Weak network under the network energy storage frequency adaptive adjustment method and system

Abstract

The application relates to the technical field of frequency regulation, in particular to a weak-network-construction energy storage frequency self-adaptive regulation method and system, and the following scheme is provided: a node state diagram reflecting device state and coupling relationship is constructed in an island operation stage, main network phase characteristics are extracted based on a common coupling point measurement signal when a grid-connected condition is met, a logical main network node carrying a phase corridor attribute is introduced, and controlled injection of a main network phase reference is realized. Further, a phase convergence diagram is formed according to traction compatibility judgment and diagram reconstruction mechanism, a traction node is subjected to competitive inhibition update through a consistency evaluation index, a grid-connected structure diagram is generated, and thus coordinated regulation of frequency parameters of multiple network-construction energy storage devices under a weak network condition is realized, and stability and controllability of a regrid process are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of frequency regulation, in particular to a method and system for adaptive frequency regulation of grid-forming energy storage under weak grid conditions. BACKGROUND

[0002] With the large-scale access of new energy and energy storage devices in distribution networks and microgrids, grid-forming energy storage is increasingly widely used in island operation and reintegration scenarios. In the prior art, grid-forming energy storage devices usually establish voltage and frequency references through local control strategies, and rely on pre-set droop characteristics or equivalent synchronous control to achieve power distribution and frequency support when multiple devices are operating in parallel. However, under weak grid conditions, the above technical solutions have limitations in the reintegration stage.

[0003] On the one hand, existing methods often introduce the main grid phase or frequency reference into grid-forming energy storage devices in the form of a direct signal, and each device adjusts the phase or frequency simultaneously after receiving the main grid reference, which can easily cause multiple sources to adjust in parallel in a short period of time, leading to uncontrollable phase convergence paths and large power fluctuations during reintegration. On the other hand, existing technologies generally lack a structured description of the collaborative behavior between multiple grid-forming devices, making it difficult to distinguish between the dominant and subordinate roles of different devices during reintegration, which can lead to traction conflicts or mutual cancellation, affecting reintegration stability.

[0004] To solve the above problems, the present application designs a method and system for adaptive frequency regulation of grid-forming energy storage under weak grid conditions. SUMMARY

[0005] The technical problem to be solved by the present application is to address the shortcomings of the prior art and provide a method and system for adaptive frequency regulation of grid-forming energy storage under weak grid conditions. During island operation, a node state diagram reflecting the state and coupling relationship of the devices is constructed, and the main grid phase characteristics are extracted based on the measurement signals of the common coupling point when the reintegration conditions are met. By introducing a logical main grid node carrying phase corridor attributes, controlled injection of the main grid phase reference is achieved. Further, a phase convergence diagram is formed according to the traction compatibility judgment and graph reconstruction mechanism, and a consistent evaluation index is used to implement competitive suppression and update of the traction nodes, generating a reintegration structure diagram, thereby achieving coordinated regulation of the frequency parameters of multiple grid-forming energy storage devices under weak grid conditions and improving the stability and controllability of the reintegration process.

[0006] To achieve the above purpose, the present application provides the following technical solutions:

[0007] A method for adaptive frequency regulation of grid-forming energy storage under weak grid conditions is applied to a distributed grid-forming energy storage system, which includes at least two grid-forming energy storage devices and a common coupling point connected in parallel with the grid-forming energy storage devices. The method comprises:

[0008] In the island operation phase, an island node state graph is constructed by obtaining a node state vector of each grid-forming energy storage device;

[0009] When the grid-connected trigger condition is met, a main grid phase feature is extracted based on a measurement signal of the common coupling point, a logical main grid node is added to the island node state graph, so that the logical main grid node serves as a phase reference of the island node state graph, wherein the adding of the logical main grid node includes determining a traction node corresponding to the logical main grid node, and obtaining a phase convergence graph;

[0010] According to the consistency evaluation index of the phase convergence graph, the traction node is updated to suppress competition, and a grid-connected structure graph is obtained, and the frequency parameters of each grid-forming energy storage device are adjusted according to the grid-connected structure graph.

[0011] The node state vector at least includes one of the following:

[0012] Phase state information for characterizing the internal phase reference of the corresponding grid-forming energy storage device;

[0013] Power state information for characterizing the active power output of the corresponding grid-forming energy storage device;

[0014] Reactive power state information for characterizing the reactive power output of the corresponding grid-forming energy storage device;

[0015] Margin state information for characterizing the operation margin of the corresponding grid-forming energy storage device, wherein the margin state information includes current limit trigger state, power limit trigger state, DC side voltage margin, and modulation saturation degree;

[0016] Parameter state information for characterizing the current value of the frequency parameter of the corresponding grid-forming energy storage device, wherein the frequency parameter includes droop coefficient, virtual inertia parameter and damping parameter.

[0017] The construction of the island node state graph includes:

[0018] Obtaining the node identifier of each grid-forming energy storage device and the node state vector corresponding to the node identifier, and establishing an island node set with each grid-forming energy storage device;

[0019] According to the access position of each grid-forming energy storage device, the coupling relationship between the island nodes is determined, and the island edge set is established with the coupling relationship as the edge;

[0020] Each edge in the island edge set is configured with edge attribute information, wherein the edge attribute information includes coupling weight representing the coupling strength between the island nodes, and the coupling weight is calculated according to the node state vector;

[0021] Generate the island node state graph based on the island node set, the island edge set, and the edge attribute information of each edge, and periodically update the edge attribute information of the island node state graph in the island running phase.

[0022] Extract the main grid phase feature based on the measurement signal of the common coupling point, including:

[0023] Obtain the voltage measurement signal and the current measurement signal at the common coupling point, wherein the voltage measurement signal includes three-phase voltage instantaneous values, phase voltages, and line voltages, and the current measurement signal includes three-phase current instantaneous values;

[0024] Perform signal preprocessing on the voltage measurement signal and the current measurement signal to obtain stable measurement signals for phase analysis, wherein the signal preprocessing includes filtering, fundamental component extraction, harmonic suppression, and abnormal mutation suppression;

[0025] Based on the stable measurement signals, extract the instantaneous phase information and the phase change trend information corresponding to the common coupling point as the phase candidate feature of the current main grid;

[0026] Obtain the historical operation data of the main grid corresponding to the common coupling point, wherein the historical operation data of the main grid at least includes at least one of historical phase information, historical frequency information, and historical voltage operation information;

[0027] Based on the historical operation data of the main grid, verify the phase candidate feature, and when the verification is adapted to the preset adaptation condition, determine that the phase candidate feature is the main grid phase feature, and if not, iteratively modify the phase candidate feature based on the historical operation data of the main grid until it is adapted to the adaptation condition.

[0028] Add a logical main grid node to the island node state graph, including:

[0029] According to the main grid phase feature, construct a phase corridor object, wherein the phase corridor object is used to represent the allowed range and validity state of the main grid phase reference, and the validity state is determined by the consistency verification result of the main grid phase feature and the historical operation data of the main grid;

[0030] Based on the phase corridor object and the node state vector of each grid-forming energy storage device, perform traction compatibility determination on each grid-forming energy storage device to obtain a traction compatible node set;

[0031] When the traction compatible node set is not empty, generate a logical main grid node in the island node state graph based on the phase corridor object, wherein the node attribute of the logical main grid node includes the phase corridor object;

[0032] generating a graph reconstruction plan based on the set of traction-compatible nodes, wherein the graph reconstruction plan is used to reconstruct the edge set and edge attribute information of the island node state graph;

[0033] reconstructing the island node state graph containing the logical main grid node according to the graph reconstruction plan, to obtain the phase convergence graph.

[0034] performing traction compatibility determination on each grid-forming energy storage device based on the phase corridor object and the node state vector of each grid-forming energy storage device, including:

[0035] transmitting the node state vector of each grid-forming energy storage device to the public coupling node, and extracting phase state information, margin state information and parameter state information through the public coupling node;

[0036] determining the relative position relationship between the internal phase reference of the grid-forming energy storage device to be determined and the phase corridor object based on the phase state information, wherein the relative position relationship includes two states of being within the allowed range of the phase corridor object and being outside the allowed range of the phase corridor object;

[0037] when the relative position relationship represents being within the allowed range of the phase corridor object, determining that the grid-forming energy storage device to be determined is a traction-compatible node;

[0038] when the relative position relationship represents being outside the allowed range of the phase corridor object, determining whether the grid-forming energy storage device to be determined satisfies the operating margin constraint condition based on the margin state information;

[0039] when the operating margin constraint condition is satisfied, determining a phase convergence capability parameter based on the parameter state information, and determining the reachability of the internal phase reference of the grid-forming energy storage device to be determined into the allowed range of the phase corridor object according to the phase convergence capability parameter, wherein the phase convergence capability parameter is positively correlated with the current value of the frequency parameter;

[0040] if it is determined that it is reachable within a preset time window, the grid-forming energy storage device to be determined is a traction-compatible node;

[0041] when the operating margin constraint condition is not satisfied and it is determined that it is not reachable within a preset time window, the grid-forming energy storage device to be determined is a non-traction-compatible node.

[0042] generating a graph reconstruction plan based on the set of traction-compatible nodes, including:

[0043] establish a traction edge between the logical main grid node and each grid-forming energy storage device node in the traction-compatible node set, and determine the traction-compatible node as a traction node;

[0044] obtain non-traction nodes connected to each traction node in the island node state graph, and determine an island edge between a traction node and the non-traction node to generate a to-be-reconstructed edge set;

[0045] configure updated edge attribute information for each edge in the to-be-reconstructed edge set to obtain a traction propagation edge set, wherein the updated edge attribute information is determined by calculating a phase convergence amount between a node state vector of a corresponding traction node and the main grid phase feature;

[0046] correct the edge attribute information of the traction propagation edge set based on the node state vector of the non-traction node to obtain a graph reconstruction plan.

[0047] perform competitive suppression update on the traction nodes according to the consistency evaluation index of the phase convergence graph to obtain a grid-connected structure graph, including:

[0048] obtain node state vectors corresponding to each traction node in the phase convergence graph, and determine the consistency evaluation index based on the node state vectors;

[0049] determine a competition state of each traction node based on the consistency evaluation index and a preset competition criterion, wherein the competition state includes a competition suppression triggering state and a competition suppression release state;

[0050] in the competition suppression triggering state, perform competitive suppression update on the traction node, and the competitive suppression update includes traction strength limitation on edge attribute information of a traction edge between the traction node and a logical main grid node, wherein the traction strength limitation includes lowering a traction strength upper limit, limiting a traction strength change rate, freezing traction strength update, and reverting traction strength;

[0051] update the edge set and edge attribute information of the phase convergence graph based on the traction node after competitive suppression update to generate the grid-connected structure graph.

[0052] the consistency evaluation index includes:

[0053] a phase convergence consistency index of a traction node relative to the logical main grid node, the phase convergence consistency index being used to represent a coordination degree of phase convergence of each traction node to the main grid phase feature;

[0054] a power consistency index between traction nodes, the power consistency index being used to represent a low-frequency swing trend and a same-direction consistency of active power output of each traction node.

[0055] The weak network under the network energy storage frequency adaptive adjustment system comprises:

[0056] A data acquisition module is configured to acquire a node state vector of each network energy storage device and construct an island node state graph based on the node state vector in an island operation stage;

[0057] A main network phase feature extraction module is configured to extract a main network phase feature based on a measurement signal of a common coupling point and perform consistency verification in combination with historical operation data of the main network to construct a phase corridor object;

[0058] A graph reconstruction module is configured to perform traction compatibility determination based on the phase corridor object and the node state vector to obtain a traction compatible node set and obtain a phase convergence graph;

[0059] A frequency parameter adjustment module is configured to perform competitive inhibition update on a traction node based on a consistency evaluation index of the phase convergence graph to obtain a grid-connected structure graph and adaptively adjust frequency parameters of each network energy storage device according to the grid-connected structure graph.

[0060] Compared with the prior art, the application has the following beneficial effects:

[0061] The application unifies the description of the phase and frequency adjustment behavior of the distributed network energy storage device in the island operation and re-grid connection process by introducing a graph structured modeling method, and realizes the constraint and management of the reference introduction process by taking a logical main network node to bear the main network phase reference. Through traction compatibility determination and graph reconstruction mechanism, the main network phase reference is propagated along the controlled path, avoiding the phase pulling and power conflict caused by the simultaneous strong traction of multiple devices. At the same time, combined with consistency evaluation and competitive inhibition update, the traction nodes form a coordinated division of labor in the grid connection process, effectively suppressing the frequency fluctuation and circulating current risk under the weak network condition, thereby ensuring the stability of grid connection while improving the continuity of phase convergence and the overall coordination of frequency adjustment. BRIEF DESCRIPTION OF DRAWINGS

[0062] Other features, objects and advantages of the application will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings:

[0063] Figure 1 An exemplary application scenario provided for the embodiments of the application is shown in the following figure:

[0064] Figure 2 A module schematic diagram of the weak network under the network energy storage frequency adaptive adjustment system provided for the embodiments of the application is shown in the following figure:

[0065] Figure 3 A flowchart of the weak network under the network energy storage frequency adaptive adjustment method provided for the embodiments of the application is shown in the following figure:

[0066] Figure 4 A structure diagram of an island node state graph provided for an embodiment of the present application is shown in FIG. 1.

[0067] Figure 5 A principle diagram of a relative position relationship provided for an embodiment of the present application is shown in FIG. 2.

[0068] Figure 6 A structure diagram of an updated island node state graph provided for an embodiment of the present application is shown in FIG. 3. DETAILED DESCRIPTION

[0069] The technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments.

[0070] In this document, the term "embodiment" means that a specific feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessarily mutually exclusive of other embodiments. It will be explicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0071] The weak network construction energy storage frequency adaptive adjustment method involved in the present application is not aimed at an ideal grid-connected, parameter-known, single-machine independent laboratory scene, but a typical operating state closer to the engineering site:

[0072] In a weak power grid environment with low short-circuit capacity of the distribution network, high line impedance ratio, and insufficient inertia and damping, multiple grid-constructed energy storage devices are connected in parallel in a distributed manner to the same public coupling point for operation, and enter island power supply when the upper-level power grid fails or is planned to be disconnected, and then perform re-grid connection under the requirements of power grid restoration or operation strategy.

[0073] During such operation, the grid-constructed energy storage device not only assumes the responsibility of establishing a voltage and frequency reference, but also needs to achieve stable, controllable, and verifiable frequency support behavior under the combined action of power grid strength mutation, load step, parallel system topology change, and device current limiting / amplitude limiting non-linear constraints.

[0074] It can be understood that common scenarios of the present application include but are not limited to:

[0075] Campus microgrid, data center and industrial enterprise self-provided power system, distributed energy storage group control at the transformer area, local power grid with grid-constructed new energy parallel connection, etc.

[0076] Such scenarios often exhibit several common characteristics:

[0077] Firstly, the electrical coupling relationship of multiple grid-connected energy storage devices is significantly affected by the access location and line parameters, which dynamically drift with the switching of switches and changes in ring network structure.

[0078] Secondly, during the grid-connected window, the voltage / current measurement signals at the point of common coupling may be disturbed by temporary drops, distortions, phase angle jumps, or noise, causing the phase characteristics of the main grid to be unstable for a short time or even intermittently unavailable.

[0079] Thirdly, energy storage converters are easily limited in current, power, or modulation saturation due to device capacity and protection strategy limitations, causing sudden changes in the equivalent control structure of the system.

[0080] Fourthly, devices on site may come from different manufacturers or different control versions, lacking unified scheduling and high-speed communication conditions, making it common for each device to adaptively adjust based on local observations.

[0081] Under the existing technical framework, common practices for reconnection or weak grid stability problems usually rely on fixed parameter setting, empirical selection of single-machine droop and virtual inertia, or phase-locked loop / synchronization loop-based following grid synchronization strategies. In the case of multiple parallel machines, there are also schemes for achieving consistency through master-slave control, centralized secondary control, or communication coordination.

[0082] However, when the system is in a weak grid, distributed, multi-machine parallel, and frequently changing topology condition, the aforementioned methods often expose engineering contradictions that are difficult to avoid:

[0083] On the one hand, if each device simultaneously enhances the traction or synchronization strength of the main grid phase during reconnection, it is easy to cause phase pulling and power circulation, further inducing low-frequency power swings and triggering amplitude limiting, and thus forming a chain of disturbances between devices.

[0084] On the other hand, if each device conservatively sets the synchronization strength or excessively relies on a single device to bear the synchronization responsibility, it may lead to insufficient phase convergence speed, prolonged grid-connected window, delayed frequency recovery, or reduced stability margin under load disturbance.

[0085] It is important to note that under weak grid conditions, the equivalent parameters of the main grid cannot be directly obtained and dynamically change. Frequency adaptive regulation based on unreliable phase / frequency characteristics as a driving force is prone to parameter chattering and repeated switching. When current limiting / saturation is involved, traditional linear stability intuition and setting experience often fail, making field debugging rely on a large number of trial and error, and it is difficult to form a reusable engineering methodology.

[0086] Based on the foregoing problems, the core logic of the method of the present application is not based on pre-dividing fixed control roles or assuming that the main grid phase is always reliable, but abstracts the frequency adaptive process in the re-grid connection stage as an evolving state-coupling relationship update problem:

[0087] In the island operation stage, by obtaining the node state vector of each network-forming energy storage device, an island node state graph reflecting the running state of the device and the coupling relationship between nodes is constructed, so that the group behavior of the system can be expressed and updated at the data structure level; when the grid connection trigger condition is met, the main grid phase feature is extracted based on the common coupling point measurement signal, and the consistency of the main grid historical operation data is checked to obtain a phase reference representation with validity constraints, so as to avoid directly using the instantaneous unreliable phase point as the whole network synchronization target.

[0088] On this basis, the present application does not directly insert the logical main grid node into the island node state graph and force all devices to synchronize, but introduces a phase corridor object to describe the allowed range of the main grid phase reference and its validity state, and accordingly performs traction compatibility judgment on each device to form a node set with traction qualification; then the logical main grid node with node attributes containing the phase corridor object is generated in the island node state graph, and the edge attribute information between the traction nodes and the non- traction nodes is rearranged through the graph reconstruction plan, realizing the structured propagation of phase convergence by a few traction and majority cooperation, rather than the competitive behavior caused by the strong traction of multiple devices to the main grid phase at the same time.

[0089] Further, the present application updates the traction nodes through consistency evaluation index to convert the engineering phenomena of whether the phase convergence causes circulating current, power swing or amplitude limiting risk into triggerable structure update conditions, finally forms a grid connection structure graph, and adjusts the frequency parameters of each network-forming energy storage device according to the structure graph, so that the frequency support strategy and the parallel coupling structure evolve synchronously, thereby balancing the phase convergence, power impact control and overall stability under the nonlinear constraint of the device in the key window of weak grid reconnection.

[0090] It should be noted that the method of the present application does not rely on accurate identification of the equivalent parameters of the power grid, nor does it take the fixed device structure division or the pre-configured master-slave relationship as a prerequisite; its applicable object can be a distributed network type energy storage system with variable number of energy storage devices, variable access location, and diverse device models and control versions. Especially in the case of limited communication conditions in the engineering field, grid-connected point measurement quality fluctuations, and inevitable device current limiting / amplitude limiting, the idea proposed in the present application makes the frequency adaptive adjustment in the re-grid stage no longer rely on single machine setting experience, but can realize structured convergence guided by system level consistency and stability margin, providing an implementable technical path for the long-term difficulty of simultaneous consideration of grid connection, stability and long duration in weak grid scenarios for those skilled in the art. The point of common coupling can be a physical bus measurement point, or a logical measurement point formed by multiple measurement points through electrical equivalent mapping; when there are multiple grid-connected switches or multi-feed lines, the point of common coupling measurement signal can be obtained by impedance weighted fusion of multiple measurement points.

[0091] Reference Figure 1 , Figure 1 An exemplary application scenario graph is provided for the embodiments of the present application.

[0092] Figure 1 The application scenario is shown to include a main grid and a networked energy storage system, which is connected to the main grid in a distributed manner, and includes a plurality of networked energy storage devices (e.g., energy storage device one, energy storage device two, and energy storage device N) inside.

[0093] It can be understood that in the case of normal operation or fault splitting of the main grid, the networked energy storage system can enter an island operation mode, and the system voltage and frequency reference can be established and maintained by the networked energy storage devices. When the main grid is restored or meets the grid connection conditions, the networked energy storage system needs to rejoin the main grid without interrupting the power supply to the load or minimizing the disturbance.

[0094] It should be noted that Figure 1 The networked energy storage system shown in the above does not require that each networked energy storage device has the same rated capacity, control parameters or access location, and there is no need for centralized master-slave control or high-speed communication connection between the energy storage devices.

[0095] In actual application, the networked energy storage devices can be distributed in different feeders and different electrical distance locations, and the equivalent impedance, coupling relationship and operating state of the grid-connected system will change dynamically with switch switching, load changes and main grid operating conditions.

[0096] In the above application scenario, the networked energy storage system usually faces the following two problems during island operation and re-grid:

[0097] On the one hand, multiple grid-connected energy storage devices participate in frequency and phase regulation at the same time. If there is a lack of a reasonable coordination mechanism, low-frequency oscillations or inter-device circulating currents may be caused by phase pull or power competition during the grid reconnection phase.

[0098] On the other hand, the phase and frequency characteristics of the main grid in the early stage of recovery may have transient fluctuations or measurement uncertainties. Directly using them as the synchronization reference for all grid-type energy storage devices can easily lead to frequency adaptive regulation instability or frequent triggering of device current limiting protection.

[0099] Based on this, this application is... Figure 1 Based on the application scenario shown, the grid-connected energy storage system is abstracted as a parallel system composed of multiple grid-connected energy storage device nodes. By constructing a node state diagram during the islanded operation phase and introducing logical main grid nodes and phase corridor constraints during the reconnection phase, the coupling relationship and frequency regulation behavior between grid-connected energy storage devices are structurally reconstructed, thereby enabling the system to achieve orderly phase convergence and adaptive frequency regulation during the main grid connection process.

[0100] refer to Figure 2 , Figure 2 A schematic diagram of the module of the weak grid under-grid energy storage frequency adaptive adjustment system provided in the embodiments of this application.

[0101] Figure 2 The system shown includes:

[0102] The data acquisition module is used to acquire the node state vectors of each grid-type energy storage device, and to construct an island node state diagram based on the node state vectors during the island operation phase.

[0103] The main network phase feature extraction module extracts the main network phase features based on the measurement signals of the common coupling point, and performs consistency verification by combining the main network historical operation data to construct the phase corridor object;

[0104] The graph reconstruction module performs traction compatibility determination based on the phase corridor object and the node state vector to obtain a set of traction compatible nodes, and obtains a phase convergence graph.

[0105] The frequency parameter adjustment module updates the traction nodes to suppress competition based on the consistency evaluation index of the phase convergence diagram to obtain the grid connection structure diagram, and adaptively adjusts the frequency parameters of each grid-type energy storage device according to the grid connection structure diagram.

[0106] Next, with reference to the accompanying drawings, the adaptive frequency adjustment method for energy storage in a weak grid understructure provided in this application will be further described. Figure 3 The method shown is applied to a distributed grid-type energy storage system, which includes at least two grid-type energy storage devices and a common coupling point connected in parallel with the grid-type energy storage devices.

[0107] It should be noted that the weak network in the present application can be understood as a power grid environment with relatively low short-circuit capacity, relatively large equivalent impedance, insufficient system inertia and damping support capability, and obvious changes in equivalent parameters during operation due to changes in topology, load or power supply access state; the weak network can be a distribution network, a microgrid, an islanded grid or a local grid, or a transient weak grid state of the main grid under fault recovery, split operation or high proportion of power electronic power supply access, which is not limited in the present application.

[0108] At the same time, the point of common coupling in the present application can be understood as a common connection node of multiple grid-forming energy storage devices in the grid-forming energy storage system and the main grid or other power sources, which is electrically connected and interacts with each other. The point of common coupling can be a bus node, a grid-connected switch connection point, a collection node or an electrically equivalent node, or a measurement point or a logical mapping node for obtaining common voltage, current and frequency information, which is not limited in the present application.

[0109] S1: In the island operation phase, the island node state graph is constructed by acquiring the node state vector of each grid-forming energy storage device;

[0110] In the present embodiment, when the grid-forming energy storage system is split from the main grid and enters the island operation phase, the running state of each grid-forming energy storage device is collected to form a corresponding node state vector, and an island node state graph is constructed based on the node state vector. The node state vector is used to represent the key operating characteristics of the grid-forming energy storage device in the island operation phase, which at least reflects the phase operating state, power output state and operating margin state of the device.

[0111] Those skilled in the art can understand that the acquisition of the key operating characteristics can be realized based on real-time sampling data, operating record data or state estimation results of the internal controller of the device, and the specific data source and collection method can be flexibly determined according to actual engineering conditions, as long as it can reflect the current operating state of the device to the minimum extent, which is not limited in the present embodiment.

[0112] S2: When the grid-connected trigger condition is met, the main grid phase feature is extracted based on the measurement signal of the point of common coupling;

[0113] In the present embodiment, when it is detected that the grid-connected trigger condition is met, the main grid phase feature is extracted based on the voltage and current measurement signals collected at the point of common coupling. The grid-connected trigger condition can include that the main grid voltage is restored to the allowable range, the grid-connected switch is in a closeable state, or the upper-level dispatching instruction allows grid connection, etc., which can be set by those skilled in the art according to the specific application scenario.

[0114] It should be noted that the present application does not simply equate the public coupling point measurement signal to a stable phase reference, but rather combines historical operation data to perform consistency verification and correction on the extracted phase candidate features, so as to reduce the influence of transient fluctuations, measurement noise or reflection disturbances on phase judgment under weak network conditions.

[0115] S3: adding a logical main network node to the island node state graph, so that the logical main network node serves as a phase reference of the island node state graph, wherein the adding of the logical main network node comprises determining a traction node corresponding to the logical main network node, to obtain a phase convergence graph;

[0116] In the present embodiment, after the main network phase feature extraction is completed, the main network is not directly injected into the island node state graph as a forced synchronization reference, but a phase corridor is constructed based on the main network phase feature, and the logical main network node is generated in the island node state graph by taking the phase corridor as a constraint condition. Further, by matching the phase corridor with the node state vector of each grid-forming energy storage device, traction compatibility determination is performed, so as to screen out the grid-forming energy storage devices with traction capability as traction nodes.

[0117] As understood by those skilled in the art, the traction nodes are not fixedly designated, but are dynamically determined according to the current operation state of the device, the margin condition and the phase accessibility. Subsequently, by reconstructing the edge relationship of the island node state graph, the logical main network node only participates in the phase convergence process through the traction nodes, to form a phase convergence graph.

[0118] S4: performing competitive suppression update on the traction nodes according to the consistency evaluation index of the phase convergence graph, to obtain a grid-connected structure graph, and adjusting the frequency parameters of each grid-forming energy storage device according to the grid-connected structure graph;

[0119] In the present embodiment, based on the phase convergence graph, the phase convergence behavior and power response characteristics between the traction nodes are analyzed to form a consistency evaluation index, which is used to represent whether there is a competitive trend or an adverse coupling effect in the traction process. When the consistency evaluation index represents that the phase convergence between the traction nodes is out of synchronization, the low-frequency power swing is enhanced, or the circulating current risk is increased, competitive suppression update is performed on the traction nodes, so as to obtain a grid-connected structure graph. The competitive suppression update can be embodied as adjustment of the traction relationship strength, the propagation path or the number of participating nodes.

[0120] Before the specific technical content corresponding to the expansion step is described, the present embodiment needs to be emphasized again:

[0121] The grid-connected energy storage system under weak grid conditions is not always in an ideal state of clear reference, negligible coupling and static parameter setting. On the contrary, during the transition period from island operation to re-grid, the changes in frequency and phase are often not caused by a single source, but by the control actions of multiple grid-connected energy storage devices superimposed under network impedance coupling.

[0122] Especially when there are differences in device capacity, access location or control parameters, the phase traction action of a certain device will project similar disturbance patterns in the power and frequency response of other devices through parallel coupling, so that the frequency deviation or power swing observed from the local measurement does not completely correspond to the real grid phase change, and may contain pseudo-synchronous characteristics generated by inter-machine coupling propagation.

[0123] As the skilled person in the art can understand, the pseudo-synchronous characteristics are often similar to the real synchronous demand in performance, such as also presenting as phase difference reduction, frequency deviation convergence or power response acceleration, but the causes and propagation paths are different. If the adaptive adjustment is directly based on the traditional phase error or frequency deviation, the disturbance caused by coupling propagation is likely to be misjudged as a signal that needs to be further enhanced, thereby triggering a competitive behavior of multiple machines simultaneously straining, and making the system enter a state of rapid amplification of low-frequency swing or circulating current risk.

[0124] The processing method proposed in the embodiment is not critical to the aggressive optimization of the frequency control loop of a single device, but to the injection of reference and structural coupling as equally important constraint objects, and by introducing a phase reference expression with effectiveness constraint, the instantaneous phase point is avoided as a forced target for the whole grid.

[0125] Specifically, the phase characteristics of the main grid participate in the subsequent calculation in the form of an allowed range in the embodiment, and the effectiveness is constrained by the consistency of the historical operation data, so that the phase reference has natural anti-jitter ability.

[0126] It can be understood that the historical operation data does not require a certain fixed sampling frequency or fixed data type, and can be derived from public coupling point measurement records, or from internal phase, frequency, power and other operation records of the device, or from online state estimation, as long as it can meet the minimum requirement of consistency judgment on the stability of the phase characteristics, and the embodiment does not make more limitations.

[0127] Under the constraint, the subsequent traction action is no longer around a certain instantaneous phase point, but around the phase allowed range to achieve bounded convergence. The embodiment converts the coordination problem of the parallel system into a process management of structural propagation, that is, the introduction of the main grid reference is limited to propagation through a few qualified traction nodes, rather than direct participation in traction by all devices.

[0128] It should be noted that the determination of the traction node is not dependent on prior designation, nor does it require accurate identification of the equivalent parameters of the main grid. Instead, it is based on the availability and constraints of the device's own state, making the set of traction nodes implementable in an engineering sense.

[0129] For example, when the device is at the current limit edge, the power limit has been triggered, or the DC side margin is insufficient, it may be close to the main grid reference in phase, but it is not suitable as a traction node. On the contrary, devices with sufficient operating margins and whose phase state can enter the allowed range within a preset window are more suitable as traction nodes to undertake reference injection tasks. Through this screening process, the formation of the traction node set has an inherent safety constraint property, thereby reducing the probability of triggering saturation nonlinearity and causing control structure mutations.

[0130] Further, after the formation of the traction node set, the handling of the traction propagation path in the parallel system in this embodiment is also different from the conventional approach.

[0131] Traditional methods often treat parallel coupling as uncontrollable background conditions and passively suppress oscillations by reducing control bandwidth or increasing damping.

[0132] This embodiment takes the coupling edge property between traction nodes and non-traction nodes as a controllable object, and through updating the edge property of the propagation path, the propagation of the reference convergence presents a controllable hierarchical relationship.

[0133] As understood by those skilled in the art, the edge property information is not limited to a single weight concept, but can also represent a propagation strength upper limit, a change rate constraint, a freeze / rollback condition, or other equivalent control constraints, which can be carried by parameter mapping, amplitude limiting logic, or gain scheduling mechanisms within the controller. As long as the propagation strength and propagation speed of the traction impact in the network can be constrained, the effects described in this embodiment can be achieved, and this embodiment is not further limited.

[0134] By structurally constraining the propagation path, the traction action of the traction node can be avoided to form a symmetrical mutual projection in the coupled network, reducing the formation conditions of inter-machine circulating current and low-frequency power swing.

[0135] Next, the technical content of the island node state diagram of the method of the present application is further expanded.

[0136] In one example, the node state vector includes at least one of the following:

[0137] Phase state information for representing the internal phase reference of the corresponding grid-forming energy storage device;

[0138] Power state information for representing the active power output of the corresponding grid-forming energy storage device;

[0139] reactive state information for characterizing reactive power output of the corresponding grid-forming energy storage device;

[0140] margin state information for characterizing operation margin of the corresponding grid-forming energy storage device, wherein the margin state information comprises current limit triggering state, power limit triggering state, DC side voltage margin, and modulation saturation degree;

[0141] parameter state information for characterizing current value of frequency parameter of the corresponding grid-forming energy storage device, wherein the frequency parameter comprises droop coefficient, virtual inertia parameter, and damping parameter.

[0142] It can be understood that the node state vector is not used to exhaustively describe all operation information of the grid-forming energy storage device, but is used to abstractly express the current operation state of the grid-forming energy storage device under the premise of meeting the frequency adaptive adjustment and coordinated control of the interconnected system.

[0143] As can be appreciated by those skilled in the art, the above-mentioned phase state information, power state information, reactive state information, margin state information, and parameter state information can be obtained through real-time sampling data, operation record data, or state estimation results of the device internal controller. The specific collection period, data accuracy, and update frequency can be set according to actual engineering needs, and the present application does not limit this.

[0144] Further, the various types of state information contained in the node state vector do not necessarily exist at the same time. In different application scenarios or different operation stages, a part of them can be selectively included according to the control target and obtainable information of the grid-forming energy storage device.

[0145] For example, during the grid-connected window period, the phase state information and the margin state information can be used as the main basis for traction compatibility determination; during the frequency support stage, the power state information and the parameter state information can be used to evaluate the influence of frequency parameter adjustment on system response.

[0146] As can be appreciated by those skilled in the art, as long as the node state vector can minimally reflect the operation characteristics of the grid-forming energy storage device at the current stage and support the judgment of the coupling relationship between devices and the frequency adjustment behavior, the technical effects of the present application can be achieved, and the present application does not further limit the specific construction method of the node state vector.

[0147] In another example, the island node state graph is constructed, comprising:

[0148] S1.1: obtaining node identifiers of each grid-forming energy storage device and node state vectors corresponding to the node identifiers, and establishing an island node set with each grid-forming energy storage device;

[0149] Specifically, by assigning a unique node identifier to each grid-connected energy storage device and associating it with its corresponding node state vector, each node not only represents the physical device itself in the graph structure, but also carries key control and operation information of the device at the current running moment. This avoids storing device states in multiple independent control modules, which is beneficial for subsequent overall analysis of the relative state between devices in the grid-connected stage. The form of the node identifier can be a device number, a communication address, or a logical number, and the update frequency of the node state vector can be consistent with the device control period or monitoring period, as long as the state information is reasonably consistent with the actual running state.

[0150] S1.2: Determine the coupling relationship between the island nodes according to the access positions of the grid-connected energy storage devices, and establish an island edge set with the coupling relationship as the edge;

[0151] Specifically, the access position does not only refer to the physical geographical position, but also refers to the connection relationship of the device in the electrical network, such as whether it is connected to the same bus, whether it is in the same feeder, or whether it is coupled through line impedance. The island edge set established in this way is used to depict the mutual influence path between the grid-connected energy storage devices due to electrical connection. The coupling relationship can be obtained through the system wiring diagram, configuration parameters, or online topology identification result, and does not require accurate line parameter modeling level, as long as it can reflect whether there is a direct or indirect coupling relationship between the devices, which can meet the subsequent analysis needs.

[0152] S1.3: Configure edge attribute information for each edge in the island edge set, wherein the edge attribute information includes a coupling weight representing the coupling strength between the island nodes, and the coupling weight is calculated according to the node state vector;

[0153] Specifically, the coupling weight is not a fixed constant, but is dynamically calculated according to the operating state of the corresponding node, so that the edge attribute can reflect the equivalent coupling strength under the current operating condition.

[0154] For example, when the power output of a certain grid-connected energy storage device is large or the frequency regulation parameter is more aggressive, its influence on the adjacent node is relatively higher, and the coupling weight of the corresponding edge can be increased accordingly; on the contrary, when the device is in the state of current limiting, power limiting, or modulation saturation, its regulation ability to the system is limited, and the coupling weight can be reduced accordingly.

[0155] As can be appreciated by those skilled in the art, the calculation of the coupling weight can be realized by integrating the phase state information, power state information, or margin state information in the node state vector, and the specific calculation form can be set according to engineering experience or simulation results, and the present embodiment does not limit the specific function form.

[0156] S1.4: generating the island node state graph based on the island node set, the island edge set and the edge attribute information of each edge, and periodically updating the edge attribute information of the island node state graph in the island running stage;

[0157] Specifically, periodic updating does not require high-frequency real-time execution, but can be consistent with the device state update rhythm, so that node state changes can be gradually reflected in the graph structure. The island node state graph constructed in this way can continuously reflect the overall operation situation and coupling characteristics of the multi-network type energy storage device in the island running stage without introducing centralized control or complex communication, thereby providing a stable and evolving data basis for subsequent logical main network node injection, traction node screening and frequency parameter adaptive adjustment in the grid-connected stage.

[0158] Reference Figure 4 , Figure 4 The structure diagram of the island node state graph provided by the embodiments of the present application is shown.

[0159] As shown in Figure 4 , the island node state graph is composed of nodes corresponding to a plurality of network type energy storage devices. The distribution of each node in the graph is not a regular array or a full connection structure, but a distributed and non-fully coupled topological form according to the actual electrical access relationship.

[0160] Specifically, there is a coupling relationship between node one and node two, corresponding to island edge one, there is a coupling relationship between node one and node three, corresponding to island edge two, there is a coupling relationship between node four and node five, corresponding to island edge three, there is a coupling relationship between node five and node six, corresponding to island edge four, and node seven exists as a single network type energy storage device, without island edge directly connected thereto.

[0161] It can be understood that in actual engineering scenarios, each network type energy storage device is often scattered and accessed in different feeders, different buses or different electrical distance positions. Whether there is a coupling relationship between nodes depends on the electrical connection path and line impedance condition, and not all nodes have direct or equivalent coupling relationship. Therefore, in the island node state graph shown in Figure 4 , only the corresponding edge connection is established between the nodes with electrical coupling effect, and no direct connection relationship is established for the nodes with far electrical distance or through multi-stage switch isolation, and the coupling effect can be ignored.

[0162] Those skilled in the art can understand that Figure 4 only one illustrative structure example, the number of nodes, node distribution form and coupling relationship can be adjusted according to the scale and access mode of the specific network type energy storage system, and the present application does not limit this.

[0163] Next, the technical content of the method of the application on the main grid phase characteristics is further developed.

[0164] In one example, the main grid phase characteristics are extracted based on the measurement signals of the common coupling point, comprising:

[0165] S2.1: Obtain voltage measurement signals and current measurement signals at the common coupling point, wherein the voltage measurement signals include three-phase voltage instantaneous values, phase voltages and line voltages, and the current measurement signals include three-phase current instantaneous values;

[0166] Specifically, the voltage and current signals at the common coupling point are often in a non-steady state during the weak grid reconnection window: the voltage amplitude may fluctuate for a short time during the recovery process of the upper grid, and the current signal may have distorted components due to the current limiting action of the network device, load impact or grid impedance coupling. Relying solely on a single type of signal for phase inference is prone to phase jumps or discontinuities when voltage sag, harmonic rise or voltage measurement is affected by noise.

[0167] In this embodiment, voltage sensors and current sensors are configured at the common coupling point to collect three-phase voltage instantaneous values and three-phase current instantaneous values; when the hardware conditions permit, the phase voltages and line voltages are synchronously collected so that the line voltages can be used as backup measurement quantities when the phase voltages deviate due to wiring or sensor saturation. The sampling is synchronized with the controller clock to ensure that the voltage and current samples can be aligned at the same sampling time for phase analysis. To avoid the influence of instantaneous glitches caused by electromagnetic interference on subsequent processing, an anti-aliasing filter and over-voltage / over-current measurement protection can be provided in the measurement link. The collected original sequence is buffered to form a sliding window with a fixed length. Those skilled in the art can determine the window length according to the power frequency, sampling rate and control period, as long as the minimum data amount requirement for subsequent fundamental extraction and trend judgment is met.

[0168] S2.2: Signal preprocessing is performed on the voltage measurement signals and current measurement signals to obtain stable measurement signals for phase analysis, wherein the signal preprocessing includes filtering, fundamental component extraction, harmonic suppression and abnormal mutation suppression;

[0169] Specifically, under weak grid conditions, the waveform at the common coupling point often deviates from the ideal sine wave, and there are harmonics, interharmonics, unbalance, transient spikes and phase angle transient transitions; if phase calculation is directly performed on the original waveform, the phase trajectory will swing with high-frequency noise, and even produce discontinuous jumps when the waveform distortion is strong. As the reference input of the subsequent logic main grid node, the phase will be converted into a power shock source if it jumps, so the signal needs to be stabilized before phase calculation so that the phase can be used as a component representing the fundamental behavior of the main grid.

[0170] In the embodiment, the signal preprocessing is performed online in a sliding window manner, and the preprocessing link at least includes: band-pass or low-pass filtering of three-phase voltage and three-phase current respectively to suppress sampling noise and high-frequency spikes; after filtering, fundamental component extraction is performed on the signal, and the fundamental extraction can be realized by using synchronous reference extraction, sliding correlation or adaptive filtering, and the output is a three-phase signal equivalent to the fundamental or a quadrature component representation. For harmonic suppression, in addition to filtering and fundamental extraction, harmonic energy evaluation can also be combined to perform weighted processing on the window: when the harmonic proportion in the window exceeds the preset threshold, the contribution weight of the window to the subsequent phase candidate feature is reduced, or the window is marked as a historical correction window. The abnormal mutation suppression can be realized by combining the combination of amplitude clipping and slope constraint: the samples with instantaneous change amplitude exceeding the threshold are clipped, and the upper limit of the fundamental phase change rate of adjacent windows is set, so that the sudden spikes will not be misinterpreted as phase changes.

[0171] S2.3: based on the stable measurement signal, extracting the instantaneous phase information and the phase change trend information corresponding to the point of common coupling as the phase candidate feature of the current main grid;

[0172] In the embodiment, the instantaneous phase information of the point of common coupling is calculated based on the stable measurement signal, and the instantaneous phase information is preferentially extracted from the positive sequence fundamental voltage; in the window with low voltage effective value or strong voltage distortion, the phase information of the positive sequence fundamental current can be introduced to judge the consistency of power direction and phase evolution. The phase change trend information is maintained in a time sequence manner, and at least includes the continuous trajectory of the phase with time, the trend of the phase change rate, and the consistency mark of the phase change direction.

[0173] S2.4: obtaining the historical operation data of the main grid corresponding to the point of common coupling, wherein the historical operation data of the main grid at least includes at least one of historical phase information, historical frequency information and historical voltage operation information;

[0174] Specifically, the phase of the main grid in the weak grid reintegration stage may appear a recovery swing or a measurement interruption in a short time, and the grid-forming energy storage device needs to maintain a continuous understanding of the phase reference of the main grid in this stage. Only relying on the phase candidate feature of the current window will make the reference swing with the measurement quality fluctuation, resulting in unnecessary repeated adjustment of the traction node in the phase convergence process. The significance of introducing the historical operation data of the main grid is to establish a continuous reference trajectory, so that the current candidate phase can be verified, corrected or replaced under the constraint of the historical trajectory, thereby ensuring the time continuity of the phase reference.

[0175] S2.5: verifying the phase candidate feature based on the main grid historical operation data, determining the phase candidate feature as the main grid phase feature when the verification adapts to a preset adaptation condition, if not, iteratively modifying the phase candidate feature based on the main grid historical operation data until the adaptation condition is adapted;

[0176] Specifically, the main grid phase feature, as the basis for subsequent phase corridor construction and logical main grid node injection, needs to meet both current measurement reliability and reasonable time evolution constraints:

[0177] If only the instantaneous error threshold is used to determine the adaptation condition, it may be rejected when the phase candidate feature has a short-time deviation but the trend is correct. If only the trend consistency is used, it may be accepted when the measurement is contaminated by noise. Therefore, the adaptation condition is set as a combined condition in this embodiment, which is used to verify the phase candidate feature in multiple dimensions, and obtain a usable main grid phase feature through iterative modification when it is not adapted.

[0178] In this embodiment, the verification process includes at least three levels:

[0179] First, continuity verification, which determines whether the change of the phase candidate feature relative to the historical phase trajectory remains continuous or whether there is a jump that does not conform to the allowed change rate;

[0180] In this embodiment, the continuity verification is performed in a sliding window manner. A current window (e.g., 40 ms or 80 ms) outputs a phase candidate value, and the historical phase trajectory of the last several windows (e.g., the trajectory of the last 2 s for verification) is retained. After each phase candidate value is generated, the phase branch alignment is performed first: compare the candidate phase with the confirmed phase at the last time, if there is a possibility of boundary folding, then expand the candidate phase to the same direction as the last time according to the continuous branch, so as to avoid the numerical jump caused by crossing the boundary. Then the change amplitude constraint and the change rate constraint are performed:

[0181] Taking the change amplitude constraint as an example, under 50 Hz power frequency, if the window update period is 10 ms, the reasonable change range of the phase within 10 ms under normal working conditions should not exceed several degrees; the embodiment can be set to that the phase change between adjacent two windows does not exceed 30 degrees as the continuity threshold.

[0182] Second, trend consistency verification, which determines whether the phase change trend is consistent with the evolution direction corresponding to the historical frequency information, avoiding the situation that the phase candidate change direction contradicts the frequency evolution;

[0183] In this embodiment, the trend consistency check is performed in three steps: trend extraction, direction comparison, and anomaly determination. The trend extraction obtains two sequences: one is the trend of the phase candidate sequence (e.g., aggregate the phase changes of the last 5-10 windows to obtain the phase advancing direction and advancing strength), and the other is the trend of the historical frequency information (e.g., smooth the frequency sequence of the last 200 ms to 1 s to obtain the trend label of frequency rising, falling, or stable). The basic rule of direction comparison is: if the frequency trend is rising, the phase advance should show acceleration or at least maintain faster advance; if the frequency trend is falling, the phase advance should show deceleration or at least not appear sustained acceleration; if the frequency trend is stable, the phase advance should show stable advance rather than large fluctuation.

[0184] Thirdly, waveform quality constraint check, combined with historical voltage operation information and current window credibility label, to determine whether the current window is in a credible interval that can be used as a phase reference;

[0185] In one example, based on the rated voltage, when the effective value of the point of common coupling voltage is less than 0.85 times the rated value, the window is marked as limited; and when it is less than 0.70 times the rated value, it is marked as unusable.

[0186] As can be appreciated by those skilled in the art, the present application does not limit the specific values, and the above values are only used to illustrate one example that can be implemented.

[0187] When the phase candidate feature meets the adaptation condition, it is directly output as the main grid phase feature; when it does not meet the adaptation condition, it enters the iterative correction process, which can use one or a combination of the following methods:

[0188] Smooth fusion of the candidate phase based on the historical phase trajectory to suppress jumps and maintain trends; perform phase alignment processing when boundary jumps are detected in the candidate phase to make it return to the continuous branch of the historical trajectory; increase the weight of the historical trajectory when the credibility of the candidate phase is low, output a phase reference closer to the historical prediction, and keep the observation of the new measurement waiting for the next credible window to update.

[0189] Next, the technical content of the method of the present application related to the logical main grid node is further expanded.

[0190] It should be noted that the logical main grid node in the present application can be understood as a virtual node, which does not correspond to a newly added physical bus or a newly added switching device in the electrical network, but is a logical carrier for controllable introduction of the main grid phase reference during the reconnection window period.

[0191] Specifically, under weak network conditions, the phase information at the public coupling point often contains the real phase evolution of the main network, measurement noise, waveform distortion, and disturbance components caused by the coupling and propagation of parallel network devices. If this phase information is directly used as the forced reference input of each network device, the phase traction behavior may be amplified synchronously under multi-machine parallel conditions, and power confrontation and circulating current risks may be formed between devices. Based on this, the logical main network node in this embodiment does not directly replace the main network or directly specify the reference function, but converts the main network phase reference from a continuous waveform signal to a graph model element with constraint conditions, which can be verified and structured, so that the subsequent phase convergence behavior can be managed at the network structure level.

[0192] In one example, a logical main network node is added to the island node state diagram, including:

[0193] S3.1: According to the main network phase feature, a phase corridor object is constructed, wherein the phase corridor object is used to represent the allowed range and validity state of the main network phase reference, and the validity state is determined by the checking result of the main network phase feature and the historical operation data of the main network;

[0194] Specifically, during the weak network reconnection window, the main network phase feature measured at the public coupling point may still present characteristics of short-term drift, intermittent credibility, or local jump even after filtering and historical verification. If this phase feature is input to the subsequent phase convergence process in the form of a single phase point, the traction node will respond quickly to the instantaneous phase change when the phase is pulled, and then convert the change into a power impact or inter-machine confrontation under the condition of multi-machine parallel coupling. The phase corridor object is used to convert the main network phase reference from a point to a range, and the reliability of the phase reference is carried in the form of an effectiveness state, so that the subsequent processing can distinguish between the interval where the reference is available but needs to be released and the interval where the reference is not available and needs to be maintained.

[0195] In this embodiment, the phase corridor object at least contains corridor boundary information and validity state information. The corridor boundary information is used to depict the allowed range of the main network phase reference, which can be determined by the main network phase feature and the historical phase trajectory:

[0196] When the main network phase feature has high credibility and the waveform quality meets the conditions, the allowed range gives a relatively narrow bandwidth around the current phase candidate feature; when voltage distortion, temporary drop or abnormal mutation markers appear, the allowed range gives a relatively wide bandwidth mainly based on the historical trajectory, or maintains the corridor boundary of the last moment and only updates the trend direction.

[0197] The validity state information is used to represent the applicability of the corridor boundary in the current window, and the generation basis is the consistency verification result of the phase candidate feature and the historical operation data of the main network:

[0198] When the consistency meets the preset adaptation condition, the validity state is marked as available; when the consistency does not meet but there is still a correctable space, it is marked as limited available and the limited reason is recorded, such as waveform distortion limitation, voltage effective value limitation and phase jump limitation; when the consistency continues to not meet or the measurement is lower than the minimum trust threshold, it is marked as unavailable and a corridor maintenance strategy is triggered, such as maintaining the last corridor or falling back to the historical corridor.

[0199] As can be appreciated by those skilled in the art, the width of the corridor boundary, the maintenance duration and the limited reason category can be set according to the sampling frequency, the device control bandwidth and the field waveform quality, as long as the phase reference can be continuously output in the form of a range and with an effectiveness mark.

[0200] For example, a numerical example is given below to facilitate understanding of the construction and updating process of the phase corridor object. The example is only used to explain the operation link and dimension relationship, and the selected parameters and values are illustrative and do not represent real calibration results or engineering recommended values. Those skilled in the art can adjust and replace them according to the grid connection strategy, sampling conditions and field waveform quality.

[0201] In this embodiment, the sampling frequency at the point of common coupling is 10 kHz, the main grid phase candidate feature is output once every 80 ms sliding window, and the historical phase trajectory of the last two seconds is maintained as a verification reference. Assuming that in a certain window, the phase candidate feature obtained through fundamental positive sequence extraction and phase tracking processing is 20 degrees, the window trustworthiness mark is high, and the waveform quality indicators meet the preset conditions, such as the voltage effective value being above 0.95 of the rated value, the total harmonic distortion rate being not more than 3%, the negative sequence component proportion being not more than 2%, and the number of abnormal mutation suppression trigger being zero. At this time, the allowed range is generated using a narrow bandwidth strategy: taking the phase candidate feature as the center, an allowed bandwidth of 8 degrees above and below is given, and the corridor boundaries can be set to 12 degrees to 28 degrees; at the same time, the phase change dispersion interval of the last two seconds is read from the historical phase trajectory, assuming that it corresponds to a short-time fluctuation bandwidth of about 10 degrees above and below under similar working conditions, the smaller one is taken as the corridor bandwidth, the corridor boundaries remain 12 degrees to 28 degrees, and the validity state is marked as available. The phase corridor object formed thereby at least includes: center phase 20 degrees, corridor lower boundary 12 degrees, corridor upper boundary 28 degrees, validity state available, and update timestamp current window end time.

[0202] S3.2: Based on the phase corridor object and the node state vector of each grid-forming energy storage device, perform traction compatibility determination on each grid-forming energy storage device to obtain a traction compatible node set;

[0203] Specifically, whether the parallel grid-connected device is suitable to undertake the traction in the grid-connected stage does not depend on the single phase proximity; under the weak grid condition, the margin constraint, the amplitude limiting state and the frequency parameter configuration of the device directly affect its phase traction ability and power impact risk. If all devices are allowed to simultaneously generate traction coupling to the logical main grid node without screening, it is easy to form phase pulling and power circulation among devices; if the screening logic is too rough, only the phase difference is screened, and the device at the edge of current limiting may be selected, which triggers saturation and causes control structure mutation in the traction process.

[0204] In one example, based on the phase corridor object and the node state vector of each grid-forming energy storage device, the traction compatibility of each grid-forming energy storage device is determined, including:

[0205] S3.2.1: transmitting the node state vector of each grid-forming energy storage device to the public coupling node, extracting phase state information, margin state information and parameter state information through the public coupling node;

[0206] Specifically, the traction compatibility determination needs to compare the states of multiple grid-forming energy storage devices in the same reference frame, otherwise even if each device gives its own phase, margin and other state quantities, it may also cause determination deviation due to inconsistent time stamps, inconsistent sampling windows or inconsistent definition of the same quantity, and further introduce unnecessary fluctuations in the traction node selection stage. The calculation position related to the aggregation of the node state vector to the public coupling node is beneficial to the parallel determination of multiple devices based on the unified main grid phase characteristics and the unified phase corridor object, and makes the determination process have a traceable state snapshot, thereby facilitating the subsequent graph reconstruction plan to configure the edge properties based on the state results at the same time.

[0207] In this embodiment, each grid-forming energy storage device periodically generates a node state vector and sends it to the aggregation calculation position on the public coupling point side through the on-site communication link, which can be realized by station Ethernet, industrial bus or other equivalent communication methods.

[0208] S3.2.2: determining the relative position relationship between the internal phase reference of the grid-forming energy storage device to be determined and the phase corridor object based on the phase state information, wherein the relative position relationship includes two states of being within the allowed range of the phase corridor object and being outside the allowed range of the phase corridor object;

[0209] Reference Figure 5 , Figure 5 The principle diagram of the relative position relationship provided by the embodiment of the present application.

[0210] As Figure 5As shown, the internal phase references of different grid-forming energy storage devices present a relatively dispersed distribution state in the same phase reference space, while the phase corridor object forms an allowed interval with upper and lower boundary constraints in the phase reference space, for characterizing the acceptable main grid phase reference range in the current grid-connected window period.

[0211] It can be understood that, Figure 5 The phase state information of at least two grid-forming energy storage devices is given in the figure, for example, the phase state information one corresponding to the grid-forming energy storage device one and the phase state information two corresponding to the grid-forming energy storage device two; the position of each phase state information on the phase axis reflects the current value of the internal phase reference of the corresponding grid-forming energy storage device. The phase corridor object is represented by a strip region in the figure, and the upper and lower boundaries thereof correspond to the boundary positions of the allowed range. The strip region does not require a fixed width, but can be dynamically adjusted according to the effectiveness state of the main grid phase characteristics and the verification result of the historical operation data.

[0212] Further, when the internal phase reference of a certain grid-forming energy storage device falls within the allowed range defined by the phase corridor object, it can be determined that the grid-forming energy storage device and the main grid phase reference have consistency at the current time, and the phase state thereof does not need to undergo a cross adjustment to keep in coordination with the main grid phase; correspondingly, when the internal phase reference of a certain grid-forming energy storage device is located outside the allowed range of the phase corridor object, it indicates that there is still a difference to be converged between the current phase state of the device and the main grid phase reference, and it needs to be further combined with the operation margin and the parameter state to determine whether it has the ability to enter the allowed range.

[0213] Specifically, the phase corridor object gives the allowed range of the main grid phase reference, and the compatibility determination needs to first determine the spatial relationship between the device internal phase reference and the allowed range, in order to distinguish between the two states of being able to directly undertake traction and keep stable and being able to enter the corridor to have the qualification of traction. If the relative position relationship is not divided and the subsequent accessibility is directly judged, it will cause the devices already in the corridor to still be subjected to additional approaching process, increase unnecessary phase adjustment actions, and may convert the originally stable phase state into a power disturbance source.

[0214] In this embodiment, the public coupling node reads the phase state information of the device to be determined, obtains the current value of the internal phase reference of the device, and matches it with the allowed range of the phase corridor object. The matching process considers the periodic characteristics of the phase.

[0215] As can be appreciated by those skilled in the art, the relative position relationship is generally classified into two categories, i.e., inside the corridor and outside the corridor, and further includes auxiliary markers such as boundary risk, approach and reverse drift at the implementation level, for enhancing the stability of the determination, which does not change the basic classification of the two categories of relative position relationship.

[0216] S3.2.3: determining that the network-forming energy storage device to be determined is a traction-compatible node when the relative position relationship represents that the position is within the allowed range of the phase corridor object;

[0217] Specifically, when the internal phase reference of the device is within the allowed range of the corridor, the device has consistency with the logical main network node at the phase reference level; at this time, if the operation margin allows, it can participate in the subsequent phase convergence process with a smaller traction adjustment amount, thereby reducing the phase distance that needs to be crossed in the traction process and reducing the power impact that may occur during the grid connection window period. Including such nodes directly into the traction-compatible node set is conducive to forming a low-action traction source in the traction node set, so that the reference of the logical main network node has a smooth starting point.

[0218] S3.2.4: determining whether the network-forming energy storage device to be determined meets the operation margin constraint condition based on the margin state information when the relative position relationship represents that the position is outside the allowed range of the phase corridor object;

[0219] Specifically, when the internal phase reference of the device is outside the allowed range of the corridor, if the device is to assume the role of traction, it needs to advance the phase into the corridor within a certain time; this process will essentially cause power adjustment and current response, and is subject to the constraints of device current limiting, amplitude limiting and modulation capacity. If the phase is forcibly advanced into the corridor in the case of insufficient margin, it is easy to trigger protection actions, causing the phase advancement to be interrupted and causing the device output to mutate, and then spreading to other nodes through parallel coupling to form low-frequency oscillation. Therefore, the margin constraint determination is performed before the reachability determination, which can exclude nodes that cannot be executed in advance, avoiding unrealistic traction configuration for these nodes in the subsequent exploration or prediction process.

[0220] In this embodiment, the public coupling node reads the margin state information of the device to be determined, and performs the operation margin constraint condition judgment. The operation margin constraint condition at least includes: the current amplitude limiting trigger state is false, the power amplitude limiting trigger state is false, the DC side voltage margin is not lower than a preset threshold, and the modulation saturation degree does not exceed a preset threshold. In addition to the above basic conditions, margin trend judgment can also be introduced:

[0221] If the DC side voltage margin is in a sustained downward trend although it meets the threshold value, or the modulation saturation degree is close to the threshold value and in an upward trend, it is determined as a margin criticality, and a more stringent time window and a more conservative traction capacity estimation are used in the subsequent reachability determination. The setting of the margin threshold value can be determined by the skilled person in the art according to the rated parameters of the device, the protection setting and the grid connection strategy, which only needs to meet the condition that the uncontrollable limiting is not triggered in the traction propulsion stage, and does not require a uniform fixed value.

[0222] S3.2.5: When the operating margin constraint condition is met, a phase convergence capability parameter is determined based on the parameter state information, and reachability of the internal phase reference of the grid-forming energy storage device to be determined into the allowed range of the phase corridor object is determined according to the phase convergence capability parameter, wherein the phase convergence capability parameter is positively correlated with the current value of the frequency parameter;

[0223] S3.2.6: If it is determined that it is reachable within a preset time window, the grid-forming energy storage device to be determined is a traction compatible node;

[0224] Specifically, even if the margin of the node outside the corridor meets the condition, it does not necessarily have the ability to advance the phase to the corridor range within the grid connection window period. The phase advance capability is related to the frequency parameter configuration of the device, such as the droop coefficient, virtual inertia and damping, which will affect the response speed, response smoothness and anti-oscillation ability of the device to the phase deviation. If the reachability is directly inferred according to the size of the phase deviation without considering the parameter state, it is easy to cause the situation that the phase deviation is not large but the parameter is blunt, resulting in too long time to enter the corridor, or the parameter is aggressive, resulting in power swing during the entering process.

[0225] In the embodiment, the phase convergence capability parameter is generated by the parameter state information, and at least reflects the upper bound of the phase advance rate and the advance smoothness constraint of the device under the current frequency parameter value. The generation process can use rule mapping:

[0226] When the droop coefficient is in the strong support interval and the damping parameter meets the stability constraint, the phase advance capability is marked as strong; when the virtual inertia is large and the damping is insufficient, the phase advance capability is marked as limited and the entering time is relaxed but the advance speed is tightened in the reachability determination; when the parameter is in the locked or prohibited update state, the phase advance capability is processed according to a fixed gear.

[0227] Further, according to the boundary of the phase corridor object and the current phase position, the reachability is determined in combination with the phase advance capability and the preset time window:

[0228] The reachability determination can be achieved by online trial, which applies limited approach adjustment to the device under the condition of meeting the margin constraint, and observes whether the phase change trend remains towards the corridor and meets the corridor entry constraint in a plurality of continuous control periods; when the phase is continuously approached and the margin deterioration mark is not triggered, it is determined that it is reachable within a preset time window, and the device is added to the traction compatible node set, while recording its corridor outside reachable category mark for subsequent traction edge initial value and change rate constraint configuration.

[0229] As understood by those skilled in the art, the preset time window can be set according to the grid closing in time, the relay protection cooperation time limit or the scheduling strategy, and the amplitude of the approach adjustment can be limited by the margin constraint and the change rate constraint to ensure that the trial process does not introduce too large disturbance.

[0230] S3.2.7: When the operating margin constraint condition is not met and it is determined that it is not reachable within a preset time window, the network-forming energy storage device to be determined is a non-traction compatible node.

[0231] S3.3: When the traction compatible node set is not empty, a logical main grid node is generated in the island node state graph based on the phase corridor object, wherein the node attribute of the logical main grid node includes the phase corridor object.

[0232] Specifically, the main grid phase reference is directly mapped to a single reference node in the graph structure, which is easy to inject short-term drift in the phase candidate feature into the phase convergence process of the entire graph; at the same time, the effectiveness of the main grid phase reference in different window periods is not constant, and if the reference node in the graph does not carry effectiveness information, subsequent edge attribute update cannot distinguish whether it should be promoted to convergence or should be kept on standby. Therefore, the logical main grid node is generated in the form of a virtual node with a corridor attribute in this embodiment, so that the reference input has a state carrier that can be queried, constrained and maintained in the graph structure.

[0233] In this embodiment, in addition to the corridor boundary and effectiveness state, the node attribute can also include a corridor retention strategy mark and a corridor update timestamp, so that when the reference is unavailable in subsequent processing, the retention or rollback can be directly performed according to the node attribute without the need to retrace the historical data. The logical main grid node is given a different type identifier from the ordinary device node in the graph structure, so that the graph reconstruction plan can identify the node as a reference source node when establishing a traction edge, and different constraint rules are adopted for the out-edge attribute of the node, such as more stringent configuration of the change rate constraint and the freezing condition of the out-edge.

[0234] As understood by those skilled in the art, the type identifier, attribute field and storage form of the logical main grid node can be implemented by using a structure, an object or a key-value pair storage according to the implementation platform, as long as it can be read and participate in the edge attribute update in the subsequent steps.

[0235] Reference Figure 6 , Figure 6 The structure diagram of the updated island node state graph provided by the embodiment of the application.

[0236] It can be understood that, Figure 6 The logical main grid node in the island node state graph does not correspond to any physical access point or newly added electrical node in the island network, but is used to carry and constrain the main grid phase reference at the graph structure level, so that the main grid phase reference can participate in subsequent graph reconstruction and phase convergence processes in the form of an allowed range and validity state.

[0237] As shown in Figure 6 , nodes one to seven still correspond to each grid-forming energy storage device node in the island operation phase, and the edge relationship between the nodes still reflects the distributed coupling network formed between the devices according to the access position; the introduction of the logical main grid node does not change the physical meaning of the above-mentioned distributed coupling relationship, but additionally provides a phase reference node attribute carrier that can be queried in the graph structure.

[0238] Figure 6 Further shown is an example situation in which nodes one, six and seven are determined as traction compatible nodes after traction compatibility determination, and are further connected with the logical main grid node through a dashed line.

[0239] S3.4: generating a graph reconstruction plan based on the traction compatible node set, wherein the graph reconstruction plan is used to reconstruct the edge set and edge attribute information of the island node state graph;

[0240] Specifically, the node state graph formed in the island operation phase mainly reflects the coupling relationship between the devices based on the access position, but when the logical main grid node is introduced, if the original coupling relationship is not adjusted, the approaching action of the traction node to the main grid reference will spread to the whole network in an uncontrolled manner through the original coupling edge, which is easy to form a state in which the phase propagation is too fast and the power exchange is intensified in a local area, especially when the non- traction node is in a tight margin or has poor waveform quality, the propagation process can trigger clipping and cause secondary swing at the system level. The graph reconstruction plan is used in the embodiment to reorganize the reference injection path and propagation constraint in the graph without changing the physical wiring relationship, so that the phase convergence propagates along a controllable path, and the propagation strength is matched with the node state.

[0241] In one example, generating a graph reconstruction plan based on the traction compatible node set includes:

[0242] establishing a traction edge between the logical main grid node and each grid-forming energy storage device node in the traction compatible node set, and determining the traction compatible node as a traction node;

[0243] obtain non-pulling nodes connected with each pulling node in the island node state graph, and determine island edges between the pulling nodes and the non-pulling nodes to generate a to-be-reconstructed edge set;

[0244] configure updated edge attribute information for each edge in the to-be-reconstructed edge set to obtain a pulling propagation edge set, wherein the updated edge attribute information is determined by calculating a phase approaching amount between a node state vector of a corresponding pulling node and the main grid phase feature;

[0245] correct edge attribute information of the pulling propagation edge set based on a node state vector of the non-pulling node to obtain a graph reconstruction plan.

[0246] It can be understood that the generation of the graph reconstruction plan is not simply superimposing a number of connection relationships on the original node state graph, but selectively rearranging the original coupling structure around the core problem of how the main grid phase reference propagates in the island network in a controllable manner. Since the nodes in the pulling compatible node set have the ability to undertake phase pulling in the current running state, the relative position relationship between the internal phase reference and the phase corridor object is clear, and the running margin meets the constraint, so it is confirmed as a pulling node in the graph structure level, and a pulling edge (i.e. Figure 6 the dashed line shown) is established between the logical main grid node and each pulling node. The pulling edge is used to depict the direct path of the main grid phase reference injection into the island network, and its existence does not change the physical wiring relationship, but provides a clear reference propagation starting point for subsequent phase convergence and frequency regulation, so that the reference injection is no longer randomly diffused through implicit coupling, but is along an identifiable and manageable path.

[0247] Specifically, after determining the pulling node, it is necessary to identify the non-pulling nodes in the island node state graph that have original coupling relationship with each pulling node, and take the island edges between the pulling nodes and the non-pulling nodes as the to-be-reconstructed edge set, based on the recognition that:

[0248] After the traction node undertakes the traction task of the main grid phase reference, the phase and power behavior of the traction node will change. If the edge attribute configuration of the island operation stage is still used, the traction action of the traction node may be transmitted to the non-traction node in an uncontrolled manner through the original coupling edge, thereby amplifying the local phase adjustment or power disturbance. Therefore, in the set of to-be-reconstructed edges, the edge attribute information of each edge is reconfigured to reflect the traction strength currently undertaken by the traction node and the degree of the phase approaching the main grid reference. Specifically, the updated edge attribute information is determined by the phase approaching amount between the node state vector of the traction node and the main grid phase characteristics, so that the greater the phase adjustment amplitude of the traction node, the greater the influence of the traction node on the adjacent non-traction node in the graph structure, thereby suppressing the too fast or too strong transmission at the structure level.

[0249] S3.5: Reconstruct the island node state graph containing the logical main grid node according to the graph reconstruction plan, to obtain the phase convergence graph;

[0250] Specifically, if the graph reconstruction plan only stays at the logical configuration level without landing in the graph structure, the constraints of the traction edge and the transmission edge cannot participate in the subsequent consistency evaluation and competition suppression update. At the same time, in the weak grid reintegration stage, the topology adjustment and edge attribute update need to be executable online to avoid introducing too many mutations in one reconstruction. The phase convergence graph in this embodiment is the graph structure obtained after reconstruction, and the edge set and edge attribute information can directly support the online update of the traction behavior and traction transmission behavior of the traction node, thereby making the phase convergence process have a traceable structure foundation.

[0251] Next, the technical content of the method of the present application on competition suppression is further expanded.

[0252] It can be understood that after the logical main grid node is introduced and the graph reconstruction is completed, the island node state graph has evolved from a structure that simply reflects the physical coupling relationship to a phase convergence graph that simultaneously carries the main grid phase reference, traction path, and transmission constraint. At this stage, although multiple traction nodes have traction compatibility, the differences in phase promotion ability, parameter configuration, and electrical position may still cause implicit competition behavior in the phase convergence process, that is, multiple traction nodes simultaneously try to promote the phase to approach the main grid reference at their own pace or intensity, thereby forming unnecessary phase pulling or power interaction within the integration system.

[0253] In this embodiment, competition suppression is not achieved by directly specifying a unique dominant node or forcibly closing part of the traction nodes, but by continuously evaluating the collaborative behavior between the traction nodes in the phase convergence graph, so that the traction action naturally differentiates at the graph structure level.

[0254] Specifically, the consistency evaluation index is used to describe the cooperation of the traction nodes in the phase convergence process, which at least reflects whether the phase approaching direction of the traction nodes relative to the logical master grid node is consistent, whether there is a significant difference in the phase advancing rhythm, and whether the traction behavior causes power swing or circulating current risk at the public coupling point.

[0255] Those skilled in the art can understand that the above-mentioned consistency evaluation index does not require a fixed mathematical form, but can be formed based on multi-dimensional information such as phase change trend, power response consistency and measurement signal stability, as long as it can distinguish between cooperative promotion and mutual restraint.

[0256] In one example, the consistency evaluation index includes:

[0257] a phase convergence consistency index of the traction nodes relative to the logical master grid node, the phase convergence consistency index being used to represent the cooperation degree of the phase of each traction node approaching the phase feature of the master grid;

[0258] a power consistency index between the traction nodes, the power consistency index being used to represent the low-frequency swing trend and consistency of the active power output of each traction node.

[0259] In another example, the traction nodes are updated according to the consistency evaluation index of the phase convergence graph to obtain a grid-connected structure graph, which includes:

[0260] S4.1: obtaining the node state vector corresponding to each traction node in the phase convergence graph, and determining the consistency evaluation index based on the node state vector;

[0261] Specifically, after the phase convergence graph is formed, the traction edges have been established between the traction nodes and the logical master grid node, and multiple traction nodes will simultaneously respond to the master grid phase reference in the same time window. In the parallel structure with high weak grid impedance proportion, the phase advancing and power response of the traction nodes will project on each other through the coupling path, resulting in that the "phase deviation reduction" observed by a single traction node from the local observation does not necessarily correspond to global cooperative convergence. In order to distinguish between cooperative promotion and mutual restraint without relying on fixed master-slave relationship, an index that can represent group consistency needs to be extracted from the state data of the traction nodes, so that the subsequent competition state judgment is not limited to the local error of a certain node, but can reflect the mutual relationship between the traction nodes and the comprehensive performance at the public coupling point.

[0262] In the embodiment, the node state vector corresponding to each traction node is read from the phase convergence diagram, and the phase state information, power state information, margin state information and parameter state information are aligned. The alignment processing includes aligning the state snapshots in the same control period according to the timestamp, and performing a holding or rejection strategy on the missing field:

[0263] When a certain traction node is missing phase state information or margin state information in the current period, the holding value of the last valid period is used and marked as a holding input to avoid index mutation caused by single communication loss; when the continuous loss exceeds the allowed duration, the traction node in the period is rejected from the index calculation set and the reason is recorded.

[0264] S4.2: Based on the consistency evaluation index, determine the competition state of each traction node according to a preset competition criterion, wherein the competition state includes a competition inhibition triggering state and a competition inhibition release state;

[0265] Specifically, when multiple traction nodes are parallel traction, competition does not necessarily occur, and the competition state has a phased feature:

[0266] In the early stage of grid connection, a certain degree of parallel traction is allowed when the phase deviation is large to speed up convergence; when approaching the corridor boundary or the closing window period, too strong parallel traction is more likely to cause phase pulling and power circulation. If a fixed threshold is used to directly determine competition, it will lead to inconsistent judgment standards at different stages, resulting in excessive inhibition in the early stage leading to slow convergence or late inhibition not timely enough to cause fluctuations before closing. Therefore, the competition criterion in the embodiment uses a determination logic associated with the phase convergence process, so that the competition state determination can adjust the sensitivity according to the convergence stage, while maintaining the hysteresis characteristics of triggering and release to avoid frequent state switching.

[0267] In the embodiment, the preset competition criterion includes at least the following executable contents:

[0268] One type of criterion is used to determine the competition inhibition triggering state. When the consistency evaluation index meets at least one of the following combination conditions: the phase approaching direction differentiation, the phase advancing pace difference continues to expand, the power response shows an antagonistic split, and the public coupling point disturbance relevance continues to increase, it is determined that the competition inhibition triggering state is in a state;

[0269] Another type of criterion is used to determine the competition inhibition release state. When the consistency evaluation index meets at least one of the following combination conditions: the phase approaching direction is consistent again, the phase advancing pace difference falls within the allowed range, the power response restores the same direction cooperation, and the public coupling point disturbance risk falls and remains stable for several periods, it is determined that the competition inhibition release state is in a state.

[0270] The combination conditions of the competition criterion can be realized in the form of a rule table:

[0271] Different convergence stages correspond to different trigger thresholds and release thresholds, for example, when the traction node phase has entered the corridor allowed range or is close to the corridor boundary, the trigger criterion is more sensitive to the rhythm difference and power confrontation; when the phase deviation is still large, the trigger criterion allows larger rhythm difference but is more sensitive to the common coupling point disturbance risk.

[0272] As can be appreciated by those skilled in the art, the convergence stage can be determined according to the boundary information of the phase corridor object and the relative position relationship of the traction node, without separately introducing additional stage variables.

[0273] It should be noted that the competition suppression trigger state can be understood as a running decision state, which is used to represent that the comprehensive behavior formed by the parallel traction of multiple traction nodes has deviated from the cooperative convergence interval, i.e., the phase advancement, power response or common coupling point disturbance characteristics between the traction nodes show a mutual restraint or superposition amplification trend. In this state, continuing to advance the phase convergence according to the original traction intensity may lead to the accumulation of risks such as phase pulling, power circulation or voltage fluctuation, so it is necessary to restrict the update of the traction edge attribute and structure the constraint of the traction behavior, so that the phase convergence process returns to the controllable interval. This state is not equivalent to traction failure or grid connection failure, but represents a transition stage in which the traction behavior needs to be adjusted from active advancement to limited cooperation in the current window period.

[0274] Correspondingly, the competition suppression release state can be understood as a recovery decision state, which is used to represent that after experiencing the competition suppression update, the phase approaching direction, advancement rhythm and power response between the traction nodes return to the running interval of cooperative consistency, and the measured signal at the common coupling point does not continue to show the disturbance amplification characteristics caused by traction superposition. In this state, the traction intensity limitation, freezing or rollback measures applied to the traction edge can be gradually released, so that the traction node regains greater phase advancement freedom, thereby improving the phase convergence efficiency.

[0275] The determination of the competition suppression release state usually requires that the consistency evaluation index remains stable for a certain number of control periods, so as to avoid frequent switching of traction limitation states due to short-time disturbance, thereby maintaining the structural stability and convergence continuity of the grid connection structure diagram in the later stage of phase convergence.

[0276] S4.3: In the competition suppression trigger state, performing a competition suppression update on the traction node, the competition suppression update comprising: limiting the traction intensity of the edge attribute information of the traction edge between the traction node and the main network logical node, wherein the traction intensity limitation comprises reducing the upper limit of the traction intensity, limiting the traction intensity change rate, freezing the traction intensity update and rolling back the traction intensity;

[0277] Specifically, the competition inhibition update needs to be able to quickly reduce the antagonism between the traction nodes without changing the physical wiring relationship, while avoiding the complete cancellation of traction leading to the loss of the main grid reference constraint in the phase convergence process. The traction edge between the traction node and the logical main grid node in the phase convergence graph is the direct path of the main grid reference injection. Imposing constraints on the edge properties of the traction edge can adjust the traction strength of the traction node at the structural level, so that the traction node changes from active promotion to limited following, thereby inhibiting the mutual traction behavior without changing the identity of the traction node. The traction strength limit adopts a combination of upper limit, speed limit, freezing and rollback, so that the inhibition action has a hierarchical nature.

[0278] The competition inhibition update needs to be able to quickly reduce the antagonism between the traction nodes without changing the physical wiring relationship, while avoiding the complete cancellation of traction leading to the loss of the main grid reference constraint in the phase convergence process. The traction edge between the traction node and the logical main grid node in the phase convergence graph is the direct path of the main grid reference injection. Imposing constraints on the edge properties of the traction edge can adjust the traction strength of the traction node at the structural level, so that the traction node changes from active promotion to limited following, thereby inhibiting the mutual traction behavior without changing the identity of the traction node. The traction strength limit adopts a combination of upper limit, speed limit, freezing and rollback, so that the inhibition action has a hierarchical nature.

[0279] In the present embodiment, after determining that the traction node is in the competition inhibition trigger state, the edge property information of the traction edge of each traction node is updated, and the edge property information at least includes a traction strength parameter, a traction strength change rate constraint parameter, and a freeze / rollback control flag. The process of reducing the upper limit of the traction strength includes tightening the maximum traction effect allowed by the traction edge of the traction node to a limited interval around the current value, so that it will not continue to increase the traction effect even if there is a phase deviation in the subsequent control period; the process of limiting the traction strength change rate includes setting a maximum change step for the increase and decrease of the traction strength, so that the adjustment of the traction effect of the traction node presents a gradual characteristic, avoiding power impact caused by the jump of the traction strength in a single period; the process of freezing the traction strength update includes keeping the traction strength unchanged during the preset freezing period, so that the traction node only maintains the current state without participating in the competition promotion for a short time; the process of rolling back the traction strength includes restoring the traction strength to the traction strength record value in the historical stable window, which can be the interval in which the consistency evaluation index meets the cooperation condition in a certain number of periods before the competition trigger, so that the edge property of the traction edge returns to the stable zone that has been verified.

[0280] As can be appreciated by those skilled in the art, the traction strength record value can be cached as edge property information over time, and the cached value can be directly called when rolling back, without the need for additional calculation.

[0281] S4.4: updating the edge set and edge property information of the phase convergence graph based on the traction node updated by the competition inhibition, to generate the grid-connected structure graph.

[0282] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present application.

Claims

1. A frequency adaptive adjustment method for grid-connected energy storage under weak grid conditions, applied to distributed grid-connected energy storage systems, characterized in that... The grid-type energy storage system includes at least two grid-type energy storage devices and a common coupling point connected in parallel with the grid-type energy storage devices. The method includes: During the islanded operation phase, the node state vector of each grid-type energy storage device is obtained to construct the islanded node state diagram; When the grid connection triggering condition is met, the main grid phase features are extracted based on the measurement signal of the common coupling point, and a logical main grid node is added to the island node state diagram so that the logical main grid node serves as the phase reference of the island node state diagram. The addition of the logical main grid node includes determining the traction node corresponding to the logical main grid node and obtaining a phase convergence diagram. The competition suppression update of the traction node is performed based on the consistency evaluation index of the phase convergence diagram to obtain the grid connection structure diagram, and the frequency parameters of each grid-type energy storage device are adjusted according to the grid connection structure diagram. The traction node is updated by competition suppression based on the consistency evaluation index of the phase convergence diagram to obtain the grid connection structure diagram, including: Obtain the node state vector corresponding to each traction node in the phase convergence graph, and determine the consistency evaluation index based on the node state vector; Based on the consistency evaluation index, the competition status of each traction node is determined by combining the preset competition criteria, wherein the competition status includes the competition suppression triggering state and the competition suppression release state. In the contention suppression triggered state, a contention suppression update is performed on the traction node. The contention suppression update includes: limiting the traction strength of the edge attribute information of the traction edge between the traction node and the logical main network node. The traction strength limit includes reducing the upper limit of traction strength, limiting the rate of change of traction strength, freezing the traction strength update, and rolling back the traction strength. The edge set and edge attribute information of the phase convergence graph are updated based on the updated traction nodes after competition suppression to generate the grid connection structure graph; The consistency evaluation indicators include: The phase convergence consistency index of the traction node relative to the logical main network node is used to characterize the degree of coordination between the phases of each traction node and the phase characteristics of the main network. The power consistency index between traction nodes is used to characterize the low-frequency oscillation trend and unidirectional consistency of the active power output of each traction node.

2. The adaptive frequency adjustment method for energy storage under weak grid conditions according to claim 1, characterized in that, The node state vector includes at least one of the following: Phase state information used to characterize the internal phase reference of a grid-type energy storage device; Power state information used to characterize the active power output of grid-type energy storage devices; Reactive state information used to characterize the reactive power output of grid-type energy storage devices; Margin state information used to characterize the operating margin of grid-type energy storage devices, wherein the margin state information includes current limiting trigger state, power limiting trigger state, DC side voltage margin, and modulation saturation degree; The parameter status information used to characterize the current values ​​of the frequency parameters of the corresponding grid-type energy storage device, wherein the frequency parameters include droop coefficient, virtual inertia parameter and damping parameter.

3. The adaptive frequency adjustment method for energy storage under weak grid conditions according to claim 2, characterized in that, The construction of the isolated node state diagram includes: Obtain the node identifier of each grid-type energy storage device and the node state vector corresponding to the node identifier, and establish an island node set based on each grid-type energy storage device; The coupling relationship between island nodes is determined based on the access location of each grid-type energy storage device, and the island edge set is established using the coupling relationship as the edge. Configure edge attribute information for each edge in the island edge set, wherein the edge attribute information includes coupling weights that characterize the coupling strength between island nodes, wherein the coupling weights are calculated based on the node state vectors; Based on the set of isolated nodes, the set of isolated edges, and the edge attribute information of each edge, a state graph of the isolated nodes is generated, and the edge attribute information of the state graph of the isolated nodes is periodically updated during the operation phase of the isolated island.

4. The adaptive frequency adjustment method for energy storage under weak grid conditions according to claim 1, characterized in that, Extracting main network phase features based on the measurement signal at the common coupling point includes: Acquire voltage and current measurement signals at the common coupling point, wherein the voltage measurement signals include instantaneous values ​​of three-phase voltages, phase voltages, and line voltages, and the current measurement signals include instantaneous values ​​of three-phase currents; The voltage measurement signal and the current measurement signal are preprocessed to obtain a stable measurement signal for phase analysis. The preprocessing includes filtering, fundamental component extraction, harmonic suppression, and abnormal change suppression. Based on the stable measurement signal, the instantaneous phase information and phase change trend information corresponding to the common coupling point are extracted as phase candidate features of the current main network; Obtain the main network historical operation data corresponding to the common coupling point, wherein the main network historical operation data includes at least one of historical phase information, historical frequency information and historical voltage operation information; The phase candidate features are verified based on the historical operation data of the main network. If the verification matches the preset adaptation conditions, the phase candidate features are determined to be the main network phase features. If they do not match, the phase candidate features are iteratively corrected based on the historical operation data of the main network until they match the adaptation conditions.

5. The adaptive frequency adjustment method for energy storage under weak grid conditions according to claim 4, characterized in that, Add logical master network nodes to the isolated node state diagram, including: Based on the main network phase characteristics, a phase corridor object is constructed, wherein the phase corridor object is used to characterize the allowable range and validity status of the main network phase reference, and the validity status is determined by the consistency verification results between the main network phase characteristics and the main network historical operation data; Based on the phase corridor object and the node state vectors of each grid-type energy storage device, traction compatibility determination is performed on each grid-type energy storage device to obtain a set of traction-compatible nodes. When the traction-compatible node set is not empty, a logical main network node is generated in the island node state diagram based on the phase corridor object, wherein the node attributes of the logical main network node include the phase corridor object; A graph reconstruction plan is generated based on the traction-compatible node set, wherein the graph reconstruction plan is used to reconstruct the edge set and edge attribute information of the isolated node state graph; The state graph of the isolated node containing the logical main network node is reconstructed according to the graph reconstruction plan to obtain the phase convergence graph.

6. The adaptive frequency adjustment method for energy storage under weak grid conditions according to claim 5, characterized in that, Based on the phase corridor object and the node state vectors of each grid-type energy storage device, a traction compatibility determination is performed on each grid-type energy storage device, including: The node state vectors of each grid-type energy storage device are transmitted to the common coupling node, and phase state information, margin state information and parameter state information are extracted through the common coupling node. Based on the phase state information, the relative positional relationship between the internal phase reference of the grid-type energy storage device to be judged and the phase corridor object is determined, wherein the relative positional relationship includes two states: being within the allowable range of the phase corridor object and being outside the allowable range of the phase corridor object. When the relative positional relationship is within the allowable range of the phase corridor object, the grid-type energy storage device to be determined is identified as a traction-compatible node; When the relative positional relationship indicates that it is outside the allowable range of the phase corridor object, the margin status information is used to determine whether the grid-type energy storage device to be determined meets the operating margin constraint conditions. When the operating margin constraint is met, the phase convergence capability parameter is determined based on the parameter status information, and the reachability of the internal phase reference of the grid-type energy storage device to be judged to enter the allowable range of the phase corridor object is judged according to the phase convergence capability parameter. The phase convergence capability parameter is positively correlated with the current value of the frequency parameter. If it is determined that the energy storage device to be determined is reachable within a preset time window, then the grid-type energy storage device is a traction-compatible node. When the operating margin constraint is not met and the determination is made that the device is unreachable within the preset time window, the grid-type energy storage device to be determined is a non-traction compatible node.

7. The adaptive frequency adjustment method for energy storage under weak grid conditions according to claim 5, characterized in that, A graph reconstruction plan is generated based on the aforementioned traction-compatible node set, including: A traction edge is established between the logical main network node and each grid-type energy storage device node in the traction compatible node set, and the traction compatible node is determined as the traction node. Obtain the non-traction nodes connected to each traction node in the state graph of the isolated node, determine the isolated edges between the traction nodes and the non-traction nodes, and generate a set of edges to be reconstructed; The updated edge attribute information is configured for each edge in the set of edges to be reconstructed to obtain the traction propagation edge set, wherein the updated edge attribute information is determined by calculating the phase convergence between the node state vector of the corresponding traction node and the phase feature of the main network. Based on the node state vector of the non-traction node, the edge attribute information of the traction propagation edge set is corrected to obtain the graph reconstruction plan.

8. A frequency adaptive adjustment system for grid-connected energy storage under weak grid conditions, used to implement the frequency adaptive adjustment method for grid-connected energy storage under weak grid conditions as described in any one of claims 1-7, characterized in that, The system includes: The data acquisition module is used to acquire the node state vectors of each grid-type energy storage device, and to construct an island node state diagram based on the node state vectors during the island operation phase. The main network phase feature extraction module extracts the main network phase features based on the measurement signals of the common coupling point, and performs consistency verification by combining the main network historical operation data to construct the phase corridor object; The graph reconstruction module performs traction compatibility determination based on the phase corridor object and the node state vector to obtain a set of traction compatible nodes, and obtains a phase convergence graph. The frequency parameter adjustment module updates the traction node to suppress competition based on the consistency evaluation index of the phase convergence diagram to obtain the grid connection structure diagram, and adaptively adjusts the frequency parameters of each grid-type energy storage device according to the grid connection structure diagram.

Images