Multi-agent based voltage stability control method for ac microgrid

Abstract

The application discloses a Multi-Agent-based alternating current micro-grid voltage stability control method and relates to the technical field of alternating current micro-grid operation control, which comprises the following steps: dividing a light storage and a load into different Agents, calculating voltage support demand and a transient executable support package containing multiple constraints and version identification after data collection, obtaining a candidate support share through projection limiting consistency update, generating executable instructions through secondary projection, deducting external disturbance voltage by relying on electrical distance weighting, solving multi-channel causal effective support quantity, combining double confidence coefficient to calculate a physical consistency effective factor to modify the support share, completing channel secondary redistribution through a band leak and anti-saturation account, and finally synchronously updating VSG control parameters, distinguishing large and small disturbances to execute hierarchical support and fault slope limiting exit strategy. The application solves the problems of voltage disturbance misattribution and demand accumulation overcompensation, and improves the transient and steady-state voltage stability of the micro-grid.

Description

Technical Field

[0001] This invention relates to the field of AC microgrid operation control technology, and more specifically to a multi-agent-based AC microgrid voltage stability control method. Background Technology

[0002] Existing technologies such as multi-agent voltage control, energy storage control based on virtual synchronous motors, and hybrid energy storage control of lithium batteries / supercapacitors have been widely disclosed. In particular, physical information multi-agent voltage control, VSG-type hybrid energy storage distributed control, and resilient secondary control under FDI attacks can already cover ordinary multi-agent coordination, VSG voltage compensation, and SOC protection.

[0003] However, in actual AC microgrids, voltage changes within a given measurement window are often simultaneously affected by local agent support, adjacent agent support, load surges, and voltage fluctuations from the upstream grid. Directly using local voltage changes as the supporting effect can easily lead to misattribution of external disturbances.

[0004] Furthermore, when a channel remains unavailable, the instantaneous remaining demand may accumulate repeatedly at the coordination layer, leading to overcompensation after the subsequent available channel is restored. Additionally, when the external disturbance estimation sample is insufficient or the sample dispersion is too high, the physical consistency effectiveness factor will also be contaminated. Therefore, it is necessary to further constrain the external disturbance estimation confidence, anti-saturation ledger, and freezing logic.

[0005] Therefore, how to propose a voltage stability control method for AC microgrids based on Multi-Agent and overcome the shortcomings of existing technologies is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] In view of this, the present invention provides a voltage stability control method for AC microgrids based on Multi-Agent. The photovoltaic-storage system acts as the photovoltaic-storage agent, employing virtual synchronous motor control. The energy storage system consists of supercapacitors and lithium batteries, and adopts a charge-discharge protection strategy based on SOC segmented optimization control. The photovoltaic-storage agent generates a transient executable support package containing command event number, measurement window number, validity period, topology event number, constraint activation flag, SOC protection status, sensitivity model version number, and external disturbance estimation version number. Candidate support commands are projected into this support package and bound to command events. After execution, the effective support quantity is obtained by estimating by electrical distance weight and subtracting the external disturbance voltage component. The effective factor is then calculated by combining the sensitivity confidence coefficient, external disturbance estimation confidence coefficient, and physical consistency residual. The coordination process uses this effective factor to correct the support share and performs secondary redistribution by channel using a residual demand ledger with a discharge coefficient and anti-saturation term. This improves the voltage stability of the AC microgrid under small and large disturbances and suppresses misattribution, SOC boundary impacts, and redistribution drift. To achieve the above objectives, the present invention adopts the following technical solution: A multi-agent-based AC microgrid voltage stability control method includes: Divide the photovoltaic storage agent into optical storage agents and equipment agents, and collect data from the optical storage agents; Based on the collected data, calculate the AC voltage support requirements and transient executable support package for the photovoltaic-storage agent; Based on the consistency update of AC voltage support requirements and transient executable support package execution projection limit, the AC voltage candidate support share is obtained. The candidate support instructions are projected into the transient executable support package to obtain the executable support instructions; Calculate the causal effective support quantity decomposed by channel within the dead zone delay and measurement window corresponding to the command event number; The physical consistency effectiveness factor is calculated based on the executable support instructions and the effective support quantity, and the AC voltage candidate support share for the next coordination cycle is corrected based on the physical consistency effectiveness factor, and a secondary reallocation is performed. The parameter values ​​in the virtual synchronous motor control are updated synchronously based on the reassigned executable support instructions.

[0007] Optionally, the AC voltage support requirements for the computational photovoltaic energy storage agent include: ; in, This is a fault indicator. For inverter current limiting, , , and Non-negative weights This means restricting x to [0,1]. Per-unit voltage deviation, This represents the per-unit voltage change rate.

[0008] Optionally, the transient executable support package includes: ; in, This represents the upper limit of positive-sequence reactive power support. The upper limit is supported by the negative sequence current. This is the upper limit of energy storage discharge. This represents the upper limit of charging power, where i is the Agent number and x represents the lithium battery (bat) or supercapacitor (s). Assign a command event number, Number the topology event. This is the sensitivity model version number. To estimate the version number for external disturbances, Bind identifier to command-model-measurement. These are the support package validity period, constraint activation state vector, SOC protection state vector, supercapacitor maximum discharge power, supercapacitor maximum charging power, lithium battery maximum discharge power, and lithium battery maximum charging power, respectively.

[0009] Optionally, the consistency update based on AC voltage support requirements and transient executable support package projection limiting to obtain AC voltage candidate support share includes: Consistent updates include: ; The candidate support share is:

[0010] in, Indicates projection limiting. Let i be the set of its neighbors. Here, α, β, η, and γ represent the communication adjacency weights, and α, β, η, and γ represent the coordination gain. To normalize voltage support capability, To prevent small positive numbers with a denominator of zero, For the first l The coordinating agent in the first k Consistency variables for each coordination cycle.

[0011] Optionally, the step of projecting candidate support instructions into the transient executable support package to obtain executable support instructions includes: The candidate support instructions are: ; The executable support commands after projection are: ; in, Candidate support instructions, The executable support instructions are defined by W, which is a positive definite weight matrix. These are the total positive-sequence reactive power demand, the total negative-sequence current demand, the total high-frequency active power support demand of supercapacitors, and the total low-frequency active power support demand of lithium batteries.

[0012] Optionally, the step of calculating the causal effective support quantity decomposed by channel within the dead-time delay and measurement window corresponding to the command event number includes: Based on the electrical distance weighting, the external disturbance voltage components in unparticipated or low-support nodes are estimated and subtracted to obtain the causal effective support quantity decomposed by channel: The electrical distance weight is: ; The external disturbance voltage component is: ; The causal voltage change after deducting external disturbances is: ; The effective causal support of the channel is: ; in, Let i be the electrical distance between nodes i and j. This is the distance attenuation coefficient. Let be the per-unit voltage of node j. Used to unify the sign direction of boost support and depressurization absorption. To measure noise margin, 'c' represents positive-sequence reactive power, negative-sequence current, supercapacitor, or lithium battery channel. To map causal voltage changes and channel execution quantities into functions of effective support quantities, For channel command requirements, Agent i The actual positive-sequence reactive power change, negative-sequence current change, supercapacitor power change, and lithium battery power change performed within the measurement window.

[0013] Optionally, the step of calculating the physical consistency effectiveness factor based on the executable support instructions and the causal effective support quantity includes: ; in, This is the sensitivity confidence coefficient. To estimate the confidence coefficient for external disturbances, The event number is the current network topology event number. This is the residual decay threshold. To support the validity period of the package, For a set of blocking constraints, To prevent division by zero from regularizing small positive numbers, This serves as the support and acceptance point for the current coordination cycle. This is an indicator function; it takes the value 1 if the condition is true, and 0 otherwise.

[0014] Optionally, the step of correcting the AC voltage candidate support share for the next coordination cycle with a physical consistency effectiveness factor and performing a secondary reallocation includes: The weights for the secondary redistribution are:

[0015] The increment for the second redistribution is:

[0016] in, This is a channel qualification gate, where 'c' is the channel number. For channel requirements, For the remaining demand ledger of the channel, This indicates that the channel is available. The leakage coefficient of the ledger. This is a function that calculates channel availability based on remaining margin. To prevent small positive numbers with a denominator of zero.

[0017] Optionally, the step of synchronously updating the parameter values ​​in the virtual synchronous motor control according to the redistributed executable support instructions includes: superimposing the obtained redistribution increment onto the executable support instructions, and updating the VSG voltage / reactive power reference, virtual damping, excitation support gain, and hybrid energy storage power limit.

[0018] Optional features include: prioritizing reactive current under heavy disturbance conditions, current circular limiting, high-frequency rapid support for supercapacitors, low-frequency power limiting support for lithium batteries, and exiting without drift slope limiting after a fault.

[0019] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a multi-agent-based AC microgrid voltage stability control method, which has the following beneficial effects: This invention provides a voltage stability control method for AC microgrids based on Multi-Agent. A photovoltaic-storage system (PV-SGC) is connected to the AC microgrid as a PV-SGC Agent and exchanges status messages in parallel with other device Agents. Internally, the PV-SGC Agent utilizes virtual synchronous motor control, lithium battery / supercapacitor hybrid energy storage power allocation, SOC segmented protection, and transient executable support package projection to collectively form voltage support capabilities.

[0020] Compared to ordinary multi-agent voltage control, this invention binds command events, topology events, sensitivity model versions, and external disturbance estimation versions in a unified manner, so that each support acceptance has a clear causal correspondence.

[0021] This invention uses an electrical distance weighting method to estimate the external disturbance voltage component and forms the external disturbance estimation confidence coefficient through the sample size and sample variance. When the external disturbance estimation confidence is insufficient, the external disturbance estimation update is frozen and the corresponding effective factor is reduced to avoid misjudging load disturbances or other agent actions as contributions of this agent.

[0022] This invention also designs the remaining support requirements as a ledger with an anti-saturation term. Hysteresis switching and limited slope exit are used between small disturbance, large disturbance and post-fault recovery states; when a channel does not have enough available agents, the ledger no longer accumulates indefinitely, but is discharged through the anti-saturation term and transferred to other channels, throttling or controllable load reduction processes, thereby suppressing redistribution drift and overcompensation. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0024] Figure 1 A schematic diagram of the process for the voltage stabilization control method of AC microgrid based on Multi-Agent provided by the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] This invention discloses a multi-agent-based AC microgrid voltage stability control method, such as... Figure 1 As shown, it includes: Divide the photovoltaic storage agent into optical storage agents and equipment agents, and collect data from the optical storage agents; Based on the collected data, calculate the AC voltage support requirements and transient executable support package for the photovoltaic-storage agent; Based on the consistency update of AC voltage support requirements and transient executable support package execution projection limit, the AC voltage candidate support share is obtained. The candidate support instructions are projected into the transient executable support package to obtain the executable support instructions; Calculate the causal effective support quantity decomposed by channel within the dead zone delay and measurement window corresponding to the command event number; The physical consistency effectiveness factor is calculated based on the executable support instructions and the effective support quantity, and the AC voltage candidate support share for the next coordination cycle is corrected based on the physical consistency effectiveness factor, and a secondary reallocation is performed. The parameter values ​​in the virtual synchronous motor control are updated synchronously based on the reassigned executable support instructions.

[0027] In a specific implementation, the multi-agent-based AC microgrid voltage stability control method is applied to an AC microgrid that includes an AC bus, grid connection point, load, distributed power source, and at least one photovoltaic-storage system. The specific steps are as follows: S1, classify the photovoltaic-storage system as a photovoltaic-storage agent, and classify the load, grid connection point, grid interface or other distributed power source as a device agent; S2, the photovoltaic-storage agent collects AC voltage, current, active power, reactive power, frequency, positive / negative sequence voltage components, positive / negative sequence current components, DC bus voltage, inverter current, lithium battery SOC and supercapacitor SOC according to the local fast control cycle. S3, calculate the AC voltage support requirement based on the per-unit voltage deviation, per-unit voltage change rate, fault signs and inverter current limiting signs at the grid connection point or AC bus. Specifically, the calculation of AC voltage support requirements includes: The inputs to S3 are the actual voltage, rated voltage, sampling period, fault flag, and inverter current limiting flag at the grid connection point or AC bus. First, the per-unit voltage deviation and per-unit voltage change rate are calculated, then a normalized voltage support requirement is formed.

[0028] The per-unit voltage deviation is: ; The per-unit voltage change rate is: ; AC voltage support requirements are: ; in, This represents the effective value of the grid connection point or AC bus voltage at the k-th sampling time. Rated voltage, The sampling period is This is a fault indicator. For inverter current limiting, , , and Non-negative weights This means that x is restricted to [0,1]. As a requirement input for S5 consistency updates and S9 remaining requirement ledger.

[0029] S4, the optical storage agent generates a transient executable support package with command event number, measurement window number, validity period, topology event number, constraint activation flag, SOC protection status, sensitivity model version number and external disturbance estimation version number; Specifically, a transient executable support package is generated based on the inverter current circle, DC bus power margin, SOC segment protection factor, and DC / DC rated power, including: The transient executable support package is: ; The upper limit of positive sequence reactive power support is: ; The upper limit of negative sequence current support is: ; The upper limits for energy storage discharge and charging power are as follows: ; ; Where i is the Agent number, and x represents the lithium battery (bat) or supercapacitor (sc). This is the upper limit of reactive power capacity. This represents the current reactive power. This is the positive sequence voltage amplitude. This refers to the inverter current limit. and This is the positive sequence dq-axis current. This represents the negative sequence current amplitude. To allow an upper limit for negative sequence current, and Rated discharge and charge power for energy storage. and This refers to the available discharge and charge power margin for the DC bus. This corresponds to the rated power of the DC / DC converter. The upper limits in the equation are used as S6 projection constraints.

[0030] The SOC discharge protection factor is: ; , ; ; The SOC charging protection factor is: ; ; ; The continuous transition function is:

[0031] in, Let x be the state of charge of the energy storage. The threshold for prohibiting discharge, To fully allow the discharge threshold, To fully allow charging threshold, The charging threshold is set to prevent charging from being stopped. The charging and discharging thresholds use different hysteresis settings for the entry and exit thresholds to avoid frequent switching at the SOC boundary.

[0032] S5, each Agent exchanges status messages containing voltage support requirements and transient executable support packages, and performs a consistent update of projection limiting to obtain the AC voltage candidate support share; Specifically, it receives reports from each Agent. and First update the consistency variables Then, candidate support shares are obtained based on feasibility and timeliness weights, including: Consistency variables are updated as follows: ; The voltage support capability indicators are: ; The candidate support share is: ; Where Proj represents projection limiting. For Agenti's neighbor set, Here, α, β, η, and γ represent the communication adjacency weights, and α, β, η, and γ represent the coordination gain. To normalize voltage support capability, , , and As a normalized baseline value, , , and Non-negative weights To prevent small positive numbers with a denominator of zero. As the basis for allocating S6 candidate support instructions. These are the lower and upper bounds of the projection of consistency variables, respectively. , These are the upper limits for positive-sequence reactive power support and negative-sequence current support generated by S4, respectively.

[0033] S6, the optical storage agent projects the candidate support instructions into the transient executable support package to obtain the executable support instructions, and binds the executable support instructions with the command event number, topology event number, sensitivity model version number and external disturbance estimation version number; Specifically, binding executable support instructions with command event numbers, topology event numbers, sensitivity model version numbers, and external disturbance estimation version numbers means that a unique binding identifier is generated for each instruction after projection. Subsequent S7 and S8 can only use the same Acceptance testing should be conducted using the corresponding measurement window data to avoid mixing data from different commands, topologies, or model versions, including: The candidate support instructions are: ; The executable support commands after projection are: ; The command-model-measurement binding identifier is: ; in, Candidate support instructions, The executable support instructions are defined by W, which is a positive definite weight matrix. Assign a command event number, Number the topology event. This is the sensitivity model version number. To estimate the version number for external disturbances, The time when the instruction was issued. As the starting point of the measurement window, To measure the length of the window. Enter S7 for execution and acceptance.

[0034] S7, within the dead time and measurement window corresponding to the command event number, estimate and subtract the external disturbance voltage component in the unparticipated support node or low support node according to the electrical distance weight, and obtain the causal effective support quantity decomposed by channel. Specifically, the first set of nodes that have never participated in this support or whose support amount is below the threshold. Estimate the external disturbance voltage component, then subtract this external disturbance component from the voltage change measured by this Agent to obtain the causal voltage change attributable to the current executable support command, including: The electrical distance weight is: ; The external disturbance voltage component is: ; The causal voltage change after deducting external disturbances is: ; The effective causal support of the channel is: ; in, Let i be the electrical distance between nodes i and j. This is the distance attenuation coefficient. Let be the per-unit voltage of node j. Used to unify the sign direction of boost support and depressurization absorption. To measure noise margin, 'c' represents positive-sequence reactive power, negative-sequence current, supercapacitor, or lithium battery channel. To map causal voltage changes and channel execution quantities into functions of effective support quantities, This is for the channel's instruction requirements. Output. Enter the S8 effective factor and S9 ledger redistribution.

[0035] S8 calculates the physical consistency validity factor based on the executable support instruction, causal effective support quantity, local voltage sensitivity matrix, sensitivity confidence coefficient, external disturbance estimation confidence coefficient, validity period, topology event number and constraint activation flag, and freezes the corresponding update when the sensitivity confidence is insufficient, the external disturbance estimation confidence is insufficient or the topology changes. Specifically, the causal effective support quantity obtained from S7 is compared with the predicted value of the local voltage sensitivity model, while simultaneously considering sensitivity model confidence, external disturbance estimation confidence, command validity period, topology matching, and constraint activation state, including: The confidence coefficient for external disturbance estimation is: ; The normalized physical residual is: ; The sensitivity confidence coefficient is: ; The sensitivity matrix is ​​recursively updated as follows: ; Negative contribution gating is: ; The physical consistency effectiveness factor is: ; in, To estimate the confidence coefficient for external disturbances, To allow an upper limit on the variance of samples with external disturbances, To minimize the number of external disturbance samples, For version The corresponding local voltage sensitivity matrix, Depend on The components consist of reactive power, negative sequence current, supercapacitor power, and lithium battery power changes. To estimate the covariance matrix for sensitivity, For the covariance threshold, As an effective incentive indicator, The event number is the current network topology event number. This is the residual decay threshold. To support the validity period of the package, For a set of blocking constraints, The threshold for negative contribution. When or Below the threshold, topology event change, fault flag appears or If the conditions are not met, the account will be frozen. or Update and reduce . Proceed to S9 share adjustment and ledger calculation.

[0036] S9, the coordination layer corrects the support share of the next coordination cycle with the physical consistency effectiveness factor, and uses the remaining demand ledger with discharge coefficient and anti-saturation term to carry out secondary redistribution according to the positive sequence reactive power, negative sequence current, supercapacitor high-frequency power and lithium battery low-frequency power channels. Specifically, based on S8 Revise the support share for the next coordination cycle and establish a ledger with venting and anti-saturation items for unmet support needs.

[0037] The effective adjusted support share is: ; The available channel indicator is: ; The remaining demand log is as follows: ; The access qualification gate is as follows: ; The weights for the secondary redistribution are: ; The increment for the second redistribution is: ; Where c is the channel number. For channel requirements, For the remaining demand ledger of the channel, The leakage coefficient of the ledger. The inverse saturation coefficient, This is the upper limit of the ledger. The threshold for available channels. This is the channel constraint activation flag. This is the SOC direction permission flag. This represents the minimum margin that can be added to the channel. This is a function that calculates channel availability based on remaining margin. To prevent small positive numbers with a denominator of zero. When When this happens, no more commands will be added to unavailable channels. Instead, the unsaturation item will release the data from the ledger and transfer it to other channels, or the global load will be reduced or the controllable load will be cut. Enter S10 to execute.

[0038] S10, the photovoltaic-storage agent synchronously updates the voltage reference, reactive power reference, excitation support gain, virtual damping, and hybrid energy storage active power limit in the virtual synchronous motor control according to the redistributed executable support instructions, and performs reactive current priority, current circle limit, supercapacitor high-frequency fast support, lithium battery low-frequency power limit support, and drift-free slope limit exit after a fault under large disturbance conditions.

[0039] Specifically, the redistribution increment obtained from S9 is superimposed on the executable support command to update the VSG voltage / reactive power reference, virtual damping, excitation support gain, and hybrid energy storage power limit.

[0040] The VSG active-frequency loop is: ; The virtual excitation voltage is: ; The initial frequency division allocation for hybrid energy storage is as follows: ; The reactive power priority current circular limit is: ; ; After the fault, the slope is limited to exit as follows: ; The goal of drift-free recovery is: ; in, For virtual inertia, VSG angular frequency, The rated angular frequency, and For reference and actual measurement purposes, For virtual damping, For virtual internal potential, and For reactive power and voltage support gain, and For reference of per-unit reactive power and actual measured reactive power, and For per-unit voltage reference and measured voltage, For DC-side power support, LPF stands for low-pass filter. and These represent the low-frequency power of lithium batteries and the high-frequency power of supercapacitors, respectively. and For dq axis current reference, sat indicates saturation limiting. It can be any of the following: virtual damping, excitation support gain, voltage reference, reactive power reference, or hybrid energy storage power limiting. This is the baseline value for normal operation. To retain values ​​during the fault period To restore the cumulative time after the voltage dead zone, For the regression time constant, The slope is limited.

[0041] Furthermore, under conditions of minimal disruption, reduce the remaining demand ledger. The updated weights enable each photovoltaic storage agent to perform VSG voltage closed-loop and reactive power fine-tuning according to the local fast control cycle; Under conditions of significant disruption, improve the remaining demand ledger. The update weights prioritize scheduling the positive-sequence reactive power channel and the supercapacitor high-frequency power channel, and reduce the low-frequency active power support command of the lithium battery when the inverter current approaches the limit.

[0042] Furthermore, when communication is interrupted, communication delay exceeds the threshold, timestamp is abnormal, command event number is discontinuous, measurement window number is mismatched, topology event number is mismatched, or external disturbance estimation confidence coefficient is lower than the threshold, the corresponding Agent stops participating in consistency updates, maintains the most recently valid executable support command, and performs local autonomous control according to local voltage deviation, SOC protection status, and current circle constraint; after communication is restored, command events are renumbered, status is synchronized, projection limiting and slope transition are performed first, and then the distributed coordination process is restored.

[0043] In a specific embodiment, the local fast control cycle of the optical storage agent is 0.1ms to 2ms, and the coordination communication cycle is 5ms to 100ms. Each optical storage agent exchanges voltage, current, SOC, constraint activation flag, and support package information with neighboring device agents; each executable support command carries a command event number. The coordination process only involves... Topology event number Sensitivity model version number External disturbance estimation version number The effectiveness of the support was verified within the matched measurement window.

[0044] The transient executable support package is: ; The external disturbance voltage component consists of the set of nodes that did not participate in this support or whose support amount is below the threshold. The given values, weighted by electrical distance attenuation, are more suitable for radial or multi-feeder AC microgrids than the simple median estimate.

[0045] The external disturbance voltage component is: ; like If the sample size is insufficient or the sample variance is too large, the confidence coefficient for external disturbance estimation will be affected. As the physical consistency factor decreases, the sensitivity matrix update is frozen in the event of topology changes, large disturbances due to faults, or insufficient effective excitation.

[0046] The confidence coefficient for external disturbance estimation is: ; The physical consistency effectiveness factor is: ; For remaining support demand, this invention employs a ledger with an anti-saturation term. If the total available capacity of a certain channel falls below a threshold, the ledger releases this demand through the anti-saturation term, preventing the long-term accumulation of demand for unavailable channels. After the fault is cleared and the voltage enters the dead zone, the virtual damping, excitation support gain, voltage reference, reactive power reference, and hybrid energy storage power limiting are released towards the normal operating baseline at a limited slope.

[0047] The residual demand ledger with anti-saturation terms is as follows: .

[0048] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0049] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A voltage stability control method for AC microgrids based on Multi-Agent, characterized in that, include: Divide the photovoltaic storage agent into optical storage agents and equipment agents, and collect data from the optical storage agents; Based on the collected data, calculate the AC voltage support requirements and transient executable support package for the photovoltaic-storage agent; Based on the consistency update of AC voltage support requirements and transient executable support package execution projection limit, the AC voltage candidate support share is obtained. The consistency update based on AC voltage support requirements and transient executable support package projection limiting to obtain AC voltage candidate support share includes: Consistent updates include: , To meet the requirements for AC voltage support; The candidate support share is: ; in, Indicates projection limiting. Let i be the set of its neighbors. Here, α, β, η, and γ represent the communication adjacency weights, and α, β, η, and γ represent the coordination gain. To normalize voltage support capability, To prevent small positive numbers with a denominator of zero, For the first l The coordinating agent in the first k Consistency variables for each coordination cycle; The candidate support instructions are projected into the transient executable support package to obtain the executable support instructions; The step of projecting candidate support instructions into the transient executable support package to obtain executable support instructions includes: The candidate support instructions are: ; The executable support commands after projection are: ; in, Candidate support instructions, The executable support instructions are defined by W, which is a positive definite weight matrix. These are the total positive-sequence reactive power demand, the total negative-sequence current demand, the total high-frequency active power support demand of supercapacitors, and the total low-frequency active power support demand of lithium batteries, respectively. Calculate the causal effective support quantity decomposed by channel within the dead zone delay and measurement window corresponding to the command event number; The calculation of the causal effective support quantity decomposed by channel within the dead-zone delay and measurement window corresponding to the command event number includes: Based on the electrical distance weighting, the external disturbance voltage components in unparticipated or low-support nodes are estimated and subtracted to obtain the causal effective support quantity decomposed by channel: The electrical distance weight is: ; The external disturbance voltage component is: ; The causal voltage change after deducting external disturbances is: ; The effective causal support of the channel is: ； in, Let i be the electrical distance between nodes i and j. This is the distance attenuation coefficient. Let be the per-unit voltage of node j. Used to unify the sign direction of boost support and depressurization absorption. To measure noise margin, 'c' represents positive-sequence reactive power, negative-sequence current, supercapacitor, or lithium battery channel. To map causal voltage changes and channel execution quantities into functions of effective support quantities, For channel command requirements, Agent i The actual positive-sequence reactive power change, negative-sequence current change, supercapacitor power change, and lithium battery power change performed within the measurement window; The physical consistency effectiveness factor is calculated based on the executable support instructions and the effective support quantity, and the AC voltage candidate support share for the next coordination cycle is corrected based on the physical consistency effectiveness factor, and a secondary reallocation is performed. The parameter values ​​in the virtual synchronous motor control are updated synchronously based on the reassigned executable support instructions.

2. The AC microgrid voltage stability control method based on Multi-Agent according to claim 1, characterized in that, The AC voltage support requirements for the computational photovoltaic energy storage agent include: ； in, This is a fault indicator. For inverter current limiting, , , and Non-negative weights This means restricting x to [0,1]. Per-unit voltage deviation, This represents the per-unit voltage change rate.

3. The AC microgrid voltage stability control method based on Multi-Agent according to claim 1, characterized in that, The transient executable support package includes: ； in, This represents the upper limit of positive-sequence reactive power support. The upper limit is supported by the negative sequence current. This is the upper limit of energy storage discharge. This represents the upper limit of charging power, where i is the Agent number and x represents the lithium battery (bat) or supercapacitor (s). Assign a command event number, Number the topology event. This is the sensitivity model version number. To estimate the version number for external disturbances, Bind identifier to command-model-measurement. , These are the support package validity period, constraint activation state vector, SOC protection state vector, supercapacitor maximum discharge power, supercapacitor maximum charging power, lithium battery maximum discharge power, and lithium battery maximum charging power, respectively.

4. The AC microgrid voltage stability control method based on Multi-Agent according to claim 1, characterized in that, The calculation of the physical consistency effectiveness factor based on executable support instructions and causal effective support quantity includes: ； in, This is the sensitivity confidence coefficient. To estimate the confidence coefficient for external disturbances, The event number is the current network topology event number. This is the residual decay threshold. To support the validity period of the package, For a set of blocking constraints, To prevent division by zero from regularizing small positive numbers, This is the support and acceptance time for the current coordination cycle. This is an indicator function.

5. The AC microgrid voltage stability control method based on Multi-Agent according to claim 1, characterized in that, The process of correcting the AC voltage candidate support share for the next coordination cycle using a physical consistency effectiveness factor and performing a secondary reallocation includes: The weights for the secondary redistribution are: The increment for the second redistribution is: in, This is a channel qualification gate, where 'c' is the channel number. For channel requirements, For the remaining demand ledger of the channel, This indicates that the channel is available. The leakage coefficient of the ledger. This is a function that calculates channel availability based on remaining margin. To prevent small positive numbers with a denominator of zero.

6. The AC microgrid voltage stability control method based on Multi-Agent according to claim 1, characterized in that, The step of synchronously updating the parameter values ​​in the virtual synchronous motor control according to the redistributed executable support instructions includes: superimposing the obtained redistribution increment onto the executable support instructions, and updating the VSG voltage / reactive power reference, virtual damping, excitation support gain, and hybrid energy storage power limit.

7. The AC microgrid voltage stability control method based on Multi-Agent according to claim 1, characterized in that, It also includes: reactive current priority under large disturbance conditions, current circular limiting, high-frequency fast support for supercapacitors, low-frequency power limiting support for lithium batteries, and drift-free slope exit after a fault.

Images