Method for comprehensively evaluating strength of large-scale new energy power system

Abstract

The invention discloses a comprehensive assessment method for the strength of a large-scale new energy power system. The method comprises the following steps: S1, constructing a power grid hybrid model; s2, constructing and updating an extended impedance matrix: on the basis of the power grid hybrid model, enabling a direct current converter station to be equivalent to an additional impedance branch, constructing the extended impedance matrix of the system by adopting an additional branch method, and dynamically updating the extended impedance matrix on the basis of real-time measurement data; s3, quantitative calculation of system strength: calculating a multi-infeed short-circuit ratio based on the updated extended impedance matrix, and extracting Thevenin equivalent parameters from the multi-infeed short-circuit ratio; and S4, comprehensively evaluating the strength of the system: fusing the multi-infeed short-circuit ratio, the Thevenin equivalent parameter and the static and dynamic stability analysis result, and generating a comprehensive grading map of the strength of the system. According to the invention, a scientific decision basis is provided for power grid planning and operation control, and safe and stable operation of a power system under high-proportion new energy and direct current access is effectively supported.

Description

Technical Field

[0001] This invention relates to a comprehensive evaluation method for the strength and weakness of large-scale new energy power systems. Background Technology

[0002] The high proportion of renewable energy integration and the large-scale application of power electronic equipment have given the power system a "dual-high" characteristic (high proportion of renewable energy and high proportion of power electronic devices), posing a severe challenge to the stability of the traditional power grid. The intermittency and volatility of renewable energy generation lead to increased frequency / voltage fluctuations in the power grid. For example, sudden changes in wind / solar power output can cause power imbalances and increase the risk of local grid voltage exceeding limits. The increase in power electronic equipment has led to problems such as broadband oscillations and harmonic pollution, threatening the safe operation of equipment.

[0003] In new energy power systems, traditional single-infeed short-circuit ratio assessment methods cannot quantify the interactive effects of multiple DC infeed scenarios and do not fully consider the dynamic characteristics of new energy power plants (such as the impact of photovoltaic inverter control strategies on short-circuit current). Especially in multi-infeed DC systems, the system's equivalent strength depends not only on a single DC system but also on a complex interplay of factors such as the electrical distance between DC infeed points and reactive power compensation configuration. Note: The short-circuit ratio represents the system's short-circuit capacity divided by the equipment capacity. The short-circuit capacity, under unit voltage conditions, is numerically equal to the system admittance, which is the reciprocal of the system's Thevenin equivalent impedance.

[0004] Currently, existing technologies still have many problems: the dynamic interaction mechanism between new energy power plants and DC systems is unclear, and the accuracy of traditional equivalent models is insufficient; there is a lack of fast calculation methods for short-circuit ratio in multi-infeed scenarios, and the real-time performance is difficult to meet operational requirements; the time-varying characteristics of system parameters (such as fluctuations in new energy output and grid topology adjustments) lead to deviations in static evaluation results. Summary of the Invention

[0005] This invention proposes a comprehensive evaluation method for the strength and weakness of large-scale new energy power systems. This method constructs a full-process analysis framework centered on the impedance matrix. First, by expanding the impedance matrix, the electrical coupling relationship between AC and DC nodes is accurately quantified, forming a set of nonlinear high-dimensional external characteristic equations. This accurately characterizes the equivalent support of the external system to the DC landing point, and the impedance matrix is ​​dynamically updated using the branch addition method. At the modeling level, polymer equivalent units are adopted for new energy power plants, and their models integrate reactive power compensation equipment and maximum power point tracking control characteristics. Simultaneously, the parameters of the Thevenin equivalent model are optimized based on impedance matrix reduction technology. In the parameter identification and update stage, particle swarm optimization is used for offline global calibration, and online rolling updates are performed based on real-time PMU data using an extended Kalman filter algorithm, forming a closed-loop control system where the model and data mutually verify each other. In the evaluation stage, the critical short-circuit ratio is analyzed using the generalized Nyquist criterion to reveal the DC interaction-dominant mode. Finally, by constructing a static-dynamic collaborative evaluation system, integrating quantitative indicators such as multi-infeed short-circuit ratios with qualitative analyses such as transient stability and small-disturbance stability, a strength and weakness hierarchy map from local nodes to the global system is formed. This method provides a scientific basis for power grid planning and operation control, and effectively supports the safe and stable operation of the power system under the conditions of high proportion of new energy and DC access.

[0006] A comprehensive evaluation method for the strength and weakness of a large-scale new energy power system includes: S1, constructing a hybrid power grid model: constructing a hybrid power grid model including synchronous generators, AC lines, DC lines, and new energy power plants using a hierarchical strategy; S2, constructing and updating the extended impedance matrix: based on the hybrid power grid model, constructing the system's extended impedance matrix by equating DC converter stations with additional impedance branches using the additional branch method, and dynamically updating the extended impedance matrix based on real-time measurement data; S3, quantifying system strength: calculating the multi-infeed short-circuit ratio based on the updated extended impedance matrix, and extracting Thevenin equivalent parameters from it; S4, comprehensively evaluating system strength and weakness: integrating the multi-infeed short-circuit ratio, Thevenin equivalent parameters, and static and dynamic stability analysis results to generate a comprehensive system strength and weakness classification map.

[0007] In some examples, the hierarchical strategy in S1 includes: synchronous generators using a sixth-order practical model; AC lines using a π-type equivalent circuit; DC lines using the International Organization of Large Grids (IOG) standard model; and new energy power plants using polymer equivalent units, whose models integrate reactive power compensation equipment and maximum power point tracking control characteristics.

[0008] In some examples, the extended impedance matrix in S2 is constructed using the method of adding branches, including: starting from the node impedance matrix of the pure AC system; equating each DC converter station to an impedance branch added to the corresponding AC node; updating the entire network node impedance matrix for each added equivalent DC branch; until all DC branches have been added, the extended impedance matrix is ​​obtained.

[0009] In some examples, step S2 dynamically updates the extended impedance matrix based on real-time measurement data, including: updating the nodal admittance matrix elements, load model parameters, and renewable energy reactive power compensation parameters that constitute the basis of the extended impedance matrix using an adaptive filtering algorithm based on real-time synchronous phasor measurement unit data. Specifically, a preferred implementation of the adaptive filtering algorithm is the extended Kalman filter algorithm, the specific implementation process of which is described in the corresponding part of step one in the specific embodiments of this invention.

[0010] In some examples, a dynamic correction step for sensitive boundary parameters is performed between S2 and S3: when the rate of change of Thevenin equivalent parameters, nodal admittance matrix elements, or reactive power compensation coefficients of new energy power plants exceeds their corresponding set thresholds, the joint algorithm of least squares method and extended Kalman filter is used to perform online joint estimation and optimization of the above parameters.

[0011] In some examples, in S3, extracting the Thevenin equivalent parameters from the extended impedance matrix includes: for a single node of interest, directly extracting the self-impedance corresponding to that node in the extended impedance matrix as its Thevenin equivalent impedance; for multiple DC landing points, obtaining the equivalent multi-port Thevenin impedance at the boundary node level from the extended impedance matrix through matrix shrinking techniques.

[0012] In some examples, the formula used in S3 to calculate the multi-feed short-circuit ratio is:

[0013] In the formula, Converter station i The short-circuit capacity of the AC bus is determined by the Thevenin equivalent impedance extracted from the extended impedance matrix. Converter station i Local reactive power compensation of AC busbar; and DC system i and j Rated transmission power; DC landing point j landing point i The multi-feed mutual influence factor is determined based on the mutual impedance in the extended impedance matrix.

[0014] In some examples, the method further includes a critical short-circuit ratio analysis step: constructing a frequency domain impedance model of the AC network and converter based on the extended impedance matrix; forming a loop gain matrix; and applying the generalized Nyquist criterion to analyze the Nyquist curves of its eigenvalues ​​to determine the system stability boundary and the corresponding critical short-circuit ratio.

[0015] In some examples, the static and dynamic stability analysis in S4 includes: static strength quantification assessment: based on the calculation results of the multi-infeed short-circuit ratio, the static strength of the system is quantified and graded according to a preset grading standard; transient and dynamic stability assessment: based on the power grid hybrid model, time-domain simulation is performed to evaluate the stability characteristics of the system under extreme fault conditions; small disturbance stability analysis: based on the frequency domain impedance model derived from the extended impedance matrix, eigenvalue analysis is performed to identify the low-frequency oscillation modes and damping characteristics of the system.

[0016] In some examples, the grading criteria for the static strength quantification assessment are as follows: a multi-infeed short-circuit ratio less than 1.6 is judged as an extremely weak system, between 1.6 and 2.5 is judged as a weak system, and greater than 2.5 is judged as a strong system.

[0017] This invention is the first to incorporate Thevenin equivalent parameters, admittance matrix, and reactive power compensation from new energy sources into the external characteristic equation. By extending the impedance matrix, it accurately quantifies the electrical coupling between AC and DC nodes, solving the problem of isolated parameter adjustment in traditional methods.

[0018] This invention introduces a dynamic correction mechanism to update the equivalent impedance parameters in real time, ensuring model accuracy. It employs a dual threshold mechanism to avoid unnecessary computational overhead and combines the offline calibration capability of LS with the online filtering characteristics of EKF to achieve a balance between high accuracy and low latency.

[0019] This invention combines the generalized Nyquist criterion to analyze the critical stability boundary of the system, unifying the evaluation framework for traditional DC and new energy scenarios. This evaluation method demonstrates excellent adaptability to various application scenarios and exhibits strong engineering application potential. Attached Figure Description

[0020] Figure 1 This is a flowchart of a comprehensive evaluation method for the strength and weakness of a new energy power system according to an embodiment of the present invention.

[0021] Figure 2 This is a flowchart of short-circuit ratio analysis in one embodiment of the present invention.

[0022] Figure 3 This is a flowchart of the dynamic correction of sensitive boundary parameters in one embodiment of the present invention.

[0023] Figure 4 This is a flowchart of a comprehensive assessment of the strength and weakness of a power system according to an embodiment of the present invention. Detailed Implementation

[0024] This invention proposes a comprehensive evaluation method for the strength and weakness of new energy power systems, constructing a full-process analysis framework centered on the impedance matrix. This method expands the impedance matrix to quantify the electrical coupling relationship between AC and DC nodes, forming a nonlinear high-dimensional equation system. This accurately characterizes the equivalent support of the external system to the DC landing point and dynamically corrects the equivalent impedance. During the calculation process, the impedance matrix is ​​dynamically updated using the branch addition method, and the critical short-circuit ratio is analyzed using the generalized Nyquist criterion to reveal the DC interaction-dominant mode. Finally, by constructing a static-dynamic collaborative evaluation system that integrates quantitative indicators and qualitative assessments, a system strength and weakness classification map is formed. Figure 1 A brief overview of the method is presented, and it will be explained in detail below.

[0025] Step 1: Building the power grid model.

[0026] Power grid modeling is fundamental for system strength assessment. First, a multi-condition training set is constructed based on historical data, and the data is cleaned using the 3σ criterion. Then, the Kalman filter algorithm is used to impute missing values ​​to ensure data integrity. A hierarchical modeling strategy is employed. The synchronous generator adopts a sixth-order practical model, and the static parameters are calibrated by a subsequent genetic algorithm.

[0027] The AC line uses a π-type equivalent circuit, and its impedance is determined by the formula ( )calculate.

[0028] The DC line adopts the International Large Electric Networks (CIGRE) standard model and is matched with the control mode.

[0029] New energy power plants are aggregated into equivalent units, and their models integrate the control characteristics of SVG (Static Var Compensator) and Maximum Power Point Tracking (MPPT).

[0030] Model parameter identification is implemented in two phases: The first stage (offline global calibration): Using historical data, static parameters such as the generator's inertial constant are calibrated using the Particle Swarm Optimization (PSO) algorithm. The core idea of ​​the PSO algorithm is to find the optimal solution by simulating the cooperative behavior of electrical components as a group. In the PSO algorithm, each particle is determined based on its historical optimal position (…). ) and the group's global optimal position ( Update its speed ( ) and location ( The updated formula is as follows:

[0031]

[0032] In the formula, Inertial weights are the ability of particles to maintain their current motion trend. and The contact coefficient and the group coefficient respectively adjust the tendency of particles to move towards individual optima and global optima; and This is a Monte Carlo variable used to increase the randomness of the search.

[0033] The next stage (online global update): Based on real-time high-precision synchronous phasor measurement unit (PMU) data, the extended Kalman transform (EKF) algorithm is used to continuously update the ZIP (impedance, current, power) model coefficients of the load. Simultaneously, a time-varying factor is introduced to dynamically correct the Thevenin equivalent impedance and the reactive power compensation susceptance of new energy sources. The EKF algorithm, through local linearization of the nonlinear system, is particularly suitable for the dynamic estimation of ZIP coefficients. Its core update formula is as follows:

[0034]

[0035]

[0036] In the formula, Kalman gain; For the state estimates of each element; Here is the error covariance matrix; For the observation equation h ( The Jacobian matrix of ) needs to be calculated at each time step.

[0037] Ultimately, the model constructed using the above method can be simulated at multiple time scales, providing a high-precision and dynamically adaptable hybrid power grid model for subsequent short-circuit ratio calculations.

[0038] Step 2: Real-time data acquisition.

[0039] Real-time data acquisition is a core component of system strength assessment. By deploying PMUs and smart terminals, and sampling at a frequency of 500Hz, real-time data is acquired on the voltage of each node in the AC system, branch power, generator output, and environmental parameters of the renewable energy power plants. Data transmission uses standard protocols to ensure low latency. Simultaneously, historical operating data is retrieved from the Energy Management System (EMS) to jointly construct a multi-dimensional dataset encompassing both steady-state and transient operating conditions.

[0040] Dynamic quality control is implemented during data acquisition: First, the 3σ criterion is applied to identify and remove abnormal data; for missing data, a Kalman filter algorithm is used for smooth interpolation to ensure data integrity and continuity. To adapt to the fluctuating characteristics of new energy power output, this invention designs a sliding data window mechanism to update the characteristic statistics within the data window in real time, enabling the system to track dynamic changes in the system state. An embedded dynamic verification module performs secondary verification of data quality by calculating the residual between the PMU measurement value and the power grid model prediction value in step one. When the residual exceeds a preset adaptive threshold, the system will automatically trigger the parameter re-identification process of the model in step one, thus forming a closed-loop control system in which the model and data mutually verify and dynamically update each other.

[0041] Finally, this step outputs a high-quality real-time data stream covering both spatial and temporal dimensions, providing high-precision, low-latency data support for subsequent Thevenin equivalent parameter identification, time-varying admittance matrix update, and short-circuit ratio dynamic calculation, thus achieving closed-loop control for model verification.

[0042] Step 3: Impedance matrix and short-circuit ratio analysis.

[0043] This step focuses on multi-infeed direct current (MIDC) systems, constructing a short-circuit ratio analysis framework centered on the impedance matrix. It covers the entire process of dynamic modeling, parameter optimization, and stability assessment. A simplified flowchart is as follows: Figure 2 As shown in the diagram. Specifically, firstly, the AC-DC coupling relationship is described using multivariable external characteristic equations, and the electrical connections between nodes are quantified using an extended impedance matrix. Secondly, DC control characteristics are embedded in the power flow calculation, and node voltage sensitivity is analyzed using the impedance matrix to assist in the power flow solution. Furthermore, the Thevenin equivalent model is optimized based on the impedance matrix reduction technique to accurately characterize the impact of the external system on the DC landing point. Subsequently, the impedance matrix is ​​dynamically updated using the appended branch method to achieve rapid calculation of the multiple infeed short-circuit ratio (MISCR), avoiding the high computational cost of the traditional successive short-circuit method. Finally, combined with the generalized Nyquist criterion, the critical short-circuit ratio is analyzed through the impedance ratio trajectory in the complex frequency domain, revealing the DC interaction-dominant mode. This step, using the impedance matrix as a link, systematically solves the modeling complexity and stability quantification problems under multiple DC access, providing theoretical support for power grid planning and operation. Its main process is as follows: (1) Constructing multivariable external characteristic equations considering AC / DC coupling The purpose of constructing multivariable external characteristic equations is to describe the interaction of AC / DC hybrid systems. Its core lies in using an extended impedance matrix to represent the DC system as an additional branch on the AC node, thereby quantifying its impact within a unified framework.

[0044] First, define the internal state variables of the DC system as follows: (in It is direct current. The firing angle on the rectifier side. The inverter-side arc extinction angle, and These represent the turns ratios of the converter transformers on the rectifier and inverter sides, respectively, and the converter bus voltage at the sending and receiving ends. and By combining the steady-state control equations of the DC converter with the power balance relationship, the following set of external characteristic equations can be established to characterize the power exchange at the AC / DC interface:

[0045] In the formula, , and , These represent the active and reactive power exchanged between the AC and DC systems as seen from the converter buses on the rectifier and inverter sides, respectively.

[0046] The key role of the impedance matrix in this process is as follows: First, the node admittance matrix is ​​formed from the AC system network topology, and the basic node impedance matrix is ​​obtained by inversion. Based on this, by equating the DC converter station to the additional admittance or impedance on the AC nodes, the influence of the DC system is systematically embedded into the AC framework, thereby constructing a hybrid impedance matrix (or extended impedance matrix) characterizing the complete AC-DC coupling relationship. The elements of this hybrid impedance matrix directly quantify the electrical coupling strength between any nodes (including DC landing points). Furthermore, through the characteristic equation and the hybrid impedance matrix (denoted as...), the... The joint solution of the two methods can quantify the impact of multiple DC feeds on the voltage distribution of the AC system, providing a basis for subsequent short-circuit ratio calculations.

[0047] (2) Construct a power flow Jacobian matrix that considers AC and DC characteristics The core of power flow calculation is the Jacobian matrix, whose structure directly determines the convergence of the Newton-Raphson method. For systems with multiple DC inlets, the influence of the DC system needs to be embedded into the traditional AC power flow Jacobian matrix. Specifically, based on the external characteristic equations of the AC and DC systems, and according to the alternating iterative calculation method of power flow, partial derivatives are taken with respect to the phase angle and voltage of each node to obtain a Jacobian matrix that includes the characteristics of the DC system, forming the following extended form:

[0048] In the formula: ΔP and ΔQ are the unbalance vectors of active and reactive power at the node; Δθ and ΔU are the correction vectors of the phase angle and magnitude of the node voltage. , , which are Jacobian submatrices formed by the partial derivatives of active power with respect to voltage phase angle and voltage amplitude, respectively; , , which are Jacobian submatrices formed by the partial derivatives of reactive power with respect to voltage phase angle and voltage amplitude, respectively.

[0049] The key lies in the Jacobian matrix elements (such as...) of the DC converter bus. and The corresponding position in the equation needs to be added to the traditional AC power partial derivatives, along with new partial derivative terms obtained from the linearization of the DC external characteristic equation (such as...). .

[0050] The role of the impedance matrix in this process is to quickly calculate the changes in the overall network voltage distribution caused by changes in node injection current, using the extended impedance matrix established in Part (1). This provides key auxiliary information for quickly evaluating system strength characteristics (such as voltage stability) related to the Jacobian matrix, thereby enhancing the depth and efficiency of the entire analysis process.

[0051] (3) Optimize the parameters of the Thevenin equivalent model Thevenin equivalent aims to simplify a complex external system into one consisting of an equivalent voltage source. and equivalent impedance The constructed model is suitable for analyzing the stability of local areas (such as the DC landing point). The key to optimization is accurate extraction. Traditional methods struggle to account for the complex interactions between DC landing points in multi-infeed scenarios. This invention, based on an extended impedance matrix, directly and accurately obtains equivalent parameters through matrix operations. Its main principle is as follows: The Thevenin equivalent complex form equation is:

[0052] in, To focus on the voltage phasors of nodes (ports), This is the phasor of the current injected into the node. The equivalent parameters can be directly derived from the system extended node impedance matrix constructed in Part (1) (here denoted by the symbol). Extracted from (representation). Specifically, for a port of interest, its Thevenin equivalent impedance That is, the node is in The corresponding self-impedance in the matrix.

[0053] When focusing on multiple DC landing points (i.e., multi-port networks), it is necessary to use matrix contraction techniques to represent the external system as an equivalent multi-port Thevenin model. Let the set of boundary nodes (DC landing points) be... B The internal node set is I Then the global impedance matrix can be divided into blocks as follows:

[0054] In the formula: This is the self-impedance matrix of the boundary nodes, representing the electrical coupling relationship between the boundary nodes themselves; and The mutual impedance matrix between boundary nodes and internal nodes ( This characterizes the mutual influence between internal and external systems; This is the self-impedance matrix of the internal nodes.

[0055] By matrix contraction, the equivalent multiport impedance matrix at the boundary node level can be directly obtained. This matrix is ​​the desired multiport Thevenin equivalent impedance. .

[0056] The equivalent impedance directly extracted from the impedance matrix as described above ( ) as a high-precision initial value, and combined with real-time measurement data from the PMU to calculate the equivalent voltage source ( These factors, together, constitute the initial conditions for optimization. Subsequently, online rolling optimization is performed using the least squares method to further reduce the impact of measurement errors. This process ensures that the impedance characteristics of the Thevenin equivalent model in key frequency bands such as power frequency and subsynchronous frequency are highly matched with those of the actual system, providing a reliable equivalent network foundation for the subsequent accurate calculation of the short-circuit ratio.

[0057] (4) Apply the branch addition method to quickly calculate the multi-infeed short-circuit ratio The multiple-infeed short-circuit ratio (MISCR) is a core indicator for evaluating the strength of multi-DC systems. Traditional methods require individual short-circuit calculations for each DC input point, resulting in a large computational burden. This invention employs the append-branch method, dynamically updating the system impedance by progressively adding equivalent DC branches to the initial AC system impedance matrix, thus achieving efficient MISCR calculation.

[0058] The calculation process is as follows: Step 1: Construct the node impedance matrix of the pure AC system without DC feed. This matrix can be used to calculate the single-feed short-circuit ratio of each node.

[0059] Step 2: Each DC converter station is equivalent to an impedance branch added to the corresponding AC node. This impedance can be derived from the equivalent model of the converter.

[0060] Step 3: For each equivalent DC branch added, update the entire network node impedance matrix according to the branch addition method. The general update formula is:

[0061] In the formula: ; For the updated node , Mutual impedance between them; , , , , , These are the nodes before the update. , ,node , ,node , ,node , ,node , ,node , Mutual impedance between them; , These are the nodes before the update. , Self-impedance.

[0062] Step 4: Repeat step 3 until all DC branches have been added, obtaining the final system equivalent impedance matrix. .

[0063] Step 5: Calculate the multi-infeed short-circuit ratio based on the updated impedance matrix. This is to account for the converter station's own reactive power compensation. In response to the effects of multi-infeed interaction, this invention calculates the multi-infeed short-circuit ratio (MISCR) based on the impedance matrix using the following formula:

[0064] In the formula: (that is (for converter stations) i The short-circuit capacity of the AC bus is calculated based on the final system equivalent impedance matrix ( The Thevenin equivalent impedance of the node obtained in ) ); (that is (This is a converter station) i The reactive power provided by reactive power compensation equipment such as filters and parallel capacitors on the AC bus; To correspond to reactive power compensation equipment The equivalent capacitive impedance; DC system i Rated conveying capacity; DC system j Rated conveying capacity; To account for the interaction effects of other DC landing points, the DC system i The equivalent power; and These are the system equivalent impedance matrices. Middle node i Self-impedance and nodesi , j Mutual impedance between them; DC landing point j landing point i The multi-feed interaction factor is approximately calculated as follows: .

[0065] This formula clearly elucidates the physical essence of MISCR: the numerator term ( - This represents the net support capacity of the network itself for the DC landing point after deducting local reactive power compensation; the denominator ( This represents the total equivalent load of the DC system after considering the interactive effects of multiple infeeds. This invention uses the impedance matrix to uniformly obtain the key parameters in the formula, ensuring the accuracy and consistency of the calculation.

[0066] It should be noted that the extended impedance matrix involved in this invention... With the system's equivalent impedance matrix ( Both are physically identical, representing a complete AC / DC hybrid system. The difference lies in their underlying structure: the extended impedance matrix is ​​the theoretical definition and complete description of the system, while the system equivalent impedance matrix is ​​obtained through a specific calculation method (additional branch method) to implement the extended impedance matrix with concrete data, used for efficient calculation of specific indicators such as multi-infeed short-circuit ratio. Thevenin equivalent impedance... In principle, it is extracted from the extended impedance matrix. Specifically, in the calculation process, it is to extract the self-impedance of the corresponding node from the system equivalent impedance matrix.

[0067] (5) Critical short-circuit ratio analysis based on the generalized Nyquist criterion The purpose of this step is to determine the boundary conditions for the system's transition from stable to unstable, i.e., the critical short-circuit ratio. This analysis is based on the generalized Nyquist stability criterion, which is applicable to analyzing dynamic interactions between systems such as inverter-connected grid systems and weak grids.

[0068] The analysis process is as follows: 1. Construct a frequency domain impedance model: First, the entire system is considered as a feedback interconnected system: the AC network is equivalent to a multi-port impedance model. (Its elements can be obtained from the extended impedance matrix in part (1) through frequency domain transformation), and each DC / new energy converter is equivalent to its output impedance model. The system's loop gain matrix Constructed from this, it is usually defined as = (in = (This is the admittance model for the converter).

[0069] 2. Applying the generalized Nyquist criterion: Calculate the loop gain matrix All eigenvalues ( i =1,2,...). For each eigenvalue... Plot the Nyquist curve in the complex frequency domain. The generalized Nyquist criterion states that if the Nyquist curves of all eigenvalues ​​do not encircle the critical point, then the closed-loop system is stable.

[0070] 3. Determine the critical short-circuit ratio: The critical short-circuit ratio is the short-circuit ratio value corresponding to the system being at the stability boundary. By substituting the short-circuit ratio as a variable parameter into the model from step 1 and repeating the stability assessment from step 2, a parameter scan can be performed. When the Nyquist curve of a certain eigenvalue is found to pass exactly through the critical point, the corresponding short-circuit ratio is the critical short-circuit ratio. By analyzing the dominant interaction modes near the critical point (i.e., the modes corresponding to that eigenvalue), the dominant factors causing system instability can be identified.

[0071] Step 4: Dynamic Correction of Sensitive Boundary Parameters Dynamic correction of sensitive boundaries is a core element in ensuring the real-time performance and accuracy of system strength assessment. It involves dynamically adjusting and optimizing key boundary conditions in short-circuit ratio calculation, taking into account the time-varying characteristics of grid parameters, the fluctuations in renewable energy output, and the interactions of multiple infeeders. The core technology lies in constructing an adaptive correction mechanism that tracks system state changes in real time and ensures the accuracy of the short-circuit ratio index through multi-dimensional parameter iteration, providing dynamically adaptable boundary support for system strength assessment.

[0072] The targets of dynamic correction cover three key parameters: first, Thevenin equivalent parameters (equivalent voltage source) With equivalent impedance The accuracy of the short-circuit ratio calculation results is directly affected by several factors. Secondly, the node admittance matrix elements in multi-feed systems need to be dynamically updated to reflect the interaction between DC landing points. Thirdly, the reactive power compensation coefficient of new energy power plants (such as the dynamic susceptance of wind farms / photovoltaic power plants) needs to be adjusted in real time according to the output changes to correct the MRSCR index.

[0073] It directly determines the short-circuit capacity. The size, and It is the numerator in the short-circuit ratio calculation formula. If Inaccuracy will cause the entire short-circuit ratio metric to be inaccurate. Changes in system topology and operating conditions (such as line switching and load fluctuations) will alter the nodal admittance matrix. Step four, dynamically updating it, means that it is used to construct the extended impedance matrix. The basis is updated in real time, thus ensuring that any impedance parameters extracted from it (including...) are accurate and up-to-date. The accuracy of the MISCR formula needs to be considered. In the numerator of the MISCR formula, the short-circuit capacity needs to be determined. Deducting local reactive power compensation Fluctuations in the output of new energy sources directly affect their reactive power output; this is corrected in real time using a dynamic susceptance model. This makes the calculated net support capacity ( - This is more in line with reality.

[0074] A joint algorithm combining least squares and extended Kalman filtering is used to achieve parameter estimation and tracking. A simplified flowchart of the algorithm is as follows: Figure 3 As shown, the specific steps are as follows: (1) Initial parameter estimation Based on historical operating data, the initial values ​​of Thevenin equivalent parameters and admittance matrix are calibrated offline using the least squares method. The objective function is:

[0075] In the formula, These are the model's predicted values. These are PMU measurements.

[0076] (2) Real-time data fusion An extended Kalman filter is introduced to correct the parameters online. The following state equation and observation equation are established:

[0077] In the formula, For the parameter vector to be corrected (e.g.) , wait, These are system node admittance matrix elements that need to be tracked and updated in real time. For process noise, It is a nonlinear observation function. For measuring noise.

[0078] (3) Sliding data processing A sliding time window mechanism (configurable window length) is employed, applying weighted least squares to the data within the window, assigning higher weights to more recent data. Weights are allocated according to a time decay factor.

[0079] In the formula, The attenuation coefficient is taken as 0.1~0.3. This refers to the current moment.

[0080] (4) Reactive power correction of new energy sources Establish a dynamic susceptance model to adaptively adjust the reactive power output of renewable energy power plants:

[0081] in, This is a reactive power instruction; This represents the actual reactive power value; Δ Q ( t This represents reactive power deviation. This is the initial susceptance; The compensation coefficient can be obtained by fitting historical data (typical value 0.8~1.2).

[0082] The parameter correction process is automatically triggered when any of the following conditions are met: the rate of change of the Thevenin equivalent impedance exceeds a set threshold (e.g., The rate of change of key elements in the system admittance matrix exceeds a set threshold (e.g., When the output or active / reactive power compensation of the new energy power station fluctuates significantly (e.g., the output fluctuation exceeds 10% or the reactive power compensation capacity adjustment exceeds 15%).

[0083] Through the above mechanism, this step ensures that the boundary parameters on which the short-circuit ratio calculation depends can respond to changes in system state in real time and accurately, providing a dynamic and adaptive data foundation for real-time assessment of system strength.

[0084] Step 5: Comprehensive Assessment of Power System Strength and Weakness This step involves a systematic assessment of system strength by combining multi-dimensional quantitative indicators with static characteristics and dynamic responses. Figure 4 The presentation outlines a simplified process for comprehensive evaluation. First, the static strength of the DC system is quantified based on the multiple infeed short-circuit ratio (MISCR), and the boundary conditions from stability to oscillation are defined using the optimized equivalent impedance parameters from the Thevenin equivalent model. Second, the instability critical point is determined using the generator power angle curve, and the synchronization capability of the multi-machine system is assessed using the Lyapunov exponent. Simultaneously, low-frequency oscillation modes are identified through small-disturbance stability analysis, and the system's damping characteristics are determined using the impedance matrix eigenvalue trajectory. Finally, a comprehensive evaluation system is constructed using a weighted fusion of multiple indicators or the analytic hierarchy process (AHP), encompassing both quantitative indicators such as the short-circuit ratio and qualitative assessments such as voltage stability and transient stability, forming a strength hierarchy map from local nodes to the global system.

[0085] (1) Quantitative assessment of static strength This part aims to quantitatively classify the static voltage support capability of the system. Directly applying the results of the multiple-infeed short-circuit ratio (MISCR) calculated in step 3 (4), the following static strength classification criteria based on MISCR are established: When MISCR < 1.6, the system can be judged as an extremely weak system; when 1.6 < MISCR < 2.5, the system can be judged as a weak system; when MISCR > 2.5, the system can be judged as a strong system.

[0086] Meanwhile, combining operation data such as node voltage deviation and power transfer margin, and considering the potential impact of the dynamic regulation capabilities of DC control systems (such as power modulation and voltage support) on the system equivalent impedance (this system equivalent impedance is the Thevenin equivalent impedance optimized by expanding the impedance matrix in step 3 (3) ), auxiliary verification and correction are carried out on the classification results, and finally a static-control collaborative evaluation matrix is formed to accurately depict the spatial distribution of the static strength of the system.

[0087] (2) Transient and dynamic stability assessment This link aims to evaluate the stable recovery ability of the system after suffering a large disturbance and its own dynamic characteristics. Specifically, what is used in this part is a dynamically updated power grid model that is maintained and updated jointly by step 2 and step 4 based on the framework constructed in step 1.

[0088] Transient stability assessment: Based on the high-precision power grid hybrid model constructed in step 1, extreme fault conditions such as DC blocking and large-scale disconnection of new energy are designed for electromagnetic transient simulation. Through the simulation results, the post-fault system behavior is quantitatively evaluated, including voltage stability characteristics (recovery time and drop depth of key node voltages), power angle stability characteristics (change trajectory of power angles between leading generator sets), and power oscillation characteristics (attenuation rate of power oscillation of key tie lines).

[0089] Dynamic stability assessment: To quantitatively analyze the inherent stability trend of the system after being disturbed, the Lyapunov exponent is introduced as an auxiliary criterion. By calculating the Lyapunov exponent (λ) of the system's disturbed trajectory, the dynamic behavior of the system is quantitatively classified: When , it indicates that the system is dynamically divergent, showing chaotic characteristics and there is a risk of instability; when , it indicates that the system is in a critical oscillation state; when , it indicates that the system is dynamically convergent, showing stable characteristics.

[0090] Correlation analysis is carried out on the stability indicators and the key structural parameters obtained in step 3. For example: Analyze the correlation between the equivalent short-circuit ratio (MISCR) of each region and the post-fault voltage recovery speed. Study the correlation between the magnitude of the Thevenin equivalent impedance and the damping characteristics of the system.

[0091] (3) Stability analysis under small disturbances Based on the frequency domain impedance model established in step three (5), a small-disturbance stability analysis is performed: the eigenvalues ​​of the system state matrix are calculated using the QR decomposition method to identify the dominant low-frequency oscillation mode. By participating in factor analysis, the key generator / inverter and propagation path that cause the oscillation are located. Combining the distribution of the impedance matrix eigenvalue trajectory, the system damping characteristics are evaluated, and the vibration suppression effects of devices such as the power system stabilizer (PSS) and the additional damping controller (SEDC) are analyzed.

[0092] (4) Wide-area measurement and adaptive evaluation This section utilizes wide-area measurement data from the Power Management Unit (PMU) to perform real-time, online assessments of the dynamic stability characteristics of the power system using algorithms such as Prony transform. Specifically, it applies algorithms like Prony transform to identify real-time modal parameters in the PMU measurement data. The core of this approach is to decompose the signal into a superposition of damped oscillating modes.

[0093] This enables online tracking of the frequency and damping of the dominant oscillation mode. In the formula, For amplitude, The damping factor, For frequency, This represents the phase angle. Compared to the Fourier transform, the Prony transform can more accurately capture the characteristics of transient signals, making it suitable for non-stationary signal processing, suppressing noise interference, and improving parameter estimation accuracy.

[0094] (5) Construct a comprehensive evaluation system Based on the above multi-dimensional assessment results, an index for comprehensive system strength assessment is constructed by integrating indicators using methods such as the analytic hierarchy process (AHP). The final output includes a system strength classification map and planning and operation recommendations. For example, the strength levels of different regions are clearly marked on the power grid geographical wiring map. This provides priority recommendations for grid reinforcement to power grid planning departments and dynamic operation control strategies to operation departments, such as prioritizing the acceptance of new energy sources in strong grid areas and restricting DC power in weak grid areas.

Claims

1. A comprehensive evaluation method for the strength and weakness of a large-scale new energy power system, characterized in that, include: S1. Constructing a hybrid power grid model: A hierarchical strategy is adopted to construct a hybrid power grid model that includes synchronous generators, AC lines, DC lines, and new energy power plants; S2. Construction and updating of the extended impedance matrix: Based on the power grid hybrid model, the extended impedance matrix of the system is constructed using the additional branch method, and the extended impedance matrix is ​​dynamically updated based on real-time measurement data. S3. System strength quantification calculation: Based on the updated extended impedance matrix, calculate the multi-feed short-circuit ratio and extract Thevenin equivalent parameters from it; S4. Comprehensive evaluation of system strength and weakness: Integrate the results of multi-infeed short-circuit ratio, Thevenin equivalent parameters and static and dynamic stability analysis to generate a comprehensive classification map of system strength and weakness.

2. The method according to claim 1, characterized in that, The hierarchical strategy in S1 includes: The synchronous generator adopts a sixth-order practical model; The AC circuit adopts a π-type equivalent circuit; The DC lines adopt the standard model of the International Organization of Large Electric Systems; The new energy power station adopts a polymer equivalent unit, whose model integrates reactive power compensation equipment and maximum power point tracking control characteristics.

3. The method according to claim 1, characterized in that, The extended impedance matrix in S2 is constructed using the method of adding branches, including: Start with the nodal impedance matrix of a pure AC system; Each DC converter station is equivalent to an impedance branch added to the corresponding AC node; Each time an equivalent DC branch is added, the impedance matrix of all nodes in the network is updated. The extended impedance matrix is ​​obtained after all DC branches have been added.

4. The method according to claim 1, characterized in that, The dynamic updating of the extended impedance matrix based on real-time measurement data in step S2 includes: Based on real-time synchronous phasor measurement unit data, the nodal admittance matrix elements, load model parameters, and renewable energy reactive power compensation parameters that constitute the basis of the extended impedance matrix are updated in a rolling manner through an adaptive filtering algorithm.

5. The method according to claim 1, characterized in that, A dynamic correction step for sensitive boundary parameters is performed between S2 and S3: when the rate of change of Thevenin equivalent parameters, nodal admittance matrix elements or reactive power compensation coefficients of new energy power plants exceeds their corresponding set thresholds, the joint algorithm of least squares method and extended Kalman filter is used to perform online joint estimation and optimization of the above parameters.

6. The method according to claim 1, characterized in that, In step S3, the Thevenin equivalent parameters are extracted from the extended impedance matrix, including: For a single node of interest, the self impedance corresponding to that node in the extended impedance matrix is ​​directly extracted as its Thevenin equivalent impedance. For multiple DC landing points, the equivalent multi-port Thevenin impedance at the boundary node level is obtained from the extended impedance matrix using matrix shrinking techniques.

7. The method according to claim 1, characterized in that, The formula used in S3 to calculate the multi-infeed short-circuit ratio is: In the formula, Converter station i The short-circuit capacity of the AC bus is determined by the Thevenin equivalent impedance extracted from the extended impedance matrix. Converter station i Local reactive power compensation of AC busbar; and DC system i and j Rated transmission power; DC landing point j landing point i The multi-feed mutual influence factor is determined based on the mutual impedance in the extended impedance matrix.

8. The method according to claim 1, characterized in that, The method also includes a critical short-circuit ratio analysis step: Based on the extended impedance matrix, a frequency domain impedance model of the AC network and converter is constructed. A loop gain matrix is ​​formed, and the Nyquist curve of its eigenvalues ​​is analyzed using the generalized Nyquist criterion to determine the system stability boundary and the corresponding critical short-circuit ratio.

9. The method according to claim 1, characterized in that, The static and dynamic stability analysis in S4 includes: Static strength quantitative assessment: Based on the calculation results of the multi-infeed short-circuit ratio, the static strength of the system is quantitatively graded according to the preset grading standard; Transient and dynamic stability assessment: Time-domain simulation is performed based on the aforementioned power grid hybrid model to evaluate the stability characteristics of the system under extreme fault conditions; Small disturbance stability analysis: Based on the frequency domain impedance model derived from the extended impedance matrix, eigenvalue analysis is performed to identify the low-frequency oscillation modes and damping characteristics of the system.

10. The method according to claim 9, characterized in that, The classification criteria for the static strength quantitative assessment are as follows: a multi-infeed short-circuit ratio less than 1.6 is judged as an extremely weak system, between 1.6 and 2.5 is judged as a weak system, and greater than 2.5 is judged as a strong system.

Images