained according to step 2-2 p Determining the number of modes and a global optimal solution of penalty factors;
step 2-4: and repeating the steps 2-2 to 2-3 until the iteration times set in the step 2-1 are met, wherein the global historical optimal solution after the iteration is finished is the global optimal value of the number of modes and the penalty factor of the particle swarm optimization algorithm.
Further, in the step 2, a variation mode decomposition is adopted for the broadband oscillation voltage signal obtained in the step 1, so as to obtain decomposed mode components, which specifically include:
step 2-2-1: decomposing the broadband oscillation voltage signal obtained in the step 1 into K time domain modal components u k (t) for each time-domain modal component u k (t) performing Hilbert (Hibert) transformation to obtain a single-side frequency spectrum of each time domain modal component, wherein the solving formula of the step 2-2-1 is as follows:
[δ(t)+j/(πt)]*u k (t)
wherein, is convolution symbol, delta (t) is pulse function, satisfyingAnd->t represents a time variable in the time domain, j is a complex unit in the complex frequency domain, and pi is a circumference ratio.
Step 2-2-2: adding an index term to the single-side frequency spectrum of the time domain modal component obtained in the step 2-2-1, modulating the single-side frequency spectrum to adjust the single-side frequency spectrum to a corresponding baseband, and solving the formula in the step 2-2-2 as follows:
wherein omega k For the kth time domain modal component u k A center frequency of (t);
step 2-2-3: the constraint variation problem of the variation modal decomposition VMD can be expressed as follows, according to the gaussian smoothing calculation of the demodulation signal for each modal component bandwidth:
where min { } represents the minimum,indicating deviation of time t +.>Representing the square of the two norms of x, s.t. () representing that the constraint condition must be met, f (t) being the broadband oscillating voltage signal obtained in step 1 (representing the broadband oscillating voltage signal on the wind turbine side if the wind turbine side is calculated; representing the broadband oscillating voltage signal on the boost low-voltage side if the boost low-voltage side is calculated);
step 2-2-4: in order to ensure the accuracy of the reconstructed signal and the strictness of constraint conditions, a penalty factor alpha and a Lagrange factor lambda are introduced, the constraint variation problem of the step 2-2-3 is converted into an unconstrained variation problem, and the solving formula of the step 2-2-4 is as follows:
wherein f (t) is the voltage signal obtained in the step 1 (if the calculation is the wind turbine generator side, the voltage signal represents the broadband oscillation voltage signal of the wind turbine generator side; if the calculation is the boost low-voltage side, the voltage signal represents the broadband oscillation voltage signal of the boost low-voltage side), and lambda (t) and lambda are both Lagrangian factors;
step 2-2-5: and (2) introducing an alternate direction multiplier method to calculate a time domain modal component and a center frequency updating expression, wherein the solving formula of the step (2-2-5) is as follows:
where n represents the number of iterations,representing the ith temporal modal component u i (t) Fourier transformTransformed frequency domain modal component, < >>Representing the kth frequency-domain modal component +.>N+1th iteration value, +.>Representing the value of the broadband oscillation voltage signal f (t) obtained in step 1 after Fourier transform, < >>Represents the n+1st iteration value of λ after fourier transformation, ++>Represents ω k N+1th iteration value of (a);
step 2-2-6: carrying out updating calculation on the time domain modal components and the center frequency in the step 2-2-5 on each time domain modal component in the step 2-2-1; and updating the Lagrangian factor lambda, the solution formula of step 2-2-6 is as follows:
wherein lambda is n+1 Represents the n+1th iteration value of the lagrangian factor, τ represents the noise tolerance,representing the kth temporal modal component u k N+1th iteration value of (t);
step 2-2-7: setting iteration termination conditions, and ending the loop if the iteration termination conditions are met, so as to obtain L time domain modal components after the variation modal decomposition is optimal; otherwise, repeating the steps 2-2-5 to 2-2-6, wherein the iteration termination condition of the step 2-2-7 is as follows:
where ζ represents the set iteration termination condition.
Step 3: and (3) decomposing the optimal values of the modal number and the penalty factors obtained in the step (2) by adopting a variation mode to obtain a group of optimal time domain modal components, and then denoising and reconstructing based on a K-means clustering method, wherein the method comprises the following specific steps of:
step 3-1: obtaining the amplitude-frequency characteristic of each optimal time domain modal component according to Fourier transformation; for each optimal time-domain modal component, determining a frequency component with an amplitude greater than or equal to 10% of the maximum amplitude as an effective frequency component of each time-domain modal component;
step 3-2: taking the amplitude value of the effective frequency component of each optimal time domain modal component as initial data, and carrying out K-means clustering on the amplitude value of the effective frequency component according to the Euclidean distance to obtain a clustering center of each optimal time domain modal component, wherein K clustering centers are taken as first clustering centers; the calculation expression of the Euclidean distance adopted in the clustering process is as follows:
wherein D is kj For Euclidean distance, x, between the j-th data in the k-th optimal time domain modal component and the clustering center k Is the clustering center of the kth optimal time domain modal component, x j For the normalized value of the initial data, the expression is calculated as follows:
x i ＝u i /max(u 1 ,u 2 ,...,u n )
wherein u is i For the magnitude of the effective frequency component in each of the optimal time domain modal components, u i :＝{u 1 ,u2,…,u n }；
Step 3-3: and (3) clustering the K first clustering centers of each optimal time domain modal component obtained in the step (3-2) again according to the Euclidean distance to obtain a second clustering center. At this time, the Euclidean distance adopted in the clustering process is updated in weight coefficient based on the step 3-2, and the calculation expression is as follows:
D k ′＝D k /ρ k
wherein D is k ' is D k The updated value of the first cluster center and the second cluster center represents the Euclidean distance after the weight coefficient is updated; d (D) k The Euclidean distance between the kth first clustering center and the second clustering center; ρ k For the weight coefficient of the kth first cluster center, the expression is calculated as follows:
wherein x is k And K is the number of the optimal time domain modal components and is the clustering center of the kth optimal time domain modal component.
Step 3-4: and (3) comparing the Euclidean distance obtained by the step (3-3) after the weight coefficient update of the K first clustering centers and the second clustering centers with a set threshold value g. If the updated Euclidean distance is smaller than or equal to a threshold value g, judging the Euclidean distance as an effective modal component, and if the updated Euclidean distance is larger than the threshold value g, judging the Euclidean distance as an ineffective modal component; and reconstructing the effective modal component to obtain the denoising broadband oscillation voltage signal x (t).
Step 4: analyzing the denoising broadband oscillation voltage signal x (t) obtained in the step 3 as an input signal of a Prony algorithm, and extracting characteristic parameters of each mode of broadband oscillation; the method comprises the following specific steps:
step 4-1: converting x (t) into N equidistant sampling points to formThen constructing a constant coefficient linear differential equation, solving the linear differential equation to obtain a parameter a k ；
Wherein,,is->Matrix composed of n-k points at the middle and front, P is model order, a k The coefficients to be solved are P multiplied by 1 order matrixes;
step 4-2: solving the parameter a obtained in the step 4-1 k Carry-inRoot finding is carried out to obtain a parameter z k Is a P multiplied by 1 order matrix;
step 4-3: according to Euler's formulaSolving parameter b based on least square method k The solution formula of step 4-3 is as follows:
wherein,,representing parameter z k K times the i-th number of (a);
step 4-4: according toThe information of each mode of broadband oscillation is calculated, and the calculation formula is as follows:
wherein A is k For the amplitude, f, of each modal component k For the frequency, theta, of each modal component k For the phase sum alpha of the modal components k Is the attenuation factor of each modal component.
Step 5: analyzing and judging the identification result in the step 4, and if two signals meet the following conditions: (1) The absolute value of the error between the sum of the frequencies of the two oscillation modes smaller than 100Hz and 100Hz is smaller than 1Hz; (2) The condition that the absolute value of the frequency error of the corresponding modes of the two oscillation modes is smaller than 0.5Hz and the absolute value of the error of the attenuation factor is smaller than 0.05, and the amplitude of the boosting low-voltage side is larger than that of the wind turbine generator 690V side can be judged that the new energy station has the broadband oscillation problem, and the two oscillation modes smaller than 100Hz are the dominant oscillation modes of the broadband oscillation, otherwise, the two oscillation modes are not dominant oscillation modes of the broadband oscillation.
According to another aspect of the embodiment of the present invention, there is provided a broadband oscillation identification system for a new energy station, including: the acquisition module is used for acquiring broadband oscillation voltage signals of the wind turbine generator system side and the boosting low-voltage side; the first obtaining module is used for optimizing parameters in variation modal decomposition by adopting an optimization algorithm aiming at broadband oscillation voltage signals at two sides to obtain the optimal values of the modal number and the penalty factor; the second obtaining module is used for obtaining a group of optimal time domain modal components by adopting variation modal decomposition according to the obtained modal number and the optimal value of the penalty factor, and carrying out denoising reconstruction based on a clustering method to obtain a denoising broadband oscillation voltage signal; the extraction module is used for extracting characteristic parameters of each mode of broadband oscillation for the denoising broadband oscillation voltage signal; and the third obtaining module is used for identifying the characteristic parameters of each mode of broadband oscillation extracted by the two-side denoising broadband oscillation voltage signal to obtain an identification result.
According to another aspect of the embodiment of the present invention, there is provided a processor for running a program, wherein the program executes any one of the above methods for identifying broadband oscillation of a new energy station.
According to another aspect of the embodiment of the present invention, there is provided a computer readable storage medium, where the computer readable storage medium includes a stored program, and when the program runs, controls a device in which the computer readable storage medium is located to execute any one of the foregoing broadband oscillation identification methods for a new energy station.
Still further, the present invention will be described with reference to the experimental data shown in fig. 1 to 8:
s1: voltage transformers are respectively arranged on the 690V side and the boost low-voltage side of the wind turbine generator, as shown in fig. 2, the sampling frequency is set to be 500Hz, broadband oscillation voltage signals generated on two sides are collected, and the signals are respectively f 1 (t) and f 2 (t), FIG. 3 shows an original broadband oscillating voltage signal f at 690V side of wind turbine generator 1 (t) schematic diagram, FIG. 4 shows the original broadband oscillation voltage signal f at the low-voltage side of the boost converter 2 (t) schematic;
s2: optimizing the number of modes and penalty factors in the variation mode algorithm by using a particle swarm optimization algorithm:
s2.1: setting initial values of parameters of a particle swarm algorithm, such as iteration times of 50 times, population sizes of 20, and learning a factor c 1 And c 2 Set to 1.4995, noise tolerance 0;
s2.2: setting the variation range of the penalty factor of the parameter 1 as [1000,80000], and setting the variation range of the mode number of the parameter 2 as [2,8];
s2.3: randomly initializing to obtain initial values of parameters 1 and 2, calculating the fitness function envelope entropy of the initial values by calculating the variation modal decomposition of an objective function, updating the historical optimal value and the global historical optimal value of an individual, and iterating for 50 times;
s2.4: the global history optimal solution after iteration is the optimal value of the optimal mode number and penalty factor optimized by the particle swarm algorithm; voltage signal f on 690V side of wind turbine generator set 1 The optimal values of the modal number and the penalty factor of (t) are 4 and 52341 respectively, and the voltage signal f at the low-voltage side is boosted 2 The optimal values of the modal number and penalty factor of (t) are 4 and 54583, respectively;
s3: respectively bringing the two groups of mode numbers obtained in the step S2 and the optimal value of the penalty factor into a variation mode decomposition algorithm to obtain decomposed optimal time domain mode components; FIG. 5 is a schematic diagram of a time domain of a wind turbine generator 690V-side voltage signal decomposed into modal components by a variational modal decomposition; FIG. 7 is a schematic diagram of a time domain of the boost low-voltage side voltage signal decomposed into modal components by a variational modal decomposition;
s4: obtaining the amplitude-frequency characteristic of each modal component through Fourier transformation for each modal component obtained in the step S3; FIG. 6 is a schematic diagram of a frequency domain of a wind turbine generator 690V-side voltage signal decomposed into modal components by a variational modal decomposition; FIG. 8 is a schematic diagram of the frequency domain of the voltage signal at the boost low-voltage side decomposed into modal components by the variational modal decomposition;
s5: and clustering and calculating each time domain modal component of the voltage signal at the V side of the wind turbine generator 690 and each time domain modal component of the voltage signal at the low-voltage side of the boosting transformer obtained in the step S3. As can be seen from fig. 6 and 8 in S4, the effective component of each modal component of the set of data is one, and then the normalized values of each modal component of wind turbine 690V side and high-voltage low-voltage side are shown in tables 1 and 2; the clustering weight coefficients of each modal component of the wind turbine generator 690V side and the boosting low-voltage side are shown in tables 3 and 4; the cluster center of the wind turbine generator 690V side is 0.4052, the cluster center of the boosting low-voltage side is 0.4031, the updated Euclidean distance of the wind turbine generator 690V side is shown in Table 5, and the updated Euclidean distance of the boosting low-voltage side is shown in Table 6; compared with the set threshold value 1, the mode components 1,2 and 3 are effective components, and the mode 4 is an ineffective component, so that the modes 1,2 and 3 are respectively reconstructed into denoising reconstruction signals;
table 1690V side voltage signal decomposition mode normalization value
| Modality 1 | Modality 2 | Modality 3 | Modality 4 | |
| Normalized value | 0.4222 | 1 | 0.1370 | 0.0616 |
TABLE 2 normalization of the boost to low side voltage signal decomposition modes
| Modality 1 | Modality 2 | Modality 3 | Modality 4 | |
| Normalized value | 0.4108 | 1 | 0.1411 | 0.0605 |
Table 3690V side voltage signal decomposition modal clustering weight coefficient
| Modality 1 | Modality 2 | Modality 3 | Modality 4 | |
| Clustering weight coefficient | 0.2605 | 0.6170 | 0.0845 | 0.0380 |
Table 4 clustering weight coefficient of boost low-voltage side voltage signal decomposition mode
| Modality 1 | Modality 2 | Modality 3 | Modality 4 | |
| Clustering weight coefficient | 0.2548 | 0.6202 | 0.0870 | 0.0375 |
Updated euclidean distance of table 5690V side voltage signal decomposition mode
| Modality 1 | Modality 2 | Modality 3 | Modality 4 | |
| Updating Euclidean distance | 0.0011 | 0.5734 | 0.8513 | 3.1069 |
Table 6 update euclidean distance of step-up low-voltage side voltage signal decomposition mode
| Modality 1 | Modality 2 | Modality 3 | Modality 4 | |
| Updating Euclidean distance | 0.0003 | 0.5775 | 0.8124 | 3.0888 |
S6: carrying out Prony identification on denoising reconstruction signals of a 690V side and a boost low-voltage side of the wind turbine generator system obtained in the step S5 to obtain amplitude values, frequencies, initial phase angles and attenuation factors of all modes, wherein the amplitude values, the frequencies, the initial phase angles and the attenuation factors are shown in a table 7 and a table 8 respectively;
table 7690V side voltage signal prony identification results
| amplitude/V | frequency/Hz | Primary phase/rad | Attenuation factor | |
| Modality 1 | 681.5518 | 50.0001 | -1.0830 | 0.0062 |
| Modality 2 | 269.7348 | 11.7716 | 1.5377 | 0.0027 |
| Modality 3 | 91.6883 | 88.2386 | 1.3979 | -0.0305 |
TABLE 8 identification of the boost low side-to-low voltage signal prony
| amplitude/kV | frequency/Hz | Primary phase/rad | Attenuation factor | |
| Modality 1 | 35.5982 | 49.9873 | 0.1603 | 0.0162 |
| Modality 2 | 13.881 | 11.7188 | 0.9347 | 0.0032 |
| Modality 3 | 4.3156 | 88.2811 | -0.7963 | -0.0315 |
S7: as can be seen from tables 7 and 8, the sum of the frequencies of the two oscillation modes on 690V side is 100.0102Hz, and the absolute value of the error between the sum and 100Hz is 0.0102Hz, which is less than 1Hz; the sum of the frequencies of the two oscillation modes at the low-voltage side of the boosting low-voltage side is 99.9999Hz, and the absolute value of the error between the two oscillation modes and 100Hz is 0.0001Hz and is smaller than 1Hz; the absolute values of frequency errors of the corresponding modes of the two oscillation modes are 0.0528Hz and 0.0425Hz and are smaller than 0.5Hz respectively; and the absolute values of damping factor errors of the two oscillation modes at 690V side and high-voltage and low-voltage side are 0.0005 and 0.001, which are smaller than 0.05, and the amplitude of the voltage signal at the boosting and low-voltage side is larger than that of the voltage signal at the 690V side, so that the new energy station can be judged to generate broadband oscillation.
While the present invention has been described in detail with reference to the drawings, the present invention is not limited to the above embodiments, and various changes can be made without departing from the spirit of the present invention within the knowledge of those skilled in the art.
Claims (10)
1. A broadband oscillation identification method for a new energy station is characterized by comprising the following steps:
broadband oscillation voltage signals of a wind turbine generator side and a boosting low-voltage side are obtained;
aiming at broadband oscillation voltage signals at two sides, respectively adopting an optimization algorithm to optimize parameters in variation modal decomposition to obtain the optimal values of the modal number and the penalty factors;
according to the obtained optimal values of the modal number and the penalty factors, adopting variation modal decomposition to obtain a group of optimal time domain modal components, and carrying out denoising reconstruction based on a clustering method to obtain a denoising broadband oscillation voltage signal;
extracting characteristic parameters of each mode of broadband oscillation from the denoising broadband oscillation voltage signal;
and identifying the characteristic parameters of each mode of broadband oscillation extracted according to the two-side denoising broadband oscillation voltage signals to obtain an identification result.
2. The method for identifying broadband oscillation of a new energy station according to claim 1, wherein optimizing parameters in the variation modal decomposition by using an optimization algorithm to obtain an optimal value of a modal number and a penalty factor comprises:
setting an initial value of a parameter in a particle swarm optimization algorithm; setting a change interval of the number of modes and penalty factors;
decomposing the broadband oscillation voltage signal through variation modal decomposition to obtain decomposed time domain modal components; calculating fitness function envelope entropy according to the time domain modal components;
determining the overall optimal solution of the number of modes and the penalty factor according to the fitness function envelope entropy;
repeating until the termination condition is met, and taking the global history optimal solution after iteration is finished as the optimal value of the modal number and the penalty factor of the particle swarm optimization algorithm.
3. The method for identifying broadband oscillation of a new energy station according to claim 1, wherein the denoising reconstruction is performed based on a clustering method to obtain a denoising broadband oscillation voltage signal, comprising:
obtaining the amplitude-frequency characteristic of each optimal time domain modal component according to Fourier transformation, and extracting the effective frequency component of each optimal time domain modal component; for each optimal time-domain modal component, determining a frequency component with an amplitude greater than or equal to 10% of the maximum amplitude as an effective frequency component of each time-domain modal component;
taking the amplitude value of the effective frequency component of each optimal time domain modal component as initial data, and carrying out K-means clustering on the amplitude value of the effective frequency component according to the Euclidean distance to obtain a clustering center of each optimal time domain modal component, wherein K clustering centers are taken as first clustering centers;
clustering the K first clustering centers of each optimal time domain modal component again according to the Euclidean distance to obtain a second clustering center;
comparing Euclidean distances after the weight coefficients of the obtained K first clustering centers and the obtained K second clustering centers are updated with a set threshold value: if the updated Euclidean distance is smaller than or equal to the threshold value, judging the Euclidean distance as an effective modal component, and if the updated Euclidean distance is larger than the threshold value, judging the Euclidean distance as an ineffective modal component;
and reconstructing the effective modal component to obtain the denoising broadband oscillation voltage signal.
4. The method for identifying broadband oscillation of a new energy station according to claim 3, wherein the euclidean distance obtained by updating the weight coefficients of the K first clustering centers and the second clustering centers is expressed as follows:
D k |＝D k /ρ k
wherein D is k ' is D k The updated value of the first cluster center and the second cluster center represents the Euclidean distance after the weight coefficient is updated; d (D) k For the kth first cluster center and the second cluster center x 0 A Euclidean distance; ρ k And the weight coefficient of the kth first cluster center.
5. The method for identifying broadband oscillation of a new energy station according to claim 4, wherein the weight coefficient is expressed as follows:
wherein x is k The K is the clustering center of the K optimal time domain modal components, namely the K first clustering center, and K is the number of the optimal time domain modal components.
6. The method for identifying broadband oscillation of a new energy station according to claim 1, wherein the characteristic parameters of each mode of broadband oscillation are extracted from the de-noised broadband oscillation voltage signal, specifically: and analyzing the denoising broadband oscillation voltage signal as an input signal of the Prony algorithm, and extracting characteristic parameters of each mode of broadband oscillation.
7. The method for identifying broadband oscillation of a new energy station according to claim 1, wherein the identifying is performed according to characteristic parameters of each mode of broadband oscillation extracted from two-sided denoising broadband oscillation voltage signals, so as to obtain an identification result, specifically:
if the conditions that the absolute value of the error between the sum of the frequencies of two oscillation modes smaller than 100Hz in the signals at two sides and the frequency of the vibration modes at two sides is smaller than 1Hz and the absolute value of the error of the attenuation factor at the corresponding modes of the signals at two sides is smaller than 0.5Hz are met, the amplitude of the low-voltage side of the boosted voltage is larger than that of the wind turbine generator set side are met, the problem of broadband oscillation of the new energy station is judged, and the two analyzed oscillation modes smaller than 100Hz are dominant oscillation modes of the broadband oscillation, otherwise, the problem is not solved.
8. A broadband oscillation identification device for a new energy station is characterized by comprising:
the acquisition module is used for acquiring broadband oscillation voltage signals of the wind turbine generator system side and the boosting low-voltage side;
the first obtaining module is used for optimizing parameters in variation modal decomposition by adopting an optimization algorithm aiming at broadband oscillation voltage signals at two sides to obtain the optimal values of the modal number and the penalty factor;
the second obtaining module is used for obtaining a group of optimal time domain modal components by adopting variation modal decomposition according to the obtained modal number and the optimal value of the penalty factor, and carrying out denoising reconstruction based on a clustering method to obtain a denoising broadband oscillation voltage signal;
the extraction module is used for extracting characteristic parameters of each mode of broadband oscillation for the denoising broadband oscillation voltage signal;
and the third obtaining module is used for identifying the characteristic parameters of each mode of broadband oscillation extracted by the two-side denoising broadband oscillation voltage signal to obtain an identification result.
9. A processor, wherein the processor is configured to run a program, and wherein the program is configured to perform the method for identifying broadband oscillations of a new energy station according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored program, wherein the program, when run, controls a device in which the computer-readable storage medium is located to execute the broadband oscillation identification method for a new energy station according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310825537.0A CN116826735A (en) | 2023-07-06 | 2023-07-06 | Broadband oscillation identification method and device for new energy station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310825537.0A CN116826735A (en) | 2023-07-06 | 2023-07-06 | Broadband oscillation identification method and device for new energy station |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116826735A true CN116826735A (en) | 2023-09-29 |
Family
ID=88112473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310825537.0A Pending CN116826735A (en) | 2023-07-06 | 2023-07-06 | Broadband oscillation identification method and device for new energy station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116826735A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117725394A (en) * | 2024-02-18 | 2024-03-19 | 浙江浙能技术研究院有限公司 | Wind power plant broadband oscillation identification method based on hierarchical embedded modal decomposition |
-
2023
- 2023-07-06 CN CN202310825537.0A patent/CN116826735A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117725394A (en) * | 2024-02-18 | 2024-03-19 | 浙江浙能技术研究院有限公司 | Wind power plant broadband oscillation identification method based on hierarchical embedded modal decomposition |
| CN117725394B (en) * | 2024-02-18 | 2024-05-07 | 浙江浙能技术研究院有限公司 | Wind power plant broadband oscillation identification method based on hierarchical embedded modal decomposition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fu et al. | A hybrid approach for multi-step wind speed forecasting based on two-layer decomposition, improved hybrid DE-HHO optimization and KELM | |
| CN111193256A (en) | Power load prediction method based on variational modal decomposition and gated cyclic unit | |
| CN106933778A (en) | A kind of wind power combination forecasting method based on climbing affair character identification | |
| CN112651290A (en) | Hydroelectric generating set vibration trend prediction method and system | |
| Shao et al. | Short‐term wind speed forecasting using wavelet transformation and AdaBoosting neural networks in Yunnan wind farm | |
| CN105391083A (en) | Wind power range short-term prediction method based on variation mode decomposition and relevant vector machine | |
| CN109615152B (en) | Wind power generation prediction method and product based on cost-oriented gradient ascending regression tree | |
| CN112200038B (en) | CNN-based quick identification method for oscillation type of power system | |
| CN107844849A (en) | A kind of new energy output short term prediction method returned based on experience wavelet transformation with improving Gaussian process | |
| CN110826803A (en) | Electricity price prediction method and device for electric power spot market | |
| CN110991725B (en) | RBF ultra-short-term wind power prediction method based on wind speed frequency division and weight matching | |
| CN116826735A (en) | Broadband oscillation identification method and device for new energy station | |
| CN104966161A (en) | Electric energy quality recording data calculating analysis method based on Gaussian mixture model | |
| CN108959704B (en) | Rewarding and punishing weight type simulation sequence similarity analysis method considering morphological change | |
| CN111563631A (en) | Wind power generation power prediction method and device based on variational modal decomposition | |
| CN116707331A (en) | Inverter output voltage high-precision adjusting method and system based on model prediction | |
| Xu et al. | Interval prediction method for wind power based on VMD-ELM/ARIMA-ADKDE | |
| Liao et al. | Scenario generations for renewable energy sources and loads based on implicit maximum likelihood estimations | |
| CN113094887B (en) | Optimization method and device for frequency-shifting electromagnetic transient simulation and electronic equipment | |
| CN114298132A (en) | Wind power prediction method and device and electronic equipment | |
| CN117454768A (en) | Data and physical combined driven seismic wave generation method, system and storage medium | |
| CN106529075A (en) | Non-linear wind speed simulation method considering divided period | |
| CN116226620A (en) | Subsynchronous oscillation mode identification method based on SVM and LSTM | |
| CN116384586A (en) | Wind speed prediction method based on VMD-SVR | |
| CN110098610A (en) | The real-time identification method and system of power system oscillation dominant pattern under fault disturbance |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |