Power System Node Inertia Evaluation Method and System Based on Improved Vector Matching Method

Abstract

The present invention discloses a method and system for evaluating the inertia of power system nodes based on an improved vector matching method, which relates to the field of power system operation control, and includes: S1: Obtain the time-series data of each node in the power system and perform normalization processing on the data; S2: Represent the transfer function between the node frequency deviation and the active power deviation by a rational function containing unknown parameters, and use the vector matching method for parameter identification to calculate the unknown parameters of the rational function, so as to obtain the equivalent active power-efficiency transfer function; S3: Use the particle swarm optimization algorithm to improve the vector matching method, take the final fitting mean square error obtained by the vector matching method as the particle fitness, take the minimum fitness as the optimization goal, and find the target initial poles through particle swarm iteration. After the iteration ends, the optimal active power-frequency transfer function is obtained; S4: Obtain the inertia evaluation value according to the mathematical relationship between the optimal active power-frequency transfer function and the node inertia. The method improves the accuracy of inertia evaluation.

Description

Technical Field

[0001] The present invention relates to the technical field of power system operation control, and particularly relates to a method and system for evaluating the inertia of power system nodes based on an improved vector matching method. Background Art

[0002] The large-scale access of new energy has significantly changed the inertia characteristics of the new power system. In the area with concentrated access of new energy, the low-inertia characteristics are obvious, which is in sharp contrast to the high-inertia state in the synchronous power source cluster distribution area, resulting in a certain spatial distribution characteristic of inertia. When a fault occurs in the system, a large frequency deviation and rate of change of frequency will occur in the inertia-weak area, which is more likely to trigger the operation of under-frequency load shedding relays or (Rate of Change of Frequency, RoCoF) relays, and then exacerbate the risk of system power outage. In summary, accurately evaluating the inertia level of power system nodes is beneficial to guiding the grid connection of new energy and planning inertia compensation measures, and avoiding the system falling into the dilemma of ultra-low inertia operation.

[0003] At present, the methods for evaluating the inertia of power system nodes can be divided into two categories: offline evaluation and online evaluation methods:

[0004] Offline inertia evaluation generally deeply mines the system operation data after an accident, estimates the RoCoF based on the polynomial fitting of the frequency curve, and estimates the inertia in combination with the swing equation. However, it is difficult for such methods to eliminate the influence of the oscillation component of the frequency signal, so the evaluation accuracy is generally not high. At the same time, these methods also have some limitations: they usually assume that the magnitude of the active power disturbance and the accident occurrence time are known, which may introduce uncertainties in practical applications; the evaluation results highly depend on the quality of the measurement data, and low-quality data may lead to misjudgment; real-time monitoring and dynamic tracking cannot be achieved.

[0005] Online evaluation mostly adopts the idea of data-driven, and uses a dynamic identification model or a prediction model update based on machine learning to fit the actual system frequency dynamic response characteristics. However, such methods usually require accurate and timely data acquisition, rich measurement data, or parameter settings such as a relatively high model order for fitting, and may simultaneously have problems such as low accuracy, poor robustness, and large computational complexity. In addition, the complex power network structure and the interference of various factors may also affect the accuracy of the model.

[0006] Generally speaking, most of the existing node inertia evaluation methods have problems such as relying on large disturbance information, insufficient adaptability, and large data requirements. Therefore, there is an urgent need for a high-precision inertia evaluation method that is easier to implement and applicable to more working conditions. Summary of the Invention

[0007] The object of the present invention is to provide a method, system, device and medium for evaluating the inertia of power system nodes based on an improved vector matching method to solve the technical problems existing in the prior art.

[0008] The present invention is realized by the following technical solutions:

[0009] In the first aspect, a method for evaluating the inertia of power system nodes based on an improved vector matching method provided by an embodiment of the present invention includes:

[0010] Step S1: Obtain the frequency and power time series data of each node in the power system, and perform normalization processing on the time series data to obtain a frequency standard data set and a power standard data set respectively;

[0011] Step S2: Represent the transfer function between the node frequency deviation and the active power deviation by a rational function containing unknown parameters, and perform parameter identification using the vector matching method according to the time series data to calculate the unknown parameters of the rational function, and obtain an equivalent active-power - efficiency transfer function;

[0012] Step S3: Improve the vector matching method using the particle swarm optimization algorithm, take the final fitting mean square error obtained by the vector matching method as the particle fitness, take the minimum fitness as the optimization goal, and find the target initial poles through continuous iteration of the particle swarm. After the iteration ends, obtain the optimal active-power - frequency transfer function;

[0013] Step S4: Obtain the inertia evaluation value according to the mathematical relationship between the optimal active-power - frequency transfer function and the node inertia.

[0014] Further, the specific method for performing parameter identification using the vector matching method according to the time series data in step S2 to calculate the unknown parameters of the rational function includes:

[0015] Step S21: Set the maximum number of iterations and initialize the current number of iterations to 1;

[0016] Step S22: Set the initial poles and construct an auxiliary function;

[0017] Step S23: Set the vector matching conditions;

[0018] Step S24: Use a recursive filter and trapezoidal numerical integration to calculate the frequency convolution integral and power convolution integral of the node respectively;

[0019] Step S25: According to the set vector matching conditions, use the least squares method to fit and solve to obtain the unknown residues of the auxiliary function, and calculate the identified poles of the rational function by constructing a matrix;

[0020] Step S26: Perform residue identification of the rational function according to the identified poles;

[0021] Step S27: Determine whether the current iteration count is less than the maximum iteration count. If it is less, increment the current iteration count by 1, use the obtained identified poles as the new initial poles, and repeat Steps S22 - S27. If it is greater than or equal to, obtain the approximate expression of the rational function based on the currently calculated identified poles and residues, use this expression to calculate the fitting value of the node frequency with the node power, and calculate the mean square error based on the difference between the fitting value and the actually obtained node frequency.

[0022] Further, the specific method for improving the vector matching method using the particle swarm optimization algorithm in Step S3 includes:

[0023] Step S31: Set parameters such as the particle population size, maximum iteration count, individual learning factor, and social learning factor, and randomly set the initial positions and velocities of the particles.

[0024] Step S32: Calculate the initial fitness, and record the initial individual best fitness, individual best position, global best fitness, and global best position of the particles. Use the current initial positions of each particle as the initial poles of the vector matching method, and use the mean square error as the particle fitness.

[0025] Step S33: Record the individual best positions and the global best position of each particle at the current moment according to the fitness of each particle, and update the individual best position and the global best position of the particles.

[0026] Step S34: Determine whether the iteration count at the current moment is less than the maximum iteration count. If it is less, perform another iteration, update the position information and velocity information of each particle, and repeat Steps S32 - S34. If it is greater than or equal to, record the globally best position of the currently calculated particle swarm as the best position, and use the transfer function obtained from the best position as the optimal active - frequency transfer function.

[0027] Further, when calculating the initial fitness, the initial individual best fitness is the fitness of the particle at the initial position, the initial position of each particle is the individual best position, the global best fitness is the minimum value among the fitnesses of all particles, and the position of the particle with the minimum fitness is the globally best position of the particles.

[0028] Further, in Step S33, recording the individual best positions and the global best position of each particle at the current moment according to the fitness of each particle is specifically as follows: Compare the fitness corresponding to the current particle with the fitness of the particle in the past. The position corresponding to the particle with the minimum fitness is the individual best position of the particle. By comparing the individual optimal fitnesses of all particles, the one with the minimum fitness is the globally best position of the particle swarm.

[0029] In a second aspect, an inertia evaluation system for power system nodes based on an improved vector matching method provided by an embodiment of the present invention includes: an acquisition module, a calculation module, an optimization module, and an evaluation module;

[0030] The acquisition module is configured to acquire the frequency and power time-series data of each node in the power system, and perform normalization processing on the time-series data to obtain a frequency standard data set and a power standard data set respectively;

[0031] The calculation module is configured to represent the transfer function between the node frequency deviation and the active power deviation by a rational function containing unknown parameters, and perform parameter identification using the vector matching method according to the time-series data to calculate the unknown parameters of the rational function, so as to obtain an equivalent active-power - efficiency transfer function;

[0032] The optimization module improves the vector matching method by using the particle swarm optimization algorithm, takes the final fitting mean square error obtained by the vector matching method as the particle fitness, takes the minimum fitness as the optimization goal, and finds the target initial poles through continuous iteration of the particle swarm. After the iteration ends, an optimal active-power - frequency transfer function is obtained;

[0033] The evaluation module obtains an inertia evaluation value according to the mathematical relationship between the optimal active-power - frequency transfer function and the node inertia.

[0034] Further, the calculation module includes a vector matching calculation unit. The vector matching calculation unit is configured to set the maximum number of iterations, the initial poles, construct an auxiliary function and vector matching conditions, initialize the current number of iterations to 1, calculate the frequency convolution integral and the power convolution integral of the node respectively by using a recursive filter and trapezoidal numerical integration, solve for the unknown residues of the auxiliary function by using the least squares method fitting according to the set vector matching conditions, calculate the approximate poles of the rational function by constructing a matrix, perform residue identification of the rational function according to the identified poles, determine whether the current number of iterations is less than the maximum number of iterations. If it is less, the current number of iterations is incremented by 1, and the obtained identified poles are used as the new initial poles, and the calculation of the identified poles and residues is repeated. If it is greater than or equal to, an approximate expression of the rational function is obtained according to the currently calculated identified poles and residues, the fitting value of the node frequency is calculated by using this expression and the node power, and the mean square error is calculated according to the difference between the fitting value and the actually obtained node frequency.

[0035] Further, the optimization module includes an optimization processing unit, which sets parameters such as the particle population size, the maximum number of iterations, the individual learning factor, and the social learning factor, randomly sets the initial positions and velocities of the particles, calculates the initial fitness, and records the initial individual best fitness, the individual best position of the particle, the global best fitness of the particle, and the global best position of the particle. The current initial positions of each particle are used as the initial poles of the vector matching method, and the mean square error is used as the particle fitness. According to the fitness of each particle, the individual best position of each particle and the global best position of the particle at the current moment are recorded, the individual best position of the particle and the global best position of the particle are updated, and it is judged whether the number of iterations at the current moment is less than the maximum number of iterations. If it is less, then iterate again, update the position information and velocity information of each particle, and repeat the calculation of the position information and velocity information of each particle; if it is greater than or equal to, then record the globally best position of the particle swarm calculated currently as the best position, and use the transfer function obtained from the best position as the optimal active - frequency transfer function.

[0036] In a third aspect, an electronic device provided by an embodiment of the present invention includes a processor, an input device, an output device, and a memory. The processor is respectively connected to the input device, the output device, and the memory. The memory is used to store a computer program, and the computer program includes program instructions. The processor is configured to call the program instructions to execute the method described in the above - mentioned embodiment.

[0037] In a fourth aspect, a computer - readable storage medium provided by an embodiment of the present invention stores a computer program, and the computer program includes program instructions. When the program instructions are executed by a processor, the processor is caused to execute the method described in the above - mentioned embodiment.

[0038] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0039] A method, system, device, and medium for evaluating the inertia of power system nodes based on an improved vector matching method provided by an embodiment of the present invention regard the number and positions of the initial poles as the dimensions and positions of the particles. Through the cooperation of the particle swarm, it widely explores in the search space, significantly enhancing the global optimization ability of the algorithm. At the same time, the mean square error calculated by the vector matching method is used as the fitness index of the particle, and the minimization of the fitness is used as the optimization rule. Through continuous iteration, a better initial pole configuration is found, thereby effectively improving the fitting accuracy of the vector matching method. This improvement not only improves the accuracy of inertia evaluation but also enables the algorithm to have stronger robustness and adaptability when dealing with complex power system data. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings. In the drawings:

[0041] Figure 1 It is a flowchart of a method for evaluating the inertia of power system nodes based on an improved vector matching method provided by the first embodiment of the present invention;

[0042] Figure 2 It is a schematic diagram of the geographical location distribution of nodes in a provincial power grid in an embodiment of the present invention;

[0043] Figure 3(a) is the frequency deviation response curve of each node under the unit tripping fault scenario in an embodiment of the present invention;

[0044] Figure 3(b) is the power deviation response curve of each node under the unit tripping fault scenario in an embodiment of the present invention;

[0045] Figure 4(a) is the frequency deviation response curve of node 1 obtained by using a method for evaluating the inertia of power system nodes based on an improved vector matching method provided by an embodiment of the present invention and the least squares identification algorithm based on the ARMAX model in an actual application scenario;

[0046] Figure 4(b) is the frequency deviation response curve of node 3 obtained by using a method for evaluating the inertia of power system nodes based on an improved vector matching method provided by an embodiment of the present invention and the least squares identification algorithm based on the ARMAX model in an actual application scenario;

[0047] Figure 5(a) is the frequency deviation response curve of node 1 obtained by using a method for evaluating the inertia of power system nodes based on an improved vector matching method provided by an embodiment of the present invention and the vector matching method in an actual application scenario;

[0048] Figure 5(b) is the frequency deviation response curve of node 3 obtained by using a method for evaluating the inertia of power system nodes based on an improved vector matching method provided by an embodiment of the present invention and the vector matching method in an actual application scenario;

[0049] Figure 5(c) is the frequency deviation response curve of node 5 obtained by using a method for evaluating the inertia of power system nodes based on an improved vector matching method provided by an embodiment of the present invention and the vector matching method in an actual application scenario;

[0050] Figure 6 It is a structural block diagram of a system for evaluating the inertia of power system nodes based on an improved vector matching method provided by another embodiment of the present invention;

[0051] Figure 7 A structural block diagram of an electronic device provided in another embodiment of the present invention. Detailed implementation manners

[0052] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with embodiments and drawings. The illustrative embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention. Embodiment 1

[0053] As Figure 1 shown, a method for evaluating the inertia of power system nodes based on an improved vector matching method provided in the first embodiment of the present invention includes the following steps:

[0054] Step S1: Obtain the time-series data such as the frequency and power of each node in the system, and perform normalization processing on the data to obtain a frequency standard data set and a power standard data set respectively;

[0055] Step S2: Represent the transfer function between the node frequency deviation and the active power deviation by a rational function with unknown parameters, and use the vector matching method to perform parameter identification according to the above measurement data to obtain various parameters. The parameters include the poles and residues of the rational function. According to the poles and residues, an approximate expression of the rational function is obtained, and then an equivalent "active power - frequency" transfer function is obtained;

[0056] Step S3: Introduce a particle swarm optimization algorithm to improve the vector matching method. Take the final fitting mean square error obtained by the vector matching method as the particle fitness, take the minimum fitness as the optimization goal, and find the target initial poles through continuous iteration of the particle swarm. After the iteration ends, an optimal "active power - frequency" transfer function is obtained;

[0057] Step S4: Obtain the inertia evaluation value according to the mathematical relationship between the optimal "active power - frequency" transfer function and the node inertia.

[0058] The working principle of the embodiment of the present invention is as follows:

[0059] Most of the existing node inertia evaluation methods rely on large disturbance information or a large amount of data, and may have problems such as low accuracy, poor robustness and large calculation amount at the same time, and it is difficult to adapt to different working conditions. Although the inertia evaluation method based on the vector matching method is easy to implement and has strong adaptability, the initial poles are usually roughly distributed in a certain data segment and lack precise selection, resulting in limited fitting accuracy and being prone to falling into local optimal solutions, which affects the evaluation results.

[0060] Therefore, to solve this problem, the embodiment of the present invention provides a method for evaluating the inertia of power system nodes by combining the vector matching method and the particle swarm optimization algorithm with global convergence characteristics. The number and position of the initial poles are regarded as the dimensions and positions of the particles, and through the cooperation of the particle swarm, extensive exploration is carried out in the search space to enhance the global optimization ability. At the same time, the mean square error obtained by the vector matching method is used as the fitness index of the particles, and the minimization of the fitness is used as the optimization rule. By continuously iterating, a better initial pole configuration is found, thereby significantly improving the fitting accuracy of the vector matching method and realizing high-precision evaluation of the inertia of power system nodes.

[0061] Further, the formula for data normalization processing in step S1 is:

[0062]

[0063] Wherein, is k the standard data obtained after frequency normalization processing of the node i at time is k the standard data obtained after power normalization processing of the node i at time f i ( k ) is k the frequency of the node i at time f i ( k 0) is the frequency of the node i at the initial time f 0 is the rated frequency of the system, generally 50Hz; P i ( k ) is k the instantaneous value of the active power of the node i at time P i ( k 0) is the steady-state value of the active power of the node i at the initial time; The standard data set obtained after the above normalization processing is:

[0064]

[0065] Wherein, F i and P i are respectively the frequency standard data set and the power standard data set of the node i , and L is the data length.

[0066] Furthermore, the transfer function between the node frequency deviation and the active power deviation in step S1 is expressed as a rational function with unknown parameters as follows:

[0067]

[0068] where, p n are the poles of the rational function, generally complex poles; c n are the residues corresponding to the poles; d is the constant term, N is the fitting order, and s is the Laplace transform parameter.

[0069] Furthermore, the sub-steps of the specific method for obtaining the unknown parameters of the above rational function according to the vector matching method in step S2:

[0070] Step S21: Set the maximum number of iterations C max , and initialize the current number of iterations to 1. In this embodiment, C max is set to 100.

[0071] Step S22: Set the initial poles and construct an auxiliary function. First, given a set of starting poles , and construct an auxiliary function such that it satisfies the following conditions:

[0072]

[0073] where, k n are the unknown residues, are the zeros and poles of respectively.

[0074] Step S23: Establish the vector matching condition. Multiply the expression in step S22 by the input signal , and perform the inverse Laplace transform to obtain:

[0075]

[0076] In the above formula, is the frequency convolution integral and power convolution integral of node i .

[0077] Step S24: Solve the convolution integral. Use a recursive filter (IIR) and trapezoidal numerical integration to calculate the frequency convolution integral and power convolution integral, and their expressions are as follows:

[0078] ,

[0079] ,

[0080] Among them, α and λ are the coefficients of each item of the filter obtained by trapezoidal numerical integration, and Δ t is the sampling time.

[0081] Step S25: Pole identification of Based on the vector matching condition in step S23, the unknown residues k n of are obtained by solving using least squares fitting, and the identified poles of are further calculated by constructing the matrix H. The expressions of the matrix H and the identified poles

[0082]

[0083] are as follows: where A and B are block matrices containing the initial poles k n and residues are submatrices of the corresponding parts of the matrices A and B respectively, are the real and imaginary parts of the initial poles respectively, are respectively the real and imaginary parts of the residues k n ; is the identified pole of , is the conjugate complex number of

[0084] Step S26: Residue identification of Based on the identified poles c n obtained in step S25, the residue

[0085] can be calculated by solving the following least squares problem:

[0086] In the above formula, is calculated by substituting the identified poles into step S24.

[0087] Step S27: Judge whether the current iteration number is less than the maximum iteration number. If it is less, the current iteration number is incremented by 1, and the obtained identified poles As the new initial poles, and repeat steps S22 - S27; if it is greater than or equal to, then according to the currently calculated identified poles and residues c n Obtain the approximate expression of the rational function Use this expression to calculate the fitting value of the node frequency with the node power, and calculate the mean square error according to the difference between the fitting value and the actually obtained node frequency.

[0088] Furthermore, the improvement of the vector matching method by introducing the particle swarm optimization algorithm in step S3 includes the following sub - steps:

[0089] Step S31: Initialize the particle parameters, position, velocity, and dimension. Set the particle population size, maximum number of iterations, inertia weight, individual learning factor, and social learning factor parameters; randomly set the initial positions and velocities of the particles; where the positions and velocities of each particle in the particle swarm are expressed as:

[0090]

[0091] In the above formula, X j 、 V j are respectively the position and velocity of the j th particle in the particle swarm, d is the particle dimension, which represents the number of initial poles here.

[0092] Furthermore, the particle population size in step S31 is S P = 20, that is, there are 20 particles in the particle swarm; the maximum number of iterations I = 10, the initial number of iterations is 1; the inertia weight w = 0.5, the individual learning factor c1 = 2, the social learning factor c2 = 2; the dimension d = 5; the initial positions and velocities of the particles are expressed as follows:

[0093]

[0094] In the above formula, are respectively the initial position and velocity of the particle j ; x jm 、 y jm are both random numbers between 0 and 1; j is the imaginary symbol.

[0095] Step S32: Calculate the initial fitness, and record the initial particle individual best fitness, individual best position, global best fitness, and global best position. Take the current initial positions of each particle as the initial poles of the vector matching method, and obtain a series of mean square error values through steps S21 - S27, and use this as the particle fitness, and its expression is:

[0096]

[0097] In the above formula, is the particle corresponding to the node i ; j is the fitness of the particle ; j is the frequency prediction value and transfer function of the node obtained by the particle i through steps S21 - S27; L aplace -1 represents taking the inverse Laplace transform.

[0098] Furthermore, initially, the individual best fitness is the fitness of the particle at the initial position, and the initial position of each particle is the individual best position; the global best fitness is the minimum value among the fitnesses of all particles, and the corresponding initial position of the particle is the global best position.

[0099] Step S33: Update the individual best and the global best. Record the individual best position of each particle and the global best position of the particle swarm at the current moment according to the fitness of each particle. Specifically: Compare the fitness corresponding to the j th particle with the previous fitness corresponding to the j th particle. The position corresponding to the particle with the minimum fitness is the individual best position of the j th particle; By comparing the individual optimal fitnesses of all particles, the one with the minimum fitness is the global best position of the particle swarm.

[0100] Step S34: Determine whether the number of iterations at the current moment is less than the maximum number of iterations. If it is less, then iterate again, update the position information and velocity information of each particle, and repeat steps S32 - S34; If it is greater than or equal to, record the currently calculated global best position of the particle swarm as the best position, and use the transfer function obtained using the best position as the optimal "active power - frequency" transfer function. Update the expressions of the position information and velocity information of each particle as follows:

[0101] ,

[0102] In the above formula, V j ( k+ 1) is the velocity of the particle j after k updates; X j,best is the individual best position of the particle j ; is the particle j after kPositions after and before the update; r1 and r2 are random numbers between 0 and 1.

[0103] Furthermore, the mathematical relationship between the "active power - frequency" transfer function and the node inertia in step S4 is:

[0104] ,

[0105] In the above formula, H i , D i and F i ( s ) are the inertia, damping coefficient, and primary frequency regulation response of node i ; b ' 1,i ~ b ' n,i , a ' 1,i ~ a ' n,i are respectively the coefficients of the numerator and denominator of G i ( s ), and n is the order of the function.

[0106] The effectiveness of the power system node inertia evaluation method based on the improved vector matching method provided by the embodiments of the present invention is illustrated by the following specific application scenarios:

[0107] Based on the PSASP simulation software platform, the disturbance event simulation of a provincial power grid under the normal operation mode in the summer of 2023 is used as a test case to verify the accuracy of a power system node inertia evaluation method based on the improved vector matching method provided by the embodiments of the present invention. The geographical location distribution of the nodes of this provincial power grid is as Figure 2 shown. There are more than 100 units in the system, with a total installed capacity of about 20,000 MW, a rated frequency of 50 Hz, a base capacity of 100 MW, and a sampling time of 0.01 s. Figure 2 Among them, TS1 and TS2 represent substations in other provinces, while LR1 and LR2 are the tie lines between this provincial power grid and the power grids in other provinces. In addition, SS is a switching station, CS is a converter station, and PS represents a power plant. The disturbance fault scenario during the simulation analysis is: tAt \(t = 1s\), the units at nodes 7 and 8 tripped, resulting in a total loss of 3207 MW of output. The simulation results of the frequency and power deviation at each node are shown in Figures 3(a) and 3(b). Six generator nodes numbered 1, 2, 3, 4, 5, and 6 were selected as the test objects, and the proposed method was compared and analyzed with the least squares identification algorithm based on the ARMAX model and the vector matching method, respectively, thus indicating that the inertia evaluation method proposed in the present invention has better accuracy. The parameters of the above six generator nodes are shown in Table 1, and the comparison of the inertia evaluation results at each node and the error analysis obtained by different methods are shown in Tables 2 and 3.

[0108]

[0109]

[0110] According to the comparison results in Table 2, for the node inertia estimation values obtained by the method of the embodiment of the present invention, the minimum error is 0.455%, the maximum error is 5.622%, and the average error is about 2.3%. For the node inertia estimation values obtained by identification based on the least squares method, the minimum error is 3.947%, the maximum error is 19.974%, and the average error is about 11%. It can be seen that the node inertia evaluation method provided by the embodiment of the present invention has higher accuracy than the traditional least squares identification method. Specifically, in addition, the mean square error of frequency fitting based on the least squares method is significantly higher than the result obtained by the embodiment of the present invention. To further highlight the superiority of the method provided by the embodiment of the present invention, nodes 1 and 3 were selected as the test objects, and the frequency deviation responses were obtained by using the above two methods, as shown in Figures 4(a) and 4(b). It can be seen from Figures 4(a) and 4(b) that the frequency deviation response based on the inertia evaluation method proposed in the embodiment of the present invention is closer to the actual simulation result than the least squares identification. These test results show that compared with the traditional system identification algorithm, the method of the embodiment of the present invention performs more excellently in the estimation accuracy of node inertia, and it has lower sensitivity to the change of node position, so it has stronger robustness and applicability.

[0111]

[0112] As shown in Table 3, in the scenario of generator set tripping fault, the evaluation errors and mean square errors of Nodes 1, 3, and 5 obtained by the method according to the embodiments of the present invention are significantly lower than those of the vector matching method. The evaluation results of other nodes are close to it. The main reason is that there are differences in the frequency response characteristics of different nodes, which may cause multiple identical local optimal initial solutions to be generated when performing vector matching for some nodes, thus affecting the existence of the global optimal solution. It should be noted that in the case of generator set tripping fault, the frequency and power fluctuations of each node are small, resulting in more difficult fitting of the transfer function, which makes the error of the inertia evaluation method based on vector matching exceed 10% at individual nodes. In contrast, the method according to the embodiments of the present invention improves the selection of the initial poles by introducing the particle swarm optimization algorithm, thereby significantly reducing the errors of each node, and the evaluation errors of all nodes are kept within 6%, ensuring high accuracy. To further verify the superiority of the method of the present invention, Nodes 1, 3, and 5 are selected as the test objects, and the frequency deviation responses are calculated by different methods. The results are shown in Figures 5(a), 5(b), and 5(c). It can be seen by comparison that the frequency deviation response obtained by the method of the present invention is closer to the actual simulation result than the vector matching method. Especially in the comparison of Nodes 1 and 3, the effect is more significant, and this phenomenon is further verified by the mean square error. Combining the comparison results in Table 2, it can be seen that the power system node inertia evaluation method based on the improved vector matching method proposed in the embodiments of the present invention has significant advantages in terms of accuracy, adaptability, and stability.

[0113] In summary, the large-scale access of new energy has significantly changed the inertia characteristics of the new power system. In the area where new energy is concentrated, the low inertia characteristics are obvious, which is in sharp contrast to the high inertia state in the area where the synchronous power cluster is distributed, resulting in a certain spatial distribution characteristic of the inertia. When a system fault occurs, the weak inertia area will produce a large frequency deviation and frequency change rate, which is more likely to trigger the action of the low-frequency load reduction relay or RoCoF relay, thereby increasing the risk of system power outage. Therefore, accurately assessing the inertia level at the node level of the power system is conducive to guiding the grid connection of new energy and planning inertia compensation measures, and avoiding the system from falling into the dilemma of ultra-low inertia operation. However, most of the existing inertia evaluation methods rely on large disturbance information or a large amount of data, and may have problems of low accuracy, poor robustness and large amount of calculation at the same time, making it difficult to adapt to different working conditions. Although the inertia evaluation method based on the vector matching method is easy to implement and has strong adaptability, the initial poles are usually roughly distributed in a certain data segment, lacking precise selection, resulting in limited fitting accuracy and affecting the evaluation results. Therefore, in order to solve this problem, the present invention proposes a method for evaluating the node inertia of a power system based on an improved vector matching method, which regards the number and position of the initial poles as the dimensions and positions of the particles, and widely explores the search space through the collaboration of the particle swarm to enhance the global optimization capability. At the same time, the mean square error calculated based on the vector matching method is used as the particle fitness index, and fitness minimization is used as the optimization rule. Through continuous iteration, a better initial pole configuration is found, thereby significantly improving the fitting accuracy of the vector matching method and achieving high-precision evaluation of the node inertia of the power system. Example 2

[0114] like Figure 6 As shown, another embodiment of the present invention provides a power system node inertia evaluation system based on an improved vector matching method, including: an acquisition module, a calculation module, an optimization module and an evaluation module, the acquisition module is used to acquire the frequency and power time series data of each node in the power system, and normalize the time series data to obtain a frequency standard data set and a power standard data set respectively, the calculation module is used to represent the transfer function between the node frequency deviation and the active power deviation with a rational function containing unknown parameters, and perform parameter identification using a vector matching method according to the time series data to calculate the unknown parameters of the rational function to obtain an equivalent active power-efficiency transfer function; the optimization module uses a particle swarm optimization algorithm to improve the vector matching method, and uses the final fitting mean square error obtained by the vector matching method as the particle fitness, takes the minimum fitness as the optimization goal, finds the target initial pole through continuous iteration of the particle swarm, and obtains the optimal active power-frequency transfer function after the iteration; the evaluation module obtains the inertia evaluation value according to the mathematical relationship between the optimal active power-frequency transfer function and the node inertia.

[0115] Among them, the calculation module includes a vector matching calculation unit. The vector matching calculation unit is used to set the maximum number of iterations, the initial poles, construct an auxiliary function and vector matching conditions, initialize the current number of iterations to 1, calculate the frequency convolution integral and power convolution integral of the nodes respectively by using a recursive filter and trapezoidal numerical integration, solve for the unknown residues of the auxiliary function by fitting using the least squares method according to the set vector matching conditions, calculate the approximate poles of the rational function by constructing a matrix, identify the residues of the rational function according to the identified poles, determine whether the current number of iterations is less than the maximum number of iterations. If it is less, the current number of iterations is incremented by 1, the obtained identified poles are used as the new initial poles, and the calculation of the identified poles and residues is repeated. If it is greater than or equal to, the approximate expression of the rational function is obtained according to the currently calculated identified poles and residues.

[0116] Among them, the optimization module includes an optimization processing unit. The optimization processing unit sets parameters such as the particle population size, the maximum number of iterations, the individual learning factor, and the social learning factor, randomly sets the initial positions and velocities of the particles, calculates the initial fitness, and records the initial individual best fitness, individual best position, global best fitness, and global best position of the particles. The current initial positions of each particle are used as the initial poles of the vector matching method, and the mean square error value is used as the particle fitness. According to the fitness of each particle, the current individual best positions and particle global best positions of each particle are recorded, and the individual best position and particle global best position of the particles are updated. Determine whether the number of iterations at the current moment is less than the maximum number of iterations. If it is less, iterate again, update the position information and velocity information of each particle, and repeat the calculation of the position information and velocity information of each particle; if it is greater than or equal to, record the currently calculated global best position of the particle swarm as the best position, and use the transfer function obtained from the best position as the optimal active-frequency transfer function.

[0117] The power system node inertia evaluation system based on the improved vector matching method provided by the embodiments of the present invention has the same inventive concept and the same beneficial effects as the above-mentioned power system node inertia evaluation based on the improved vector matching method, and will not be elaborated here. Embodiment 3

[0118] As Figure 7 shown, the structural block diagram of an electronic device according to another embodiment of the present invention. The device includes a processor, an input device, an output device, and a memory. The processor is respectively connected to the input device, the output device, and the memory. The memory is used to store a computer program. The computer program includes program instructions. The processor is configured to call the program instructions to execute the method described in the above first embodiment.

[0119] It should be understood that in the embodiments of the present invention, the so-called processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0120] The input device may include a touchpad, a fingerprint acquisition sensor (for acquiring the fingerprint information and the direction information of the fingerprint of the user), a microphone, etc., and the output device may include a display (such as an LCD), a speaker, etc.

[0121] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A part of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0122] In specific implementation, the processor, input device, and output device described in the embodiments of the present invention may implement the implementation manners described in the method embodiments provided by the embodiments of the present invention, or may also implement the implementation manners of the system embodiments described in the embodiments of the present invention, which will not be elaborated herein. Embodiment 4

[0123] In another embodiment of the present invention, there is also provided an embodiment of a computer-readable storage medium. The computer-readable storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by the processor, the processor is caused to execute the method described in the above-mentioned first embodiment.

[0124] The computer-readable storage medium may be an internal storage unit of the terminal described in the foregoing embodiments, such as the hard disk or memory of the terminal. The computer-readable storage medium may also be an external storage device of the terminal, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the terminal. Further, the computer-readable storage medium may also include both the internal storage unit and the external storage device of the terminal. The computer-readable storage medium is used to store the computer program and other programs and data required by the terminal. The computer-readable storage medium may also be used to temporarily store the data that has been output or is to be output.

[0125] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0126] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described terminal and unit can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0127] In several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed couplings or direct couplings or communication connections to each other can be indirect couplings or communication connections through some interfaces, devices or units, and can also be electrical, mechanical or other forms of connection.

[0128] The specific embodiments described above further elaborate the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for evaluating node inertia of a power system based on an improved vector matching method, characterized in that: include: Step S1: Obtain frequency and power time series data of each node in the power system, and normalize the time series data to obtain a frequency standard data set and a power standard data set respectively; Step S2: The transfer function between the node frequency deviation and the active power deviation is expressed by a rational function containing unknown parameters, and the unknown parameters of the rational function are calculated by using the vector matching method according to the time series data to obtain an equivalent active power-efficiency transfer function; Step S3: The particle swarm optimization algorithm is used to improve the vector matching method, and the final fitting mean square error obtained by the vector matching method is used as the particle fitness. The minimum fitness is taken as the optimization goal, and the target initial pole is found through continuous iteration of the particle swarm. After the iteration, the optimal active power-frequency transfer function is obtained; Step S4: Obtaining an inertia evaluation value according to a mathematical relationship between an optimal active power-frequency transfer function and a node inertia; The specific method of performing parameter identification using the vector matching method according to the time series data in step S2 to calculate the unknown parameters of the rational function includes: Step S21: Set the maximum number of iterations and initialize the current number of iterations to 1; Step S22: setting the initial pole and constructing auxiliary functions; Step S23: setting vector matching conditions; Step S24: using a recursive filter and trapezoidal numerical integration to respectively calculate the frequency convolution integral and the power convolution integral of the node; Step S25: according to the set vector matching conditions, the unknown residue of the auxiliary function is obtained by fitting and solving using the least square method, and the identification pole of the rational function is obtained by constructing a matrix calculation; Step S26: performing residue identification of rational functions according to the identified poles; Step S27: Determine whether the current number of iterations is less than the maximum number of iterations. If so, add 1 to the current number of iterations, use the obtained identification pole as the new initial pole, and repeat steps S22-S27. If it is greater than or equal to, obtain an approximate expression of the rational function based on the currently calculated identification pole and residue, use the expression and the node power to calculate the fitting value of the node frequency, and calculate the mean square error based on the difference between the fitting value and the actual node frequency.

2. The method according to claim 1, characterized in that The specific method of improving the vector matching method by using the particle swarm optimization algorithm in step S3 includes: Step S31: setting the particle population size, maximum number of iterations, individual learning factor and social learning factor parameters, and randomly setting the initial position and speed of the particles; Step S32: Calculate the initial fitness, and record the individual best fitness of the particle, the individual best position of the particle, the global best fitness of the particle, and the global best position of the particle at the initial time, use the current initial position of each particle as the initial pole of the vector matching method, and use the mean square error as the particle fitness; Step S33: Record the best individual position of each particle and the best global position of the particle at the current moment according to the fitness of each particle, and update the best individual position of the particle and the best global position of the particle; Step S34: Determine whether the number of iterations at the current moment is less than the maximum number of iterations. If so, iterate again, update the position information and velocity information of each particle, and repeat steps S32-S34; if greater than or equal to, record the currently calculated global optimal position of the particle swarm as the optimal position, and use the transfer function obtained at the optimal position as the optimal active-frequency transfer function.

3. The method according to claim 2, characterized in that When calculating the initial fitness, the initial individual best fitness is the fitness of the particle at the initial position, the initial position of each particle is the individual best position, the global best fitness is the minimum value of all particle fitnesses, and the position of the particle with the smallest fitness is the global best position of the particle.

4. The method according to claim 2, characterized in that: In step S33, the best individual position of each particle and the best global position of the particle are recorded according to the fitness of each particle at the current moment. Specifically, the fitness corresponding to the current particle is compared with the fitness corresponding to the particle in the past. The position corresponding to the particle with the smallest fitness is the best individual position of the particle. By comparing the optimal fitness of all individual particles, the one with the smallest fitness is the global best position of the particle group.

5. A power system node inertia assessment system based on an improved vector matching method, characterized in that: include: Acquisition module, calculation module, optimization module and evaluation module; The acquisition module is used to acquire the frequency and power time series data of each node in the power system, and normalize the time series data to obtain a frequency standard data set and a power standard data set respectively; The calculation module is used to express the transfer function between the node frequency deviation and the active power deviation by a rational function containing unknown parameters, and to perform parameter identification using a vector matching method according to the time series data to calculate the unknown parameters of the rational function, thereby obtaining an equivalent active power-efficiency transfer function; The optimization module uses a particle swarm optimization algorithm to improve the vector matching method, takes the final fitting mean square error obtained by the vector matching method as the particle fitness, takes the minimum fitness as the optimization goal, finds the target initial pole through continuous iteration of the particle swarm, and obtains the optimal active power-frequency transfer function after the iteration is completed; The evaluation module obtains an inertia evaluation value according to a mathematical relationship between an optimal active power-frequency transfer function and a node inertia; The calculation module includes a vector matching calculation unit, which is used to set the maximum number of iterations, the initial pole, construct an auxiliary function and a vector matching condition, and initialize the current number of iterations to 1, use a recursive filter and a trapezoidal numerical integration to respectively calculate the frequency convolution integral and the power convolution integral of the node, according to the set vector matching condition, use the least squares method to fit and solve to obtain the unknown residue of the auxiliary function, and obtain the approximate pole of the rational function by constructing a matrix calculation, perform residue identification of the rational function according to the identification pole, judge whether the current number of iterations is less than the maximum number of iterations, if less than, the current number of iterations is increased by 1, the obtained identification pole is used as the new initial pole, and the identification pole and residue are repeatedly calculated, if greater than or equal to, then the approximate expression of the rational function is obtained according to the currently calculated identification pole and residue, the fitting value of the node frequency is calculated using the expression and the node power, and the mean square error is calculated according to the difference between the fitting value and the actually obtained node frequency.

6. The system according to claim 5, characterized in that The optimization module includes an optimization processing unit, which sets the particle population size, the maximum number of iterations, the individual learning factor and the social learning factor parameters, randomly sets the initial position and speed of the particles, calculates the initial fitness, and records the individual best fitness of the particles, the individual best position of the particles, the global best fitness of the particles and the global best position of the particles at the initial time, takes the current initial position of each particle as the initial pole of the vector matching method, takes the mean square error as the particle fitness, records the individual best position of each particle and the global best position of the particles at the current moment according to the fitness of each particle, updates the individual best position of the particles and the global best position of the particles, judges whether the number of iterations at the current moment is less than the maximum number of iterations, if so, iterates again, updates the position information and speed information of each particle, and repeatedly calculates the position information and speed information of each particle; if greater than or equal to, records the currently calculated global best position of the particle group as the best position, and uses the transfer function obtained at the best position as the optimal active power-frequency transfer function.

7. An electronic device, comprising a processor, an input device, an output device and a memory, wherein the processor is connected to the input device, the output device and the memory respectively, and the memory is used to store a computer program, wherein the computer program comprises program instructions, and wherein: The processor is configured to call the program instructions to execute the method according to any one of claims 1-4.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor is caused to perform the method according to any one of claims 1 to 4.

Images