Islanded microgrid multi-element cooperative frequency modulation method based on ev cluster and vsg

Abstract

The application discloses a kind of based on EV cluster and VSG's island microgrid multivariate cooperative frequency modulation method, belong to microgrid operation control technical field. Including: generating the adjustable capacity boundary of EV cluster and the multiple scenario photovoltaic output set;Build multiple scenario random optimization scheduling model and solve, generate day-ahead scheduling plan;The island microgrid operation cycle is divided into several determination intervals, and the adjustable group is screened out from EV cluster in determination interval;When island microgrid is disturbed, VSG controller and EV cluster cooperatively fill the power difference of island microgrid, obtain the multivariate cooperative frequency modulation result of island microgrid;VSG controller uses the time series prediction model trained and pre-constructed VSG active frequency modulation model to adjust the power of island microgrid;EV cluster selects EV from adjustable group, and the power required by island microgrid to be compensated is distributed to EV.The application can cope with load disturbance, improve the frequency stability of system.

Description

Technical Field

[0001] This invention relates to the field of microgrid operation and control technology, and in particular to a multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs. Background Technology

[0002] Microgrids, as a new type of power grid capable of integrating local distributed power sources, loads, and energy storage, and enabling grid-connected or islanded operation, are a key carrier for the efficient absorption of renewable energy. However, the intermittency and volatility of renewable energy sources, and their replacement of traditional synchronous generators, lead to a decrease in the system's equivalent inertia and damping. Meanwhile, the large-scale integration of EV clusters enriches the dispatch resources of islanded microgrids, but also increases the system's randomness and uncertainty. Therefore, under the dual influence of high-proportion renewable energy sources and EV cluster integration, the frequency stability of islanded microgrids will face differentiated challenges at different time scales.

[0003] Over the long term, the integration of EV clusters into a high proportion of renewable energy microgrids increases system operational complexity and makes overall system planning more difficult. The coordination of renewable energy output, EV cluster charging and discharging power, and energy storage in isolated microgrids urgently needs to be addressed, as this is crucial for both the long-term stable operation and short-term frequency stability of isolated microgrids.

[0004] On short timescales, load disturbances and fluctuations in renewable energy output place higher demands on the rapid regulation capabilities of islanded microgrids. However, the randomness of EV user travel behavior makes their charging and discharging patterns difficult to predict, posing a significant challenge to the frequency stability of islanded microgrids. Existing research generally treats EV clusters as a load resource with flexible regulation capabilities, leading to their widespread application in microgrid frequency regulation. However, how to accurately assess their adjustable capacity based on real-time vehicle status, efficiently identify dispatchable vehicles, and tap into the regulation potential of EV clusters remains a core problem to be solved.

[0005] Furthermore, virtual synchronous generator (VSG) technology, by simulating the inertia and damping characteristics of a synchronous generator, enables inverters to possess inertial response and frequency regulation functions similar to those of a synchronous generator. This effectively compensates for the low inertia of renewable energy generation, improves system damping and frequency regulation capabilities, and provides an effective technical means for the stable operation of renewable energy islanded microgrids. However, the weak inertia characteristics of islanded microgrids and the impact of EV cluster access on inertia make it difficult for existing VSG fixed or slowly adjusted parameters to match the entire transient process of frequency changes. Moreover, the power support provided by EV clusters needs to be considered during frequency regulation. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs, optimize resource allocation in multiple scenarios, give full play to the flexible adjustment capability of EV clusters, drive EV clusters and VSG controllers to work together to cope with load disturbances, and improve the frequency stability of the system.

[0007] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0008] This invention provides a multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs, comprising:

[0009] Obtain behavioral characteristic parameters, photovoltaic power output forecast, and load forecast output of the EV cluster;

[0010] Based on the behavioral characteristic parameters of the EV cluster, the adjustable capacity boundary of the EV cluster is generated by Monte Carlo simulation, and a set of photovoltaic output for multiple scenarios is generated based on the photovoltaic predicted output.

[0011] Based on the load forecast output, the adjustable capacity boundary of the EV cluster, the photovoltaic output set of multiple scenarios and the power margin coefficient, with the goal of minimizing the total expected operating cost of the system, a multi-scenario stochastic optimization scheduling model is constructed and solved to generate the day-ahead scheduling plan for energy storage and the day-ahead scheduling plan for the EV cluster.

[0012] The operation cycle of the islanded microgrid is divided into several decision intervals. Within each decision interval, dispatchable groups are selected from the EV clusters based on the pre-built EV cluster energy aggregation model and the minimum SOC setting value.

[0013] When an isolated microgrid is disturbed, the VSG controller and the EV cluster work together to fill the power difference of the isolated microgrid according to the day-ahead scheduling plan of energy storage, the day-ahead scheduling plan of EV cluster and the power margin factor, so as to obtain the multi-dimensional coordinated frequency regulation result of the isolated microgrid.

[0014] Among them, the VSG controller and EV cluster work together to fill the power gap in islanded microgrids, including:

[0015] According to the day-ahead scheduling plan for energy storage and the day-ahead scheduling plan for EV clusters, the central controller sends a power reference value to the VSG controller and sends the power difference that needs to be compensated by the EV cluster to the EV cluster. This power difference does not exceed the upper limit of the EV cluster discharge power margin reserved by the power margin coefficient. The VSG controller adjusts the power of the islanded microgrid based on the power reference value using the trained time-series prediction model and the pre-built VSG active frequency regulation model. The EV cluster selects EVs from the dispatchable group according to the power difference and allocates the power that needs to be compensated by the EV to the islanded microgrid according to the discharge allocation principle.

[0016] Optionally, based on the behavioral characteristic parameters of the EV cluster, the adjustable capacity boundary of the EV cluster is generated using Monte Carlo simulation, including:

[0017] Monte Carlo sampling was performed on the behavioral characteristic parameters of the EV cluster to obtain the sampling results; the behavioral characteristic parameters of the EV cluster include access time, disconnection time, daily driving mileage, maximum comprehensive driving range, expected state of charge when disconnected, scheduling protection threshold, charging price and discharging price;

[0018] Based on the sampling results, the charging and discharging states of EVs are divided into chargeable and dischargeable states, chargeable-only states, dischargeable-only states, and unschedulable states.

[0019] The maximum allowable charging power of EVs in each time period is calculated by summing the maximum allowable charging power of EVs in a chargeable and dischargeable state or a chargeable state, and the maximum charging power of EV clusters in each time period is obtained.

[0020] The maximum allowable discharge power of EVs in each time period is calculated by summing the maximum discharge power of the EV cluster in each time period.

[0021] The dispatchable energy of each EV is evenly distributed to each time period according to its effective parking time. The distributed energy of all EVs in dispatchable state in each time period is added up to obtain the available energy of the EV cluster in each time period.

[0022] Optionally, the formula for the combined photovoltaic output of the multiple scenarios is:

[0023] ;

[0024] in, Representing a scene During the period Photovoltaic power output; Indicates the time period The benchmark photovoltaic power output curve; Representing a scene During the period The random perturbation factor; , These represent the start and end times of the effective photovoltaic power output period, respectively. This indicates taking the maximum value.

[0025] Optionally, the formula for the multi-scenario stochastic optimization scheduling model is:

[0026] The objective function is expressed as:

[0027] ;

[0028] The constraints are expressed as follows:

[0029] ;

[0030] ;

[0031] ;

[0032] ;

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] in, Represents minimizing the objective function ; Representing a scene Quantity; Representing a scene The probability of; Indicates the number of time periods; , These represent the penalty cost coefficients for wasted light and loss of load, respectively. Representing a scene During the period The power of discarded light; Representing a scene During the period The power of the load shedding; , These represent the unit charge / discharge operating cost coefficients for energy storage and EV clusters, respectively. , These represent the time periods. Time period Energy storage charging power; , These represent the time periods. Time period The energy storage discharge power; , These represent the time periods. Time period EV charging power and EV discharging power; This indicates the number of periods of high photovoltaic output; This represents the penalty cost coefficient for insufficient charging of EV clusters. Indicates the time period Non-negative variables; Representing a scene During the period The actual output of photovoltaic power; Indicates the time period Net energy storage output; Indicates the time period EV cluster net output; Representing a scene During the period The actual load value; Representing a scene During the period Photovoltaic power output; Representing a scene During the period The load output; , These represent the time periods. Time period Energy storage capacity; , These represent the energy storage charging efficiency coefficient and the energy storage discharging efficiency coefficient, respectively. Indicates the time step; , These represent the minimum and maximum energy levels for energy storage, respectively. Indicates the rated energy of the energy storage; Indicates the time period Binary variables used to ensure mutual exclusion of energy storage charging and discharging states; Indicates the power margin factor; , These represent the time periods. The maximum charging power and maximum discharging power of the energy storage; , These represent the initial moments of the system's operating cycle. Energy storage capacity, system operation cycle end time Energy storage capacity; Indicates the time period Binary variables used to ensure mutual exclusion of EV charging and discharging states; , These represent the time periods. The maximum charging power and maximum discharging power of the EV; Indicates the percentage of safety discount; Indicates the time period The total energy can be estimated; This represents the minimum excitation power.

[0039] Optionally, the operating cycle of the islanded microgrid can be divided into several determination intervals, including:

[0040] Using a fixed duration as the judgment period, the operating cycle of the islanded microgrid is divided into... There are several decision intervals, and each decision interval is represented as follows:

[0041] ;

[0042] The EV cluster energy aggregation model is represented as follows:

[0043] ;

[0044] in, Indicates the decision interval ; , These represent the start and end times of the decision interval, respectively. Indicates the number of decision intervals; , They represent the first EVs during the time period The state of charge, in the initial period The state of charge; Indicates the first The rated capacity of the battery in an EV; , These represent the energy storage charging efficiency coefficient and the energy storage discharging efficiency coefficient, respectively. , They represent the first EVs during the time period The charging power and discharging power.

[0045] Optionally, based on a pre-built EV cluster energy aggregation model and a minimum SOC setting, schedulable groups are selected from the EV cluster, including:

[0046] For each decision interval, calculate the minimum SOC setting value required at the end of the decision interval, expressed as:

[0047] ;

[0048] Based on the pre-constructed EV cluster energy aggregation model and the minimum SOC setting, the EV state parameters are calculated and expressed as follows:

[0049] ;

[0050] Will The EV collection is used as an off-grid group. The EV group will be used as a charging support group. The set of EVs is used as a schedulable group;

[0051] in, Indicates the first EVs in the judgment range Minimum SOC setting; Indicates the first The State of Charge (SOC) of an EV when it is expected to leave the grid; Indicates the first The expected time for an EV to leave the power grid; Indicates the first The rated charging power of an EV; Indicates the first EVs in the judgment range EV state parameters; Indicates the first EVs during the time period The grid connection status.

[0052] Optionally, the VSG controller adjusts the power of the islanded microgrid based on the power reference value, using a trained time-series prediction model and a pre-built VSG active power frequency regulation model, including:

[0053] Input the dataset containing virtual inertia and damping coefficients into the trained time series prediction model, and output the adjusted virtual inertia and damping coefficients.

[0054] The adjusted virtual inertia and damping coefficient are input into the pre-built VSG active frequency regulation model to output the power of the islanded microgrid.

[0055] The VSG active frequency modulation model is represented as follows:

[0056] ;

[0057] ;

[0058] ;

[0059] in, Represents virtual inertia; , These represent the output angular frequency and reference angular frequency of the VSG controller, respectively. , These represent the mechanical power and electromagnetic power of the VSG controller, respectively. Indicates the damping coefficient; This indicates the active power command value of the VSG controller; This represents the active frequency droop coefficient of the VSG controller; This represents the Laplace operator.

[0060] Optionally, the EV cluster selects EVs from the dispatchable group based on the power difference, and allocates the power that the islanded microgrid needs to compensate for to the EVs according to the discharge allocation principle, including:

[0061] When an isolated microgrid is disturbed, obtain the discharge power indication value that the isolated microgrid requires the EV cluster to provide;

[0062] Calculate the number of EVs that need to be dispatched based on the discharge power indication value and the maximum discharge power of a single EV. ;

[0063] Select the group with the largest discharge margin from the schedulable group. When an EV participates in the discharge, the power that the islanded microgrid needs to compensate for is allocated to the EV participating in the discharge according to the discharge allocation principle.

[0064] Optionally, the discharge allocation principle is expressed as:

[0065] ;

[0066] ;

[0067] in, Indicates the first EVs during the time period The discharge power; Indicates the first EVs at all times Discharge margin; Indicates the time period The discharge power indication value; Indicates the number of EVs; Indicates the first EVs during the time period The state of charge; Indicates the first EVs in the judgment range The minimum SOC setting value.

[0068] Optional, also includes:

[0069] When the islanded microgrid is undisturbed, EVs connected to the islanded microgrid charge at their rated power, while EVs not connected to the microgrid charge at 0 power, as shown below:

[0070] ;

[0071] in, Indicates the first EVs during the time period The discharge power; Indicates the first The rated charging power of an EV; Indicates the first EVs in the judgment range EV state parameters.

[0072] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0073] This invention establishes a long-term and short-term collaborative optimized scheduling architecture based on EV clusters and VSG controllers. Considering the uncertainties of photovoltaic power output across multiple scenarios, it evaluates the adjustability of EV clusters using Monte Carlo simulation and couples the power margin into a multi-scenario stochastic optimization model. This model balances economy and stability, minimizing expected operating costs while reserving necessary adjustment resources for short-term frequency regulation. Furthermore, due to the stochastic state of EV clusters, a dynamic EV cluster grouping and aggregation mechanism is designed. Based on the long-term strategy's reserved power margin, it dynamically divides vehicles into groups based on real-time vehicle status information, quickly identifying dispatchable vehicles and decomposing and issuing aggregated power commands, making EV clusters a reliable flexible frequency regulation resource. Simultaneously, an adaptive parameter VSG control strategy driven by a time-series prediction model adaptively adjusts the virtual inertia and damping coefficient, which, together with the dispatchable EVs obtained from dynamic grouping and aggregation, enhances the system's short-term frequency regulation capability. Attached Figure Description

[0074] Figure 1 This is a partial flowchart illustrating the multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs provided in an embodiment of the present invention.

[0075] Figure 2 This is a schematic diagram of the overall process of the multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs provided in an embodiment of the present invention;

[0076] Figure 3 This is a schematic diagram of the 24-hour charge-discharge plan and power margin characteristic curve of an EV cluster provided in an embodiment of the present invention;

[0077] Figure 4 This is a schematic diagram of the changes in energy storage SOC under long-term optimized scheduling provided in an embodiment of the present invention;

[0078] Figure 5 A schematic diagram of energy storage charging and discharging power under long-term optimized scheduling provided in an embodiment of the present invention;

[0079] Figure 6 This is a schematic diagram of power balance in an islanded microgrid provided in an embodiment of the present invention;

[0080] Figure 7 This is a schematic diagram comparing the controllable number of EV clusters under different strategies provided in the embodiments of the present invention;

[0081] Figure 8 This is a schematic diagram of the VSG output frequency variation waveform under different strategies provided in the embodiments of the present invention;

[0082] Figure 9 This is a schematic diagram of the waveform of virtual inertia change provided in an embodiment of the present invention;

[0083] Figure 10 This is a schematic diagram of the waveform of the damping coefficient change provided in an embodiment of the present invention. Detailed Implementation

[0084] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0085] The term "and / or" simply describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0086] Example 1

[0087] like Figure 1 As shown in the figure, this embodiment introduces a multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs, including:

[0088] First, on a long-term scale, multi-scenario stochastic optimization scheduling is performed on islanded microgrid systems, including:

[0089] Obtain behavioral characteristic parameters, photovoltaic power output forecast, and load forecast output of the EV cluster;

[0090] Based on the behavioral characteristic parameters of the EV cluster, the adjustable capacity boundary of the EV cluster is generated by Monte Carlo simulation, and a set of photovoltaic output for multiple scenarios is generated based on the photovoltaic predicted output.

[0091] Based on the load forecast output, the adjustable capacity boundary of the EV cluster, the ensemble of photovoltaic outputs across multiple scenarios, and the power margin coefficient, a multi-scenario stochastic optimization scheduling model is constructed and solved to minimize the total expected operating cost of the system, thereby generating a day-ahead scheduling plan for energy storage. ,and and the EV cluster's recent scheduling plan ;in , These represent the time periods under the day-ahead dispatch plan for energy storage. The energy storage charging power and energy storage discharging power; This indicates the time period under the day-ahead dispatch plan for energy storage. Energy storage capacity; , These represent the time periods under the day-ahead scheduling plan of the EV cluster. EV charging power and EV discharging power.

[0092] On a short-term timescale, frequency coordinated control of islanded microgrid systems is achieved through VSG controllers and EV clusters, including:

[0093] The operation cycle of the islanded microgrid is divided into several decision intervals. Within each decision interval, dispatchable groups are selected from the EV clusters based on the pre-built EV cluster energy aggregation model and the minimum SOC setting value.

[0094] When an isolated microgrid is disturbed, the VSG controller and the EV cluster work together to fill the power difference of the isolated microgrid according to the day-ahead scheduling plan of energy storage, the day-ahead scheduling plan of EV cluster and the reserved power margin, so as to obtain the multi-dimensional coordinated frequency regulation result of the isolated microgrid.

[0095] Among them, the VSG controller and EV cluster work together to fill the power gap in islanded microgrids, including:

[0096] The central controller sends active power reference values ​​to the VSG controller. Reactive power reference value Send the power difference that needs to be compensated by the EV cluster. The VSG controller adjusts the power of the islanded microgrid based on the power reference value, using the trained time-series prediction model and the pre-built VSG active frequency regulation model; the EV cluster selects EVs from the dispatchable group based on the power difference, and allocates the power that the islanded microgrid needs to compensate to the EVs according to the discharge allocation principle.

[0097] In this embodiment, as Figure 2 As shown, a multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs includes the following steps:

[0098] Step 1: Generate the adjustable capacity boundary of the EV cluster and the photovoltaic output set for multiple scenarios, specifically:

[0099] Obtain behavioral characteristic parameters, photovoltaic power output forecast, and load forecast of electric vehicle (EV) clusters.

[0100] Based on statistics of private car travel in typical office areas, the behavioral characteristics parameters of the EV cluster include access time, off-grid time, daily mileage, maximum comprehensive range, expected state of charge when off-grid, dispatch protection threshold, charging price and discharging price.

[0101] Based on the behavioral characteristic parameters of the EV cluster, the adjustable capacity boundary of the EV cluster is generated using Monte Carlo simulation, namely:

[0102] Monte Carlo sampling was performed on the behavioral characteristic parameters of the EV cluster to obtain the sampling results.

[0103] Based on the sampling results, the charging and discharging states of EVs are divided into rechargeable and dischargeable states. Rechargeable only Dischargeable state only and unschedulable state .

[0104] Calculate the charging and discharging state at each time period. Or in rechargeable state only The maximum allowable charging power of each EV is summed, and considering engineering upper limit constraints, the maximum charging power of the EV cluster in each time period is obtained. .

[0105] Calculate the charging and discharging state at each time period. Or only in a dischargeable state The sum of the maximum allowable discharge power of the EVs is used to obtain the maximum discharge power of the EV cluster in each time period. .

[0106] The dispatchable energy of each EV is evenly distributed across time periods based on its effective parking duration. The allocated energy of all EVs in dispatchable status within each time period is summed to obtain the available energy of the EV cluster in each time period. .

[0107] Based on the predicted photovoltaic output, a set of photovoltaic output data for multiple scenarios is generated, namely:

[0108] To address the intermittency and forecasting uncertainty of photovoltaic (PV) power output, a multi-scenario stochastic optimization method is employed to generate a PV power output set for multiple scenarios. A baseline PV power output curve for the next 24 hours is obtained based on day-ahead forecasts. Generated by superimposing random perturbations Photovoltaic output in an equally probable scenario ,in The specific generation process is as follows:

[0109] During the effective output period of photovoltaic power Within, for each time period The generation follows a normal distribution random disturbance factor and restrict it to an interval This is to eliminate extreme biases; Indicates the mean squared error; This represents a normal distribution. The formula for the aggregated photovoltaic output across multiple scenarios is:

[0110] ;

[0111] in, Representing a scene During the period Photovoltaic power output; Indicates the time period The benchmark photovoltaic power output curve; Representing a scene During the period The random perturbation factor; , These represent the start and end times of the effective photovoltaic power output period, respectively. This indicates taking the maximum value.

[0112] Output in all scenarios is constrained. Within the range; This represents the rated power of the photovoltaic system. The multi-scenario set generated by this method effectively addresses the randomness of photovoltaic power output, providing crucial input for subsequent stochastic optimization models.

[0113] Step 2: Construct a multi-scenario stochastic optimization scheduling model, specifically as follows:

[0114] Based on the load forecast output, the adjustable capacity boundary of the EV cluster, the photovoltaic output set of multiple scenarios, and the power margin coefficient, a multi-scenario stochastic optimization scheduling model is constructed and solved with the goal of minimizing the total expected operating cost of the system, generating day-ahead scheduling plans for energy storage and EV clusters.

[0115] The construction of a multi-scenario stochastic optimization scheduling model includes:

[0116] To address the uncertainties in photovoltaic and load forecasting and to provide the system with reliable real-time adjustment capabilities, this model introduces a power margin factor. By limiting the planned charging and discharging power of energy storage and EV clusters, this reserved power capacity constitutes a two-way adjustment margin for rapid response in real-time operation. By coupling this power margin constraint to the optimization framework, it provides a guarantee for short-term frequency regulation of EV clusters.

[0117] The model aims to minimize the total expected operating cost of the system within the scheduling cycle, and its objective function is as follows:

[0118] ;

[0119] in, Represents minimizing the objective function ; Representing a scene Quantity; Representing a scene The probability of; Indicates the number of time periods; , These represent the penalty cost coefficients for wasted light and loss of load, respectively. Representing a scene During the period The power of discarded light; Representing a scene During the period The power of the load shedding; , These represent the unit charge / discharge operating cost coefficients for energy storage and EV clusters, respectively. , These represent the time periods. Time period Energy storage charging power; , These represent the time periods. Time period The energy storage discharge power; , These represent the time periods. Time period EV charging power and EV discharging power; This indicates the number of periods of high photovoltaic output; The penalty cost coefficient representing insufficient charging of EV clusters is defined only during periods of high photovoltaic output. (like ), used to construct soft constraints in the optimization model to incentivize EV cluster charging; Indicates the time period The non-negative variable represents the EV cluster charging power and minimum excitation power during the high-output period of photovoltaics. The difference.

[0120] The constraints include:

[0121] The system power balance constraint requires that, in each time period under each photovoltaic scenario, the net generated power equals the net load power:

[0122] ;

[0123] The specific definitions of each term in the formula are as follows:

[0124] ;

[0125] in, Representing a scene During the period The actual output of photovoltaic power; Indicates the time period Net energy storage output; Indicates the time period EV cluster net output; Representing a scene During the period The actual load value; Representing a scene During the period Photovoltaic power output; Representing a scene During the period The load output. Energy storage operation needs to reserve short-term frequency regulation capability for the system while meeting energy conservation and operational limit constraints. Introducing binary variables. To ensure mutual exclusion between charging and discharging states, and to ensure that the energy at the end of scheduling is not lower than the initial value, this, together with the power margin constraint, constitutes a dual power-energy guarantee for frequency stability.

[0126] ;

[0127] ;

[0128] ;

[0129] ;

[0130] in, , These represent the time periods. Time period Energy storage capacity; , These represent the energy storage charging efficiency coefficient and the energy storage discharging efficiency coefficient, respectively. Indicates the time step; , These represent the minimum and maximum energy levels for energy storage, respectively. Indicates the rated energy of the energy storage; Indicates the time period Binary variables used to ensure mutual exclusion of energy storage charging and discharging states; Indicates the power margin factor; , These represent the time periods. The maximum charging power and maximum discharging power of the energy storage; , These represent the initial moments of the system's operating cycle. Energy storage capacity, system operation cycle end time The energy storage capacity. For EV clusters, the scheduling plan must simultaneously meet the requirements of adjustable capacity boundaries and operational safety. Their charging and discharging power is constrained by both the adjustable boundary and the power margin coefficient, which is also addressed by introducing binary variables. To achieve mutual exclusion of charging and discharging states for the EV cluster, and to prevent over-discharging, the discharge energy is constrained to not exceed a certain percentage of the current total available energy. Specifically, the discharge energy corresponding to the discharge power of the EV cluster during a given time period must not exceed its estimated total available energy for that period. A safe discount percentage This effectively prevents over-discharge.

[0131] ;

[0132] ;

[0133] in, Indicates the time period Binary variables used to ensure mutual exclusion of EV charging and discharging states; , These represent the time periods. The maximum charging power and maximum discharging power of the EV; Indicates the percentage of safety discount; Indicates the time period The estimated total available energy. To promote the consumption of new energy, a soft constraint is set for EV cluster charging incentives during periods of high photovoltaic output:

[0134] ;

[0135] in, This represents the minimum excitation power, a non-negative constant, used to encourage the model to efficiently utilize the charging capacity of the EV cluster to absorb photovoltaic surplus, thereby reducing the deviation between the aggregated scheduling plan and the response intention of decentralized individuals at the optimization level.

[0136] This multi-scenario stochastic optimization scheduling model is ultimately constructed as a mixed-integer linear programming problem. Substituting the values ​​into the solution yields the optimal 24-hour collaborative scheduling plan for energy storage and EV clusters. and energy storage state of charge plan Among them, the day-ahead energy storage dispatch plan includes and The EV cluster's recent scheduling plan includes By designing the power margin factor The power space of energy storage and EV clusters not occupied by day-ahead scheduling in the optimization results can be calculated, i.e., the power space in each time period. These multiples of the maximum adjustable power together constitute the power regulation resources of the system to cope with real-time frequency disturbances.

[0137] Step 3: The dynamic clustering and aggregation strategy for the EV cluster is as follows:

[0138] The islanded microgrid operating cycle is divided into several judgment intervals, with a fixed duration as the judgment period. There are several judgment intervals. In this embodiment, one hour is used as a judgment period. The judgment interval is represented as follows:

[0139] ;

[0140] in, Indicates the decision interval ; , These represent the start and end times of the decision interval, respectively. This indicates the number of decision intervals. The system only considers the start time of each decision interval. The status of all EVs in the network is synchronously determined, and dynamic aggregation and grouping are completed accordingly.

[0141] To ensure the first When an EV is disconnected from the grid, it can achieve its expected state of charge (SOC). For each judgment interval... It is necessary to calculate the end time of its judgment interval. The minimum required SOC setting is expressed as:

[0142] ;

[0143] in, Indicates the first EVs in the judgment range Minimum SOC setting; Indicates the first The State of Charge (SOC) of an EV when it is expected to leave the grid; Indicates the first The expected time for an EV to leave the power grid; Indicates the first The rated charging power of an EV.

[0144] Within the decision interval, based on the pre-built EV cluster energy aggregation model and the minimum SOC setting, schedulable groups are selected from the EV clusters, i.e.:

[0145] The EV state parameters are calculated based on the pre-built EV cluster energy aggregation model and the minimum SOC setting.

[0146] An EV cluster consists of several individual EVs with varying battery capacities, initial states of charge, and charging / discharging power. To efficiently manage this heterogeneous EV cluster, an energy state-based aggregation modeling approach is adopted to construct an EV cluster energy aggregation model. The core idea of ​​this method is to ignore the dynamic details of each EV, such as its motor and body, and abstract it as a distributed energy storage unit, retaining only its essential attributes as an energy storage device, namely the battery's energy storage and release capabilities. This simplified representation significantly reduces the computational complexity of cluster simulation while preserving the most critical energy and power dynamic characteristics in grid dispatching.

[0147] Within this framework, each individual EV is represented by its battery model, the core of which is the dynamic equation describing its state of charge. The EV cluster energy aggregation model is expressed as:

[0148] ;

[0149] in, , They represent the first EVs during the time period The state of charge, in the initial period The state of charge; Indicates the first The rated capacity of the battery in an EV; , These represent the energy storage charging efficiency coefficient and the energy storage discharging efficiency coefficient, respectively. , They represent the first EVs during the time period The charging power and discharging power.

[0150] The EV state parameters are represented as follows:

[0151] ;

[0152] in, Indicates the first EVs in the judgment range EV state parameters; Indicates the first EVs during the time period The grid connection status.

[0153] Will The EV collection is used as an off-grid group. The EV group will be used as a charging support group. The set of EVs is used as a schedulable group.

[0154] Off-grid groups represent electric vehicles that have left the grid and will not be scheduled for any further purposes. EVs in charging guarantee groups must maintain a charging status to ensure they reach the expected SOC when off-grid. EVs in dispatchable groups are allowed to discharge to the grid (Vehicle-to-Grid, V2G) while ensuring their off-grid SOC, providing flexibility support to the grid.

[0155] Step 4: The VSG controller and EV cluster work together to fill the power gap in the islanded microgrid, specifically as follows:

[0156] When an isolated microgrid is disturbed, the Virtual Synchronous Generator (VSG) controller and the EV cluster work together to fill the power gap in the isolated microgrid, based on the day-ahead scheduling plan of the energy storage system, the day-ahead scheduling plan of the EV cluster, and the power margin factor, thus obtaining the multi-factor coordinated frequency regulation result of the isolated microgrid. The disturbance to the isolated microgrid refers to frequency fluctuations, and the frequency deviation is calculated. and frequency change rate The method for collaboratively filling the power gap in isolated microgrids is as follows: the proportional-integral unit in the central controller performs real-time calculations on the frequency, and sends the output result, i.e., the power gap, to the VSG controller and EV cluster according to the allocation coefficient; the VSG controller and EV cluster react quickly upon receiving the signal to fill the power deficit.

[0157] Based on the day-ahead scheduling plan for energy storage and the day-ahead scheduling plan for EV clusters, the central controller sends power reference values ​​to the VSG controller and sends the power difference that needs to be compensated by the EV cluster. power difference The discharge power margin of the EV cluster shall not exceed the upper limit of the power margin reserved by the power margin coefficient. The power reference value includes the active power reference value. and reactive power reference value .

[0158] The VSG controller adjusts the power of the islanded microgrid based on the power reference value, using a trained time-series prediction model and a pre-built VSG active power frequency regulation model.

[0159] The time series prediction model was trained using an Adaptive Moment Estimation (Adam) optimizer with a learning rate of 10. -3 The mean squared error (MSE) is used as the loss function, and an early stopping mechanism is introduced, terminating training if the loss set does not decrease after three consecutive rounds. In this embodiment, the time series prediction model is a Transformer model.

[0160] Input the dataset containing virtual inertia and damping coefficients into the trained time-series prediction model, and output the adjusted virtual inertia and damping coefficients. The time-series prediction model learns the changing patterns of virtual inertia and damping coefficients; the adjustment method for virtual inertia and damping coefficients is as follows:

[0161] Will , , , , , , The 7-dimensional variables are input into the trained Transformer model, where , These represent the change in angular frequency and the rate of change of angular frequency, respectively. , , , , These represent the state of charge of energy storage, the active power of energy storage, the active power of the EV cluster, the active power of photovoltaics, and the active power of the load, respectively, under real-time operating conditions.

[0162] The Transformer model first maps the 7-dimensional original features to a 64-dimensional embedding space through a linear embedding layer, then inputs them into a two-layer Transformer encoder. Each layer of the Transformer encoder has a single attention head dimension of 8, and a random dropout ratio of 0.1 is introduced to suppress overfitting. Finally, the output features of the encoder at the last time step are extracted and adjusted by the output of a fully connected layer to obtain the virtual inertia. With damping coefficient .

[0163] The adjusted virtual inertia and damping coefficient are input into the pre-constructed VSG active frequency regulation model, and the output is the power of the islanded microgrid. Frequency deviation, frequency change rate, and multi-source power state are used as inputs, and virtual inertia is used as the output. With damping coefficient The VSG active frequency modulation model is trained using the adjustment pattern as the learning objective, and the virtual inertia is adaptively adjusted based on the frequency deviation and the rate of frequency change. With damping coefficient To optimize frequency response characteristics.

[0164] The construction of the VSG active frequency modulation model includes:

[0165] VSG technology, by introducing the rotor motion equations of a synchronous generator, enables the inverter to possess inertial and damping characteristics similar to those of a synchronous generator:

[0166] ;

[0167] in, Represents virtual inertia; , These represent the output angular frequency and reference angular frequency of the VSG controller, respectively. The rated angular frequency of a common power grid is used to represent this. , These represent the mechanical power and electromagnetic power of the VSG controller, respectively. This represents the damping coefficient.

[0168] To simulate the frequency regulation characteristics of a synchronous generator, a speed governor is introduced, and active power frequency regulation is achieved through droop control:

[0169] ;

[0170] in, This indicates the active power command value of the VSG controller; This represents the active frequency droop coefficient of the VSG controller, which characterizes the sensitivity of VSG power regulation when a certain frequency change occurs in the system.

[0171] Combining the above two equations, we can obtain:

[0172] ;

[0173] in, This represents the Laplace operator.

[0174] The EV cluster selects EVs from the dispatchable group based on power differences. According to discharge allocation principles, it distributes the power that the islanded microgrid needs to compensate to the EVs. Upon receiving the instruction, the EV cluster that needs to discharge quickly discharges according to the allocated power. Simultaneously, it dynamically rotates vehicles when the vehicle's State of Charge (SOC) is low. That is:

[0175] When disturbances occur, the EV cluster responds quickly, providing support for short-term frequency regulation by adjusting the charging and discharging power.

[0176] First, when the islanded microgrid is disturbed, the discharge power indication value required by the islanded microgrid to be provided by the EV cluster is obtained.

[0177] Then, based on the discharge power indication value and the maximum discharge power of a single EV, the number of EVs that need to be dispatched is calculated. .

[0178] Finally, the group with the largest discharge margin is selected from the schedulable groups. When an EV participates in the discharge, the power that the islanded microgrid needs to compensate for is allocated to the EV participating in the discharge according to the discharge allocation principle.

[0179] The discharge allocation principle is expressed as follows:

[0180] ;

[0181] ;

[0182] in, Indicates the first EVs during the time period The discharge power; Indicates the first EVs at all times Discharge margin; Indicates the time period The discharge power indication value, i.e. the power difference that needs to be compensated by it; This indicates the number of EVs.

[0183] When the islanded microgrid is undisturbed, to prioritize user charging needs, all EVs connected to the islanded microgrid are charged at their rated power, while EVs not connected to the microgrid have a charging power of 0, represented as:

[0184] .

[0185] The EV cluster provides rapid power support through charging piles. The VSG controller, after adjusting the power, adaptively adjusts through the DC / AC inverter to suppress frequency drops. The EV cluster and VSG controller work together through the DC bus and AC bus to fill the power difference in the grid and restore the grid frequency to the rated value.

[0186] This embodiment addresses the issue of decreased frequency stability caused by the high proportion of renewable energy and EV clusters integrated into isolated microgrids, and the current situation where VSG technology is receiving significant attention in renewable energy power systems due to its ability to simulate the inertia and damping characteristics of synchronous generators. It proposes the following: In the long term, based on Monte Carlo simulation to estimate the adjustable capacity of EV clusters, the power and energy of energy storage and EV clusters are coordinated in an optimization model that comprehensively considers the costs of curtailment and load shedding, thereby improving the renewable energy absorption capacity. Simultaneously, a certain power margin is reserved to accommodate short-term frequency regulation needs. In the short term, firstly, a dynamic clustering and aggregation strategy is used to adjust the real-time charging and discharging power of individual EVs within the reserved power margin. Secondly, VSG collaborative frequency regulation is achieved through adaptive parameters of the EV clusters and the Transformer model. This effectively addresses grid frequency fluctuations caused by power imbalances and significantly improves the frequency stability of the system.

[0187] Example 2

[0188] Based on Example 1, this example introduces an experimental case of a multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs:

[0189] In the short term, to verify the economy and robustness of the method of the present invention in long-term operation, Monte Carlo simulation was used to evaluate the aggregated adjustable capability of 200 EVs, and the day-ahead 24-hour optimized scheduling simulation was carried out in combination with multi-scenario photovoltaic output prediction.

[0190] Changes in planned charge / discharge power and power margin of the EV cluster within 24 hours are as follows: Figure 3 As shown, from Figure 3 As can be seen, the power of the EV cluster is strictly limited within the reserved power margin coefficient. Moreover, in the short time scale, when the discharge power margin of the EV cluster is relatively insufficient and load disturbance occurs, the EV cluster can quickly provide power support. This indicates that while pursuing economy, the optimization model has reserved the necessary adjustment capacity for short-time frequency regulation.

[0191] Furthermore, the operating trajectory of the energy storage system under optimized scheduling is as follows: Figure 4 and Figure 5 As shown in the figure, it can be seen that the change in energy storage SOC is related to... Figure 3 The charging and discharging power of the EV cluster is highly coupled with the photovoltaic output and load demand in terms of timing.

[0192] During peak daytime solar PV periods, the energy storage system continuously charges to store surplus energy, while the EV cluster charges at a higher power to jointly enhance solar PV consumption. During peak nighttime load periods, the energy storage system discharges to support the system, while the EV cluster provides appropriate discharge power during certain periods, forming complementary support.

[0193] Throughout the entire scheduling cycle, the energy storage SOC remained within a safe range. This ensured that both the EV cluster and energy storage power constraints were met. Figure 6 The power balance results show that the established model can minimize the overall cost of curtailment and load loss by coordinating "source-storage-load" resources under fluctuating conditions, while taking into account the reliability and economy of the system's long-term operation.

[0194] Under a robust optimization framework that simultaneously considers multiple possible photovoltaic (PV) output scenarios, the system achieved zero load shedding, fully guaranteeing power supply reliability, and controlling the PV curtailment rate at 7.7%. It should be noted that this portion of curtailed light sources was achieved through stochastic optimization across multiple scenarios, a conservative strategy adopted to address PV uncertainties, aiming to ensure the safe operation of the system in each scenario, reflecting a trade-off between robustness and economy. Theoretically, zero curtailment could be achieved if deterministic optimization were used and predictions were completely accurate.

[0195] To verify the superiority of the dynamic clustering strategy for EV clusters proposed in this invention, a time period of 0-300 minutes was selected. Simulations were performed on three strategies: a traditional strategy using the expected off-grid SOC as a threshold, a K-means clustering algorithm (K-means), and the strategy proposed in this invention. The number of schedulable EVs per minute was obtained. To highlight the differences between the three strategies, 30 data points were grouped together, and the number of schedulable EVs per half-hour was compared. The results are as follows: Figure 7 As shown, during the experimental period, the strategy proposed in this invention outperformed the other two strategies for a greater proportion of the time. That is, within the same time frame, the strategy proposed in this invention is more likely to gain the lead, demonstrating the superiority of the strategy proposed in this invention.

[0196] A simulation model of an islanded photovoltaic-storage microgrid covering EV clusters and Transformer-driven parameter adaptive VSGs was built. The model was compared with four typical schemes: transformer model + EV frequency regulation (this scheme), fixed parameter VSG without EV frequency regulation (Scheme 1), fixed parameter VSG with EV frequency regulation (Scheme 2), fuzzy control VSG without EV frequency regulation (Scheme 3), and fuzzy control VSG with EV frequency regulation (Scheme 4) to verify the effectiveness of the strategy proposed in this invention.

[0197] In the simulation, the system initially had an 85 kW active load, and a 47.85 kW active load was connected at 1.2 s. The system simulation operating parameters are shown in Table 1 below. Figure 8 The waveform changes of the VSG output frequency are shown, along with the VSG's virtual inertia. Damping coefficient The changes are as follows: Figure 9 and Figure 10 As shown.

[0198] Table 1 System Simulation Operation Parameters

[0199]

[0200] contrast Figure 8 The frequency modulation waveforms under different schemes show that this scheme exhibits excellent frequency modulation performance and dynamic response characteristics. The schedulable EV cluster obtained by dynamic grouping and aggregation achieves precise matching of charging and discharging power in the face of load disturbances. In the schemes involving frequency modulation, the frequency drop is significantly reduced, indicating that it can quickly provide power support to the system during frequency modulation and alleviate regulation pressure. Meanwhile, according to... Figure 9 , Figure 10 The waveform changes, and when the system experiences load disturbances, VSG, through the Transformer model, can quickly and dynamically adjust the virtual inertia based on the frequency deviation. and damping coefficient This effectively suppresses frequency drops and shortens frequency fluctuation periods, significantly improving the speed and stability of frequency response compared to fixed-parameter VSG and fuzzy-controlled VSG. Therefore, the proposed collaborative frequency modulation scheme achieves efficient collaboration between the EV cluster and the Transformer-driven parameter adaptive VSG on a short-time scale, ensuring rapid frequency smoothing and stable recovery of the system frequency.

[0201] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0202] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0203] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0204] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0205] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs, characterized in that, include: Obtain behavioral characteristic parameters, photovoltaic power output forecast, and load forecast output of the EV cluster; Based on the behavioral characteristic parameters of the EV cluster, the adjustable capacity boundary of the EV cluster is generated by Monte Carlo simulation, and a set of photovoltaic output for multiple scenarios is generated based on the photovoltaic predicted output. Based on the load forecast output, the adjustable capacity boundary of the EV cluster, the photovoltaic output set of multiple scenarios and the power margin coefficient, with the goal of minimizing the total expected operating cost of the system, a multi-scenario stochastic optimization scheduling model is constructed and solved to generate the day-ahead scheduling plan for energy storage and the day-ahead scheduling plan for the EV cluster. The operation cycle of the islanded microgrid is divided into several decision intervals. Within each decision interval, dispatchable groups are selected from the EV clusters based on the pre-built EV cluster energy aggregation model and the minimum SOC setting value. When an isolated microgrid is disturbed, the VSG controller and the EV cluster work together to fill the power difference of the isolated microgrid according to the day-ahead scheduling plan of energy storage, the day-ahead scheduling plan of EV cluster and the power margin factor, so as to obtain the multi-dimensional coordinated frequency regulation result of the isolated microgrid. Among them, the VSG controller and EV cluster work together to fill the power gap in islanded microgrids, including: According to the day-ahead scheduling plan for energy storage and the day-ahead scheduling plan for EV clusters, the central controller sends a power reference value to the VSG controller and sends the power difference that needs to be compensated by the EV cluster to the EV cluster. This power difference does not exceed the upper limit of the EV cluster discharge power margin reserved by the power margin coefficient. The VSG controller adjusts the power of the islanded microgrid based on the power reference value using the trained time-series prediction model and the pre-built VSG active frequency regulation model. The EV cluster selects EVs from the dispatchable group according to the power difference and allocates the power that needs to be compensated by the islanded microgrid to the EVs according to the discharge allocation principle.

2. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 1, characterized in that, Based on the behavioral characteristic parameters of the EV cluster, the adjustable capacity boundary of the EV cluster is generated using Monte Carlo simulation, including: Monte Carlo sampling was performed on the behavioral characteristic parameters of the EV cluster to obtain the sampling results; the behavioral characteristic parameters of the EV cluster include access time, disconnection time, daily driving mileage, maximum comprehensive driving range, expected state of charge when disconnected, scheduling protection threshold, charging price and discharging price; Based on the sampling results, the charging and discharging states of EVs are divided into chargeable and dischargeable states, chargeable-only states, dischargeable-only states, and unschedulable states. The maximum allowable charging power of EVs in each time period is calculated by summing the maximum allowable charging power of EVs in a chargeable and dischargeable state or a chargeable state, and the maximum charging power of EV clusters in each time period is obtained. The maximum allowable discharge power of EVs in each time period is calculated by summing the maximum discharge power of the EV cluster in each time period. The dispatchable energy of each EV is evenly distributed to each time period according to its effective parking time. The distributed energy of all EVs in dispatchable state in each time period is added up to obtain the available energy of the EV cluster in each time period.

3. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 1, characterized in that, The formula for the combined photovoltaic power output across multiple scenarios is as follows: ； in, Representing a scene During the period Photovoltaic power output; Indicates the time period The benchmark photovoltaic power output curve; Representing a scene During the period The random perturbation factor; , These represent the start and end times of the effective photovoltaic power output period, respectively. This indicates taking the maximum value.

4. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 1, characterized in that, The formula for the multi-scenario stochastic optimization scheduling model is: The objective function is expressed as: ； The constraints are expressed as follows: ； ； ； ； ； ； ； ； ； in, Represents minimizing the objective function ; Representing a scene Quantity; Representing a scene The probability of; Indicates the number of time periods; , These represent the penalty cost coefficients for wasted light and loss of load, respectively. Representing a scene During the period The power of abandoned light; Representing a scene During the period The power of the load shedding; , These represent the unit charge / discharge operating cost coefficients for energy storage and EV clusters, respectively. , These represent the time periods. Time period Energy storage charging power; , These represent the time periods. Time period The energy storage discharge power; , These represent the time periods. Time period EV charging power and EV discharging power; This indicates the number of periods of high photovoltaic output; This represents the penalty cost coefficient for insufficient charging of EV clusters. Indicates the time period Non-negative variables; Representing a scene During the period The actual output of photovoltaic power; Indicates the time period Net energy storage output; Indicates the time period EV cluster net output; Representing a scene During the period The actual load value; Representing a scene During the period Photovoltaic power output; Representing a scene During the period The load output; , These represent the time periods. Time period Energy storage capacity; , These represent the energy storage charging efficiency coefficient and the energy storage discharging efficiency coefficient, respectively. Indicates the time step; , These represent the minimum and maximum energy levels for energy storage, respectively. Indicates the rated energy of the energy storage; Indicates the time period Binary variables used to ensure mutual exclusion of energy storage charging and discharging states; Indicates the power margin factor; , These represent the time periods. The maximum charging power and maximum discharging power of the energy storage; , These represent the initial moments of the system's operating cycle. Energy storage capacity, system operation cycle end time Energy storage capacity; Indicates the time period Binary variables used to ensure mutual exclusion of EV charging and discharging states; , These represent the time periods. The maximum charging power and maximum discharging power of the EV; Indicates the percentage of safety discount; Indicates the time period The total energy can be estimated; This represents the minimum excitation power.

5. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 1, characterized in that, The operating cycle of the islanded microgrid is divided into several determination intervals, including: Using a fixed duration as the judgment period, the operating cycle of the islanded microgrid is divided into... There are several decision intervals, and each decision interval is represented as follows: ； The EV cluster energy aggregation model is represented as follows: ； in, Indicates the decision interval ; , These represent the start and end times of the decision interval, respectively. Indicates the number of decision intervals; , They represent the first EVs during the period The state of charge, in the initial period The state of charge; Indicates the first The rated capacity of the battery of an EV; , These represent the energy storage charging efficiency coefficient and the energy storage discharging efficiency coefficient, respectively. , They represent the first EVs during the period The charging power and discharging power.

6. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 5, characterized in that, Based on the pre-built EV cluster energy aggregation model and minimum SOC setting, schedulable groups are selected from the EV clusters, including: For each decision interval, calculate the minimum SOC setting value required at the end of the decision interval, expressed as: ； Based on the pre-constructed EV cluster energy aggregation model and the minimum SOC setting, the EV state parameters are calculated and expressed as follows: ； Will The EV collection will be used as an off-grid group. The EV group will be used as a charging support group. The set of EVs is used as a schedulable group; in, Indicates the first EVs in the judgment range Minimum SOC setting; Indicates the first The State of Charge (SOC) of an EV when it is expected to leave the grid; Indicates the first The expected time for an EV to leave the power grid; Indicates the first The rated charging power of an EV; Indicates the first EVs in the judgment range EV state parameters; Indicates the first EVs during the period The grid connection status.

7. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 1, characterized in that, The VSG controller adjusts the power of the islanded microgrid based on the power reference value, using a trained time-series prediction model and a pre-built VSG active frequency regulation model, including: Input the dataset containing virtual inertia and damping coefficients into the trained time series prediction model, and output the adjusted virtual inertia and damping coefficients. The adjusted virtual inertia and damping coefficient are input into the pre-built VSG active frequency regulation model to output the power of the islanded microgrid. The VSG active frequency modulation model is represented as follows: ； ； ； in, Represents virtual inertia; , These represent the output angular frequency and reference angular frequency of the VSG controller, respectively. , These represent the mechanical power and electromagnetic power of the VSG controller, respectively. Indicates the damping coefficient; This indicates the active power command value of the VSG controller; This represents the active frequency droop coefficient of the VSG controller; This represents the Laplace operator.

8. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 1, characterized in that, The EV cluster selects EVs from the dispatchable group based on power differences, and allocates the power that the islanded microgrid needs to compensate for to the EVs according to the discharge allocation principle, including: When an isolated microgrid is disturbed, obtain the discharge power indication value that the isolated microgrid requires the EV cluster to provide; Calculate the number of EVs that need to be dispatched based on the discharge power indication value and the maximum discharge power of a single EV. ; Select the group with the largest discharge margin from the schedulable group. When an EV participates in the discharge, the power that the islanded microgrid needs to compensate for is allocated to the EVs participating in the discharge according to the discharge allocation principle.

9. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 8, characterized in that, The discharge allocation principle is expressed as follows: ； ； in, Indicates the first EVs during the period The discharge power; Indicates the first EVs at all times Discharge margin; Indicates the time period The discharge power indication value; Indicates the number of EVs; Indicates the first EVs during the period The state of charge; Indicates the first EVs in the judgment range The minimum SOC setting value.

10. The multi-element coordinated frequency regulation method for islanded microgrids based on EV clusters and VSGs according to claim 1, characterized in that, Also includes: When the islanded microgrid is undisturbed, EVs connected to the islanded microgrid charge at their rated power, while EVs not connected to the microgrid charge at 0 power, as shown below: ； in, Indicates the first EVs during the period The discharge power; Indicates the first The rated charging power of an EV; Indicates the first EVs in the judgment range EV state parameters.

Images