Method and system for tracing sub-hyper synchronous oscillation based on response data and diffusion energy

Abstract

The application discloses a method and system for tracing a sub-synchronous oscillation source based on response data and diffusion energy, and the method comprises the following steps: obtaining multi-node measurement response data and network topology, performing time-frequency analysis on branch oscillation response sequences after preprocessing, and identifying dominant oscillation modes in a target frequency band; constructing a branch mode oscillation energy model based on branch active power, reactive power, voltage amplitude and phase angle, calculating branch oscillation energy sharing degree, screening high-load branches, and traversing the topology to divide an oscillation high-risk cut set; further constructing an oscillation diffusion energy model based on instantaneous energy theory, calculating the oscillation energy diffusion rate of cut set elements under each dominant mode, and positioning the oscillation source according to the sign; and finally outputting the tracing results such as the oscillation source, the dominant mode, the high-risk cut set and the diffusion rate. The application takes into account the frequency coupling characteristics, clearly represents the oscillation energy flow direction and diffusion intensity through the energy diffusion rate, realizes the positioning of the oscillation source with physical meaning, and improves the tracing accuracy and interpretability.

Description

Technical Field

[0001] This invention relates to the field of power system oscillation monitoring technology, and more specifically to a method and system for tracing the source of sub-supersynchronous oscillations based on response data and diffused energy. Background Technology

[0002] Currently, with the large-scale integration of converter-based renewable energy sources such as wind power and photovoltaics into the power grid, as well as power electronic equipment, significant changes have occurred in the grid structure, power supply structure, and control structure. Under conditions such as weak grids, control parameter mismatches, enhanced grid coupling, or external disturbances, new energy grid-connected systems may experience subsynchronous, supersynchronous, or broadband oscillations. These oscillations are characterized by random time-varying, wide-area propagation, and multi-source, multi-operating-condition coupling, which poses a significant challenge to the online source tracing of novel subsynchronous / supersynchronous oscillations.

[0003] Existing methods for locating oscillations in power systems mainly include model-based methods, signal data-based methods, and artificial intelligence-based methods. Model-based methods require analysis of the system and component structures, and suffer from problems such as difficulty in accurate modeling and the curse of dimensionality, making them difficult to apply online. Signal data-based methods often use wavelet filtering, variational mode decomposition combined with transient energy flow or power flow analysis to locate oscillation sources, but their data requirements, applicable scenarios, computational efficiency, and accuracy vary. Artificial intelligence-based methods are highly dependent on the size and quality of the training dataset, and generally suffer from problems such as the need to verify adaptability to engineering environments and insufficient interpretability.

[0004] Current research on oscillation source localization mainly focuses on traditional synchronous generators and low-frequency oscillation problems. However, for multi-source and multi-operating-condition coupled oscillations caused by the grid connection of new energy sources, existing methods are difficult to simultaneously achieve quantitative characterization of oscillation propagation paths and accurate localization of oscillation sources under multi-mode and frequency coupling conditions. There is a lack of a unified means to compress the search space from the perspective of network structure and to distinguish oscillation sources from energy dissipation units with clear physical criteria.

[0005] Therefore, how to design a method and system for the propagation and source tracing of subsynchronous / supersynchronous oscillations that can adapt to multi-source, multi-condition coupled oscillation scenarios without relying on detailed system models and control parameters is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] In view of the above problems, the present invention proposes a sub-supersynchronous oscillation tracing method and system based on response data and diffusion energy to overcome or at least partially solve the above problems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] In a first aspect, embodiments of the present invention provide a method for tracing the source of sub-supersynchronous oscillations based on response data and diffusion energy, comprising the following steps: S1. Acquire the measurement response data and network topology connection relationship of multiple monitoring nodes in the power system, and preprocess the wide-area measurement response data; S2. Based on the oscillation response sequences of each node and branch obtained after preprocessing, the oscillation mode parameters of the oscillation response sequences are identified using time-frequency analysis methods, and the dominant oscillation mode in the target sub-supersynchronous frequency band is selected from the identified oscillation modes. S3. Based on the active power, reactive power, voltage amplitude and voltage phase angle at both ends of each branch, construct a branch modal oscillation energy model and calculate the branch oscillation energy sharing degree of each branch; sort the branches according to the branch oscillation energy sharing degree, select the branches that carry oscillation energy higher than a preset threshold, and traverse the network topology based on the selected branches to divide the high-risk cut set of oscillation. S4. Based on the instantaneous energy theory, construct an oscillation diffusion energy model. For each dominant oscillation mode, calculate the oscillation energy diffusion rate of each cut set element in the high-risk cut set of the oscillation, and locate the oscillation source according to the sign of the oscillation energy diffusion rate. S5, the source of the output oscillation, the dominant oscillation mode, the high oscillation risk cut set, and the oscillation energy diffusion rate.

[0009] Further, in step S1, the measurement response data includes at least one of the following: voltage, branch current, active power, reactive power, and voltage phase angle information of each bus and branch node. The monitoring nodes include at least one of the following: wind farm grid connection point, photovoltaic power station grid connection point, synchronous generator endpoint, flexible transmission device interface, and nodes at both ends of transmission line; The measurement response data originates from at least one of the following: phasor measurement unit, wide-area measurement system, field monitoring system, or fault recorder.

[0010] Furthermore, the oscillation mode parameter identification in step S2 includes: performing time-frequency domain analysis on the oscillation response sequence using windowed interpolation-type fast Fourier transform, and identifying the frequency, amplitude, and phase parameters of each oscillation mode by combining improved least squares-rotation invariant technique.

[0011] Further, in step S3, the method for constructing the branch modal oscillation energy model is as follows: for any branch, the oscillation energy flowing into the branch from both ends of the branch is expressed as the sum of the integral of active power with respect to the voltage phase angle and the integral of reactive power with respect to the logarithm of the voltage amplitude. The electrical quantities at both ends of the branch are decomposed into steady-state components and oscillation components of each oscillation mode. The oscillation energy of the dominant oscillation mode is much greater than that of the non-dominant mode and the steady-state value of the electrical quantity is much greater than that of the oscillation quantity. The oscillation energy of the branch is then expressed as the superposition of the energy components of each dominant mode and a constant term.

[0012] Furthermore, the calculation method for the branch oscillation energy sharing is as follows: the ratio of the branch oscillation energy in a certain dominant mode to the sum of the oscillation energy of all branches in the entire network in that mode is defined as the branch oscillation energy sharing in that mode.

[0013] Further, in step S4, the method for constructing the oscillation diffusion energy model includes: defining the oscillation diffusion energy flowing out of the node from the external system as the instantaneous energy integral of the branch current and the node voltage in a two-phase rotating coordinate system, and stipulating that the oscillation energy flowing out of the node is positive and the oscillation energy flowing into the node is negative; By substituting the voltage and current into the oscillation diffusion energy expression after coordinate transformation, the coefficient of the non-periodic energy term is extracted and defined as the oscillation energy diffusion rate.

[0014] Furthermore, when subsynchronous frequency coupling is not considered, the oscillation energy diffusion rate is determined by the voltage amplitude, current amplitude, and initial phase of the oscillation component; When considering subsynchronous and supersynchronous frequency coupling, the oscillation energy diffusion rate incorporates the voltage amplitude, current amplitude, and initial phase of both the subsynchronous and supersynchronous oscillation components to characterize the effect of the coupling between the subsynchronous and supersynchronous oscillation components on energy diffusion.

[0015] Further, in step S4, the specific criterion for locating the oscillation source based on the sign of the oscillation energy diffusion rate is as follows: When the oscillation energy diffusion rate is positive, the element is determined to be the oscillation source under the corresponding dominant oscillation mode; When the oscillation energy diffusion rate is negative, the element is determined to be an energy dissipation unit; when the oscillation energy diffusion rate approaches zero, the element is determined to be a neutral unit.

[0016] Furthermore, step S5 also includes: determining the differences in oscillation sources under different modes based on the differences in the sign and amplitude of the oscillation energy diffusion rate of each cut-set element under each dominant mode, and revealing the differences in the damping effect of the synchronous generator under different oscillation modes; The tracing results of oscillation source, dominant oscillation mode, high-risk cut set of oscillation, and oscillation energy diffusion rate are output through graphical interface, data file, or monitoring platform interface.

[0017] Secondly, embodiments of the present invention provide a sub-supersynchronous oscillation tracing system based on response data and diffusion energy, comprising: The data acquisition module is used to acquire measurement response data and network topology connections from multiple monitoring nodes in the power system. The data preprocessing and mode identification module is used to preprocess the measurement response data to obtain an oscillation response sequence, and to identify the dominant oscillation mode using time-frequency analysis methods. The branch energy analysis and cut set partitioning module is used to construct the branch modal oscillation energy model, calculate the branch oscillation energy sharing degree of each branch, and partition the high-risk cut set of oscillation based on the branch oscillation energy sharing degree. The oscillation source localization module is used to construct an oscillation diffusion energy model, calculate the oscillation energy diffusion rate of each cut set element within the high-risk oscillation cut set, and locate the oscillation source according to the sign of the oscillation energy diffusion rate. The results output module is used to output the source of oscillation, dominant oscillation mode, high oscillation risk cut set, and oscillation energy diffusion rate.

[0018] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a method and system for tracing the source of sub-supersynchronous oscillations based on response data and diffusion energy, which has the following beneficial effects: 1. This invention uses wide-area measurement response data as the main information source, and does not require a complete understanding of complex system models and converter control parameters, which can reduce the threshold for engineering deployment; 2. This invention constructs a branch mode oscillation energy expression form and defines the branch oscillation energy sharing degree, realizing the quantitative characterization of the propagation path of oscillation in complex power networks. It can identify key branches and high-risk cut sets with concentrated oscillation energy from the perspective of network structure, effectively compressing the oscillation source search space. 3. This invention proposes an oscillation energy diffusion rate criterion that takes into account frequency coupling characteristics. The oscillation energy diffusion rate clearly characterizes the flow direction and diffusion intensity of oscillation energy in the network, thereby realizing the physical meaning of oscillation source localization. 4. This invention can not only accurately identify the dominant secondary / supersynchronous oscillation modes, but also accurately locate the oscillation source under multi-source interactive oscillation conditions, and reveal the differences in the damping effect of synchronous generators under different modes, with good accuracy and robustness. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1This is a flowchart of the oscillation source tracing method provided in the embodiments of the present invention; Figure 2 This is a schematic diagram of the oscillation response characteristics provided in an embodiment of the present invention; Figures 3a-b are schematic diagrams of response spectrum analysis provided in the embodiments of the present invention; Figures 4a-b are schematic diagrams of branch oscillation energy sharing under different dominant modes provided in the embodiments of the present invention; Figures 5a-b are schematic diagrams of high-risk oscillation cut sets provided in the embodiments of the present invention; Figure 6 This is a structural diagram of the oscillation tracing system provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] This invention discloses a method for tracing the source of sub-supersynchronous oscillations based on response data and diffusion energy, such as... Figure 1 As shown, it includes the following steps: S1. Acquire the measurement response data and network topology connection relationship of multiple monitoring nodes in the power system, and preprocess the wide-area measurement response data; S2. Based on the oscillation response sequences of each node and branch obtained after preprocessing, the oscillation mode parameters of the oscillation response sequences are identified using time-frequency analysis methods, and the dominant oscillation mode in the target sub-supersynchronous frequency band is selected from the identified oscillation modes. S3. Based on the active power, reactive power, voltage amplitude and voltage phase angle at both ends of each branch, construct a branch modal oscillation energy model and calculate the branch oscillation energy sharing degree of each branch; sort the branches according to the branch oscillation energy sharing degree, select the branches that carry oscillation energy higher than a preset threshold, and traverse the network topology based on the selected branches to divide the high-risk cut set of oscillation. S4. Based on the instantaneous energy theory, construct an oscillation diffusion energy model. For each dominant oscillation mode, calculate the oscillation energy diffusion rate of each cut set element in the high-risk cut set of the oscillation, and locate the oscillation source according to the sign of the oscillation energy diffusion rate. S5, the source of the output oscillation, the dominant oscillation mode, the high oscillation risk cut set, and the oscillation energy diffusion rate.

[0023] The following is a detailed description of each of the above steps: Step S1: Data acquisition and preprocessing; Acquire wide-area measurement response data and network topology connections of multiple monitoring nodes in the power system; New energy grid-connected power systems include wind farms, photovoltaic power stations, flexible AC transmission equipment, flexible DC transmission equipment, synchronous generators, converter grid-connected equipment, or combinations thereof.

[0024] Acquire at least one of the following information for each bus and branch node in the power system: voltage, branch current, active power, reactive power, and voltage phase angle. The response data comes from at least one of the following sampling data: phasor measurement unit, wide-area measurement system, station monitoring system, or fault recorder. Acquire the topological connections and node numbers of wind farms, photovoltaic power plants, synchronous generators, and transmission lines.

[0025] In this embodiment, the busbar, wind farm grid connection point, photovoltaic power station grid connection point, synchronous generator endpoint, and other dynamic equipment are equivalent to nodes, and the coupling relationship or equivalent electrical connection relationship between the transmission lines, transformers, and busbars between the nodes are used as edges, thereby forming a system topology structure for oscillation propagation and source tracing analysis.

[0026] Data from multiple time scales, including phasor measurement units, wide-area measurement systems, fault recorders, site monitoring systems, or converter local controllers, are unified to the same time base to achieve time alignment.

[0027] Missing value imputation and moving median filtering were used to remove outliers, and least squares detrending was performed on the data to eliminate the interference of steady-state components on oscillation analysis.

[0028] Based on the event start time and the preset data window length, an oscillation response data block is formed for online calculation, resulting in the oscillation response sequence of each node and branch.

[0029] In this embodiment, the preprocessed data is cached in the form of data blocks, supporting online sliding window updates to meet the real-time requirements of online oscillation monitoring and tracing.

[0030] like Figure 2 As shown, the oscillation characteristics of the three-phase current in the wind farm's grid-connected tie line before and after the disturbance are illustrated. The red, blue, and green curves represent the current signals of phases A, B, and C, respectively, with the horizontal axis representing time and the vertical axis representing the current amplitude.

[0031] As shown in the figure, before 1.5s, the system was in a stable operating state. The three-phase current waveforms of the wind farm tie line exhibited periodic changes, maintaining good symmetry among the three phase currents, with relatively stable amplitudes and no obvious oscillations or abrupt amplitude changes, indicating that the wind farm grid-connected system was operating smoothly at this time. After the 1.5s disturbance, the three-phase current waveforms changed rapidly, disrupting the original symmetrical operating state. The current amplitude increased significantly in a short period of time, accompanied by a violent fluctuation process.

[0032] Subsequently, the three-phase current signals continued to exhibit irregular oscillation characteristics, with increased waveform distortion and significant shifts in the amplitude and phase relationship of the interphase currents, indicating that the disturbance caused a dynamic instability trend in the wind farm grid-connected system. The magnified area in the figure further shows that after the disturbance, the oscillation amplitude and waveform differences of the three-phase currents significantly increased, and the system transitioned from a stable operating state before the disturbance to a strongly oscillating operating state after the disturbance.

[0033] Therefore, the pre-processed oscillation response sequence can intuitively reflect the transient response process and oscillation development characteristics of the three-phase current of the tie line after the wind field is subjected to external disturbance.

[0034] Step S2: Dominant oscillation mode identification; Based on the oscillation response sequences of each node and branch obtained after preprocessing, the oscillation mode parameters of the oscillation response sequences are identified using time-frequency analysis methods, and the dominant oscillation modes in the target sub / supersynchronous frequency band are selected from the identified oscillation modes.

[0035] A windowed interpolation-based fast Fourier transform was used to perform time-frequency domain analysis on the response sequence, and an improved least squares-rotation invariant technique was combined to identify the parameters of each oscillation mode.

[0036] In this implementation example, the windowed interpolation fast Fourier transform is specifically as follows: The Kaiser window achieves a balance between leakage suppression and frequency resolution by adjusting the main lobe width and side lobe attenuation. Its discretization expression is as follows:

[0037] in, For the zeroth-order modified Bessel function of the first kind, through power series Calculation, denominator To achieve window function normalization, β is the shape parameter of the Kaiser window, n is the number of network nodes, and N is the total number of sampling points contained in the current data window.

[0038] The signal after weighting by the window function is The FFT result needs to be compensated for the amplitude attenuation caused by the energy correction window function. The correction factor is defined as:

[0039] In addition, the double-peak spectral line interpolation algorithm is used to perform weighted calculations on multiple spectral lines in the frequency domain of the harmonic signal, thereby obtaining a value closer to the theoretical value of the measured signal. This can improve the picket fence effect and further reduce spectral leakage and noise interference.

[0040] For peak frequency is Amplitude Phase is The corresponding frequency point is The signal has spectral lines on its left and right sides, respectively. and The corresponding amplitude is and The corrected formulas for frequency, amplitude, and phase are derived using polynomial approximation:

[0041] in, The spectrum is the Fast Fourier Transform. This is the frequency domain form of the window function; The sampling frequency for measuring the response sequence; The frequency resolution of the Fast Fourier Transform; relative to the peak frequency The dimensionless frequency offset correction.

[0042] The electrical quantity during oscillation is decomposed into the sum of steady-state quantity and oscillation components of different oscillation modes. Since the energy of the dominant oscillation mode is much greater than the oscillation energy of other non-dominant modes, the oscillation frequency with the dominant energy is determined as the dominant oscillation mode.

[0043] Based on the complementary characteristics of subsynchronous / supersynchronous oscillations, namely that the sum of the frequency components of the oscillation current and the frequency components of the oscillation power equals the power frequency, the identified dominant oscillation mode is verified.

[0044] Figure 3 shows the results of frequency domain analysis of electrical quantities and verification of the dominant oscillation mode during the oscillation process. Figure 3(a) shows the oscillation current spectrum analysis results, and Figure 3(b) shows the oscillation power spectrum analysis results. As shown in Figure 3(a), in addition to the fundamental component near the power frequency, several obvious oscillation frequency components exist in the spectrum. Specifically, a current oscillation component with a large amplitude appears at approximately 15Hz, with a current amplitude of approximately 1.33kA; another current oscillation component appears at approximately 57Hz, with a current amplitude of approximately 330.6A; simultaneously, a power frequency component exists at approximately 60Hz, with a current amplitude of approximately 962A. The amplitude of the current oscillation component at 15Hz is significantly higher than that of the other non-power frequency oscillation components. Therefore, the mode corresponding to this frequency can be identified as the energy-dominant oscillation mode in the current signal, i.e., one of the dominant oscillation modes.

[0045] As shown in Figure 3(b), after spectral decomposition of the oscillating power signal, multiple frequency components are also present in the power spectrum, with relatively obvious power oscillation components appearing at approximately 3.04 Hz and 45.06 Hz. Specifically, the power amplitude at 3.04 Hz is approximately 71.52 MW, and the power amplitude at 45.06 Hz is approximately 51.39 MW. It can be seen that the current oscillation frequency and the power oscillation frequency satisfy the typical complementary characteristics of subsynchronous oscillation, that is, the sum of the frequency components of the oscillating current and the power is the power frequency of 60 Hz, indicating a consistent physical correspondence between the identified current oscillation frequency and the power oscillation frequency.

[0046] Step S3: Branch mode oscillation energy analysis and high-risk cut set partitioning; Based on the active power, reactive power, voltage amplitude, and voltage phase angle at both ends of each branch, a branch modal oscillation energy model is constructed to calculate the branch oscillation energy sharing of each branch. The branches are sorted according to the branch oscillation energy sharing, and the cumulative energy contribution rate is used to determine the branch screening threshold. Branches carrying oscillation energy higher than the preset threshold are screened out, and the network topology is traversed based on the screened branches to divide the high-risk cut set of oscillation.

[0047] Construct a branch modal oscillation energy model; power grid branches are the main channels for the transmission and exchange of oscillation energy in the network. For any branch... If the oscillating energy flows into the branch from the node and both sides of the node, then the branch oscillating energy can be expressed as:

[0048] in, , These represent the active power and reactive power flowing from node i to node j in the branch, respectively. , These represent the active power and reactive power flowing from node j to node i in the branch, respectively. , Let be the voltage vector angles at nodes i and j, respectively; , Let be the voltage vector magnitudes at nodes i and j, respectively.

[0049] To compare the relative magnitudes of oscillatory energy carried by different branches under the same dominant mode, a branch is defined. In modality The energy sharing of the branch oscillation is:

[0050] Where n is the number of network nodes.

[0051] To verify the applicability of the above-mentioned branch oscillation energy sharing method in power systems of different scales and with different new energy access forms, simulation analysis was conducted on a new energy grid-connected four-machine two-zone system and a new energy grid-connected 10-machine 39-node system.

[0052] In one exemplary embodiment, the rated frequency of the four-unit, two-zone grid-connected new energy system is 60Hz. A wind farm consisting of 30 doubly-fed wind turbines with a single rated capacity of 5MW is connected at node 10. An aggregate model is used to perform equivalent modeling of the wind farm, and the wind speed is set to 11m / s. When the system has been running stably for 1.5s, the series compensation degree of the wind farm grid connection line is adjusted to 46%, thereby forming a subsynchronous oscillation condition lasting about 2.5s.

[0053] The rated frequency of the 10-unit, 39-node renewable energy grid-connected system is 50Hz. A wind farm consisting of 100 doubly-fed wind turbines with a single unit rated capacity of 5MW is connected at node 10, and a photovoltaic power station consisting of 600 photovoltaic power generation units with a single unit rated capacity of 0.45MW is connected at node 21. After the system has been operating stably for 2 seconds, the series compensation of the wind farm's grid-connected lines is adjusted to 42%, and the series compensation of the photovoltaic power station's related transmission lines is adjusted to 50%, thus forming a multi-modal sub-supersynchronous oscillation interactive coupling condition lasting approximately 2 seconds.

[0054] For each dominant oscillation mode, calculate the branch oscillation energy sharing of each branch in the whole system.

[0055] Figure 4 shows the oscillation energy sharing of each branch in the 4-unit, 2-zone system and the 10-unit, 39-node system for new energy grid connection. Figure 4(a) shows the oscillation energy sharing of each branch in the 4-unit, 2-zone system for new energy grid connection, and Figure 4(b) shows the oscillation energy sharing of each branch in the 10-unit, 39-node system for new energy grid connection. The horizontal axis represents different branches in the system, and the vertical axis represents the oscillation energy sharing of each branch. The figure uses a bar chart to represent the index values ​​of different branches under the same mode. The larger the bar, the stronger the index value of that branch. The larger the value, the better. As shown in Figure 4(a), the values ​​for branches 10-4 and 9-10 are relatively large, approximately 0.235 and 0.223 respectively, significantly higher than most other branches, indicating that these two branches have a strong influence in this mode. Branches 10-11 and 11-3... The values ​​are also relatively high; in comparison, the values ​​of branches 7-8, 8-9, 5-6, 1-5, and 2-6 are... The smaller the value, the lower the participation of these branches in this mode. As shown in Figure 4(b), there are significant differences in the oscillation energy carrying capacity of different branches for the dominant oscillation mode; most branches... The value is close to zero, indicating that it participates in the propagation of oscillatory energy to a low extent in this mode; while some branches... The significantly increased values ​​indicate that these branches bear a greater share of the oscillatory energy transfer. The branches, through their interactions with each other... By comparing values, key branches can be screened and identified.

[0056] Branches are ranked from largest to smallest based on their oscillation energy share, and key branches with larger oscillation energy shares and greater oscillation energy loads are identified as high-risk branches. Specifically, branches with oscillation energy share exceeding a preset threshold are designated as high-risk branches.

[0057] The network is traversed based on high-risk branches to form oscillatory high-risk cut sets, thereby compressing the search space for oscillation sources.

[0058] Figure 5 illustrates the partitioning results of the high-risk cutsets for the entire system's oscillations. Based on the topological connections of the renewable energy grid-connected power system, this figure designates branches with high energy contribution rates in the branch mode oscillations as initial high-risk branches. A network traversal is then performed along their adjacent nodes, connecting branches, and the direction of oscillation energy transmission, forming a high-risk cutset composed of key branches, associated nodes, and potential oscillation energy transmission boundaries. Branches included in this cutset indicate that they carry a significant amount of oscillation energy in the dominant oscillation mode or have a strong topological connection with high-contribution branches, making them key areas for concentrated oscillation energy transmission, diffusion, and coupling. Branches not included in the cutset typically correspond to smaller branch oscillation energy contributions, exhibiting relatively lower participation in oscillation propagation under this mode, and can be given lower priority in subsequent source tracing analysis.

[0059] By constructing a high-risk cut set for oscillations, the search range for oscillation sources, which originally covered all nodes and branches of the entire network, can be compressed to a local network region with a higher energy share. In the example of a four-machine two-zone system for grid-connected new energy, without search space partitioning, it would be necessary to perform oscillation propagation and source tracing discrimination one by one according to the nine candidate branches in Figure 5(a). These candidate branches include branches 10-4, 9-10, 10-11, 11-3, 7-8, 8-9, 5-6, 1-5, and 2-6. After adopting the method of this invention, the candidate branches are first sorted according to the oscillation energy share of each branch under the dominant oscillation mode, and only the four branches with higher energy share are retained as high-risk oscillation branches, namely 10-4, 9-10, 10-11, and 11-3. Thus, compared to searching one by one of the nine candidate branches before partitioning, only the four high-risk branches need to be considered after partitioning, reducing the search space by 55.56%.

[0060] In the 10-unit, 39-node grid-connected new energy system modal 1 example, without search space partitioning, oscillation source identification would require 39 node-level candidate objects. Using the method of this invention, high-risk propagation branches are screened based on the branch modal oscillation energy sharing, further forming a high-oscillation-risk cutset including WF, adjacent synchronous generator units, and node 11. At this point, subsequent oscillation energy diffusion rate calculations only need to be performed within 4 cutset elements, reducing the number of candidate objects from 39 to 4, compressing the search space by 89.74%, reducing the number of objects in the subsequent oscillation diffusion energy model and oscillation energy diffusion rate calculations, and improving online analysis efficiency. Simultaneously, this cutset can also reflect the topological connectivity between key branches and possible oscillation propagation paths, providing a candidate range for subsequent identification of oscillation sources, energy dissipation units, neutral units, and extraction of dominant propagation paths within the high-risk cutset.

[0061] Step S4: Construction of oscillation diffusion energy model and localization of oscillation source; An oscillation diffusion energy model is constructed based on instantaneous energy theory. For each dominant oscillation mode, the oscillation energy diffusion rate of each cut set element is calculated within the high-risk cut set of the oscillation, and the oscillation source is located according to the sign of the oscillation energy diffusion rate.

[0062] The oscillatory diffusion energy flowing out of the node from the external system is defined as the instantaneous energy integral of the branch current and the node voltage in a two-phase rotating coordinate system. It is agreed that the oscillatory energy flowing out of the node is positive and the oscillatory energy flowing into the node is negative. When frequency coupling is not considered, the voltage and current are substituted into the oscillatory diffusion energy expression after coordinate transformation, and the coefficient of the non-periodic energy term is defined as the oscillatory energy diffusion rate without considering frequency coupling. When frequency coupling is considered, the voltage and current are substituted into the oscillatory diffusion energy expression after coordinate transformation, and the coefficient of the non-periodic energy term is defined as the oscillatory energy diffusion rate considering frequency coupling.

[0063] When grid connection of new energy sources triggers novel sub-supersynchronous oscillations, the component generating oscillation energy contributes negatively to oscillation decay, while the component consuming energy contributes positively. Oscillation energy diffuses from the oscillation source component to the external system. Therefore, the location of the oscillation source can be identified based on the diffused oscillation energy. The diffused oscillation energy flowing from the node to the external system can be expressed as:

[0064] in, For the branch road The current; For nodes The voltage, where all variables in the xy coordinate system can be transformed from instantaneous value variables in the abc coordinate system.

[0065] To standardize the representation of three-phase electrical quantities, let The three-phase phase displacement parameters are as follows:

[0066] Rotation reference angle:

[0067] When frequency coupling is neglected, voltage and current can be decomposed into a fundamental component and a single dominant oscillatory component. For any phase... It can be uniformly written as:

[0068]

[0069] Where: the subscripts 0 and s correspond to the fundamental component and the oscillating component, respectively; the relevant parameters represent the amplitude and initial phase of the corresponding components, respectively.

[0070] Assume the frequency offset is as follows. The three-phase instantaneous quantities are transformed to two-phase coordinates using a three-phase to two-phase rotating coordinate transformation. Coordinate system, rearranged to obtain:

[0071]

[0072]

[0073]

[0074]

[0075] Substituting the dominant oscillation component into the oscillation diffusion energy model and rearranging it using the sum-to-product relationship, we obtain:

[0076] Oscillatory diffusion energy includes periodic and aperiodic diffusion components. The periodic component varies with oscillation and has a weak relationship with energy direction determination; the aperiodic diffusion component characterizes the net diffusion effect in the network after disturbance injection. The oscillatory energy diffusion rate can be obtained and expressed as:

[0077] in, , These represent the voltage amplitude and initial phase of the oscillation component, respectively. , These represent the current amplitude and initial phase of the oscillating component, respectively.

[0078] When frequency coupling is considered, each electrical quantity simultaneously contains a fundamental component, a subsynchronous oscillation component, and their complementary supersynchronous oscillation component. For any phase p, we have:

[0079]

[0080] Complementary frequencies satisfy:

[0081] Using the aforementioned fundamental synchronous rotating reference angle, after rotational coordinate transformation, the voltage and current components are:

[0082]

[0083]

[0084]

[0085] Substituting the dominant oscillation component into the oscillation diffusion energy model, we obtain:

[0086] The energy diffusion rate of the oscillation considering frequency coupling is:

[0087] in, , and , These represent the voltage amplitude and initial phase of the subsynchronous and supersynchronous oscillation components, respectively. , and , These represent the current amplitude and initial phase of the subsynchronous and supersynchronous oscillation components, respectively. This index fully incorporates the coupling effect between the subsynchronous and supersynchronous oscillation components, reflecting both the energy interaction characteristics between different frequency components and accurately quantifying the diffusion intensity and direction of oscillation energy.

[0088] Within a high-oscillation-risk cut set, for each dominant oscillation mode, the oscillation energy diffusion rate of each cut set element is calculated by combining the improved least squares-rotation invariant technique with the oscillation component extraction results. When the oscillation energy diffusion rate of a certain cut set element is positive, it indicates that oscillation energy flows out of the element, and the element provides negative damping to the external system, thus being identified as an oscillation source. When the oscillation energy diffusion rate is negative, it indicates that oscillation energy flows into the element, and the element provides positive damping to the external system, thus being identified as an energy dissipation unit. When the oscillation energy diffusion rate is approximately zero, it is identified as a neutral unit of the corresponding mode. Based on the differences in the sign and amplitude of the oscillation energy diffusion rate of each cut set element under different dominant modes, the oscillation source of each mode is determined, and the differences in the damping effect of the synchronous generator under different modes are revealed, as shown in Tables 1 and 2.

[0089]

[0090] Table 1. Oscillation energy diffusion rate under the dominant mode

[0091] Table 2. Oscillation energy diffusion rate under the dominant mode considering frequency coupling Table 1 shows the oscillation energy diffusion rate of a four-unit, two-zone renewable energy grid-connected system under the dominant mode, used for locating the system oscillation source. As can be seen from the figure, under the 15Hz dominant mode, the wind field is: According to the definition of energy diffusivity, This indicates that oscillating energy flows out from WF, meaning WF provides negative damping to the system and is the oscillation source in the 15Hz mode; generators G4, G3, and node 11: All values ​​are negative (-0.664, -0.0007, -0.156), indicating that oscillating energy flows into these elements, corresponding to positive damping characteristics. They are energy dissipation units for oscillating energy. The oscillating energy diffusion rate of G3 is approximately -0.0007, indicating that its damping contribution to the 15Hz mode is extremely weak, approximating a "neutral" unit. Wind field under the dominant 57Hz mode: It also exhibits energy outflow characteristics and is an oscillation source in the 57Hz mode, and its The amplitude is greater than 15Hz in the mode, indicating that WF has a stronger negative damping contribution in this mode; generators G4, G3 and node 11: All values ​​are negative (-0.448, -0.296, -0.394), indicating energy inflow and acting as a positive damping unit, representing the energy dissipation unit of this mode. Through... The signs and amplitudes of the data clarify that the wind field (WF) is the common oscillation source of the two dominant modes at 15Hz and 57Hz, while the conventional generators (G3, G4) and node 11 are the damping units of the system; under different modes... The difference in amplitude reflects the different negative damping contribution strength of WF in each oscillation mode.

[0092] Table 2 shows the oscillation energy diffusion rate of the dominant mode considering frequency coupling, used to locate the oscillation source in a 10-unit, 39-node large-scale system for new energy grid connection. As can be seen from the figure, the oscillation energy diffusion rate of WF is... This indicates that it outputs oscillatory energy to the system in the sub-supersynchronous oscillation of mode 1, and is the core oscillation source; while the synchronous generator unit , The oscillation energy diffusion rate of node 11 is negative, indicating energy inflow and positive damping, making it an energy dissipation element. The oscillation energy diffusion rate of node 11 is approximately 0.0036, indicating its extremely weak damping contribution to this mode, approximating a "neutral" element. This yields oscillation source localization results consistent with simulation settings and the sign of the diffusion rate. Compared to judging solely by a single frequency, considering frequency coupling avoids splitting paired oscillation energies into independent source points, improving the localization integrity and engineering reliability under complex coupling conditions.

[0093] Step S5: Output the source tracing results; Output the source of oscillation, dominant oscillation mode, high oscillation risk cut set, propagation path, and oscillation energy diffusion rate.

[0094] The identification results can be output through a graphical interface, data file, monitoring platform interface, or online analysis platform on the dispatch side, and can be used for oscillation source tracing, operation risk assessment, and decision support in new energy grid-connected power systems.

[0095] An alarm mechanism based on preset rules is triggered: when the oscillation energy diffusion rate of the located oscillation source exceeds the preset alarm threshold, an alarm signal is triggered to notify the operators to intervene.

[0096] Based on the same inventive concept, embodiments of the present invention also provide a sub-supersynchronous oscillation tracing system based on response data and diffusion energy, such as... Figure 6 As shown, it includes: The data acquisition module is used to acquire measurement response data and network topology connections from multiple monitoring nodes in the power system. The data preprocessing and mode identification module is used to preprocess the measurement response data to obtain an oscillation response sequence, and to identify the dominant oscillation mode using time-frequency analysis methods. The branch energy analysis and cut set partitioning module is used to construct the branch modal oscillation energy model, calculate the branch oscillation energy sharing degree of each branch, and partition the high-risk cut set of oscillation based on the branch oscillation energy sharing degree. The oscillation source localization module is used to construct an oscillation diffusion energy model, calculate the oscillation energy diffusion rate of each cut set element within the high-risk oscillation cut set, and locate the oscillation source according to the sign of the oscillation energy diffusion rate. The results output module is used to output the source of oscillation, dominant oscillation mode, high oscillation risk cut set, and oscillation energy diffusion rate.

[0097] Since the principle behind the problem solved by this system is similar to that of the aforementioned methods, the implementation of this system can be found in the implementation of the aforementioned methods, and the repetitive parts will not be repeated.

[0098] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for tracing the source of sub-supersynchronous oscillations based on response data and diffusion energy, characterized in that, Includes the following steps: S1. Acquire the measurement response data and network topology connection relationship of multiple monitoring nodes in the power system, and preprocess the wide-area measurement response data; S2. Based on the oscillation response sequences of each node and branch obtained after preprocessing, the oscillation mode parameters of the oscillation response sequences are identified using time-frequency analysis methods, and the dominant oscillation mode in the target sub-supersynchronous frequency band is selected from the identified oscillation modes. S3. Based on the active power, reactive power, voltage amplitude and voltage phase angle at both ends of each branch, construct a branch modal oscillation energy model and calculate the branch oscillation energy sharing degree of each branch; sort the branches according to the branch oscillation energy sharing degree, select the branches that carry oscillation energy higher than a preset threshold, and traverse the network topology based on the selected branches to divide the high-risk cut set of oscillation. S4. Based on the instantaneous energy theory, construct an oscillation diffusion energy model. For each dominant oscillation mode, calculate the oscillation energy diffusion rate of each cut set element in the high-risk cut set of the oscillation, and locate the oscillation source according to the sign of the oscillation energy diffusion rate. S5, the source of the output oscillation, the dominant oscillation mode, the high oscillation risk cut set, and the oscillation energy diffusion rate.

2. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 1, characterized in that, In step S1, the measurement response data includes at least one of the following: voltage, branch current, active power, reactive power, and voltage phase angle information of each bus and branch node. The monitoring nodes include at least one of the following: wind farm grid connection point, photovoltaic power station grid connection point, synchronous generator endpoint, flexible transmission device interface, and nodes at both ends of transmission line; The measurement response data originates from at least one of the following: phasor measurement unit, wide-area measurement system, field monitoring system, or fault recorder.

3. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 1, characterized in that, The oscillation mode parameter identification in step S2 includes: performing time-frequency domain analysis on the oscillation response sequence using windowed interpolation-type fast Fourier transform, and identifying the frequency, amplitude, and phase parameters of each oscillation mode by combining improved least squares-rotation invariant technique.

4. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 1, characterized in that, In step S3, the method for constructing the branch modal oscillation energy model is as follows: for any branch, the oscillation energy flowing into the branch from both ends of the branch is expressed as the sum of the integral of active power with respect to the voltage phase angle and the integral of reactive power with respect to the logarithm of the voltage amplitude. The electrical quantities at both ends of the branch are decomposed into steady-state components and oscillation components of each oscillation mode. The oscillation energy of the dominant oscillation mode is much greater than that of the non-dominant mode and the steady-state value of the electrical quantity is much greater than that of the oscillation quantity. The oscillation energy of the branch is then expressed as the superposition of the energy components of each dominant mode and a constant term.

5. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 4, characterized in that, The calculation method for the branch oscillation energy sharing is as follows: the ratio of the branch oscillation energy in a certain dominant mode to the sum of the oscillation energy of all branches in the entire network in that mode is defined as the branch oscillation energy sharing in that mode.

6. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 1, characterized in that, In step S4, the method for constructing the oscillation diffusion energy model includes: defining the oscillation diffusion energy flowing out of the node from the external system as the instantaneous energy integral of the branch current and the node voltage in a two-phase rotating coordinate system, and stipulating that the oscillation energy flowing out of the node is positive and the oscillation energy flowing into the node is negative; By substituting the voltage and current into the oscillation diffusion energy expression after coordinate transformation, the coefficient of the non-periodic energy term is extracted and defined as the oscillation energy diffusion rate.

7. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 6, characterized in that, When subsynchronous frequency coupling is not considered, the oscillation energy diffusion rate is determined by the voltage amplitude, current amplitude and initial phase of the oscillation component; When considering subsynchronous and supersynchronous frequency coupling, the oscillation energy diffusion rate incorporates the voltage amplitude, current amplitude, and initial phase of both the subsynchronous and supersynchronous oscillation components to characterize the effect of the coupling between the subsynchronous and supersynchronous oscillation components on energy diffusion.

8. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 1, characterized in that, In step S4, the specific criterion for locating the oscillation source based on the sign of the oscillation energy diffusion rate is as follows: When the oscillation energy diffusion rate is positive, the element is determined to be the oscillation source under the corresponding dominant oscillation mode; When the oscillation energy diffusion rate is negative, the element is determined to be an energy dissipation unit; When the oscillation energy diffusion rate approaches zero, the element is determined to be a neutral unit.

9. The sub-supersynchronous oscillation source tracing method based on response data and diffusion energy as described in claim 1, characterized in that, Step S5 also includes: determining the differences in oscillation sources under different modes based on the differences in the sign and amplitude of the oscillation energy diffusion rate of each cut-set element under each dominant mode, and revealing the differences in the damping effect of the synchronous generator under different oscillation modes; The tracing results of oscillation source, dominant oscillation mode, high-risk cut set of oscillation, and oscillation energy diffusion rate are output through graphical interface, data file, or monitoring platform interface.

10. A sub-supersynchronous oscillation tracing system based on response data and diffusion energy, characterized in that, include: The data acquisition module is used to acquire measurement response data and network topology connections from multiple monitoring nodes in the power system. The data preprocessing and mode identification module is used to preprocess the measurement response data to obtain an oscillation response sequence, and to identify the dominant oscillation mode using time-frequency analysis methods. The branch energy analysis and cut set partitioning module is used to construct the branch modal oscillation energy model, calculate the branch oscillation energy sharing degree of each branch, and partition the high-risk cut set of oscillation based on the branch oscillation energy sharing degree. The oscillation source localization module is used to construct an oscillation diffusion energy model, calculate the oscillation energy diffusion rate of each cut set element within the high-risk oscillation cut set, and locate the oscillation source according to the sign of the oscillation energy diffusion rate. The results output module is used to output the source of oscillation, dominant oscillation mode, high oscillation risk cut set, and oscillation energy diffusion rate.

Images