Reactive power control method and device

Abstract

The invention relates to the technical field of power dispatching in a power distribution network, in particular to a reactive power control method and device. The method comprises the steps of obtaining output reactive power of each converter, and determining a reactive power distribution error of each converter according to preset reactive power distributed by each converter according to a rated proportion and the output reactive power; determining an adaptive virtual impedance value of each converter according to the reactive power distribution error and the obtained output voltage of the corresponding converter; and carrying out droop control on the converters in the micro-grid system based on the self-adaptive virtual impedance value to obtain the current output voltage of each converter, and skipping to the step of obtaining the output reactive power of each converter until the current output voltage meets the first mode switching condition. According to the invention, the problems of reactive power distribution imbalance and circulating current risk during multi-machine parallel connection in the prior art can be solved, and the system stability and the power supply reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of power dispatching technology in power distribution networks, and particularly to a method and apparatus for controlling reactive power. Background Technology

[0002] Under the current trend of energy transition, renewable energy sources, represented by distributed photovoltaics, are being integrated into the distribution network on a large scale. This integration has changed the traditional operation mode of the power grid. In off-grid operation scenarios such as islanded operation, microgrids, and post-disaster reconstruction, the traditional operation mode that relies on centralized grid dispatch is gradually becoming unsuitable, and new technical means are urgently needed to ensure the stable operation of the system.

[0003] Grid-type converters, leveraging their voltage source characteristics, play a crucial role in maintaining system frequency and voltage stability. Meanwhile, there is a wealth of theoretical research on reactive power optimization scheduling, providing a theoretical foundation for power control in power grid operation.

[0004] However, traditional droop control strategies have obvious limitations. When multiple units are connected in parallel, reactive power distribution imbalance and circulating current risks may occur, which weaken system stability and power supply reliability. Summary of the Invention

[0005] This invention provides a reactive power control method and apparatus to solve the problems of reactive power imbalance and circulating current risk that occur when multiple machines are connected in parallel in the prior art, which weakens system stability and power supply reliability.

[0006] In a first aspect, embodiments of the present invention provide a reactive power control method, applied to a microgrid system in which multiple converters operate in parallel and are connected to an AC bus to supply power to loads, including: Obtain the output reactive power of each converter, and determine the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power. Based on the reactive power distribution error and the obtained output voltage of the corresponding converter, determine the adaptive virtual impedance value of each converter; Based on the adaptive virtual impedance value, droop control of the converters is performed in the microgrid system to obtain the current output voltage of each converter, and then jumps to the "obtain the output reactive power of each converter" step until the current output voltage meets the first mode switching condition.

[0007] Secondly, embodiments of the present invention provide a reactive power control device, applied in a microgrid system in which multiple converters operate in parallel and are connected to an AC bus to supply power to loads, comprising: The acquisition module is used to acquire the output reactive power of each converter; The processing module is used to determine the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power. The processing module is also used to determine the adaptive virtual impedance value of each converter based on the reactive power allocation error and the obtained output voltage of the strain converter. The processing module is also used to perform droop control of the converters in the microgrid system based on the adaptive virtual impedance value, obtain the current output voltage of each converter, and jump to the "obtain the output reactive power of each converter" step until the current output voltage meets the first mode switching condition.

[0008] This invention provides a reactive power control method and apparatus. It acquires the output reactive power of each converter and determines the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to its rated ratio and the output reactive power. Based on the reactive power allocation error and the acquired corresponding output voltage of the converter, it determines the adaptive virtual impedance value of each converter. Based on the adaptive virtual impedance value, it performs droop control of the converters in the microgrid system to obtain the current output voltage of each converter. This embodiment achieves adaptive updating of the virtual impedance through the reactive power allocation error of the converters, effectively offsetting line impedance differences, enabling each converter to allocate load according to its capacity, and realizing dynamic reactive power allocation of the converters. This solves the problem of reactive power imbalance and circulating current risk that occurs when multiple converters are connected in parallel in the prior art, improving system stability and power supply reliability. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating the implementation of the reactive power control method provided in this embodiment of the invention. Figure 2 This is a schematic diagram of the topology of a microgrid system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the multimodal reactive power control model provided in an embodiment of the present invention; Figure 4 This is a flowchart illustrating the implementation of the reactive power control method provided in this embodiment of the invention. Figure 5 This is a flowchart illustrating the implementation of the mode switching method provided in this embodiment of the invention. Figure 6 This is a schematic diagram of the reactive power control device provided in an embodiment of the present invention. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0012] To make the objectives, technical solutions, and advantages of the present invention clearer, specific embodiments will be described below in conjunction with the accompanying drawings.

[0013] Figure 1 This is a flowchart illustrating the implementation of a reactive power control method according to an embodiment of the present invention. The reactive power control method is applied to a microgrid system in which multiple converters operate in parallel and are connected to an AC bus to supply power to loads. See [link to relevant documentation]. Figure 2 The diagram shows the topology of a microgrid system. The converters are grid-forming converters (GFCs). Multiple GFCs are connected to distributed generation (DG) devices, converting the DC power generated by the DG devices into AC power, which is then connected to the bus to supply power to the loads. Multiple grid-forming converters operate in parallel to achieve load distribution even without communication. Furthermore... Figure 2 In the middle, a resistor is also installed between the GFC and the busbar. ,inductance and capacitors The filter and resistors and inductor The resulting line impedance.

[0014] In the off-grid state, the photovoltaic converters in the distribution network rationally distribute reactive power according to their respective rated capacity and line impedance, thereby effectively preventing line circulating current and voltage over-limit caused by unbalanced reactive power distribution.

[0015] In grid-based control, droop control, as one of its core technical solutions, has been widely used due to its simple structure and the advantage of communication-free coordinated control. In distribution network systems containing distributed photovoltaic converters (GFCs), multiple GFCs are often connected in parallel through droop control to achieve load distribution under communication-free conditions. Figure 2 The output power of the converter can be obtained. The output voltage of the converter and common point voltage The relationship is as follows: ; In the formula, Indicates the first The active power of each converter Indicates the first The reactive power of each converter Indicates bus voltage. This indicates the resistance value of the output line. This indicates the output voltage of the converter. Indicates the phase angle of the output voltage. This indicates the inductance value of the output line.

[0016] The above formula is simplified as follows: .

[0017] Ideally, droop control can distribute load proportionally to capacity without communication. Traditional droop control correlates active power with frequency changes and reactive power with voltage amplitude changes, i.e.: ; in, Indicates the first The output voltage angular velocity of the converter Indicates the first The output voltage angular velocity reference value of each converter This represents the active power droop coefficient. Indicates the first The current output voltage of each converter Indicates the first The voltage reference value of each converter This represents the reactive power droop coefficient.

[0018] However, in actual operation, impedance mismatch at the converter output line and load changes often lead to power distribution errors and circulating current problems. Therefore, this embodiment provides a reactive power control method, deeply analyzes the power control model of the traditional droop control strategy, clarifies the essential source of power coupling and its negative impact on the dynamic and steady-state performance of the system, and finally establishes an off-grid adaptive reactive power mode.

[0019] In this embodiment, based on the reactive power allocation strategy of a multi-parallel grid converter with adaptive virtual impedance, high-precision reactive power allocation is achieved by dynamically adjusting the virtual impedance value and combining information such as reactive power allocation error and output voltage, while suppressing voltage drop and circulating current.

[0020] The methods for controlling reactive power are detailed below: Step 101: Obtain the output reactive power of each converter, and determine the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power.

[0021] Optionally, other data from the converter can also be collected for subsequent data processing, such as the converter's output current, photovoltaic output, load size, load power, bus voltage amplitude, and power factor in the microgrid system.

[0022] In one embodiment, after collecting the aforementioned data, the data is preprocessed. Preprocessing methods may include normalization, missing value handling, and data augmentation. Normalization can employ min-max normalization or Z-score standardization to eliminate dimensional differences. Missing value handling can use linear interpolation or methods to remove incomplete samples, resulting in more complete data. Data augmentation can involve adding Gaussian noise or time shifts to improve model robustness.

[0023] After data acquisition, data features can be constructed from the acquired reactive power and output voltage. For example, time series features can be established by sliding the acquired photovoltaic output and load magnitude across a time window to construct a time series feature matrix. The time window can be 20 minutes, 30 minutes, or 45 minutes, etc.

[0024] For example, spatial features can be established from the data. For instance, based on the grid topology of the microgrid system, data such as the output voltage and reactive power of adjacent nodes can be organized into a two-dimensional grid to simulate an image channel. Here, converters can be considered nodes, and adjacent nodes are adjacent converters.

[0025] For example, to construct a composite feature vector: a. Static features, such as setting line distance, transformer rated parameters, etc., as static data.

[0026] b. Dynamic features, such as setting tap position, capacitor status, photovoltaic output, load power, etc. as dynamic data.

[0027] The 'c' tag can be used to add tags to data such as bus voltage and power factor.

[0028] In one embodiment, determining the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power may include: according to Determine the reactive power distribution error of each converter; in, Indicates the first reactive power distribution error of each converter Indicates the first The output reactive power of each converter Indicates the first The preset reactive power of each converter.

[0029] Step 102: Determine the adaptive virtual impedance value of each converter based on the reactive power distribution error and the obtained output voltage of the corresponding converter.

[0030] In one embodiment, determining the adaptive virtual impedance value of each converter based on the reactive power allocation error and the obtained output voltage of the corresponding converter may include: according to Determine the adaptive virtual impedance value for each converter; in, Indicates the first The adaptive virtual impedance value of the first converter, i.e., the first... The adaptive part of the virtual impedance value of the converter This represents the virtual impedance adaptive coefficient. Indicates the first The current output voltage of the converter.

[0031] To address the reactive power distribution problem in a multi-converter parallel system operating without communication under off-grid adaptive mode, a virtual impedance method is employed. This method feeds back the reactive power distribution error of each converter in real time to its virtual inductance design, enabling adaptive adjustment and matching its equivalent output impedance with the target power distribution ratio. This allows for precise allocation without the need for communication between converters.

[0032] Step 103: Based on the adaptive virtual impedance value, perform droop control of the converter in the microgrid system to obtain the current output voltage of each converter, and jump to the "obtain the output reactive power of each converter" step until the current output voltage meets the first mode switching condition.

[0033] Virtual impedance technology involves introducing an artificially constructed "equivalent reactance" into the converter control loop to adjust the output impedance, thereby improving power distribution and suppressing circulating current. It can take the form of a virtual resistor, virtual inductor, or even a reactance network. By adding virtual inductive reactance, the inductive properties of the microgrid system can be increased without altering the physical structure, allowing for more effective decoupling of droop control and improved distribution accuracy. Its expression is: .in, Indicates the virtual impedance value. This represents a constant virtual inductance.

[0034] In one embodiment, droop control of converters in a microgrid system is performed based on adaptive virtual impedance values ​​to obtain the current output voltage of each converter, which may include: according to Obtain the current output voltage of each converter; in, Indicates the first The current output voltage of each converter Indicates the first The voltage reference value of each converter This represents the reactive power droop coefficient. Indicates the first The output voltage angular velocity of the converter Represents the first in the dq coordinate system The output current of the converter Indicates the virtual impedance value. This represents a constant virtual inductance.

[0035] The grid-type converter uses the improved droop control method described above. The output voltage and frequency are dynamically adjusted, and the process jumps to step 101 to reacquire the output reactive power of each converter in order to achieve adaptive updating of virtual impedance, effectively offset the line impedance difference, and enable each converter to distribute the load according to its capacity, thereby realizing the dynamic reactive power distribution of the converter.

[0036] Dynamically adjusting the virtual impedance can suppress circulating current between converters, avoid equipment overheating or protection malfunction caused by circulating current, and improve the system's response speed to load changes, thereby enhancing system stability.

[0037] However, under the above reactive power control strategy, there may be slight disturbances in each state variable near the steady-state operating point, and there is a serious coupling relationship between active power and reactive power. Therefore, in this embodiment, based on the above reactive power control strategy, we conduct in-depth analysis and construct its small-signal model.

[0038] Small-signal modeling, a linearization-based approach, reveals key dynamic indicators such as system stability, response speed, and damping characteristics by modeling small perturbations of each state variable near the steady-state operating point. For grid-type converter systems, regardless of whether traditional droop control is used or improved methods such as voltage compensation and adaptive virtual impedance are introduced, the system can essentially be abstracted as a multi-input multi-output system with multiple coupled channels, where complex nonlinear dynamic relationships exist between power, voltage amplitude, and phase angle. Therefore, effective small-signal modeling not only quantifies these relationships but also provides fundamental support for the optimized design of control parameters.

[0039] In one embodiment, such as Figure 4 As shown, before performing droop control of the converters in the microgrid system based on the adaptive virtual impedance value to obtain the current output voltage of each converter, step 104 is also included.

[0040] Step 104: Use a small-signal model to dynamically decouple each converter to obtain the compensation voltage.

[0041] Then step 103 is adaptively modified to step 105: Based on the adaptive virtual impedance value and compensation voltage, droop control of the converter is performed in the microgrid system to obtain the current output voltage of each converter.

[0042] In one embodiment, active power and reactive power can be expressed in the dq coordinate system as follows: ; in, This indicates the active power of the converter. This indicates the reactive power of the converter. This represents the direct-axis component of the converter's output voltage. This represents the quadrature-axis component of the converter's output voltage. This represents the direct-axis component of the converter's output current. This represents the quadrature-axis component of the converter's output current.

[0043] Linearizing the above expression yields: ; in, This represents the active power of the linearized converter. This represents the reactive power of the linearized converter. This represents the adjustment amount of the direct-axis component of the converter's output voltage. This represents the adjustment amount of the quadrature axis component of the converter's output voltage. This indicates the adjustment amount of the direct-axis component of the converter's output current. This indicates the adjustment amount of the quadrature axis component of the converter's output current.

[0044] In microgrid systems, comparing the pole distribution of the system transfer function with and without coupling reveals that the impedance ratio R / X is related to the power angle. Sensitivity to system stability. Based on this, this embodiment proposes to introduce voltage-compensated angular frequency offset to improve the dynamic stability margin of the system, while adding a feedforward compensation term to correct steady-state output voltage deviation, thereby reducing reactive power control error.

[0045] Will , Simultaneously regarded as , The function, this modeling method can form the following transfer function matrix model: , about , The relationships between them constitute a 2×2 coupled system, whose expression contains four main channel transfer functions, among which, and These represent the main control channels for active power and power angle, and reactive power and voltage under ideal conditions, respectively. and This reflects the coupling characteristics within the system. Furthermore, the active and reactive small disturbance components can be obtained. , Small perturbation components regarding power angle and converter voltage , Transfer function: ; in, This is the transfer function matrix.

[0046] A dynamic decoupling mechanism is introduced into the small-signal model. This involves designing a suitable voltage compensator and applying negative feedback compensation for cross-channel interference in the control channel to eliminate coupling effects. For example, angular velocity offset compensation related to active power disturbances is added to the reactive power control channel. , making No longer suffer This reduces the negative impact of the system, thereby improving the system's decoupling capability and stability margin.

[0047] Therefore, in one embodiment, using a small-signal model to dynamically decouple each converter to obtain the compensation voltage may include: Obtain the reference value of the output voltage angular velocity of each converter; determine the angular velocity offset based on the reference value of the output voltage angular velocity and the current output voltage angular velocity of each converter; add the angular velocity offset related to the active power disturbance to the reactive power control channel to obtain the compensation voltage.

[0048] That is, according to Determine the first Voltage compensation value for the output voltage of each converter; in, This represents the voltage compensation value of the output voltage of the i-th converter. Indicates the first The output voltage angular velocity reference value of each converter The transfer function representing the output reactive power and output voltage phase angle of a photovoltaic converter under DC conditions. The transfer function representing the output reactive power and output voltage amplitude of a photovoltaic converter under DC conditions.

[0049] In one embodiment, based on the adaptive virtual impedance value and the compensated voltage, droop control of the converters in the microgrid system is performed to obtain the current output voltage of each converter, which may include: according to Obtain the current output voltage of each converter; in, This represents the voltage compensation coefficient. Indicates the first The voltage compensation value of the output voltage of each converter.

[0050] By designing a small-signal model, this study aims to achieve two main objectives: first, to efficiently decouple the active and reactive power control of grid-connected converters; and second, to improve the coordination and accuracy of reactive power allocation among multiple converters in a microgrid system. Specifically, regarding the first aspect, addressing the power coupling problem in traditional droop control, this study proposes a dual-compensation voltage control mechanism by embedding frequency offset compensation and feedforward active power voltage regulation in the voltage control loop. This mechanism dynamically corrects reactive power disturbances caused by active power fluctuations, improving frequency stability while reducing steady-state reactive power deviation, effectively enhancing the system's dynamic robustness and static control accuracy. Compared to the traditional virtual impedance method, this method achieves continuous control of power decoupling without introducing significant response delay. Secondly, this study combines a reactive power allocation error feedback mechanism with virtual impedance technology, proposing for the first time an adaptive virtual inductance control method based on the relationship between power error and voltage response. This method does not rely on communication or require knowledge of line impedance parameters. It automatically adjusts the virtual impedance of each converter according to the output voltage difference and target reactive power ratio, thereby achieving a precise, dynamic, and decentralized reactive power allocation target.

[0051] In one embodiment, steps 101-103 above refer to the microgrid system operating in droop control mode. When the droop control based on adaptive virtual impedance makes the system stable, or when the system is stable and the output voltage deviation is controllable, the microgrid system operates with the goal of reducing network losses and energy costs. This operating mode can be called the economic operating mode. When the microgrid system operates in the economic operating mode and voltage exceeds the limit, equipment overload, or fault occurs, the microgrid system is controlled to operate in the safe operating mode to ensure system safety.

[0052] Optional, such as Figure 3The diagram shows a multi-modal reactive power control model. This model sets three different operating modes: off-grid operation mode (i.e., droop control mode), economic operation mode, and safe operation mode, comprehensively considering the different reactive power distributions in the microgrid system. In the droop control mode, the converters in the microgrid system rationally allocate reactive power output according to their rated capacity and line impedance, effectively preventing line circulating currents and voltage over-limits caused by unbalanced reactive power distribution. In the economic and safe operation modes, the converter output of the microgrid system is dispatched by the dispatching system. The power system sets economic and safe operation modes with two objective functions: optimal economy and optimal safety. An algorithm combining particle swarm optimization and the Kriging model is used to solve these modes, yielding the corresponding dispatch plans.

[0053] When switching between the three modes mentioned above, a convolutional neural network is used for state recognition, and the mode switching is automatically triggered when the mode switching conditions are met, so as to achieve automatic mode switching.

[0054] Based on adaptive virtual impedance values, droop control of converters in a microgrid system is performed to obtain the current output voltage of each converter. The process then proceeds to the "obtain the output reactive power of each converter" step. After the current output voltage meets the first mode switching condition, the process continues with the following steps (see [link to relevant documentation]). Figure 5 As shown.

[0055] Step 501: In droop control, determine the line impedance fluctuation rate in the microgrid system.

[0056] Step 502: If the voltage deviation rate between the current output voltage and the rated voltage of the strain gauge and the line impedance fluctuation rate meet the first mode switching conditions, then switch the current droop control mode to the economic operation mode.

[0057] In one embodiment, if the voltage deviation rate between the current output voltage and the rated voltage of the strain gauge is less than a first preset value and the line impedance fluctuation rate is less than a second preset value, then the current droop control mode is switched to the economic operation mode.

[0058] The first and second preset values ​​can be set based on experience. For example, the first preset value can be set to 1% and the second preset value can be set to 5%. That is, when the output voltage deviation is small and the line impedance fluctuation is small, the microgrid system can be switched to the economic operation mode.

[0059] Economic operation mode is a core component of multimodal dispatching strategies, primarily applicable to microgrid systems operating under normal conditions, with stable load changes and high accuracy in photovoltaic output prediction. In this mode, the microgrid system does not need to prioritize emergency voltage control or reactive power reserve strategies, but instead shifts its resource dispatching focus to the economic objective of minimizing network losses. The economic operation mode does not replace traditional Volt / Var optimization, but rather further integrates real-time photovoltaic converter control capabilities with intelligent optimization algorithms. By establishing a functional mapping relationship between system network losses and reactive power configuration, it seeks an optimal control solution under multidimensional variables to achieve the lowest active power loss under reactive power support. Its operational logic can be described as follows: among all reactive power dispatching schemes that satisfy voltage and equipment capacity constraints, the configuration that minimizes network losses is prioritized, and real-time dispatching instructions for the photovoltaic converters are generated accordingly.

[0060] The mathematical model of the economic operation mode should fully express the nonlinear dependence between system network losses and reactive power. According to the above model, line power loss can be expressed as the square of the current in each branch multiplied by its impedance. Reactive power injected into nodes will affect the node voltage level, thereby indirectly changing the current distribution. Therefore, the core of the optimization model for the economic operation mode is: under the hard constraints of node voltage, current limits, and converter capacity, adjust the reactive power output of each photovoltaic converter to minimize the total network loss of the entire system. Therefore, the established objective function is expressed as: ; in, Characterizing the network loss of a microgrid, Characterizes load loss.

[0061] Specific constraints include: Current flow constraints: ; Output voltage constraints of the converter: ; Power constraints: ; Node voltage constraints: ; in, Represents a node The output active power of the photovoltaic converter Represents a node Active load, Indicates from node Flowing out to the node The meritorious trend, Indicates from node Flow into node active power, Represents a node The output reactive power of the photovoltaic converter Represents a node The reactive power provided by the capacitor bank, Represents a node reactive load, Indicates from node Flowing out to the node The unproductive current, Indicates from node Flow into node reactive power, Represents a node and nodes The square value of the current between them, Represents a node and nodes The square value of the current between them, Represents a node voltage, Represents a node voltage, Represents a node and nodes The line resistance value between them, Represents a node and nodes The line resistance value between them, Represents a node and nodes The line inductance value between them, Represents a node and nodes The line inductance value between them.

[0062] The objective function faces the following challenges in its solution: First, the nonlinearity of power flow calculation makes it difficult for traditional analytical methods to converge quickly; second, the large number of photovoltaic nodes and high dimensionality of variables make traditional exhaustive strategies unsuitable for real-time computation; third, the grid loss response to converter regulation exhibits non-convex characteristics, with multiple local minima. Therefore, to address these issues, this paper introduces a data-driven approach based on the objective function, utilizing the Kriging model to fit and predict the response surface of the objective function, efficiently searching for the optimal solution in the optimization space, thereby constructing an efficient, accurate, and real-time responsive scheduling solution mechanism.

[0063] In the process of solving the objective function, the Kriging surrogate model is first trained using existing operational data samples to generate a response surface function between reactive power and network loss. This function is used to quickly predict the objective function value and filter out scheduling points that do not meet physical constraints. Secondly, the Particle Swarm Optimization (PSO) algorithm is used to update particle positions based on a global search mechanism, continuously approaching the optimal solution. In each iteration, the Kriging model predicts the target value, and a verification module is used to check physical feasibility and correct power flow calculations until the optimization termination condition is met. For example, in this embodiment, the optimization termination condition can be that the rate of change of the objective function is less than a set threshold. After optimization, the scheduling center uses the reactive power output of each converter in the optimal solution as the expected value and sends it to each control terminal through the master station to drive the photovoltaic equipment to adjust its operating status. The entire process can be updated on a minute-by-minute or 15-minute basis, achieving rolling optimization scheduling and keeping the system in an economically optimal state at all times.

[0064] To ensure that dispatch instructions can be executed in the actual system, the optimization process under the economic operation mode must embed a series of physical and operational constraints, including but not limited to: the voltage of each node must be kept within the standard specified range; the reactive power output of the photovoltaic converter must not exceed the adjustable range under its capacity constraints; the current of each line must not exceed the thermal stability limit; and the converter response time must meet the dispatch cycle requirements. During the solution process, for points where variables exceed limits, a "penalty factor mechanism" or a "boundary projection mechanism" should be adopted to ensure that the optimization path converges within the physically feasible region. Furthermore, considering that the model relies on prediction data, to address the risks caused by prediction errors, a safety margin can be set in the constraints to improve the robustness and conservatism of the optimization results and prevent the system from operating beyond its limits due to dispatch deviations.

[0065] The Kriging model, based on Gaussian process regression theory, achieves optimal approximation of the objective function by minimizing the variance of the prediction error. After training, it possesses the capability of "local weighting + global interpolation," providing real-time predictions of the objective function during particle swarm optimization, significantly improving optimization speed and reducing computational resource consumption. Kriging model construction: The core of the Kriging model is the construction of the correlation function. Considering the nonlinear relationship between the converter output power and the voltages of each node in the system, a Gaussian function is selected as the correlation function. Based on the correlation between the converter output power and the voltages of each node in the system, the changes in node output voltage are predicted. The Kriging model reflecting the reactive power and output voltage of the photovoltaic converter is constructed as follows: ; In the formula, For node output voltage, For the set of output reactive power The basis functions of the regression model, These are the basis function coefficients. To characterize the reactive power to be evaluated With known sample set The vector of stochastic processes with correlation between them This is the prediction correction factor.

[0066] Furthermore, to improve the efficiency of solving the economic operating mode, the constraints on voltage and power are linearized, as shown below: ; In the formula, For the active power of the load, For the reactive power of the load, , The active and reactive power are the operating power and reactive power in the off-grid operation mode. This represents the proportionality coefficient of the "constant power" portion of the active load. This represents the proportionality coefficient of the "constant power" portion of the reactive load. This represents the proportionality coefficient of the "current-type" portion of the active load. This represents the proportionality coefficient of the "current-type" portion of the reactive load. This represents the proportionality coefficient of the "impedance-type" portion of the active load. This represents the proportionality coefficient of the "impedance-type" portion of the reactive load. This represents the square of the voltage magnitude at node i. Indicates the power correction factor. This represents the impedance correction factor. and It needs to be determined based on the actual situation.

[0067] Step 503: In the economic operation mode, if the current output voltage meets the second mode switching condition, then switch the economic operation mode to the safe operation mode.

[0068] Unlike the economic operation mode, which focuses on reducing network losses, the safe operation mode prioritizes the stable operation of the distribution system under uncertain conditions such as disturbances and fluctuations. This mode is mainly applicable to the following scenarios: First, large fluctuations in photovoltaic output, such as frequent start-stop cycles of converters causing rapid voltage changes when cloud cover is unstable in the afternoon; second, severe load fluctuations, such as transient voltage fluctuations caused by the start-stop cycles of large motors or load mismatches; third, voltage support path reconfiguration due to topology changes; and fourth, the distribution system being in a high-risk state, such as line maintenance, unavailability of energy storage, or the shutdown of static compensation equipment. In these scenarios, if an economic-first dispatch mode is adopted, insufficient reactive power support may lead to voltage exceedances, resulting in operational risks. Therefore, the main dispatch objective of the safe operation mode is to minimize node voltage deviations, that is, to minimize the sum of the differences between the voltage of each node and its rated value, while ensuring voltage stability and considering other operational indicators.

[0069] In one embodiment, if the current output voltage of the new output voltage is less than the first voltage value and the duration exceeds the first time, or the current output voltage of the new output voltage is greater than the second voltage value and the duration exceeds the first time, then the economic operation mode is switched to the safe operation mode; the first voltage value is less than the second voltage value.

[0070] Optionally, the first voltage value, second voltage value, and first time can be set according to requirements. For example, the first voltage value can be set to 0.95 pu, where "pu" is an abbreviation for "Per unit," a dimensionless representation used to normalize physical quantities (such as voltage, current, and power) in a power system relative to a certain reference value. The second voltage value can be set to 1.05 pu. The first time can be set to 10 seconds, 15 seconds, etc.

[0071] The safe operating mode is not a fixed operating state, but should exist as a dynamic "backup mechanism" in system operation scheduling. When the system operating state meets the mode switching conditions after being judged by the neural network, the "economic → safe" mode switching is automatically triggered, and the optimized scheduling mode with the goal of minimizing voltage deviation is initiated. When the system operates stably for a certain period of time and the voltage deviation remains within the safe range (such as within ±1%), it can switch back to the economic operating mode according to the load level and resume the strategy execution with the goal of minimizing network loss.

[0072] The essence of the safe operation mode is a multivariable, multi-constraint nonlinear optimization problem. Its core is to minimize the sum of voltage deviations at each node by adjusting the reactive power output of the converter. The objective function of the optimization model can be abstracted as the sum of the squares of the voltage deviations at all nodes, thereby avoiding the problem of positive and negative deviations canceling each other out and making the model more sensitive to local voltage anomalies. In terms of decision variables, the reactive power output values ​​of all photovoltaic converters with reactive power regulation capabilities are selected as the main control variables. In terms of constraints, the following factors should be fully considered: (1) The node voltage must be strictly controlled within the range allowed by the technical specifications (generally 0.95 pu~1.05 pu); (2) The reactive power regulation capability of the converter is limited by its capacity, active power output and bus voltage; (3) The current of the distribution line should not exceed the thermal stability limit to avoid overload caused by regulation; (4) The overall power flow of the system should be kept balanced to avoid energy dispatch mismatch problems.

[0073] Therefore, in this embodiment, the constructed objective function comprehensively considers the loss reduction rate, voltage deviation, and over-limit penalty, and its mathematical expression is: ; in, The loss reduction rate is defined as the percentage change in network losses before and after optimization. This represents the loss reduction coefficient.

[0074] ; in, This indicates the optimized system line loss. This indicates the system line loss before optimization.

[0075] The voltage deviation index is calculated based on the root mean square error between the grid connection point voltage and the rated value. ; In the formula, The total number of nodes. For nodes The actual voltage value, For nodes The rated voltage value.

[0076] This is a penalty term used to constrain the converter's reactive power output from exceeding limits. ; In the formula, Indicates the penalty coefficient. It is a piecewise linear function, when the converter has reactive power output Exceeding its adjustment range When the limit is exceeded, penalties will be imposed according to the degree of violation.

[0077] Optimization model constraint design (1) Power flow balance constraints: including active and reactive power balance equations at nodes to ensure that the system operates in a steady state; (2) Equipment safety constraints: The reactive power output of the converter is limited to: ;in, This indicates the reactive power of the converter. This represents the minimum reactive power of the converter. This indicates the reactive power of the converter; (3) Voltage stability constraint: The voltage at the grid connection point must meet the following requirements: ; Indicates the voltage at the grid connection point. This indicates the minimum voltage at the grid connection point. This indicates the maximum voltage at the grid connection point.

[0078] In the aforementioned optimization framework, complex interactions exist among the control variables: the reactive power regulation of a converter not only affects the voltage of its own node but may also have positive or negative impacts on the voltage of other nodes through line coupling. Therefore, the scheduling strategy cannot be based solely on local voltage deviations but should comprehensively consider the "systematic response" characteristics of each regulation behavior across the entire network. This multidimensional coupling structure makes the optimization problem extremely complex and non-convex, making it difficult for traditional analytical algorithms to find the optimal solution within an acceptable timeframe. A combination of the Kriging model and particle swarm optimization algorithm is required for a faster solution.

[0079] Compared to other operating modes, the safe operating mode requires the optimization model to have the characteristics of fast solution. To improve the solution speed, this paper adopts an intelligent solution method based on the combination of particle swarm optimization (PSO) and the Kriging model. PSO, as a global search heuristic algorithm, can achieve fast convergence with fewer parameters in high-dimensional nonlinear problems. The Kriging model can construct a response prediction model of the objective function based on historical optimization samples, serving as a surrogate function to replace complex power flow calculations in the PSO algorithm, significantly reducing the number of objective function calls and thus accelerating the solution process.

[0080] During the model training phase, a reactive power output sample set is first generated using the Latin hypersolution method. Power flow calculations are then performed on each sample set to obtain the corresponding voltage deviation index value, which is used to construct the Kriging training set. The Kriging model generates a response surface prediction expression for the objective function by minimizing the covariance matrix of the sample errors. During the scheduling operation phase, the PSO evaluates the fitness based on the objective function value predicted by Kriging, continuously iterates and updates the population position, and finally selects the globally optimal reactive power configuration scheme as the scheduling command. This method not only significantly improves the solution efficiency but also has good generalization ability, making it suitable for scheduling tasks under various system operating states. To better characterize the system safety state, the system safety margin is defined as follows: (35) in Indicates the node voltage margin. Indicates node voltage. This represents the minimum node voltage. Indicates the angular velocity margin of the photovoltaic converter. Indicates the capacity of the photovoltaic converter. Represents a node The active power output of the photovoltaic converter, Represents a node The reactive power output of the photovoltaic converter.

[0081] Step 504: In the safe operation mode, if the current output voltage and risk assessment index meet the third mode switching conditions, the safe operation mode is switched to the droop control mode, and the process jumps to the step of obtaining the output reactive power and output voltage of each converter.

[0082] In one embodiment, if the current output voltage and risk assessment index meet the third mode switching conditions, the safe operation mode is switched to the droop control mode, including: If the current output voltage is greater than or equal to the first voltage value, less than or equal to the second voltage value, and the risk assessment index is less than the third preset value, then the safe operation mode will be switched to the droop control mode.

[0083] The third preset value can be set according to needs. For example, the third preset value can be set to 0.1, 0.15, etc.

[0084] In one embodiment, during mode transition, a convolutional neural network is used for adaptive mode transition decision-making. The design of the convolutional neural network is as follows: (1) Input structure design Spatiotemporal data transformation: The time-series data obtained after constructing data features from various collected data is mapped into a two-dimensional matrix according to the node position. For example, the node voltage at each time step constitutes an "image" channel, with the time dimension as the height. Input tensor dimensions: [Batch Size, Time Steps, Nodes, Features], and spatiotemporal correlations are extracted through convolutional layers.

[0085] (2) Network architecture Input layer: Receives a spatiotemporal feature matrix containing multiple channels.

[0086] Convolutional layers: a. Temporal Convolution: Uses 1D convolutional kernels to capture temporal dependencies; b-space convolution: 2D convolution kernels extract spatial relationships between nodes, such as line connections; Pooling layer: Max pooling reduces dimensionality while preserving key features; Fully connected layer: Flattens the output of the convolutional layer and maps it to the predicted value of the bus voltage; Output layer: Regression outputs the operating mode of the system.

[0087] (3) Hyperparameter configuration Convolution kernel size: Set the time dimension to 3-5, and adjust the spatial dimension according to the topological density.

[0088] Activation function: ReLU avoids gradient vanishing.

[0089] Optimizer: Adam optimizer, initial learning rate 0.001.

[0090] Loss function: Mean Squared Error (MSE) combined with a voltage deviation weighting term, set as follows: .

[0091] in, Represents the loss function. and For balance coefficient, This represents the mean square error function. Indicates the voltage reference value. Indicates the actual voltage value. Indicates the baseline value of the power flow in the line. This represents the actual value of the power flow along the line.

[0092] This invention obtains the output reactive power of each converter and determines the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power. Based on the reactive power allocation error and the obtained output voltage of the corresponding converter, the adaptive virtual impedance value of each converter is determined. Based on the adaptive virtual impedance value, droop control of the converters is performed in the microgrid system to obtain the current output voltage of each converter, and then jumps to the "obtain the output reactive power of each converter" step until the current output voltage meets the first mode switching condition. Through the droop control strategy based on virtual impedance, the problem of reactive power distribution imbalance and circulating current risk that occurs when multiple machines are connected in parallel in the prior art can be solved, thereby improving system stability and power supply reliability.

[0093] In this embodiment, by adding a small-signal model to the microgrid system, the active and reactive power of each converter are decoupled, the coupling effect is eliminated, the stability margin of the microgrid is improved, and thus the system stability and power supply reliability are further improved.

[0094] In this embodiment, the overall safety and economy of the system are improved by adaptively switching between three modes: off-grid operation mode, economic operation mode, and safe operation mode.

[0095] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0096] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0097] Figure 6 A schematic diagram of the reactive power control device provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below: like Figure 6 As shown, the reactive power control device 6 is applied to a microgrid system in which multiple converters operate in parallel and are connected to the AC bus to supply power to the load. It includes: an acquisition module 61 and a processing module 62.

[0098] Module 61 is used to acquire the output reactive power of each converter; Processing module 62 is used to determine the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power. The processing module 62 is also used to determine the adaptive virtual impedance value of each converter based on the reactive power distribution error and the obtained output voltage of the corresponding converter. The processing module 62 is also used to perform droop control of the converter in the microgrid system based on the adaptive virtual impedance value, obtain the current output voltage of each converter, and jump to the "obtain the output reactive power of each converter" step until the current output voltage meets the first mode switching condition.

[0099] In one possible implementation, when processing module 62 determines the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power, it is used for: according to Determine the reactive power distribution error of each converter; in, Indicates the first reactive power distribution error of each converter Indicates the first The output reactive power of each converter Indicates the first The preset reactive power of each converter.

[0100] In one possible implementation, when processing module 62 determines the adaptive virtual impedance value of each converter based on the reactive power distribution error and the obtained output voltage of the corresponding converter, it is used for: according to Determine the adaptive virtual impedance value for each converter; in, Indicates the first The adaptive virtual impedance value of the converter This represents the virtual impedance adaptive coefficient. Indicates the first The current output voltage of the converter.

[0101] In one possible implementation, based on the adaptive virtual impedance value, the processing module 62 performs droop control of the converters in the microgrid system. When obtaining the current output voltage of each converter, it is used for: according to Obtain the current output voltage of each converter; in, Indicates the first The voltage reference value of each converter This represents the reactive power droop coefficient. Indicates the virtual impedance value. Represents a constant virtual inductance. Indicates the first The output voltage angular velocity of the converter Represents the first in the dq coordinate system The output current of the converter.

[0102] In one possible implementation, before obtaining the current output voltage of each converter by performing droop control of the converters in the microgrid system based on adaptive virtual impedance values, the processing module 62 is further configured to: A small-signal model is used to dynamically decouple each converter to obtain the compensation voltage; Based on adaptive virtual impedance values, droop control of converters in a microgrid system is performed to obtain the current output voltage of each converter, including: Based on adaptive virtual impedance values ​​and compensation voltages, droop control of converters is performed in a microgrid system to obtain the current output voltage of each converter.

[0103] In one possible implementation, processing module 62 uses a small-signal model to dynamically decouple each converter, and when obtaining the compensation voltage, it is used for: Obtain the reference values ​​of the output voltage angular velocity of each converter; The angular velocity offset is determined based on the output voltage angular velocity reference value and the current output voltage angular velocity of each converter; In the reactive power control channel, an angular velocity offset related to the active power disturbance is added to obtain a compensation voltage.

[0104] In one possible implementation, based on the adaptive virtual impedance value and the compensated voltage, the processing module 62 performs droop control of the converters in the microgrid system. When obtaining the current output voltage of each converter, it is used for: according to Obtain the current output voltage of each converter; in, This represents the voltage compensation coefficient. Indicates the first The voltage compensation value of the output voltage of each converter.

[0105] In one possible implementation, based on the adaptive virtual impedance value, droop control of the converters in the microgrid system is performed to obtain the current output voltage of each converter, and the process jumps to the "obtain the output reactive power of each converter" step. After the current output voltage meets the first mode switching condition, the processing module 62 is further used to: In droop control, the line impedance fluctuation rate in the microgrid system is determined; If the voltage deviation rate between the current output voltage and the rated voltage of the strain gauge and the line impedance fluctuation rate meet the first mode switching condition, then the current droop control mode will be switched to the economic operation mode. In the economic operating mode, if the current output voltage meets the second mode switching condition, the economic operating mode will be switched to the safe operating mode. In the safe operation mode, if the current output voltage and risk assessment index meet the switching conditions of the third mode, the safe operation mode will be switched to the droop control mode, and the process will jump to the step of obtaining the output reactive power and output voltage of each converter.

[0106] In one possible implementation, if the voltage deviation between the current output voltage and the rated voltage of the strain gauge current transformer meets the first mode switching condition for the line impedance fluctuation rate, then when the processing module 62 switches the current droop control mode to the economic operation mode, it is used to: If the voltage deviation rate between the current output voltage and the rated voltage of the strain gauge is less than the first preset value, and the line impedance fluctuation rate is less than the second preset value, then the current droop control mode will be switched to the economic operation mode. If the current output voltage meets the second mode switching condition, when processing module 62 switches from economic operation mode to safe operation mode, it is used for: If the current output voltage is less than the first voltage value and the duration exceeds the first time, or if the current output voltage is greater than the second voltage value and the duration exceeds the first time, then the economic operation mode will be switched to the safe operation mode; the first voltage value is less than the second voltage value. If the current output voltage and risk assessment index meet the third mode switching conditions, then when processing module 62 switches from safe operation mode to droop control mode, it is used for: If the current output voltage is greater than or equal to the first voltage value, less than or equal to the second voltage value, and the risk assessment index is less than the third preset value, then the safe operation mode will be switched to the droop control mode.

[0107] The aforementioned reactive power control device acquires the output reactive power of each converter through an acquisition module. Then, the processing module determines the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power. Based on the reactive power allocation error and the acquired output voltage of the corresponding converter, the adaptive virtual impedance value of each converter is determined. Based on the adaptive virtual impedance value, droop control of the converters is performed in the microgrid system. The processing module obtains the current output voltage of each converter and jumps to the "acquire the output reactive power of each converter" step until the current output voltage meets the first mode switching condition. Through the droop control strategy based on virtual impedance, the problems of reactive power distribution imbalance and circulating current risk that occur when multiple machines are connected in parallel in the prior art can be solved, thereby improving system stability and power supply reliability.

[0108] In this embodiment, by adding a small-signal model to the microgrid system, the processing module decouples the active and reactive power of each converter, eliminates coupling effects, improves the stability margin of the microgrid, and thus further improves system stability and power supply reliability.

[0109] In this embodiment, the overall safety and economy of the system are improved by adaptively switching between three modes: off-grid operation mode, economic operation mode, and safe operation mode.

[0110] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for controlling reactive power, characterized in that, Applied to microgrid systems where multiple converters operate in parallel and are connected to the AC bus to supply power to loads, including: Obtain the output reactive power of each converter, and determine the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power. Based on the reactive power distribution error and the obtained output voltage of the corresponding converter, determine the adaptive virtual impedance value of each converter; Based on the adaptive virtual impedance value, droop control of the converters is performed in the microgrid system to obtain the current output voltage of each converter, and then jumps to the "obtain the output reactive power of each converter" step until the current output voltage meets the first mode switching condition.

2. The reactive power control method according to claim 1, characterized in that, The determination of the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power includes: according to Determine the reactive power distribution error of each converter; in, Indicates the first reactive power distribution error of each converter Indicates the first The output reactive power of each converter Indicates the first The preset reactive power of each converter.

3. The reactive power control method according to claim 2, characterized in that, Based on the reactive power distribution error and the obtained output voltage of the corresponding converter, the adaptive virtual impedance value of each converter is determined, including: according to Determine the adaptive virtual impedance value for each converter; in, Indicates the first The adaptive virtual impedance value of the converter This represents the virtual impedance adaptive coefficient. Indicates the first The current output voltage of the converter.

4. The reactive power control method according to claim 3, characterized in that, Based on the adaptive virtual impedance value, droop control of the converters in the microgrid system is performed to obtain the current output voltage of each converter, including: according to Obtain the current output voltage of each converter; in, Indicates the first The voltage reference value of each converter This represents the reactive power droop coefficient. Indicates the virtual impedance value. Represents a constant virtual inductance. Indicates the first The output voltage angular velocity of the converter Represents the first in the dq coordinate system The output current of the converter.

5. The reactive power control method according to claim 1, characterized in that, Before performing droop control of the converters in the microgrid system based on the adaptive virtual impedance value to obtain the current output voltage of each converter, the process further includes: A small-signal model is used to dynamically decouple each converter to obtain the compensation voltage; Based on the adaptive virtual impedance value, droop control of the converters in the microgrid system is performed to obtain the current output voltage of each converter, including: Based on the adaptive virtual impedance value and the compensation voltage, droop control of the converters is performed in the microgrid system to obtain the current output voltage of each converter.

6. The reactive power control method according to claim 5, characterized in that, The method of dynamically decoupling each converter using a small-signal model to obtain the compensation voltage includes: Obtain the reference values ​​of the output voltage angular velocity of each converter; The angular velocity offset is determined based on the aforementioned output voltage angular velocity reference value and the current output voltage angular velocity of each converter; In the reactive power control channel, the angular velocity offset related to the active power disturbance is added to obtain the compensation voltage.

7. The reactive power control method according to claim 6, characterized in that, Based on the adaptive virtual impedance value and the compensated voltage, droop control of the converters in the microgrid system is performed to obtain the current output voltage of each converter, including: according to Obtain the current output voltage of each converter; in, This represents the voltage compensation coefficient. Indicates the first The voltage compensation value of the output voltage of each converter.

8. The reactive power control method according to any one of claims 1-7, characterized in that, Based on the adaptive virtual impedance value, droop control of the converters in the microgrid system is performed to obtain the current output voltage of each converter, and the process jumps to the "obtain the output reactive power of each converter" step. After the current output voltage meets the first mode switching condition, the process further includes: In droop control, the line impedance fluctuation rate in the microgrid system is determined; If the voltage deviation rate between the current output voltage and the rated voltage of the strain gauge and the line impedance fluctuation rate meet the first mode switching condition, then the current droop control mode is switched to the economic operation mode. In the economic operation mode, if the current output voltage meets the second mode switching condition, the economic operation mode is switched to the safe operation mode. In the safe operation mode, if the current output voltage and risk assessment index meet the third mode switching conditions, the safe operation mode is switched to the droop control mode, and the process jumps to the step of obtaining the output reactive power and output voltage of each converter.

9. The reactive power control method according to claim 8, characterized in that, If the voltage deviation between the current output voltage and the rated voltage of the strain gauge current transformer meets the first mode switching condition for the line impedance fluctuation rate, then the current droop control mode is switched to the economic operation mode, including: If the voltage deviation rate between the current output voltage and the rated voltage of the strain gauge is less than a first preset value, and the line impedance fluctuation rate is less than a second preset value, then the current droop control mode will be switched to the economic operation mode. If the current output voltage meets the second mode switching condition, then the economic operation mode is switched to the safe operation mode, including: If the current output voltage is less than a first voltage value and the duration exceeds a first time, or if the current output voltage is greater than a second voltage value and the duration exceeds a first time, then the economic operation mode is switched to the safe operation mode; where the first voltage value is less than the second voltage value. If the current output voltage and risk assessment index meet the third mode switching conditions, then the safe operation mode is switched to the droop control mode, including: If the current output voltage is greater than or equal to the first voltage value, less than or equal to the second voltage value, and the risk assessment index is less than the third preset value, then the safe operation mode will be switched to the droop control mode.

10. A reactive power control device, characterized in that, Applied to microgrid systems where multiple converters operate in parallel and are connected to the AC bus to supply power to loads, including: The acquisition module is used to acquire the output reactive power of each converter; The processing module is used to determine the reactive power allocation error of each converter based on the preset reactive power allocated to each converter according to the rated ratio and the output reactive power. The processing module is also used to determine the adaptive virtual impedance value of each converter based on the reactive power allocation error and the obtained output voltage of the strain converter. The processing module is also used to perform droop control of the converter in the microgrid system based on the adaptive virtual impedance value, obtain the current output voltage of each converter, and jump to the "obtain the output reactive power of each converter" step until the current output voltage meets the first mode switching condition.

Images