A method and system for estimating the frequency of a power system

Abstract

The application discloses a power system frequency estimation method and system, which comprises the following steps: obtaining a power grid model and remote signaling data, a generator rotating speed and a PMU frequency measurement, forming a node branch model; forming an extended node admittance matrix according to the node branch model, a generator internal reactance parameter and the remote signaling data; performing node reordering on the extended node admittance matrix; forming a factor table of the reordered extended node admittance matrix by using a symbolic decomposition and a numerical decomposition method; calculating a frequency participation factor matrix F by using a multi-thread parallel technology; obtaining a reduced frequency participation factor matrix F' according to the size of element values in the frequency participation factor matrix F and a set threshold value; and calculating a node frequency by using a multi-thread parallel method. The application overcomes the shortcoming that the inertia center frequency of a power grid can only calculate an average frequency of the power grid, and has important reference significance for guiding the selection of a power generator frequency point monitored by a dispatching center.

Description

Technical Field

[0001] This invention belongs to the field of power system automation technology, and relates to a method and system for estimating power system frequency. Background Technology

[0002] With the large-scale integration of wind and solar renewable energy, centralized / decentralized wind and solar power units, which can provide frequency regulation capabilities, are increasingly replacing traditional synchronous generator units. This has led to a decrease in the inertia level of the power system, and the temporal and spatial characteristics of the grid frequency are becoming increasingly pronounced. When a system accident occurs, and the grid node frequency and voltage change significantly, the frequency characteristics of each bus in the grid need to be accurately depicted for modeling and analysis purposes. Traditional frequency simulation analysis, based on inertial center frequency analysis methods, struggles to reproduce indicators such as node frequency deviation and frequency change rate after disturbances in the new power system. The resulting frequency analysis results lack accuracy and fail to account for the grid-connected performance of renewable energy generation equipment. This approach is no longer suitable for analyzing the new characteristics of high-proportion renewable energy power systems, and power system frequency analysis and control are facing severe challenges. Simultaneously, renewable energy uses additional control methods to provide virtual inertia to the system, affecting the system's inertia response process. Therefore, analyzing the influence factors / weights of different power generation resources on the grid node frequency in the new power system is becoming increasingly urgent.

[0003] Existing frequency estimation methods include: Patent CN114552572A provides a photovoltaic-supported grid frequency method and device based on optimized control and predictive tracking, belonging to the field of grid support technology. This method is based on an output optimization model of the frequency of the grid-connected photovoltaic-supported AC system, obtains the optimal output curve of photovoltaic active power and converts it into the optimal output curve of active current, predicts the photovoltaic active current output within a time window, and selects the switching control signal when the predicted photovoltaic active current is closest to the optimal active current output curve as the optimal control quantity to achieve predictive optimization tracking of photovoltaic active current. Patent CN118281907A provides a frequency regulation method, device and computer-readable storage medium for a wind-storage system, belonging to the field of wind-storage system technology. This method establishes a frequency regulation simulation model of the wind-storage system, and determines the target control sequence at the current moment to regulate the frequency of the wind-storage system based on the frequency regulation simulation model, objective function and constraints. Analysis shows that none of the above methods describe a method for estimating the node frequency of the power system. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention discloses a method and system for estimating power system frequencies. This method estimates the frequencies of nodes in the power grid based on the susceptance parameters of power lines, transformers, and generators, real-time topology connections, generator speed measurements, and PMU frequency measurements. The method considers the electrical distance between nodes and frequency measurement points and estimates node frequencies through a linear combination of generator speed and PMU frequency measurements, yielding an analytical expression for the bus node frequency. This more accurately reflects the local frequencies at each bus node in the power grid, precisely characterizes the spatiotemporal distribution characteristics of power grid frequencies, and provides guidance for frequency monitoring, analysis, and control in new power systems.

[0005] To achieve the above-mentioned objectives, the present invention adopts the following technical solution.

[0006] On one hand, this invention discloses a method for estimating the frequency of a power system, the method comprising the following steps:

[0007] Step 1: Obtain the remote signaling status of the target power grid model and real-time monitoring system, as well as generator speed measurement data and PMU frequency measurement data. Determine whether the current remote signaling status of the target power grid has changed relative to the previous cycle. If it has changed, proceed to Step 2: Recalculate the frequency participation factor matrix of the target power grid in the current cycle. Otherwise, directly use the frequency participation factor matrix of the target power grid in the previous calculation cycle and proceed to Step 4.

[0008] Step 2: Perform topology analysis based on the power grid model and current remote signaling status to form a node branch model. Based on the node branch model and the reactance parameters of the online generators, form an extended node susceptance matrix B′. BB And decompose to obtain the extended nodal susceptance matrix B′ BB LDL factor table T Where L is a triangular matrix, L T Let L be the transpose of L, and D be a diagonal matrix;

[0009] Step 3: Calculate the frequency participation factor matrix F and obtain the reduced frequency participation factor matrix F′;

[0010] Step 4: Generate the speed vector ω based on the acquired generator speed measurement and PMU frequency measurement data. G ;

[0011] Step 5, based on ω B =F′ω G Calculate the node frequency ω in the target power grid B ;

[0012] Step 6: Based on the reduced frequency participation factor matrix F′, obtain the set of generators that have the greatest impact on the frequency of all nodes in the target power grid, and the set of nodes where the generators that have the greatest impact on the frequency of a certain node are located.

[0013] More preferably,

[0014] In step 2, an extended nodal susceptance matrix B′ is formed based on the nodal branch model and the reactance parameters of the online generator. BB ,include:

[0015] Based on the nodal susceptance matrix B formed by the inter-node branch reactances. BB Based on the total network susceptance matrix B formed by the internal reactance of the synchronous generator. BS and only contains the generator internal susceptance matrix B formed based on the generator's internal reactance. BG ;

[0016] According to B BB and B BS Forming an extended nodal susceptance matrix B′ BB , where B′ BB =B BB +B BS .

[0017] More preferably,

[0018] At nodes connected to generators, the nodal susceptance matrix B is corrected based on the generator's internal reactance. BB This directly forms B′ BB ;

[0019] When considering the frequency measurement value obtained by the PMU, the node where the PMU frequency measurement is located is calculated as the generator node. BG .

[0020] More preferably,

[0021] In step 2, the extended nodal susceptance matrix B′ is... BB Before decomposition, the extended nodal susceptance matrix B′ is... BB Reorder the node numbers to reduce the expanded node susceptance matrix B′. BB The number of non-zero elements filled in the factorization process.

[0022] More preferably,

[0023] In step 2, symbolic decomposition and numerical decomposition are used to form the extended nodal susceptance matrix B′. BB Factor table.

[0024] More preferably,

[0025] In step 3, the extended node susceptance matrix B′ is used. BB LDL factor table T and generator internal susceptance matrix B BG According to (LDL) T F = B BG The frequency participation factor matrix F is calculated using multi-threaded parallel technology based on OpenMP.

[0026] Based on the magnitude of the element values ​​in the frequency participation factor matrix F and the set threshold value, the node generators in F that have a small impact on the node frequency are ignored, resulting in the reduced frequency participation factor matrix F′.

[0027] More preferably, the reduced frequency participation factor matrix F′ is obtained by the first method, that is, arranging all elements in each row of the frequency participation factor matrix F in descending order, taking the first k largest elements of each row of the sorted matrix, setting the remaining elements to zero, and updating the values ​​in the matrix to obtain the reduced frequency participation factor matrix F′, where k is an integer value less than the number of columns in the frequency participation factor matrix F.

[0028] More preferably,

[0029] The reduced frequency participation factor matrix F′ is obtained using the second method, which involves sequentially finding the largest element F in each row of the frequency participation factor matrix F. max According to the set threshold value |f i |<αF max Rowwise, elements f in matrix F i After filtering, α is a parameter between (0, 1.0), and the values ​​in F are updated to obtain the reduced frequency participation factor matrix F′.

[0030] More preferably,

[0031] In step 4, when considering the frequency measurement value obtained from the PMU, the acquired PMU frequency measurement is added to the rotational speed vector ω. G .

[0032] More preferably,

[0033] In step 5, when it is necessary to calculate the frequency of all or multiple nodes in the power grid, the node frequency ω is calculated using matrix-vector multiplication based on OpenMP's multithreaded parallel technology. B .

[0034] More preferably,

[0035] In step 6, the number of effective non-zero elements in each column of the reduced frequency participation factor matrix F′ is counted by column, and the effective non-zero elements are averaged and arranged in descending order. The column with the larger average value corresponds to the greater influence of the generator in that column on the frequency of each node in the power grid.

[0036] In the reduced frequency participation factor matrix F′, the generator in the column with the largest value in each row has the greatest impact on the frequency of the corresponding node in that row, while the generator in the column with the smaller value has a smaller impact on the frequency of the corresponding node in that row.

[0037] On the other hand, the present invention also discloses a power system frequency estimation system based on the aforementioned estimation method, comprising:

[0038] The module includes a power grid model acquisition and parameter measurement module, a remote signaling status judgment module, an extended node susceptance matrix factor table calculation module, a reduced frequency participation factor matrix calculation module, a generator speed vector acquisition module, a power grid node frequency calculation module, and a generator impact on node frequency analysis module.

[0039] The power grid model acquisition and parameter measurement module acquires the target power grid model and the remote signaling status of the real-time monitoring system, as well as generator speed measurement data and PMU frequency measurement data.

[0040] The remote signaling status determination module determines whether the current remote signaling status of the target power grid has changed relative to the previous cycle;

[0041] If the remote signaling status changes, then,

[0042] The extended nodal susceptance matrix factor table calculation module calculates the extended nodal susceptance matrix B′ based on the nodal branch model. BB LDL factor table T ;

[0043] The reduced frequency participation factor matrix calculation module calculates the frequency participation factor matrix F and obtains the reduced frequency participation factor matrix F′.

[0044] The generator speed vector acquisition module generates a speed vector ω based on the acquired generator speed measurement and PMU frequency measurement data. G ;

[0045] The power grid node frequency calculation module calculates the node frequency ω based on the reduced frequency participation factor and the rotational speed vector. B ;

[0046] The generator impact analysis module analyzes and obtains the set of generators that have the greatest impact on the frequency of all nodes in the power grid, as well as the set of nodes where the generator with the greatest impact on the frequency of a certain node is located.

[0047] If the remote signaling status remains unchanged, then,

[0048] The power grid node frequency calculation module directly calculates the node frequency ω based on the frequency participation factor matrix obtained from the previous cycle and the generator speed vector obtained in this cycle. B Then, the generator-to-node-frequency impact analysis module analyzes and obtains the set of generators that have the greatest impact on the frequency of all nodes in the power grid, as well as the set of nodes where the generator with the greatest impact on the frequency of a certain node is located.

[0049] The present invention also claims protection for an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the computer program, when loaded onto the processor, implements the method for estimating the power system frequency.

[0050] A computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the method for estimating the frequency of a power system.

[0051] The beneficial effects of this invention are compared with those of the prior art:

[0052] 1. This invention realizes the estimation of node frequencies in power systems. The method is based on the susceptance parameters of power grid lines, transformers and generators, and the real-time topology connection relationship. It estimates the node frequencies by linearly combining generator speed and PMU frequency measurements. It takes into account the power grid structure, generator parameters and the electrical distance between the frequency measurement point and the node. It overcomes the shortcomings of the power grid inertial center frequency, which can only calculate the average frequency of the power grid. It can more accurately reflect the local frequency at each node of the power grid and accurately characterize the spatiotemporal distribution characteristics of the power grid frequency.

[0053] 2. This invention enables the calculation of the frequency participation factor matrix. By examining the values ​​of each row in the reduced frequency participation factor matrix F′, the node containing the generator that has the greatest impact on the frequency of a given node can be identified. Simultaneously, by counting the number of valid non-zero elements in each column of the reduced frequency participation factor matrix F′ and calculating the average value, the generator that has the greatest impact on the frequency of all nodes in the power grid can be obtained from the perspective of the average value. From the perspective of power grid operation monitoring, it is necessary to accurately measure the rotational speed of the generators at these nodes or accurately record the frequencies of these nodes based on PMU devices, which has significant guiding significance for power grid frequency monitoring.

[0054] 3. This invention effectively reduces the non-zero element density of the frequency participation factor matrix by reducing its size. The first method ensures the accuracy of frequency estimation at each node in the power grid, making it particularly suitable for dynamic or real-time frequency monitoring, control, and applications. The second method is suitable for simulation analysis and calculation research of large-scale power systems, offering high computational efficiency. In practical applications, the two methods can also be combined to reduce computational load while maintaining accuracy.

[0055] 4. The calculation of the frequency participation factor matrix of this invention can be extended to nodes containing asynchronous generators, such as nodes with PMU device frequency acquisition. Meanwhile, new energy power generation equipment connected to the grid via power electronics interfaces commonly uses phase-locked loop (PLL) technology to acquire its node frequency, which can also be added to the calculation of the frequency participation factor matrix. This patent has scalability. Attached Figure Description

[0056] Figure 1 This is a flowchart illustrating an embodiment of the power system frequency estimation method of the present invention;

[0057] Figure 2 This is a schematic diagram of the non-zero element structure analysis of an L-matrix;

[0058] Figure 3 This is a schematic diagram of the j-column factorization of LU. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0060] This application discloses a method for estimating the frequency of a power system, the method comprising the following steps:

[0061] Step 1: Obtain the remote signaling status of the target power grid model and the real-time monitoring system, as well as the generator speed measurement data and PMU frequency measurement data. Determine whether the current remote signaling status of the target power grid has changed relative to the previous cycle. If it has changed, proceed to Step 2 and recalculate the frequency participation factor matrix of the target power grid in the current cycle. Otherwise, directly use the frequency participation factor matrix of the target power grid in the previous cycle and proceed to Step 4.

[0062] Step 2: Perform topology analysis based on the power grid model and current remote signaling status to form a node branch model. Based on the node branch model and the reactance parameters of the online generators, form an extended node susceptance matrix B′. BB And decompose to obtain the extended nodal susceptance matrix B′ BB LDL factor table T Where L is a triangular matrix, L T Let L be the transpose of L, and D be a diagonal matrix;

[0063] An extended nodal susceptance matrix B′ is formed based on the nodal branch model and the reactance parameters of the generator operating online. BB ,include:

[0064] Based on the nodal susceptance matrix B formed by the inter-node branch reactances. BB Based on the total network susceptance matrix B formed by the internal reactance of the synchronous generator. BS and only contains the generator internal susceptance matrix B formed based on the generator's internal reactance. BG ;

[0065] According to B BB and B BS Forming an extended nodal susceptance matrix B′ BB , where B′ BB =B BB +B BS .

[0066] At nodes connected to generators, the nodal susceptance matrix B is corrected based on the generator's internal reactance. BB This directly forms B′ BB When considering the frequency measurement value obtained from the PMU, calculate B according to the generator node, where the PMU frequency measurement is located. BG .

[0067] In the extended nodal susceptance matrix B′ BB Before decomposition, the extended nodal susceptance matrix B′ is sorted using Tinney-1, Tinney-2, or an approximate minimum degree sorting algorithm. BB Reorder the node numbers to reduce the expanded node susceptance matrix B′. BB The number of non-zero elements filled in the factorization process. An extended nodal susceptance matrix B′ is formed using symbolic and numerical decomposition. BB Factor table.

[0068] Step 3: Calculate the frequency participation factor matrix F and obtain the reduced frequency participation factor matrix F′;

[0069] The extended nodal susceptance matrix B′ BB LDL factor table T and generator internal susceptance matrix BBG According to (LDL) T F = B BG The frequency participation factor matrix F is calculated using multi-threaded parallel technology based on OpenMP.

[0070] Based on the magnitude of the element values ​​in the frequency participation factor matrix F and the set threshold value, the node generators in F that have a small impact on the node frequency are ignored, resulting in the reduced frequency participation factor matrix F′.

[0071] In a preferred embodiment of the present invention, the reduced frequency participation factor matrix F′ can be obtained by using the following two methods, or a combination of the two methods:

[0072] The first method involves arranging all elements in each row of the frequency participation factor matrix F in descending order, taking the first k largest elements of each row of the sorted matrix, setting the remaining elements to zero, and updating the values ​​in the matrix to obtain the reduced frequency participation factor matrix F′, where k is an integer value less than the number of columns in the frequency participation factor matrix F.

[0073] The second method involves finding the largest element F in each row of the frequency-participating factor matrix F, row by row. max According to the set threshold value |f i |<αF max Rowwise, elements f in matrix F i After filtering, α is a parameter between (0, 1.0), and the values ​​in F are updated to obtain the reduced frequency participation factor matrix F′.

[0074] Step 4: Generate the speed vector ω based on the acquired generator speed measurement and PMU frequency measurement data. G ;

[0075] When considering the frequency measurements obtained from the PMU, the acquired PMU frequency measurements are added to the rotational speed vector ω. G .

[0076] Step 5: Based on ω B =F′ω G Calculate the node frequency ω B ;

[0077] When it is necessary to calculate the frequency of all or multiple nodes in a power grid, matrix-vector multiplication is used to calculate the node frequency ω based on OpenMP's multithreaded parallel technology. B .

[0078] Step 6: Obtain the set of generators that have the greatest impact on the frequency of all nodes in the power grid, and the set of nodes containing the generators that have the greatest impact on the frequency of a certain node;

[0079] The number of effective non-zero elements in each column of the reduced frequency participation factor matrix F′ is counted column by column, and the effective non-zero elements are averaged and arranged in descending order. From the perspective of average value, the set of generators that have the greatest impact on the frequency of all nodes in the power grid is obtained. For example, if the average value of the i-th column is the largest, then the vector ω... G The i-th generator in the grid has the greatest combined impact on the frequency of all nodes in the grid. Based on this logic, we can obtain the set of generators that have the greatest impact on the frequency of all nodes in the grid.

[0080] From a row perspective, the values ​​in each row of the reduced frequency participation factor matrix F′ determine the set of nodes containing the generators that have the greatest impact on the frequency of a given node. The generators in the nodes corresponding to the nodes with the largest values ​​in each row of F′ have the greatest impact on the node frequency, and the smaller the values, the smaller the impact of the generators in the corresponding nodes. For example, if the value in the i-th row and j-th column of F′ is the largest, it represents the vector ω. G The j-th generator has the greatest impact on the frequency of node i.

[0081] This application also protects a power system frequency estimation system based on the aforementioned estimation method, comprising:

[0082] The module includes a power grid model acquisition and parameter measurement module, a remote signaling status judgment module, an extended node susceptance matrix factor table calculation module, a reduced frequency participation factor matrix calculation module, a generator speed vector acquisition module, a power grid node frequency calculation module, and a generator impact on node frequency analysis module.

[0083] The power grid model acquisition and parameter measurement module acquires the target power grid model and the remote signaling status of the real-time monitoring system, as well as generator speed measurement data and PMU frequency measurement data.

[0084] The remote signaling status determination module determines whether the current remote signaling status of the target power grid has changed relative to the previous cycle;

[0085] If the remote signaling status changes, then,

[0086] The extended nodal susceptance matrix factor table calculation module calculates the extended nodal susceptance matrix B′ based on the nodal branch model. BB LDL factor table T ;

[0087] The reduced frequency participation factor matrix calculation module calculates the frequency participation factor matrix F and obtains the reduced frequency participation factor matrix F′.

[0088] The generator speed vector acquisition module generates a speed vector ω based on the acquired generator speed measurement and PMU frequency measurement data. G ;

[0089] The power grid node frequency calculation module calculates the node frequency ω based on the reduced frequency participation factor and the rotational speed vector. B ;

[0090] The generator impact analysis module analyzes and obtains the set of generators that have the greatest impact on the frequency of all nodes in the power grid, as well as the set of nodes where the generator with the greatest impact on the frequency of a certain node is located.

[0091] If the remote signaling status remains unchanged, then,

[0092] The power grid node frequency calculation module directly calculates the node frequency ω based on the frequency participation factor matrix obtained from the previous cycle and the generator speed vector obtained in this cycle. B Then, the generator-to-node-frequency impact analysis module analyzes and obtains the set of generators that have the greatest impact on the frequency of all nodes in the power grid, as well as the set of nodes where the generator with the greatest impact on the frequency of a certain node is located.

[0093] Example 1:

[0094] To more clearly illustrate the present invention, the following methods will be used: Figure 1 The illustrated embodiment 1 details a specific scheme for a method of estimating power system frequency. In a preferred but non-limiting embodiment of the present invention, the method includes the following:

[0095] S1. Obtain the power grid model and remote signaling status of the real-time monitoring system, as well as generator speed measurements and PMU frequency measurements. Based on the obtained model and remote signaling status, perform topology analysis to form a node-branch model. Let the number of computational nodes after the power grid topology analysis be n, the number of branches be b, and the number of generators be m.

[0096] The acquisition method is based on the unified data service of the scheduling automation system, which ensures that the acquisition method is simple and efficient.

[0097] This step first acquires the telemetry and tele-signaling status of the power grid model and real-time monitoring system, as well as generator speed measurements and PMU frequency measurements. The acquired power grid model and tele-signaling data include the power grid line reactance parameters, transformer reactance parameters, generator speed and internal reactance parameters, topological connections between equipment, and the tele-signaling status of switches required for frequency calculation. If the frequency measurements collected by the Phasor Measurement Unit (PMU) devices in the power grid are available and PMU frequency measurements are considered, then PMU frequency measurement data are also acquired.

[0098] Secondly, based on the obtained power grid model and the remote signaling status of the switches and disconnectors, power grid topology analysis is performed using either a global or local topology approach:

[0099] When the changes in the current remote signaling status and the remote signaling status used in the previous frequency calculation do not exceed the set range, topology analysis is performed using a local topology approach.

[0100] When the change between the current remote signaling status and the previous calculated remote signaling status exceeds the set range, a global topology analysis is performed to form a node branch model.

[0101] S2. Form the nodal susceptance matrix B based on the nodal branch model and the reactance parameters of the generator operating online. BB B BS and B BG B BB B is a high-dimensional sparse nodal susceptance matrix (n×n) formed based on the reactance (ignoring resistance) of the branches between nodes in power flow calculations. BS To form a high-dimensional sparse susceptance matrix (n×n dimensions) based on the internal reactance of the synchronous generator, B BG This consists only of a high-dimensional sparse susceptance matrix (n×m dimension) formed based on the generator's internal reactance. The nodal susceptance matrix B is formed using the branch superposition method based on the reactances of inter-node lines and transformer branches, and the generator's internal reactance. BB B BS and B BG .

[0102] In the process of forming the node susceptance matrix using the branch superposition method, a balanced binary red-black tree is used to store the intermediate calculation process and the final node susceptance matrix. In the specific implementation, the associative container map container provided by C++STL (Standard Template Library) is used for storage.

[0103] The structure `typedef std::map<int,float> SparseRowVal` is used to store the non-zero column index and value set of a row. The column index is represented by the key index of the int type (integer variable) in the map, and the value is represented by the float type value corresponding to the key index.

[0104] The structure is defined as: typedef std::map<int, SparseRowVal> SparseMatrixValMap is used to store the non-zero elements of the sparse matrix nodes, and the row number is represented by an int key index.

[0105] Traverse all branches in the power grid, starting from k and going up to b in sequence, based on the branch reactance parameter x. k and the first and last calculation node numbers i and j corresponding to B BB The value of .

[0106]

[0107] If we consider the frequency measurements obtained from the PMU device, calculate B. BG The node where the PMU frequency measurement is located needs to be processed as a generator node in order to include the PMU frequency measurement in the calculation.

[0108] S3, according to B BB and B BS Calculate the extended nodal susceptance matrix B′ BB , where B′ BB =B BB +B BS At nodes connected to generators, the nodal susceptance matrix B is corrected based on the generator's internal reactance. BB This directly forms B′ BB In practice, there is no need to perform the addition operation of two high-dimensional matrices.

[0109] S4. To reduce the susceptance matrix B′ of the extended nodes BB The number of non-zero elements filled in the factorization process is determined by using Tinney-1, Tinney-2, or Approximate Minimum Degree (AMD) sorting algorithms on matrix B′. BB The node numbers are then reordered. After eliminating a vertex during the node sorting process, updating the degree of the vertex's adjacent vertices is the most time-consuming stage. In the approximate minimum degree sorting algorithm, instead of calculating the vertex's precision, the upper limit of the vertex's degree, which is easier to calculate, is used instead. The approximation of the vertex is used instead of the precision to reduce the sorting time.

[0110] During the node sorting process, it is only necessary to base it on the node susceptance matrix B′. BB The zero-element structure in China and Africa does not use B′. BB The value is numerically calculated to achieve node reordering. In the implementation, a global n-dimensional permutation matrix perm[] is designed. Regardless of whether Tinney-1, Tinney-2, or AMD methods are used, a global perm[] is formed. This facilitates the use of different node reordering methods for B′ based on the power grid scale. BB Perform node reordering.

[0111] Simultaneously, two global n-dimensional node index mapping tables, oldnew[] and newold[], are designed. newold[i] represents the new index of node i after topological analysis and subsequent reordering, while oldnew[j] represents the original index of node i after topological analysis. This avoids issues with matrix B′ after sorting. BB The values ​​of each row in the middle are rearranged.

[0112] S5. The extended nodal susceptance matrix B′ is formed by symbolic decomposition and numerical decomposition. BB Factor table.

[0113] The reordered extended node susceptance matrix B′ in S4 BB Using a symbolic decomposition method based on the G / P (Gilbert-Peierls) algorithm, B′ is completed sequentially from column 1 to column n. BB =LDL T Analysis of the non-zero row number set of each column in the lower triangular matrix L. The calculation analyzes one column at a time, returning the sparse structure Lp and the elimination tree Parent of L. A schematic diagram of the non-zero row number set analysis is shown below. Figure 2 As shown. Symbolic analysis is based on matrix B′ BB The set of row numbers containing the non-zero elements in the k-th column is used to calculate the LDL. T The set of non-zero row numbers in the k-th column after decomposition.

[0114] The reordered extended node susceptance matrix B′ in S4 BB Based on the left-look numerical decomposition analysis method, combined with the sparse structure Lp obtained from symbolic decomposition, the tree Parent and matrix B′ are eliminated. BB LDL is performed by iterating through columns k = 1, 2, ..., n, looking to the left. T Numerical decomposition yields the numerical value of L and the diagonal matrix D, thus completing the decomposition of B′. BB LDL T Factorization. Looking to the left, the LU numerical factorization method decomposes each column sequentially from column 1 to column n. That is, it calculates one column vector at a time, performs numerical calculations based on the sign analysis results of that column, and thus obtains the values ​​of all non-zero elements in column k. A schematic diagram of the factorization process is shown below. Figure 3 As shown.

[0115] S6, derived from the extended nodal susceptance matrix B′ BB LDL factor table T and B BG According to (LDL) T F = B BG The frequency participation factor matrix F is calculated using multi-threaded parallel technology based on OpenMP.

[0116] Since the calculations of the columns of the frequency participation factor matrix F are independent of each other, according to (LDL) T )[F1 F2 …F m ] = [B BG1 B BG2 … B BGm In this format, each column is independent.

[0117] (LDLT )F1B BG1

[0118] (LDL T F2B BG2

[0119] …

[0120] (LDL T )F m B BGm

[0121] The elements of the participation factor matrix are quickly computed using multithreaded parallel computing based on OpenMP. Here, F1 is the first column of the frequency participation factor matrix F, and the other columns are calculated sequentially.

[0122] S7, due to The F obtained from the calculation of a dense matrix is ​​also a dense matrix. For power grids (especially large power grids), generators far from a specified node will not significantly affect the frequency of that node. Many elements in matrix F have very small or negligible values, so it is necessary to ignore those generator nodes in F that have little impact on the node frequency. One of the following two methods is used to filter the elements of F to obtain the reduced frequency participation factor matrix F′.

[0123] Method M1: Sort all elements in each row of F in descending order, take the first k largest elements of each row of the sorted matrix F, set the remaining elements to zero, and update the values ​​in F to obtain the reduced frequency participation factor matrix F′.

[0124] Method M2: Find the largest element F in each row of F sequentially. max According to the set threshold value |d i |<αF max Rowwise, elements d in matrix F i After filtering, α is a parameter between (0, 1.0), and the values ​​in F are updated to obtain the reduced frequency participation factor matrix F′.

[0125] S8. Generate the speed vector ω based on the generator speed measurement obtained in S1. G If the frequency measurements acquired by the phasor measurement unit (PMU) in the power grid are available, the acquired PMU frequency measurements can also be added to the rotational speed vector ω. G If we consider the frequency measurement obtained from the PMU device, calculate B in steps 2-3 above. BG B′ BB The node where the PMU frequency measurement is located needs to be processed as a generator node in order to include the PMU frequency measurement in the calculation.

[0126] S9, according to ω B =F′ω G Calculate node frequencies. When it is necessary to calculate the frequencies of all or multiple nodes in the power grid, the OpenMP-based multithreaded parallel method is still used to quickly obtain the node frequencies ω through matrix-vector multiplication. B .

[0127] If the remote signaling status obtained in this calculation is unchanged from the previous remote signaling status, the node frequency is calculated directly using F′ formed in the previous calculation and the latest generator speed and PMU frequency measurements.

[0128] S10. Count the number of effective non-zero elements in each column of the reduced frequency participation factor matrix F′, and average the effective non-zero elements. Arrange them in descending order to obtain the set of generators that have the greatest impact on the frequency of all nodes in the power grid from the perspective of the average value. For example, if the average value of the i-th column is the largest, then the vector ω... G The i-th generator in the network has the greatest combined impact on the frequencies of all nodes. Based on this logic, we can obtain the set of generators that have the greatest impact on the frequencies of all nodes in the network. From a row perspective, the values ​​of each row in the reduced frequency participation factor matrix F′ can determine the set of nodes containing the generators that have the greatest impact on the frequency of a given node. The largest value in the i-th row and j-th column of F′ represents the vector ω. G The j-th generator has the greatest impact on the frequency of node i.

[0129] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0130] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0131] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0132] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for estimating the frequency of a power system, characterized in that, The method includes the following steps: Step 1: Obtain the remote signaling status of the target power grid model and real-time monitoring system, as well as generator speed measurement data and PMU frequency measurement data. Determine whether the current remote signaling status of the target power grid has changed relative to the previous calculation cycle. If it has changed, proceed to Step 2 and recalculate the frequency participation factor matrix of the target power grid in the current cycle. Otherwise, directly use the frequency participation factor matrix of the target power grid in the previous cycle and proceed to Step 4. Step 2: Perform topology analysis based on the power grid model and current remote signaling status to form a node branch model. Based on the node branch model and the reactance parameters of the online generators, form an extended node susceptance matrix. And decompose to obtain the extended nodal susceptance matrix. Factor table ,in, L It is a triangular matrix. for The transpose of the matrix, D It is a diagonal matrix; Step 3: From the extended node susceptance matrix Factor table and only includes the generator internal susceptance matrix formed based on the generator's internal reactance. according to Calculate the frequency participation factor matrix F And participate in the factor matrix according to frequency. F The magnitude of the element values ​​and the set threshold value are used to ignore node generators that have a small impact on the node frequency, resulting in a reduced frequency participation factor matrix. ; Step 4: Generate a speed vector based on the acquired generator speed measurement and PMU frequency measurement data. ; Step 5, according to Calculate the frequency of each node in the target power grid ; Step 6: Based on the reduced frequency participation factor matrix The set of generators that have the greatest impact on the frequency of all nodes in the target power grid, and the set of nodes containing the generators that have the greatest impact on the frequency of a certain node.

2. The method for estimating power system frequency according to claim 1, characterized in that: In step 2, an extended nodal susceptance matrix is ​​formed based on the nodal branch model and the reactance parameters of the online generator. ,include: Based on the nodal susceptance matrix formed by the inter-node branch reactance The full network susceptance matrix formed by the internal reactance of the synchronous generator and generator internal susceptance matrix ; according to and Forming an extended nodal susceptance matrix ,in .

3. The method for estimating power system frequency according to claim 2, characterized in that: At nodes connected to generators, the nodal susceptance matrix is ​​modified based on the generator's internal reactance. Thus directly forming ; When considering the frequency measurement value obtained from the PMU, the node where the PMU frequency measurement is located is calculated as a generator node. .

4. The method for estimating power system frequency according to claim 2, characterized in that: In step 2, the extended nodal susceptance matrix is... Before decomposition, the extended nodal susceptance matrix is... Reorder the node numbers to reduce the expansion node susceptance matrix. The number of non-zero elements filled in the factorization process.

5. The method for estimating the frequency of a power system according to claim 2 or 4, characterized in that: In step 2, symbolic decomposition and numerical decomposition are used to form an extended nodal susceptance matrix. Factor table.

6. The method for estimating the frequency of a power system according to claim 2 or 3, characterized in that: In step 3, the extended node susceptance matrix is ​​used. Factor table and generator internal susceptance matrix according to The frequency participation factor matrix is ​​calculated using multi-threaded parallel technology based on OpenMP. .

7. The method for estimating power system frequency according to claim 6, characterized in that: The frequency participation factor matrix is ​​arranged row by row. All elements in each row of the matrix are arranged in descending order. The first element of each row of the sorted matrix is ​​then taken. k Find the largest element, set the remaining elements to zero, update the values ​​in the matrix to obtain the reduced frequency in the factor matrix. , k For less than the frequency participation factor matrix The integer value of the number of columns.

8. The method for estimating power system frequency according to claim 6, characterized in that: Find the frequency participation factor matrix row by row. The largest element in each row According to the set threshold value Row pairs of matrices elements in Perform the screening. Update parameters between (0, 1.0) The value in the reduced frequency participation factor matrix .

9. The method for estimating the frequency of a power system according to claim 1, 7, or 8, characterized in that: In step 4, when considering the frequency measurement value obtained from the PMU, the acquired PMU frequency measurement is added to the rotational speed vector. .

10. The method for estimating power system frequency according to claim 9, characterized in that: In step 5, when it is necessary to calculate the frequency of all or multiple nodes in the power grid, matrix-vector multiplication is used to calculate the node frequency based on OpenMP's multithreaded parallel technology. .

11. The method for estimating power system frequency according to claim 1, characterized in that: In step 6, the frequency participation factor matrix is ​​reduced by column. The number of valid non-zero elements in each column is calculated, and the average of the valid non-zero elements is calculated. The columns are then arranged in descending order of their average values. The larger the average value in a column, the greater the impact of the generator in that column on the frequency of each node in the power grid. Reduced frequency participation factor matrix The generator in the column with the largest value in each row has the greatest impact on the frequency of the corresponding node in that row, while the generator in the column with the smaller value has a smaller impact on the frequency of the corresponding node in that row.

12. A power system frequency estimation system based on the estimation method according to any one of claims 1-11, characterized in that, include: The module includes a power grid model acquisition and parameter measurement module, a remote signaling status judgment module, an extended node susceptance matrix factor table calculation module, a reduced frequency participation factor matrix calculation module, a generator speed vector acquisition module, a power grid node frequency calculation module, and a generator impact on node frequency analysis module. The power grid model acquisition and parameter measurement module acquires the target power grid model and the remote signaling status of the real-time monitoring system, as well as generator speed measurement data and PMU frequency measurement data. The remote signaling status determination module determines whether the current remote signaling status of the target power grid has changed relative to the previous cycle; If the remote signaling status changes, then, The extended nodal susceptance matrix factor table calculation module calculates the extended nodal susceptance matrix based on the nodal branch model. Factor table ; The reduced frequency participation factor matrix calculation module calculates the frequency participation factor matrix. F And obtain the reduced frequency participation factor matrix ; The generator speed vector acquisition module generates a speed vector based on the acquired generator speed measurement and PMU frequency measurement data. ; The power grid node frequency calculation module calculates the frequency of each node based on the reduced frequency participation factor and the rotational speed vector. ; The generator impact analysis module analyzes and obtains the set of generators that have the greatest impact on the frequency of all nodes in the power grid, as well as the set of nodes where the generator with the greatest impact on the frequency of a certain node is located. If the remote signaling status remains unchanged, then, The power grid node frequency calculation module directly calculates the frequency of each node based on the frequency participation factor matrix obtained in the previous cycle and the generator speed vector obtained in this cycle. Then, the generator-to-node-frequency impact analysis module analyzes and obtains the set of generators that have the greatest impact on the frequency of all nodes in the power grid, as well as the set of nodes where the generator with the greatest impact on the frequency of a certain node is located.

13. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is loaded into the processor, it implements the method for estimating the power system frequency according to any one of claims 1-11.

14. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for estimating the power system frequency according to any one of claims 1-11.

Images