Heterogeneous new energy transient voltage stabilization network following-network construction resource dynamic reconstruction method

Abstract

The invention relates to a network following-network construction resource dynamic reconstruction method for heterogeneous new energy transient voltage stability, and the method comprises the steps: building a network following type converter mathematical model and a network construction type converter model based on the operation data of a grid-connected point and a heterogeneous resource node, deducing a power voltage coupling relation, and constructing a Lyapunov function as a transient stability criterion; analyzing the influence mechanism of the core parameters on the voltage stability, and determining the voltage support boundary under the amplitude-limited operation of the constructed network type converter; and by taking minimization of the peak value of the Lyapunov function as a target, constructing an optimization model of the proportion of the optimal networking type resource capacity to the total new energy output. Mathematical models are respectively established for a network-constructing type converter and a network-following type converter, and a power voltage coupling relation is deduced; a Lyapunov function integrating converter power deviation and control parameters is constructed, theoretical support is provided for optimal networking proportion solution and parameter regulation and control, and the problems that heterogeneous system coupling characteristics are not accurately described and stability criterion adaptability is poor are solved.

Description

Technical Field

[0001] This application relates to the field of power distribution system collaborative control technology, specifically to a method for dynamic reconfiguration of grid-connected resources to stabilize transient voltages of heterogeneous new energy sources. Background Technology

[0002] As the penetration rate of renewable energy continues to deepen, the proportion of conventional power sources, represented by synchronous machines, continues to decline, resulting in a significant weakening of the system's voltage support capability. This poses a severe challenge to the voltage security and stable operation of modern power systems. Currently, large-scale renewable energy power plants generally adopt grid-connected converters that exhibit controlled current source characteristics. In recent years, grid-connected converters, which possess voltage source characteristics and can actively provide voltage support, have been applied in numerous projects. The coexistence of grid-connected and grid-connected control strategies has created a complex pattern of heterogeneous renewable energy sources fed into the power system. In grid-connected converter systems, the transient voltage response characteristics under the control switching of various heterogeneous resources are complex and variable, easily inducing new transient voltage instability problems, leading to challenges in voltage stability assessment and stability control of heterogeneous resource grid-connected systems. Furthermore, although GFMs possess dynamic voltage support capabilities, they may lose their voltage source characteristics during transient periods due to entering a limiting state. Simultaneously, both GFMs and GFLs must meet steady-state power flow constraints. Therefore, in the context of massive heterogeneous resource access, how to rationally allocate the capacity ratio between grid-connected and grid-connected resources has become a critical issue that urgently needs to be addressed. Currently, there are numerous research findings on the stability analysis and evaluation of heterogeneous renewable energy systems, analyzing the interaction mechanism between GFM and GFL in small-signal stability, indicating that GFM affects the small-signal stability of PLL integrated systems by enhancing grid strength. Regarding grid-connected capacity configuration research, existing studies mostly propose capacity configuration recommendations based on steady-state grid strength.

[0003] However, existing technologies have obvious shortcomings: the capacity allocation scheme does not take into account the dynamic voltage response characteristics of heterogeneous resources, and its applicability under transient voltage instability conditions lacks quantitative verification. Therefore, there is still a lack of heterogeneous resource capacity allocation and control methods that take into account transient voltage safety and stability constraints.

[0004] Therefore, this invention proposes a dynamic reconfiguration method for heterogeneous new energy transient voltage stability based on grid-connected resources. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this application provides a method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources, specifically adopting the following technical solution.

[0006] A method for dynamic reconfiguration of grid-connected resources with transient voltage stability of heterogeneous new energy sources includes the following steps.

[0007] Multi-dimensional data acquisition: Real-time acquisition of full operational data from the grid-connected PCC and various heterogeneous resource nodes, covering three dimensions: electrical parameters, control parameters, and energy status.

[0008] Transient disturbance identification and assessment: Based on the collected PCC voltage data, a multi-threshold disturbance monitoring mechanism is constructed to continuously monitor voltage amplitude changes. When a voltage drop or rise exceeding the preset threshold is detected, the dynamic reconstruction process is immediately triggered, and core features such as the time of disturbance occurrence, initial voltage amplitude, and rate of change are recorded.

[0009] Voltage response characteristics analysis of heterogeneous new energy systems: Based on the differences in control mechanisms of heterogeneous new energy converters, mathematical models of grid-connected converters (GFL) and grid-connected converters (GFM) are established, the coupling relationship between power and voltage is derived, and a transient stability criterion adapted to heterogeneous hybrid systems is constructed.

[0010] Influence mechanism of dominant voltage stability parameters and voltage support boundary analysis of grid-type converter: Based on mathematical model, the influence of core parameters on voltage stability is analyzed, and the voltage support boundary under GFM limited operation is clarified.

[0011] Dynamic Reconfiguration Optimization of Network-Network Resources Based on Second-Order Cone Programming (SOCP): Considering system operation constraints and control limiting conditions, an online optimization scheme for the optimal network ratio r is proposed to improve the transient voltage stability of the system. The objective function is to minimize the transient energy peak value min H(x). Since H(x) is a non-convex function, slack variables are introduced, and SOCP is used to efficiently solve the non-convex model.

[0012] Simulation Verification and Effectiveness Analysis: To verify the feasibility and effectiveness of the proposed dynamic reconfiguration method, a simulation test platform was built based on the improved IEEE 39-node system. The built-in transient stability simulation module of the digital twin was called simultaneously to conduct comparative simulations, verifying the consistency between the module's simulation results and the data of the physical test platform. This ensures that the module can replace some physical test scenarios and improve the implementation efficiency of the control scheme.

[0013] The technical solution of this application has achieved the following beneficial effects.

[0014] To address the differences in control mechanisms between GFM-based and GFL-based converters, refined mathematical models incorporating control loop characteristics, equipment impedance, and limiting constraints were established, and the power-voltage coupling relationship was derived. Furthermore, a Lyapunov function integrating converter power deviation and control parameters was constructed to quantitatively characterize the system's transient stability. The effect of increasing the GFM proportion on expanding the system's stability domain was clarified, revealing the technical mechanism of grid-based resources as the core of active voltage support. This provides solid theoretical support for solving the optimal grid ratio and parameter control, and solves the problems of inaccurate characterization of heterogeneous system coupling characteristics and poor adaptability of stability criteria in traditional methods. Attached Figure Description

[0015] Figure 1 This diagram illustrates the relationship between the system's stability profile and the Lyapunov function under different GFM capacity ratios; where... Figure 1 (a) is a schematic diagram showing the relationship between heterogeneous resource output and voltage stability when the proportion of grid-type resource capacity to total new energy output is 10%. Figure 1 (b) is a schematic diagram of the system transient response trajectory when the proportion of grid-type resource capacity to total new energy output is 10%; Figure 1 (c) is a schematic diagram showing the relationship between heterogeneous resource output and voltage stability when the proportion of grid-type resource capacity to total new energy output is 30%. Figure 1 (d) is a schematic diagram of the system transient response trajectory when the network-type resource capacity accounts for 30% of the total new energy output.

[0016] Figure 2 The diagram shows the relationship between the dominant control parameters and transient voltage stability. In this diagram, 2(a) shows the influence of the upper limit amplitude, and 2(b) shows the influence of the droop coefficient of the reactive power control loop of the grid-type converter and the damping coefficient of the active power control loop of the GFM converter.

[0017] Figure 3 This is a graph showing system operating data under severe conditions (IM=60%).

[0018] Figure 4 A schematic diagram of the topology of the IEEE 39-node system.

[0019] Figure 5 5(a) is a schematic diagram of the N-1 fault result when IM=40%; 5(b) is a graph of the output active power when IM=40%; 5(c) is a graph of the PCC voltage when IM=40%; 5(c) is a graph of the evaluation index when IM=40%. Figure 6The diagram shows the N-1 fault result when IM=40% after optimization control. 6(a) is the output active power curve when IM=40% after optimization control; 6(b) is the PCC voltage curve when IM=40% after optimization control; and 6(c) is the evaluation index curve when IM=40% after optimization control.

[0020] Figure 7 The diagram shows the N-1 fault result when IM=50%. 7(a) is the output active power curve when IM=50%; 7(b) is the PCC pressure curve when IM=50%; and 7(c) is the evaluation index curve when IM=50%.

[0021] Figure 8 The diagram shows the N-1 fault result when IM=50% after optimization control. 8(a) is the output active power curve when IM=50% after optimization control; 8(b) is the PCC voltage curve when IM=50% after optimization control; and 8(c) is the evaluation index curve when IM=50% after optimization control.

[0022] Figure 9 The diagram shows the N-1 fault result when IM=60%. 9(a) is the output active power curve when IM=60%; 9(b) is the PCC voltage curve when IM=60%; and 9(c) is the evaluation index curve when IM=60%.

[0023] Figure 10 The diagram shows the N-1 fault result when IM=60% after optimization control. 10(a) is the output active power curve when IM=60% after optimization control; 10(b) is the PCC voltage curve when IM=60% after optimization control; and 10(c) is the evaluation index curve when IM=60% after optimization control.

[0024] Figure 11 The diagram shows the control parameter data of GFM / GFL used in the simulation.

[0025] Figure 12 This is a graph showing system operating data under severe conditions (IM=40%).

[0026] Figure 13 This is a graph showing system operating data under severe conditions (IM=50%). Detailed Implementation

[0027] The present application will now be further described with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application and should not be construed as limiting the scope of protection of the present application. It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present application.

[0028] This application investigates the mechanism by which the capacity ratio and control parameters of network-type equipment affect voltage support capability, and clarifies the voltage support boundary of network-type equipment under limited operation conditions. Since the active power output of the network-type subsystem is constrained by operational limits and control loop limiting conditions, this application proposes a dynamic reconfiguration method for heterogeneous resource capacity allocation based on second-order cone programming, considering multiple constraints during operation. This method solves for the optimal network ratio and improves the system's transient voltage support capability.

[0029] like Figure 1 As shown, this invention discloses a method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources, comprising the following steps.

[0030] Step 1: Multi-dimensional data acquisition: Real-time acquisition of full operational data of the grid-connected PCC and each heterogeneous resource node, covering three dimensions: electrical parameters, control parameters, and energy status.

[0031] Step 2, Transient Disturbance Identification and Assessment: Based on the PCC voltage data collected in Step 1, a multi-threshold disturbance monitoring mechanism is constructed to continuously monitor voltage amplitude changes. When a voltage drop or rise exceeding the preset threshold is detected, the dynamic reconstruction process is immediately triggered, and core features such as the time of disturbance occurrence, initial voltage amplitude, and rate of change are recorded.

[0032] Step 3: Voltage response characteristics analysis of heterogeneous new energy system: Based on the differences in control mechanisms of heterogeneous new energy converters, mathematical models of grid-connected converter GFL and grid-connected converter GFM are established, the coupling relationship between power and voltage is derived, and a transient stability criterion adapted to the heterogeneous hybrid system is constructed to provide theoretical support for subsequent boundary analysis and optimization modeling.

[0033] Step 4: Influence mechanism of dominant voltage stability parameters and voltage support boundary analysis of grid-type converter: Based on the mathematical model in Step 3, the influence of core parameters on voltage stability is analyzed, and the voltage support boundary under GFM limiting operation is clarified.

[0034] Step 5: Dynamic reconfiguration and optimization of network resources based on second-order cone programming (SOCP).

[0035] Based on the analysis in steps 3 and 4, and considering system operation constraints and control limiting conditions, an online optimization scheme for the optimal grid ratio r is proposed to improve the transient voltage stability of the system. The objective function is to minimize the transient energy peak value min H(x). It is easy to see from the Lyapunov function that H(x) is a non-convex function. By introducing relaxation variables, the non-convex model can be solved efficiently through SOCP.

[0036] Step 6, refer to Figures 4-13Simulation Verification and Effectiveness Analysis: To verify the feasibility and effectiveness of the proposed dynamic reconfiguration method, a simulation test platform was built based on the improved IEEE 39-bus system. A comparative simulation was conducted using the built-in transient stability simulation module of the digital twin, verifying the consistency between the module's simulation results and the data from the physical test platform. This ensures that the module can replace some physical test scenarios and improve the efficiency of implementing control schemes. Through N-1 fault simulations under different operating conditions, the transient voltage response characteristics of the system before and after optimization were quantitatively analyzed, verifying the actual effect of the method on improving system stability. The error calculation method is as follows: three core indicators—voltage amplitude, active power, and reactive power—are selected, and the maximum error of each indicator is taken as the overall error, ensuring that the error is ≤3%.

[0037] Example 1.

[0038] This embodiment 1 discloses a method for dynamic reconfiguration of grid-connected resources to ensure transient voltage stability of heterogeneous new energy sources. In step 1, electrical parameters include PCC voltage amplitude and phase angle, frequency and volatility, active and reactive power output of each inverter, and terminal voltage and output current. Control parameters include the current control mode of each converter, control loop proportional gain, droop factor, virtual internal potential reference value and limit value, low voltage ride-through (LVRT) trigger threshold and adjustment strategy. Energy status parameters include energy storage state of charge, available adjustable capacity of each new energy unit, and active and reactive power reserve capacity. A high-frequency synchronous acquisition module ensures data real-time performance, and edge computing nodes are used for data preprocessing to remove abnormal noise data, ensuring data integrity and accuracy.

[0039] The digital twin system updates the virtual mapping model of the physical system in real time based on preprocessed data, achieving synchronization of the operating states of the physical and virtual layers. The physical layer refers to the hardware entities and operating states of the actual heterogeneous renewable energy grid-connected system, including physical equipment such as GFL converters, GFM converters, transmission lines, energy storage devices, and load devices, as well as the real-time electrical and control parameters of each device. The virtual layer is a digital mirror of the physical layer, consisting of a geometric model replicating the physical system topology, a data model mapping the real-time state, and a built-in simulation module. It can accurately synchronize the dynamic changes of key parameters such as converter control characteristics, line equivalent impedance, and load distribution. The system has a built-in transient stability simulation module that, based on real-time data synchronized from the virtual layer, extrapolates the voltage stability margin under the current operating state, identifies potential voltage weaknesses through sensitivity analysis, and predicts the system response trend under different disturbance scenarios, providing accurate data support and a model foundation for subsequent disturbance response and reconfiguration optimization. The digital twin model and the subsequent SOCP optimization model use Gigabit Ethernet + 5G edge gateway to achieve data interaction. The virtual model update frequency is consistent with the sampling frequency (1kHz), the data synchronization error is controlled within ≤10μs, and the clock deviation is corrected through the timestamp alignment compensation mechanism to ensure the timing consistency of millisecond-level control.

[0040] Example 2.

[0041] This embodiment 2 discloses a dynamic reconfiguration method for heterogeneous new energy transient voltage stability based on grid connection and network construction. In step 2, the criteria for determining weak grid conditions are: line short-circuit ratio (SCR) ≤ 3, new energy penetration rate ≥ 50%, or voltage fluctuation rate ≥ 2% / ms. If any one of these conditions is met, the system switches to the weak grid threshold. The disturbance exit mechanism is as follows: if the voltage of a key node recovers to the range of 0.9-1.1 pu and the stabilization time is ≥ 2 power frequency cycles, the system automatically exits dynamic reconfiguration and maintains the current optimal network construction ratio r. If the steady-state operation exceeds 5 minutes, the system gradually reverts to the benchmark r value under economic operating conditions.

[0042] Rapidly conduct quantitative analysis of disturbance characteristics, including voltage dip depth, voltage change rate, whether accompanied by frequency abrupt changes, fault type, and fault location. Simultaneously, utilize a digital twin virtual model to calculate the voltage sensitivity of each node, accurately locate the propagation path of voltage weak points, clarify the impact range and severity of the disturbance on the system, and classify the disturbance level. Based on the built-in transient stability simulation module of the digital twin system, use real-time acquired data to conduct rapid simulation and deduction, predicting the duration of the disturbance, voltage recovery trend, and potential instability risk. During rapid quantitative analysis, the system simultaneously utilizes a pre-decision strategy knowledge base. This knowledge base is generated offline, storing optimal reconstruction strategies pre-calculated for different operating conditions and typical disturbance scenarios. The knowledge base covers 10 load levels, 8 renewable energy output ranges, 5 GFM proportion ranges, and 15 typical fault scenarios. It uses the k-nearest neighbor (KNN) algorithm for online matching, with a similarity threshold set to ≥85%. A default conservative strategy is activated when no suitable strategy is found. The knowledge base is updated offline every 72 hours based on real-time operating data, supplementing with newly added operating conditions and fault cases. The online system quickly matches the current disturbance characteristics and operating status with the knowledge base to obtain basic control strategies, serving as an efficient initial solution for subsequent online optimization. Simultaneously, it invokes a lightweight digital twin virtual model to calculate the voltage sensitivity of each node, providing quantitative support for disturbance level classification and avoiding judgment lag caused by relying solely on real-time data.

[0043] Example 3.

[0044] This embodiment 3 discloses a method for dynamic reconfiguration of grid-connected resources with transient voltage stability of heterogeneous new energy sources, wherein step 3 includes the following steps.

[0045] 3.1 Mathematical Modeling of GFM Grid-Type Converter: The GFM converter adopts a virtual synchronous machine control strategy to simulate the rotor motion characteristics and voltage support capability of a synchronous generator. Its core feature is the voltage source characteristic, which can actively respond to grid voltage changes and generate reactive power to support voltage. The modeling process fully considers the combined effects of virtual internal potential regulation, reactive power-voltage droop control, filtering devices, and line impedance to accurately characterize the voltage support mechanism under transient operating conditions. The differential state equation of the grid-type converter can be expressed as follows.

[0046] Formula (1): .

[0047] In the formula, The amplitude of the AC voltage at the grid-connected bus of the grid-connected converter GFM. and These represent the active power and reactive power output of the GFM converter, respectively, with corresponding reference active power and reactive power as follows: and ; The magnitude of the virtual internal potential. The phase angle is the virtual internal potential. It is the instantaneous angular velocity of the converter's virtual rotor; It is the rated angular velocity of the power grid; the synchronous angular velocity corresponding to the rated frequency of the power grid is the reference value for frequency regulation. It is the rated voltage amplitude of the power grid; the rated line voltage amplitude (or phase voltage amplitude, which needs to be combined with the voltage reference defined by the equation) of the power grid at the converter connection point is the reference value for voltage regulation. t is a time variable. It is the droop coefficient of the GFM reactive power control loop of the grid-type converter. It is the active power reference value, the target active power output value that the GFM converter needs to track, and the given input of the active power-frequency control loop. This is the reactive power reference value, the target reactive power output value that the GFM converter needs to track, and the given input quantity of the reactive power-voltage control loop. Sources include grid voltage support demand commands, adaptively generated values ​​of grid connection point voltage deviation, or reactive power allocation commands from energy storage coordinated regulation; during transient faults, It will increase rapidly, triggering the converter to generate additional reactive power to support the voltage. It is the proportional gain coefficient of the reactive power control loop, and the proportional gain coefficient of the reactive power-voltage regulation loop of the GFM converter.

[0048] Based on the external circuit equations and power transfer equations of the virtual synchronous machine controller of the GFM converter, the output active power, reactive power, and voltage U of the GFM converter can be obtained. gfm The relationship between them is as follows.

[0049] Formula (2) . Equation (3): .

[0050] Where .

[0051] In the formula is the bus voltage at the PCC; is the line equivalent reactance including the virtual resistance, reactance and filter, is its impedance ratio. is the equivalent line impedance value of the GFM converter including the filter and transformer. is the fundamental wave equivalent reactance of the outgoing line of the GFM converter, which is the comprehensive component of the inherent reactance of the line on the GFM side of the network-forming converter, the built-in virtual reactance of the converter and the equivalent reactance of the filtering device. It is the basic core parameter that constitutes the equivalent total impedance of the GFM. Its value directly reflects the impedance characteristics of the line and auxiliary electrical devices to the current, affects the matching relationship between the output power of the GFM converter and the terminal voltage of the machine, and进而关联系统暂态电压支撑能力的强弱. is the turns ratio of the transformer on the GFM converter side, which refers to the ratio of the number of turns of the primary winding to the number of turns of the secondary winding of the transformer supporting the GFM converter.

[0052] Substituting the active power and reactive power output by the GFM converter into Equation (3), it can be seen that the terminal voltage of the GFM machine is a function of the state variables and . The evolution direction of the state variable of the GFM converter is opposite to the voltage change direction. During the voltage dip, the controller responds by increasing the reactive power output, thereby forming an active support ability for the voltage dip. [[ID=3⑥]]

[0053] The terminal voltage of the GFM machine has a strong coupling relationship with the amplitude and phase angle of the virtual internal potential. During the transient voltage dip, the controller can quickly increase the virtual internal potential, compensate for the line voltage drop to increase the reactive power, and form an active voltage support ability. This characteristic is different from the passive response mode of the GFL and is the core support for the voltage stability of the heterogeneous system.

[0054] 3.2. Mathematical modeling of the GFL network-following converter: The GFL converter adopts a current source control strategy, tracks the changes in the grid voltage and frequency to output power, and its core characteristic is the passive response characteristic. Under transient conditions, the LVRT strategy is used to preferentially adjust the reactive current to support the voltage. In the modeling process, the impacts of current tracking control, LVRT mode switching, and current limiting constraints on power output are key depicted, and the reactive response law of the GFL under different disturbance intensities is clarified. The differential state equation of the network-following (GFL) converter can be expressed as.

[0055] Equation (4): . It should be noted that there is an unclear expression "进而关联系统暂态电压支撑能力的强弱" in the original text, and the translation is made as accurately as possible based on the understanding. If there are specific requirements or corrections for this part, it can be adjusted accordingly.

[0056] In the formula, , These are the active current and reactive current output by the GFL converter, respectively. The sampling time interval, , The subscripts represent reference values ​​for the active and reactive currents of the GFL under different control modes. It is a discrete switching function. This can be derived from the equivalent circuit equations and power transfer equations of the GFL subsystem.

[0057] Formula (5): ; .

[0058] in This refers to the terminal voltage of the GFL converter. This is the equivalent line impedance value of the GFL converter, which includes filters and transformers. The equivalent reactance of the GFL output line includes virtual resistance, reactance, and filters. Its impedance ratio.

[0059] GFL terminal voltage is related to state variables. and In GFL converters, reactive current priority strategy is generally adopted during low voltage ride-through. During a downtrend, the reactive power output is increased by adjusting the reactive current to match the actual reactive power output of the grid converter (GFL). Therefore, the total reactive power response of the system can be decomposed into two parts.

[0060] Formula (6): .

[0061] In the formula, r is the proportion of grid-type resource capacity to total new energy output. It represents the total reactive power response of the heterogeneous new energy system, which is the sum of the reactive power output to the grid by all GFL and GFM converters. It is the sum of the actual reactive power output of all grid-connected converters (GFLs), and it is the total amount of reactive power support provided by the GFL cluster through the reactive current priority strategy, reflecting the passive reactive power regulation contribution of the GFLs. It is the sum of the actual reactive power output of all grid-type converters, reflecting the active voltage support contribution of GFM.

[0062] 3.3 Stability Characterization Based on Lyapunov Function: A Lyapunov function adapted to heterogeneous hybrid systems is constructed, integrating core variables such as power deviation and control parameters of GFL and GFM converters to quantify the transient stability of the system. The closer the function value is to 0, the stronger the system stability; a sudden increase in the function value indicates that the system is on the verge of instability. The expression of the Lyapunov function is as follows.

[0063] Formula (7): .

[0064] In the formula, D is the damping coefficient of the active power control loop of the GFM converter. The droop coefficient of the reactive power control loop is the Lyapunov function value, and the closer it is to 0, the more stable the system. It is the active power reference value of the grid converter GFL. This is the reactive power reference value of the grid converter (GFL). Under low voltage ride-through (LVRT) conditions, its value will adjust according to the drop in terminal voltage, guiding the GFL to generate more reactive power, suppressing voltage drops, and supporting the stability of the system's transient voltage. It is the active power reference value of the grid-type converter GFM, that is, the target active power output value that the GFM converter needs to track, and it is the core given input quantity of the GFM active-frequency control loop. It is the reactive power reference value of the grid-connected converter GFM. The sources include grid voltage support demand command, grid connection point voltage deviation adaptive generation value or energy storage collaborative control reactive power allocation command. It will increase rapidly when the transient voltage drops, triggering the GFM to generate more reactive power and form active voltage support capability. It is a Lyapunov function. The Lyapunov function selects power deviation, virtual rotor angular velocity deviation, and virtual internal potential deviation as core variables. The weighting coefficients are set based on control sensitivity analysis to ensure compliance with the Lyapunov stability criterion. It can quantitatively characterize the transient stable state of the system.

[0065] Reference Figure 1 As the proportion of GFM output increases in (a), 1(b), 1(c), and 1(d), the system's stability region and voltage stability characteristics will change. For example... Figure 1 As shown in (a), when the proportion of grid-connected renewable energy is r=10%, the Lyapunov function value increases significantly as the GFL enters the LVRT state and the total output of heterogeneous resources gradually increases, indicating that the system is in an unstable operating state. Figure 1 (b) The equipotential surface of the function value shows that the total output of heterogeneous resources corresponding to the right boundary of the stability region when r=10% is 3.18 pu. Furthermore, the transient voltage response trajectory repeatedly touches the LVRT switching boundary after fault clearance, and the Lyapunov function... The voltage fluctuated back and forth, and ultimately the trajectory failed to return to the steady-state equilibrium (SEP) point, indicating transient voltage instability. The proportion of grid-connected renewable energy was increased to r=30%, from... Figure 1 (c) It can be seen that the overall Lyapunov function value of the system decreases, and the system's instability region decreases. From... Figure 1 (d) shows that the equipotential surface of the function value indicates that the system's stable domain has shifted to the right. The total output of heterogeneous resources corresponding to the right boundary of the stable domain is 3.31 pu, which means that after increasing the proportion of grid-connected new energy capacity, the transmission capacity of heterogeneous resources has been improved. Furthermore, the transient operating trajectory of the system eventually satisfies the steady-state equilibrium condition, indicating that the system has recovered stability and no voltage instability has occurred. Figure 1 In section a, the red arrow curve on the left represents the curve approaching the static voltage stability limit; the red arrow curve on the upper right represents the low voltage ride-through curve; and the black arrow on the lower right represents the steady-state equilibrium point curve. Figure 1 In diagram c, the red arrow curve in the upper right corner represents the low-voltage ride-through curve; the black arrow in the lower right corner represents the steady-state equilibrium point curve. Figure 1 b and Figure 1 In diagram d, the red curve represents the transient response curve during the fault, and the green curve represents the transient response curve after the fault is cleared.

[0066] The above analysis shows that when the proportion of grid-connected renewable energy is low, the system voltage stability is dominated by the GFL control switching process, and the Lyapunov function value is mainly determined by the GFL output power deviation, resulting in a high value. As the proportion of GFL capacity gradually increases, the system's active support capability strengthens, and the overall Lyapunov function value decreases. Therefore, appropriately increasing the proportion of grid-connected renewable energy capacity can improve the system's voltage support capability.

[0067] 3.4 The peak value and trend of change are mainly used as online relative assessment indicators of system instability risk and optimization objective function. Normalization is applied to... The optimization objectives are transformed into quantifiable targets to ensure that the optimization direction is completely consistent with the improvement of voltage stability. To provide more intuitive engineering criteria, the following auxiliary indicators are calculated and monitored simultaneously: 1) the time for the critical node voltage to recover to 0.9 pu; 2) the maximum and minimum voltage values ​​in the first cycle after fault clearance; 3) the system oscillation damping ratio, which is approximately obtained through Prony analysis. These indicators, combined with H(x), constitute a multi-dimensional evaluation system for the transient voltage stability of the system.

[0068] Example 4.

[0069] This embodiment 4 discloses a method for dynamic reconfiguration of grid-connected resources with transient voltage stability of heterogeneous new energy sources, wherein step 4 includes the following steps.

[0070] 4.1 Analysis of the influence mechanism of the dominant parameters.

[0071] The key components affecting the transient voltage response characteristics of the GFM are the reactive power-voltage loop and the power synchronization loop. Therefore, when analyzing the dominant influencing factors of the system's transient voltage support capability, the main parameters and physical quantities in the control loop must be considered. First, the influence mechanism of the virtual internal potential is analyzed. When the active power output of the grid-connected power source increases or a fault occurs, the current flowing through the line increases. An equivalent circuit equation is established based on the grid-connected power source topology.

[0072] Formula (8): .

[0073] In the formula, , The line equivalent resistance and inductance include virtual resistance, reactance, and filters. Represent the imaginary part of a complex number; It is the output current amplitude of the GFM converter; it reflects the actual amount of current delivered by the GFM converter to the grid and is the core electrical quantity that characterizes the power output capability of the GFM. It is the voltage amplitude at the grid connection point of the GFM converter, that is, the actual voltage at the connection point between the GFM and the grid. It is a key feedback quantity for triggering virtual internal potential regulation and judging transient voltage status. It is the virtual internal potential amplitude of the GFM converter, which is the core regulation variable of the GFM controller. It will actively increase when the voltage drops in order to compensate for the line voltage drop and generate additional reactive power to support the voltage.

[0074] Due to the decrease in grid connection point voltage, the virtual internal potential in the reactive voltage loop... Increased to compensate for the voltage drop between the converter port and the grid connection point, resulting in Reaching the upper limit amplitude Substituting this into the formulas for the output active and reactive power of the GFM converter, we can obtain the solution. and Furthermore, the GFM converter simulates the rotor motion characteristics of a synchronous generator, and the sine and cosine values ​​of the voltage phase angle output by its active power control loop vary relatively little. Therefore, when the GFM reaches the limiting condition, the output active and reactive power will become voltage. The power response process is actually the process of feeding the sampled voltage value at the grid connection point into the control system, thereby realizing the iterative solution of the power, which can be expressed by the following formula.

[0075] Formula (9): .

[0076] In the formula, i = 1, 2, 3, 4...k represents the iteration step. In the... iWhen performing power flow calculations, the known quantities at the GFM grid connection point are active power and reactive power. This means the GFM's external characteristics will change from a voltage source to a power source. Therefore, the GFM's voltage support capability is affected by the control loop's limiting constraints, and thus has a certain support boundary. A smaller value means less reactive power support the GFM can provide under external power grid disturbances.

[0077] In addition, under transient conditions An increased response and a sharp increase in reactive power may cause transient overcurrent, therefore Increasing the value can improve voltage support capability, but it is not conducive to fault current limiting, such as... Figure 2 As shown in (a). When the voltage is low, the voltage support capability is insufficient, and the Lyapunov function value is high. As the number of units increases, the system tends to stabilize. A continuous increase in H(x) increases the risk of transient overcurrent and makes the system more prone to transient overvoltage. A high H(x) indicates... Too small or too large a value is detrimental to system stability.

[0078] By changing the damping coefficient D of the active power control loop and the coefficient of the reactive power control loop of the GFM converter Further analysis was conducted to examine the impact of control parameters on transient voltage stability. Figure 2 (b) It can be seen that increasing the damping coefficient D increases the equivalent damping of the system, which is beneficial for maintaining the stability of the system under large disturbance voltage. Within a certain range, increasing the reactive power control loop coefficient... It can enhance the dynamic voltage regulation capability of the GFM converter during transients and improve the transient voltage stability of the system.

[0079] In summary, to improve the system's voltage support capability, the output ratio of GFL and GFM converters can be changed by adjusting the control reference value, without needing to alter the grid structure or install new equipment to enhance grid strength. Since the active power output of the grid-connected subsystem is constrained by operational limits and control loop limiting conditions, solving for the optimal grid ratio r becomes a crucial issue.

[0080] 4.2 Dynamic parameter adjustment: While optimizing the optimal network ratio r, the core control parameters of GFM and GFL need to be adjusted in conjunction to achieve coordinated adaptation of active support and passive response, thereby improving the transient voltage stability effect.

[0081] GFM converter parameter adjustment: Based on the optimal r value, the target active power of each GFM unit is allocated, and the virtual internal potential reference value and limiting value are adjusted synchronously; when the transient voltage drops, the virtual internal potential reference value is increased synchronously according to the increase ratio of r to compensate for the line voltage drop and increase reactive power generation; the reactive power-voltage droop coefficient is dynamically adjusted for different disturbance levels, decreasing the reactive power-voltage droop coefficient to enhance voltage support sensitivity during severe disturbances, and increasing it during minor disturbances. To maintain voltage stability. At the same time, the virtual inertia time constant is optimized by adapting the r value. Under high r ratio conditions, it is set to 0.5-1.5s to improve response speed, and under low r ratio conditions, it is set to 1.5-3s to enhance system inertia.

[0082] GFL converter parameter adjustment: LVRT parameters are dynamically set based on the optimization direction of r. When r increases, the reactive current injection slope of the GFL is appropriately reduced to avoid reactive power conflict with the active support of the GFM. When r decreases, the injection slope is increased to 1.2-1.5 pu / kV to strengthen passive support capabilities. Simultaneously, the output share of the GFL cluster is allocated based on the r value. The power recovery rate during the voltage recovery phase is adjusted inversely according to the r value; the larger r is, the smoother the recovery rate, preventing secondary voltage drops and ensuring compliance with current limiting constraints. The linkage rule is that for every 10% increase in r, the reactive current injection slope of the GFL is reduced by 0.1 pu / kV, and the power recovery rate is reduced by 0.02 pu / s, ensuring seamless coordination between the two.

[0083] Example 5.

[0084] This embodiment discloses a method for dynamic reconfiguration of grid-connected resources to stabilize transient voltage of heterogeneous new energy sources, wherein in step 5...

[0085] To resolve the contradiction between the computational time required for pure online optimization and the requirement for rapid transient processes, this invention adopts a hierarchical coordinated control architecture of offline pre-decision, online fast matching, and quasi-steady-state optimization correction.

[0086] Offline pre-decision: Based on a high-fidelity digital twin simulation model, simulation calculations are performed on a massive set of typical operating conditions and anticipated faults. The optimization model used solves for the optimal network ratio r and its corresponding set of key control parameters under each scenario, forming a pre-decision strategy knowledge base.

[0087] Online rapid matching and execution (main path): After the disturbance is identified in step 2, the system immediately matches the closest preset strategy from the knowledge base based on the current operating status and disturbance characteristics, and executes it in milliseconds to provide the fastest initial stability support.

[0088] Online optimization correction (auxiliary path): After disturbance clearance, the main transient processes of the system subside, and the system enters a quasi-steady state (usually a few seconds after the fault), an online optimization model based on SOCP is initiated. At this point, the input data to the optimization model is the system's quasi-steady-state operating point. Its task is to fine-tune the executed matching strategy to accommodate subtle deviations between the actual system and the model, or to provide a backup solution when no perfectly matching strategy is available in the knowledge base. The optimization results are used to guide subsequent slow power reallocation.

[0089] A hierarchical time-sharing optimization framework is adopted.

[0090] First layer: After a disturbance occurs, immediately match and execute the preset reconstruction strategy that is closest to the current operating condition from the pre-decision strategy knowledge base. This action buys time for subsequent optimization and provides a good starting point.

[0091] The second layer: Based on the new steady-state data of the system after the disturbance, the SOCP optimization model described in step 5.1 is started for solving. At this time, the initial running point data in the model is updated to the system state after the actions of the first layer. Due to the high-quality initial solution provided by the knowledge base, the convergence speed of the SOCP solver will be greatly improved. A converter cluster aggregation and order reduction strategy is adopted to aggregate converters of the same type and with the same response characteristics into a single equivalent unit, reducing the number of variables by more than 60%, ensuring that the Cplex solver completes the solution within 50~80ms, meeting the control requirements at the hundred-millisecond level.

[0092] The third layer: During the system recovery process, the optimization model is periodically resolved based on the latest system state, and r is fine-tuned to cope with the slow changes in the system state.

[0093] It includes the following steps.

[0094] 5.1 Constructing the optimization objective and constraints: The core objective is to minimize the peak value of the Lyapunov function min H(x). As can be seen from step 1, H(x) is a non-convex function. By introducing slack variables t1, t2 and t3, the objective function is transformed into a quantifiable constraint form.

[0095] Formula (10): .

[0096] in.

[0097] Formula (11): . In the formula, It is the active power reference value of the i-th grid-connected (GFL) converter; This is the active power reference value for the j-th grid-connected (GFM) converter. It is the total active power output of the heterogeneous new energy system, which is the sum of the actual active power outputs of all GFL and GFM converters. It is the actual active power output of the i-th GFL converter, which is the real-time monitoring value during system operation. It is the actual active power output of the j-th GFM converter, which is a real-time monitoring value. It is the actual reactive power output of the j-th GFM converter. It is the actual reactive power output of the i-th GFL converter. It is the damping coefficient of the active power control loop of the j-th GFM converter. It is the droop factor of the reactive power control loop of the j-th GFM converter, used to accelerate the reactive power response speed; It is the terminal voltage of the i-th GFL converter, and is the voltage amplitude monitored in real time. It is the equivalent impedance of the output line of the j-th GFM converter, including virtual resistance, reactance, and filter impedance. This refers to the active power reference value of the j-th grid-type converter in the previous control cycle. It is the active power reference value of the i-th grid converter in the previous control cycle.

[0098] The terminal voltage and output power of each GFM and GFL are subject to algebraic constraints. In addition, the GFL must meet the current limiting constraint: Formula (12): .

[0099] In the formula, It is the maximum allowable output current of the GFL converter, i.e., the current limiting threshold. It is an inherent parameter of the equipment and is used for current limiting constraints to prevent the equipment from being damaged by overcurrent.

[0100] GFM also needs to meet the reactive power-voltage loop limiting constraint.

[0101] Formula (13): .

[0102] In the formula, This is the virtual internal potential reference value of the GFM converter; It is the core constraint parameter of the GFM voltage support boundary.

[0103] Therefore, an optimization model for dynamic reconfiguration of GFM / GFL capacity configuration can be constructed.

[0104] Formula (14): .

[0105] 5.2 Convex Relaxation and SOCP Transformation of Non-convex Models: The constraints in the optimization model are non-convex functions, making it a nonlinear and non-convex optimization problem that is difficult to solve directly. One existing approach to solving nonlinear and non-convex optimization problems is to use convex relaxation techniques to transform the original problem into a directly solvable linear convex optimization problem. This application introduces this approach to solving the GFM / GFL dynamic reconstruction optimization model, proposing a solution method based on second-order cone programming (SOCP). This method performs convex relaxation and linearization on the constraints, transforming the nonlinear and non-convex model into a second-order cone programming model, thereby enabling effective solution.

[0106] 5.2.1 First, introduce auxiliary variables. zi , rj , qj Rewrite Formula 10 in SOCP form.

[0107] Formula (15): .

[0108] Formula (16): . In the formula, It is the second-order cone conversion auxiliary variable of the i-th GFL converter, used to rewrite the GFL power deviation term into the standard form of the second-order cone, which is adapted to the SOCP model for solution. It is a second-order cone transformation auxiliary variable for the active power term of the j-th GFM converter, used to transform the GFM active power deviation term into a second-order cone form; It is a second-order cone transformation auxiliary variable for the reactive power term of the j-th GFM converter, used to transform the GFM reactive power term into a second-order cone form.

[0109] 5.2.2 Secondly, regarding the output voltage of the GFM converter... Algebraic constraints are introduced to perform convex relaxation by introducing auxiliary variables, thereby avoiding nonlinear coupling between the voltage and power square terms.

[0110] Formula (17): .

[0111] Formula (18): .

[0112] In the formula It is the squared auxiliary variable of the terminal voltage of the i-th GFL, used for the convex relaxation transformation of the constraint conditions. It is the square auxiliary variable of the terminal voltage of the j-th GFM machine, used for convex relaxation processing. It is the apparent power auxiliary variable of the i-th GFL converter, used for convex relaxation of power-voltage coupling constraints; It is the apparent power auxiliary variable of the j-th GFM converter, used for convex relaxation. It is the square auxiliary variable of the combined bus PCC voltage, which is the bus voltage at PCC and is the core evaluation node voltage of the system voltage support capability. It is an auxiliary variable for the total apparent power of the bus PCC, used for convex relaxation constraints on the total active and reactive power of the system to ensure power balance; It is the voltage amplitude of the busbar PCC, a core evaluation indicator of the system's voltage support capability; It is the total reactive power output of the heterogeneous new energy system, which is the sum of the total reactive power output of the GFL and GFM clusters.

[0113] 5.2.3 Since the square of the voltage is replaced by an auxiliary variable, the voltage amplitude constraint needs to be transformed into a linear constraint.

[0114] Formula (19): .

[0115] 5.2.4. A first-order Taylor expansion is performed on the squared terms and nonlinear product terms at the initial operating point to achieve linearization. A minimum tolerance ξ is introduced for relaxation compensation to balance linearization error and solution accuracy. Since the algebraic constraints are non-convex, a first-order Taylor expansion is performed on the squared terms and nonlinear terms at the initial operating point to linearize them, and a minimum tolerance is introduced for relaxation. The initial operating point is selected from the steady-state operating point before the disturbance. If the data before the disturbance is invalid, the steady-state point corresponding to the knowledge base matching strategy is selected. The tolerance ξ is set to 0.1 based on the simulation accuracy and dynamically adjusted according to the linearization error. When the error exceeds 5%, ξ decreases to 0.05; when the error is less than 1%, ξ increases to 0.15 to balance accuracy and solution speed.

[0116] Formula (20): .

[0117] in.

[0118] Formula (21) .

[0119] In the formula, , The terminal voltages of each GFL and GFM at the initial operating point. This is the voltage of the bus that converges at the initial operating point. The equivalent reactance for the busbar connecting to the external power grid.

[0120] It is the linearized auxiliary variable of the squared term of the squared auxiliary variable of the terminal voltage of the i-th GFL; It is the linearized auxiliary variable of the squared term of the auxiliary variable of the terminal voltage of the j-th GFM machine; It is a linearized auxiliary variable that aggregates the squared terms of the auxiliary variable of the bus voltage squared; It has a very small tolerance, with a value of 0.1. It is the square auxiliary variable of the terminal voltage of the i-th GFL converter; It is the square auxiliary variable of the terminal voltage of the j-th GFM converter.

[0121] 5.3 Model Solving and Dynamic Reconfiguration Control: The SOCP model is solved using the Cplex solver to obtain the optimal grid ratio r (i.e., the ratio of grid-type resource capacity to total new energy output); the controller limiting constraints are rewritten as constraints that active and reactive power must satisfy.

[0122] Formula (22): .

[0123] Therefore, the original model can be transformed into an SOCP model that can be solved directly.

[0124] Formula (23): .

[0125] Existing commercial solvers can be directly used to solve convex optimization models, such as Groubi and Cplex. This application uses the Cplex solver to solve formula (23). Considering the switching control of the new energy system, if the terminal voltage of a certain unit gradually drops to trigger low voltage ride-through control again after the fault is cleared as active power recovers, or if the terminal voltage fails to recover for a long time, it indicates that the system voltage has become unstable.

[0126] Therefore, after determining the system operating conditions, an N-1 traversal is performed to obtain the expected fault set. Voltage instability scenarios are then selected, and the GFL and GFM operating data corresponding to the maximum energy function amplitude during repeated low-voltage ride-throughs or continuous low-voltage ride-throughs are taken for dynamic reconfiguration optimization calculations. This yields the optimal network configuration ratio under this operating condition, which is then adjusted during actual operation. Solving for the optimal network configuration ratio... r Virtual control calculations are performed using the transient stability simulation module to verify whether the voltage response characteristics (such as voltage recovery time and oscillation amplitude) corresponding to the r value meet the constraints. If not, the model is returned for iterative optimization, forming a pre-loop of optimization-simulation verification-correction; and the capacity of GFM / GFL resources is dynamically controlled.

[0127] The solution process starts with a strategy of matching from the knowledge base to improve computation speed. After obtaining the optimal ratio r, it enters the instruction serialization and security verification stage, and at the same time, it performs parameter adjustment actions.

[0128] A. Command decomposition and parameter linkage: The overall r is decomposed into the active power reference value adjustment of each specific GFM and GFL converter. Simultaneously, according to the parameter adjustment strategy in step 4.2, control parameter adjustment commands such as virtual internal potential, droop coefficient, and LVRT injection slope corresponding to each converter are generated to ensure that the r value and parameter adjustment are coordinated and adapted.

[0129] B. Sequence Generation: Considering equipment response speed and network power flow security, a time-series power adjustment command is generated. Priority is given to rapidly increasing the active power output of fast-responding GFM units (such as GFMs with energy storage support), followed by gradual adjustments to other units to avoid power surges. Parameter adjustment commands are issued in the order of GFM first, then GFL, and transient parameters first, then steady-state parameters to ensure coordinated control. The command timing is as follows: GFM transient parameters are issued within 10ms, GFL LVRT parameters are issued within 20ms, and power adjustment commands are executed step-by-step at 50ms intervals, with a maximum single-step adjustment not exceeding 0.05 pu to avoid power surges.

[0130] C. Digital Twin Forward Simulation: The generated instruction sequence and linkage parameter adjustment values ​​are input into the digital twin simulation model for rapid closed-loop simulation to predict the system dynamics after execution. If the prediction results meet the voltage stability requirements and there is no risk of overcurrent or overvoltage, execution is approved; otherwise, the instruction sequence and parameter adjustment values ​​are smoothly corrected, such as fine-tuning the r value by ±2%, adjusting the GFM virtual internal potential increase, or reverting to a conservative strategy.

[0131] D. Command Issuance and Closed-Loop Monitoring: Verified command sequences and parameter adjustment commands are issued to each converter, and the system response is monitored in real time. If the actual response deviates from the expected response by more than 3%, or the voltage recovery target is not met, execution is suspended and an alarm is triggered. Simultaneously, a secondary adjustment mechanism for the r-value is activated, lowering the r-value by 3%-5%, increasing the GFM virtual internal potential limit, and switching to a degraded control strategy to address unexpected situations such as parameter drift and communication interruptions, ensuring system stability. During communication interruptions, a local caching strategy is activated, caching the three most recent optimal r-values ​​and parameters. In case of solver failure, the knowledge base default strategy is switched to. When the digital twin model fails, a simplified simulation model verification is used, and maintenance alarms are triggered simultaneously to ensure stable system operation under extreme scenarios.

[0132] This technical solution analyzes the influence mechanism of core parameters such as GFM virtual internal potential, damping coefficient, and droop coefficient on voltage stability, clarifying the voltage support boundary under GFM limited operation. Simultaneously, it proposes a linkage adjustment strategy for the core control parameters of GFM and GFL based on the optimal grid ratio r, achieving dynamic adaptation of parameters such as virtual internal potential, droop coefficient, LVRT injection slope, and power recovery rate. This avoids reactive power conflicts between active support and passive response, solving the problem of insufficient voltage support capability caused by fixed control parameters and poor inter-converter coordination in traditional methods. Without changing the grid structure or adding new equipment, it significantly improves the system's voltage support capability.

[0133] Example 6.

[0134] This embodiment discloses a method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources. In step 6, the simulation test system uses an improved IEEE 39-node system, replacing generator G9 at node 38 with a wind farm with a rated total capacity of 320MW. This wind farm has three direct-drive wind turbine generators, each with an initial active power of 80MW and a rated voltage of 0.69kV. Their grid-side converters are controlled by GFLs, with control parameters as follows: Figure 11 As shown. Additionally, this wind farm system is equipped with a wind turbine controlled by GFM, with control parameters as follows: Figure 11 As shown. The initial active power of the GFM equipment is 80MW. The grid-connected system topology and the equivalent impedance of the GFL and GFM transmission lines are as follows. Figure 4 As shown, the system baseline capacity is set to 100MW, and all parameters in the simulation are expressed in per-unit values.

[0135] Based on the above grid-connected system model, the dynamic response process of heterogeneous resources under transient conditions can be further analyzed. In the test system, an N-1 fault condition simulation is performed on the adjacent lines around the wind farm connection point bus38. The disconnected lines are as follows: Figure 4 As shown, the fault start time is set to t=1.5s, and the faulty line is disconnected after 0.1s. Keeping the fault settings unchanged, the proportion of induction motors in the system load model is adjusted, and all N-1 results are traversed under different system operating conditions to obtain the expected fault set.

[0136] (1) Set the proportion of induction motors in the system load model to IM=40%, and the initial active power P of the grid-connected resources in the wind farm to P. gfm =0.8 pu, with an output capacity accounting for 25.8% of the total capacity. First, under the current operating conditions, the N-1 scheme is traversed to obtain the results of the wind farm's output active power, grid connection point voltage, and system transient energy function after the fault, such as... Figure 5 As shown.

[0137] Depend on Figure 5As shown in (a) and 5(b), after the fault, as the active power recovers, the voltage at the wind farm grid connection point drops again and repeated low-voltage ride-throughs occur, causing voltage oscillations. After the fault is cleared, H(x) first decreases slowly, and then continues to oscillate as the voltage repeatedly enters low-voltage ride-throughs, indicating that the system experienced transient voltage instability and ultimately failed to return to a steady state. The maximum peak value of H(x) during the system voltage oscillation instability occurred when a fault occurred between bus26 and bus29 lines. This fault scenario was selected as the most severe operating condition, and the system operating data at the moment of the maximum peak value of H(x) were obtained as follows: Figure 12 As shown.

[0138] Inputting the current running point data into the SOCP model, the optimal resource ratio for the optimized network structure is obtained as r = 49.35%. Based on the solution results, P is adjusted under steady state. gfm Adjustments were made, and simulations were conducted under the aforementioned fault conditions to obtain the following results: Figure 6 As shown.

[0139] Depend on Figure 6 As can be seen from (a) and 6(b), regulating P gfm After achieving the optimal ratio, the voltage oscillation phenomenon disappeared under all N-1 schemes, and the wind farm's output active power and grid connection point voltage eventually returned to normal steady state. Figure 6 (c) H(x) eventually approaches zero, indicating that the system has not experienced transient voltage instability.

[0140] (2) Set the proportion of induction motors in the system load model to IM=50%, and maintain the initial active power P of the grid-connected resources in the wind farm. gfm =0.8pu remains unchanged. Under the current operating conditions, the results of the N-1 scheme are obtained to obtain the output active power, grid connection point voltage, and Lyapunov function value of the wind farm after the fault, such as... Figure 7 As shown.

[0141] Depend on Figure 7 As shown in (a) and 7(b), due to the increased IM ratio, the induction motor absorbs more reactive power during the transient period, resulting in a slower system power and voltage recovery rate. Under all N-1 schemes, the voltage at the wind farm grid connection point also drops again after the fault and experiences repeated low-voltage ride-throughs, causing voltage oscillations. Figure 7 (c) It can be seen that the maximum peak value of H(x) during the system voltage oscillation and instability occurs when a fault occurs between bus26 and bus29 lines. This fault scenario is selected as the most severe operating condition, and the system operating data at the moment of the maximum peak value of H(x) is obtained as follows: Figure 13 As shown.

[0142] Inputting the current running point data into the SOCP model, the optimal resource ratio for the optimized network structure is obtained as r = 60.32%. Similarly, based on the solution results, P is optimized under steady-state conditions. gfmAdjustments were made, and simulations were conducted under the aforementioned fault conditions to obtain the following results: Figure 8 As shown.

[0143] Depend on Figure 8 As can be seen from (a) and 8(b), regulating P gfm After achieving the optimal ratio, under all N-1 schemes, the wind farm's output active power and grid connection voltage eventually recovered to normal steady state, and Figure 8 (c) H(x) eventually approaches zero, indicating that the system has not experienced transient voltage instability.

[0144] (3) Set the proportion of induction motors in the system load model to IM=60%, and maintain the initial active power P of the grid-connected resources in the wind farm. gfm =0.8pu remains unchanged. Under the current operating conditions, the results of the N-1 scheme are obtained to obtain the output active power, grid connection point voltage, and Lyapunov function value of the wind farm after the fault, such as... Figure 9 As shown.

[0145] Depend on Figure 9 As shown in (a) and 9(b), when faults occur in the middle of the bus28-bus29 and bus26-bus28 lines, the voltage at the wind farm's grid connection point experiences repeated low-voltage ride-throughs, causing voltage oscillations. However, when faults occur in the middle of the bus26-bus29 and bus25-bus26 lines, the wind farm's output active power never recovers, and the voltage remains below 0.9 pu, indicating a continuous low-voltage ride-through state. From... Figure 9 (c) It can be seen that the maximum peak value of H(x) occurs when a fault occurs between bus25 and bus26 lines in all instability scenarios. This fault scenario is selected as the most severe operating condition, and the system operating data at the moment of the maximum peak value of H(x) is obtained as follows: Figure 3 As shown.

[0146] Inputting the current running point data into the SOCP model of equation (23), the optimal resource ratio of the optimized network structure is obtained as r = 72.54%. Similarly, based on the solution results, P is optimized under steady state. gfm Adjustments were made, and simulations were conducted under the aforementioned fault conditions to obtain the following results: Figure 10 As shown.

[0147] Depend on Figure 10 (a) and 10(b) show that regulating P gfm After achieving the optimal ratio, under all N-1 schemes, the wind farm's output active power and grid connection voltage eventually returned to normal steady state. Figure 10 (c) H(x) eventually approaches zero, indicating that the system has not experienced transient voltage instability, which verifies the effectiveness of the dynamic reconfiguration method proposed in this patent.

[0148] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for dynamic reconfiguration of heterogeneous new energy transient voltage stability based on grid connection and grid construction, characterized in that, include: Mathematical models of grid-connected and grid-connected converters are established, with real-time operating data and control parameters of the corresponding nodes as inputs. The power-voltage coupling relationship and voltage support boundary quantization results are output respectively. The power-voltage coupling relationship is derived, and the Lyapunov function is constructed as a transient stability criterion to quantify the transient stability state of the system. Based on the power-voltage coupling relationship and the quantification results of the voltage support boundary, the influence mechanism of the core parameters on voltage stability is analyzed, and the voltage support boundary under the limited operation of the grid-type converter is determined. Combining the power-voltage coupling relationship and the parameter influence mechanism, an optimization model is constructed with the goal of minimizing the peak value of the Lyapunov function. The optimal grid configuration resource capacity is then transformed into a second-order cone programming model through convex relaxation and linearization, and the optimal ratio value is output.

2. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 1, characterized in that... In establishing the mathematical model of the grid-type converter, the grid-type converter adopts a virtual synchronous machine control strategy to simulate the rotor motion characteristics and voltage support capability of the synchronous generator. The modeling process considers the comprehensive effects of virtual internal potential regulation, reactive power-voltage droop control, filtering devices and line impedance, characterizes the voltage support mechanism under transient conditions, establishes differential state equations and power transmission equations, and derives the relationship between output active power, reactive power and terminal voltage.

3. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 2, characterized in that... In establishing the mathematical model of the grid-connected converter, the grid-connected converter adopts a current source control strategy to track the output power of the grid voltage and frequency changes. In the modeling process, the influence of current tracking control, low voltage ride-through mode switching, and current limiting constraints on power output is characterized, the reactive power response law under different disturbance intensities is clarified, the differential state equation and power transmission equation are established, and the relationship between output power and terminal voltage is derived.

4. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 2, characterized in that... A Lyapunov function adapted to heterogeneous hybrid systems is constructed, integrating the power deviations and control parameters of grid-connected and grid-connected converters to quantify the transient stable state of the system. The Lyapunov function model is as follows: In the formula, D is the damping coefficient of the active power control loop of the grid-type converter. This refers to the droop coefficient of the reactive power control loop; It is the active power reference value for the grid converter; This is the reactive power reference value for the grid converter; This is the active power reference value for grid-type converters; This is the reactive power reference value for grid-type converters; It is a Lyapunov function; and These are the active power and reactive power output of the grid-type converter, respectively. This refers to the equivalent line impedance value of a grid-type converter that includes filters and transformers. The sampling time interval is... To match the terminal voltage of the grid-connected converter; and These are the active power and reactive power output by the grid converter, respectively.

5. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 4, characterized in that... The peak value and trend of the Lyapunov function are used as online relative assessment indicators and optimization objective functions for instability risk, and are transformed into quantifiable optimization objectives after normalization. The time for the voltage of critical nodes to recover to 0.9 per unit value, the voltage extreme value of the first cycle after fault clearance, and the system oscillation damping ratio are calculated simultaneously as auxiliary indicators. These are combined with the Lyapunov function to form a multi-dimensional evaluation system for the transient voltage stability of the system.

6. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 1, characterized in that... The analysis of the influencing mechanism includes the analysis of the influencing mechanism of the dominant parameter and the dynamic adjustment of the core parameter; Analysis of the Influence Mechanism of Dominant Parameters: Taking the reactive power-voltage loop and power synchronization loop of the grid-type converter as the core, this paper analyzes the influence of key parameters such as virtual internal potential, active power control loop damping coefficient and reactive power control loop coefficient on transient voltage support capability. Dynamic adjustment of core parameters: Based on the optimal resource capacity ratio of the grid-type converter, the core control parameters of the grid-type converter and the grid-following converter are adjusted in a coordinated manner to achieve coordinated adaptation.

7. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 1, characterized in that... The optimization model for determining the optimal proportion of grid-connected resource capacity to total new energy output includes: With the goal of minimizing the peak value of the Lyapunov function, relaxation variables are introduced to transform the non-convex objective function into a quantifiable constraint, which is then incorporated into the converter voltage-power algebra constraint, the grid-connected converter current limiting constraint, and the grid-connected converter reactive-voltage loop limiting constraint. By introducing auxiliary variables for convex relaxation, linearizing the first-order Taylor expansion of the nonlinear term, and introducing dynamic minimum tolerance to balance error and accuracy, the model is transformed into a second-order cone programming model. A solver is used to solve the second-order cone programming model.

8. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 7, characterized in that... : By introducing slack variables t1, t2, and t3, the objective function is transformed into a quantifiable constraint form: ； in, ; In the formula, It is the active power reference value of the i-th grid converter; This is the active power reference value for the j-th grid-type converter; Heterogeneous new energy systems always have active power output; It is the actual active power output of the i-th grid-connected converter; It is the actual active power output of the j-th grid-type converter; This is the actual reactive power output of the jth grid-type converter; It is the actual reactive power output of the i-th grid-connected converter; It is the damping coefficient of the active power control loop of the j-th grid-type converter; It is the droop coefficient of the reactive power control loop of the j-th grid-type converter; It is the terminal voltage of the i-th grid-connected converter; is the equivalent impedance of the output line of the j-th grid-connected converter, including virtual resistance, reactance and filter impedance; r is the proportion of grid-connected resource capacity to total new energy output; n represents the total number of grid-connected converters participating in power distribution; m represents the total number of grid-connected converters participating in power distribution.

9. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 1, characterized in that... : By introducing auxiliary variables for the power deviation of grid-connected converters and active and reactive power of grid-connected converters, the optimization model for the optimal ratio of grid-connected resource capacity to total new energy output is rewritten into a standard form of second-order cone programming. To address the algebraic constraint on the output voltage of grid-connected converters, auxiliary variables for the square of the terminal voltage, apparent power, and the square of the collector bus voltage and total apparent power of grid-connected and grid-connected converters are introduced to avoid the nonlinear coupling between the square terms of voltage and power and to achieve convex relaxation of the constraint conditions. Synchronization transforms voltage amplitude constraints into linear constraints.

10. The method for dynamic reconfiguration of grid-connected resources to stabilize the transient voltage of heterogeneous new energy sources according to claim 1, characterized in that... : The solution process starts with a strategy of matching from the knowledge base to improve computation speed. After obtaining the optimal ratio of network resource capacity to total new energy output, it enters the instruction serialization and security verification stage, and simultaneously performs parameter adjustment actions.

Images