Micro-grid monitoring method and system based on coordinated control and safety protection linkage

Abstract

This application relates to the field of microgrid monitoring technology, specifically disclosing a microgrid monitoring method and system based on coordinated control and safety protection linkage. The method includes collecting spatial layout parameters, control parameters, and time-series operational data streams of microgrid equipment to establish a topology model; selecting cooperative nodes strongly coupled with key monitoring points, extracting state feature vectors representing system behavior, and performing similarity calculations to identify candidate related event pairs originating from the same disturbance or control process; determining valid related event pairs through dual verification and connecting them in chronological order to form a disturbance propagation event chain; analyzing node response signals to obtain linear response growth rate and nonlinear saturation intensity parameters; dividing collaborative handling units based on state trend discrimination results, and applying fine-grained coordinated control to units containing risky operating points. This method achieves full-process monitoring of microgrid multi-dimensional perception, dynamic evaluation, and targeted optimization, improving its dynamic stability and disturbance robustness.

Description

Technical Field

[0001] This application relates to the field of microgrid monitoring technology, and in particular to a microgrid monitoring method and system based on coordinated control and security protection linkage. Background Technology

[0002] During the operation of a microgrid, various elements such as distributed power sources, energy storage devices, load centers and grid connection interfaces are widely distributed in different spatial locations, and there are complex energy flow and command transmission relationships between them. Its operating status is affected by a variety of factors such as fluctuations in new energy output, sudden load changes and network topology switching. Currently, traditional microgrid monitoring methods largely rely on single-point or local data acquisition and independent analysis, lacking the ability to uniformly model the entire network's spatial layout and energy / command transmission paths. This makes it difficult to capture the generation and propagation of cross-node disturbances in a timely manner, resulting in an inability to accurately obtain the energy and information coupling characteristics between different regions. Furthermore, existing monitoring methods often process multi-source heterogeneous time-series data streams in isolation, lacking intelligent screening and feature fusion mechanisms for strongly coupled nodes at key monitoring points. This leads to insufficient accuracy in identifying system behavior patterns and an inability to effectively distinguish between disturbances from the same source and independent anomalies. In addition, the inertia, time delay, and convergence / divergence characteristics exhibited by microgrids under small disturbances, as well as the nonlinear saturation performance under large-signal conditions constrained by resources such as energy storage capacity, inverter power, and line thermal stability limits, are difficult to characterize using conventional linearization models. This creates blind spots in judging the state development trends of operating points and makes it difficult to quantitatively predict potentially high-risk states. Summary of the Invention

[0003] Based on this, the embodiments of this application provide a microgrid monitoring method and system based on coordinated control and security protection linkage, aiming to solve the technical problems in related technologies where, in the scenario of multi-regional and multi-source heterogeneous equipment collaborative operation of microgrids, traditional monitoring methods cannot accurately capture the generation and propagation process of cross-node disturbances, lack intelligent screening and feature fusion mechanisms for strongly coupled collaborative nodes at key monitoring points, resulting in low accuracy in identifying homogeneous disturbances and independent abnormal events, and incomplete construction of disturbance propagation chains.

[0004] In a first aspect, embodiments of this application provide a microgrid monitoring method based on coordinated control and security protection linkage, including: Collect spatial layout parameters and control parameters of equipment in various areas of the microgrid, and combine them with the time-series operation data streams of multiple monitoring nodes to establish a topology model covering the entire network; Based on the aforementioned topology model, collaborative nodes strongly coupled with key monitoring points are selected. State feature vectors representing system behavior are extracted from the original data streams of each node and the fused data streams of the collaborative nodes. Similarity calculations are performed on the state feature vectors to determine candidate related event pairs originating from the same disturbance or control process. The candidate associated event pairs are double-checked to obtain valid associated event pairs, and then connected in chronological order to form a disturbance propagation event chain that reflects the transmission of disturbances or control processes across nodes. The internal response signals of each node in the disturbance propagation event chain are analyzed in segments. The energy transfer characteristics are identified by using external excitation fluctuations and internal response fluctuation signals. The real part of the transfer characteristics is obtained by phase unwinding and equivalent time delay mean correction, and converted into the linear response growth rate. The steady-state amplitude information in the disturbance propagation event chain is extracted to determine the nonlinear saturation intensity parameters. The linear response growth rate and nonlinear saturation intensity parameters are input into the state trend discrimination model. Based on the discrimination results output by the model, the microgrid is divided into several collaborative processing units according to function and control level. Fine-grained coordinated control is applied to the collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands.

[0005] In some embodiments, establishing a topology model covering the entire network includes: A basic graph structure is constructed using devices in each area of ​​the microgrid as nodes and physical wiring and logical control links between devices as edges. The spatial layout parameters and control parameters are embedded in the node attributes, and the energy transfer direction, capacity limit, command transmission protocol and communication delay index are embedded in the edge attributes. The time-series running data stream is mapped to corresponding nodes and edges to form a dynamic attribute layer, resulting in the topology model that simultaneously describes the physical location of the device, the energy transfer path, and the command transfer path.

[0006] In some embodiments, the selection of collaborative nodes strongly coupled with key monitoring points based on the topology model includes: In the topology model, identify key monitoring points that have a significant impact on the overall network operation status. These key monitoring points include at least one of the following: power exchange hubs, the beginning of important load feeders, grid connection interfaces, and the main control location of energy storage systems. For each key monitoring point, the coupling degree between it and other nodes in the network on the energy transmission path and command transmission path is calculated based on three indicators: the number of paths, the total transmission capacity of the paths, and the weighted sum of command response delay. Nodes with a coupling degree higher than a preset threshold are selected as cooperative nodes.

[0007] In some embodiments, extracting state feature vectors characterizing system behavior from the original data streams of each node and the fused data streams of collaborating nodes includes: For a single node or a collaborative node, the running data segment is extracted according to the time window, and the window length is set according to the duration of the typical disturbance or control process to be captured. Multidimensional features are extracted within the window, including power change morphology features, voltage fluctuation fingerprint features, and control response curve features. In the extraction process, configurable feature templates and sliding calculations are used to obtain state feature vectors with consistent structures under the same operating conditions.

[0008] In some embodiments, the step of performing double verification on the candidate associated event pairs to obtain valid associated event pairs, and then concatenating them in chronological order to form a disturbance propagation event chain reflecting the transmission of disturbances or control processes across nodes, includes: For each candidate associated event pair, perform temporal continuity verification and spatial topological causal reachability verification in sequence. If both verifications pass, the candidate associated event pair is confirmed as a valid associated event pair. All valid related event pairs are arranged in ascending order according to the timestamp of the first event. The connectivity of adjacent event pairs is determined in turn. Adjacent event pairs that satisfy temporal continuity and spatial topological reachability are merged to form one or more disturbance propagation event chains.

[0009] In some embodiments, the sequential execution of temporal continuity verification and spatial topological causal reachability verification on candidate associated event pairs includes: Based on the occurrence time of the first event in the candidate related event pair, a time neighborhood width is set to determine whether the occurrence time of the second event falls within the time neighborhood. At the same time, it is verified whether the two events meet the preset time sequence dependency rules. If both are met, the time continuity check passes. In the topology model, locate the two nodes to which the candidate associated event pair belongs, and explore from the source node to the target node. If there is at least one path that satisfies causal directionality and is allowed by the transmission medium and permission configuration, then the spatial topology causal reachability verification passes.

[0010] In some embodiments, the segmented analysis of the internal response signals of each node in the disturbance propagation event chain, and the identification of energy transfer characteristics using external excitation fluctuations and internal response fluctuation signals, includes: The time-series response signals of the acquisition node before and after the event are collected. The response signals are divided into initial transient segment, transition segment and steady-state segment by the segmentation boundary with the abrupt change point of the signal amplitude change rate and statistical variance. By aligning each signal segment with the corresponding external excitation fluctuation signal in the time domain, the energy transfer characteristics are obtained through cross-correlation peak detection and sliding window covariance analysis.

[0011] In some embodiments, obtaining the real part of the transfer characteristic through phase unwinding and equivalent time delay mean correction includes: Continuous correction is performed for phase jumps caused by sampling or transmission. When a phase change is detected to exceed a preset reasonable range, the compensation amount is accumulated or decreased according to the trend of phase change before and after, so that the phase curve is restored to a smooth and monotonous state. The difference distribution of the peak times of excitation and response within each segment is statistically analyzed. After removing outliers, a weighted average is taken. This average is used as the equivalent transmission delay of the node and is used as the horizontal axis reference for correcting the transfer characteristic curve, thus obtaining the real part of the corrected transfer characteristic that reflects the system inertia and transmission delay.

[0012] In some embodiments, dividing the microgrid into several collaborative processing units according to function and control levels, and applying fine-grained coordinated control authority to collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands, includes: The system continuously monitors the rate of change of key performance indicators and dynamically adjusts the frequency of control command issuance and the trigger threshold of protection actions based on the rate of change. The key performance indicators include frequency offset rate, voltage sag amplitude, and communication delay fluctuation. Specifically, when the rate of change of key performance indicators enters the preset first fluctuation range, the frequency of control command issuance is increased while the threshold for triggering protective actions is lowered; when the rate of change of key performance indicators is in the second fluctuation range, the frequency of control command issuance is reduced while the threshold for triggering protective actions is increased.

[0013] Compared with the prior art, the technical solution provided in the first aspect of this application includes at least the following beneficial effects or advantages: The method provided in this application, by constructing a space-energy coupling topology model covering the entire network, achieves unified modeling and real-time sensing of multi-source heterogeneous information such as distributed power sources, energy storage, loads, and grid connection interfaces. This overcomes the limitations of traditional single-point monitoring in capturing distributed interactive effects, making the analysis of disturbance propagation paths and causal relationships more refined and reliable. Through in-depth segmented analysis and equivalent time delay correction of signals within the event chain, it not only obtains linear response characteristics reflecting system inertia and transmission delay but also quantifies the degree of resource constraint and nonlinear saturation intensity under large signal conditions, enhancing the model's universality and identification accuracy for strong and weak disturbance scenarios. Furthermore, by fusing linear growth rate and nonlinear saturation intensity… The state trend discrimination model can predict in advance whether the operating point will slide into the limit cycle oscillation or instability, and accurately locate high-risk operating points, realizing the transformation from passive response to active early warning, and reducing the probability of cascading failures caused by sudden instability. In addition, based on the division of functional and control-level collaborative processing units and fine-grained coordinated control strategies, combined with dynamic commands and threshold adjustment mechanisms based on the rate of change of key performance indicators, the balance between the timeliness and consistency of control and protection actions is ensured. This avoids secondary disturbances caused by overly aggressive regulation and prevents the risk solidification caused by slow response, thereby improving the overall dynamic stability, disturbance resistance robustness and intelligent coordination level of security protection of the microgrid.

[0014] Secondly, embodiments of this application provide a microgrid monitoring system based on coordinated control and security protection linkage, comprising: The topology module is configured to collect spatial layout parameters and control parameters of equipment in various areas of the microgrid, and combine them with the time-series operation data streams of multiple monitoring nodes to establish a topology model covering the entire network. The associated event filtering module is configured to select collaborative nodes that are strongly coupled with key monitoring points based on the topology model, extract state feature vectors representing system behavior from the original data streams of each node and the fused data streams of the collaborative nodes, and perform similarity calculations on the state feature vectors to determine candidate associated event pairs originating from the same disturbance or control process. The verification module is configured to perform double verification on the candidate associated event pairs to obtain valid associated event pairs, and to connect them in chronological order to form a disturbance propagation event chain that reflects the transmission of disturbances or control processes across nodes. The correction module is configured to perform segmented analysis of the internal response signals of each node in the disturbance propagation event chain, identify energy transfer characteristics using external excitation fluctuations and internal response fluctuation signals, obtain the real part of the transfer characteristics through phase unwinding and equivalent time delay mean correction, convert it into a linear response growth rate, and extract steady-state amplitude information in the disturbance propagation event chain to determine nonlinear saturation intensity parameters. The dynamic adjustment module is configured to input the linear response growth rate and nonlinear saturation intensity parameters into the state trend discrimination model. Based on the discrimination results output by the model, the microgrid is divided into several collaborative processing units according to function and control level. Fine-grained coordination control is applied to the collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands.

[0015] It is understood that the beneficial effects of the technical solution provided in the second aspect above can be found in the relevant description in the first aspect above, and will not be repeated here.

[0016] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a microgrid monitoring method based on coordinated control and security protection linkage provided in an embodiment of this application; Figure 2 This is a flowchart illustrating the data processing of the spatiotemporal coupling verification model provided in an embodiment of this application. Figure 3 A flowchart illustrating the data processing in the state trend discrimination model provided in this application embodiment; Figure 4 A structural block diagram of a microgrid monitoring system based on coordinated control and security protection linkage provided in this application embodiment; Figure 5 This is a structural block diagram of a server in an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings, which illustrate several embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of this application will be thorough and complete.

[0020] It should be noted that, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0021] Please see Figures 1 to 3 This embodiment provides a microgrid monitoring method based on coordinated control and security protection linkage, specifically including steps S10 to S50.

[0022] Step S10: Collect the spatial layout parameters and control parameters of the equipment in each area of ​​the microgrid, and combine them with the time-series operation data stream of multiple monitoring nodes to establish a topology model covering the entire network; In some embodiments, key element information of the physical and logical structure of the microgrid is systematically collected from different regions. These regions include, but are not limited to, distributed power source access points, energy storage device deployment locations, load center clusters, and grid connection interface nodes. The collected data covers the spatial layout parameters and control parameters of each device within these regions. Specifically, spatial layout parameters include geographical coordinates, electrical connection port numbers, relative distances, and laying medium characteristics. Control parameters include rated power, adjustment range, response time constant, control mode identifier, and communication protocol version with the upper-level control system. The data collection process is performed according to a predefined device list and parameter table. For each device, the latest valid data is obtained through field sensor interfaces, configuration database export, or remote communication sessions. Retrying and missing data are performed for communication interruptions or data anomalies to ensure information integrity.

[0023] Furthermore, it aggregates time-series operational data streams from multiple monitoring nodes, which can be distributed across local control units of critical equipment, regional centralized monitoring stations, and the network-wide operational data center. Data types include instantaneous power, RMS voltage, current waveform characteristics, frequency deviation, energy storage state of charge, grid-connected power exchange, and control command transmission and reception records. Upon access, data from each node undergoes time-base unification processing: using a unified network clock source as a reference, deviation estimation and linear compensation are performed on the received timestamps of each node to eliminate timing misalignments caused by crystal oscillator drift or transmission delays, ensuring strict alignment of data from different sources on the same timeline.

[0024] Furthermore, based on the acquisition of spatial layout and control parameters as well as time-series operation data, a topology model covering the entire network is established. The establishment of the topology model includes: constructing a basic graph structure with devices in each area of ​​the microgrid as nodes and physical wiring and logical control links between devices as edges; embedding spatial layout parameters and control parameters in the node attributes; and embedding energy transfer direction, capacity limit, command transmission protocol, and communication delay indicators in the edge attributes; mapping the time-series operation data stream to the corresponding nodes and edges to form a dynamic attribute layer, resulting in a topology model that simultaneously describes the physical location of the devices, the energy transfer path, and the command transmission path.

[0025] In one example, the topology model allows any node or edge to be retrieved at runtime, showing its upstream and downstream relationships and current load status, thus providing a structured panoramic view for subsequent analysis. During the topology model construction process, to further clarify the attribute integration method of nodes and edges and the numerical implementation of dynamic mapping, the corresponding calculation ideas and numerical calculations are given in conjunction with an example: For instance, consider a distributed power node N1 with spatial layout parameters as geographical coordinates (120.15°E, 30.28°N), electrical connection port number P3, a relative distance of 85m to the adjacent energy storage node N2, and a cross-linked polyethylene cable as the laying medium; control parameters are rated power 20kW, adjustment range ±10kW, response time constant 0.8s, control mode identifier constant power tracking, and communication protocol version Modbus-TCPV1.2. Meanwhile, the edge E1 physically connecting N1 and N2 has an energy transfer direction from N1 to N2, a capacity limit of 25kW, a command transmission protocol of IEC61850GOOSE, and a measured communication latency of 12ms.

[0026] In the time-series data mapping stage, at a certain acquisition period t0=12:00:00.000 (unified clock source time), the instantaneous power reading of N1 is 18.6kW, the effective voltage value is 400.5V, and the state of charge of energy storage node N2 is 72%. The corresponding edge E1 has an energy flow rate of 17.9kW during this period, with 3 command interactions and an average command response delay of 11.5ms. At this time, the dynamic attribute layer update logic can be considered as real-time writing to the node attribute table and edge attribute table: the power attribute of node N1 is assigned a value of 18.6, the voltage attribute is assigned a value of 400.5, the energy flow rate attribute of edge E1 is assigned a value of 17.9, and the average command response delay attribute is assigned a value of 11.5. If it is necessary to quantify the node load rate, it can be calculated using the formula: load rate = current power / rated power × 100%. Substituting the values, the load rate is calculated as 18.6kW / 20kW × 100% = 93%. Similarly, the power flow load factor of edge E1 is 17.9kW / 25kW×100%=71.6%. Through the above calculations and assignments, the topology model not only retains the static connection and parameter characteristics of N1 and N2, but also intuitively reflects the real-time operating status of N1 near full capacity and E1 near medium-high load at time t0. This allows subsequent analysis modules to directly retrieve the upstream and downstream relationships of nodes and the current load status, forming a structured panoramic view.

[0027] Step S20: Based on the topology model, select cooperative nodes that are strongly coupled with key monitoring points, extract state feature vectors representing system behavior from the original data streams of each node and the fused data streams of the cooperative nodes; and perform similarity calculation on the state feature vectors to determine candidate related event pairs originating from the same disturbance or control process. In some embodiments, key monitoring points that significantly affect the overall network operation are identified in the topology model. These points are typically located at power exchange hubs, the beginning of important load feeders, grid connection interfaces, and the main control location of energy storage systems. For each key monitoring point, the coupling degree between it and other nodes in the network on the energy transmission path and command transmission path is analyzed. The coupling degree is calculated based on a comprehensive evaluation of three indicators: the number of paths, the total transmission capacity of the paths, and the weighted sum of command response delays. Nodes with a coupling degree higher than a preset threshold are selected as cooperative nodes to ensure that the selected nodes can generate significant linkage responses with the monitored key points under disturbances or control actions.

[0028] Subsequently, state feature vectors characterizing the system's behavior patterns are extracted from the original data streams of each node and the fused data streams of collaborating nodes. Specifically, for a single node or collaborating node, a time window is used to extract the running data segment, with the window length set according to the duration of the typical disturbance or control process to be captured. Within the window, multidimensional features are extracted, including power change morphology features (such as step, ramp, and oscillation decay profiles), voltage fluctuation fingerprint features (such as high-frequency ripple density, short-time drop slope, and harmonic distortion distribution profiles), and control response curve features (such as the interval distribution between command arrival and execution times, and the statistical values ​​of overshoot and steady-state error). The extraction process uses configurable feature templates and a sliding calculation method to ensure that different nodes can obtain structurally consistent and comparable feature vectors under the same operating conditions.

[0029] In some embodiments, a deep metric learning algorithm is used to calculate the similarity of state feature vectors from different sources to filter candidate associated event pairs originating from the same disturbance or control process. For example, the deep metric learning algorithm internally includes a feature encoding network and a similarity discriminant network. The feature encoding network takes the original feature vectors as input and maps the data to a metric space through multiple nonlinear transformations, causing features of similar processes to cluster in the space and features of dissimilar processes to separate. The similarity discriminant network calculates the distance metric between pairwise vectors based on the mapped features. The distance calculation can use a contrastive loss optimization based on spatial positional relationships to output a similarity score.

[0030] During algorithm execution, the feature vectors of all nodes are first transformed into a metric space representation through an encoding network. Then, pairing is performed within a set similarity threshold to generate a set of candidate associated event pairs, and the source node, time window, and similarity score of the paired vectors are recorded. This set is provided for subsequent stages for spatiotemporal coupling verification and event chain construction.

[0031] In one example, to more intuitively illustrate the numerical implementation of coupling degree calculation and similarity judgment in the process of selecting cooperative nodes and calculating feature vector similarity, the corresponding calculation ideas and substitution calculations are given with examples. In the topology model, a key monitoring point is the grid-connected interface node K, which has three energy transmission paths and two command transmission paths with other nodes: the transmission capacities of the three energy paths are 30kW, 25kW, and 20kW, respectively, with a corresponding weight factor of 1; the response delays of the two command paths are 8ms and 12ms, respectively, with weight factors of 1 and 1.5 (the shorter the delay, the greater the weight, which is set here to reflect the priority of fast response paths). The coupling degree is calculated based on a comprehensive evaluation of three indicators: the number of paths, the total transmission capacity of the paths, and the weighted sum of command response delays. It can be formalized as coupling degree = α·number of paths + β·(total capacity / reference capacity) + γ·(1 / weighted sum of delays), where α, β, and γ are normalization coefficients, all set to 1 for ease of example, and the reference capacity is taken as the maximum capacity of a single path, 30kW. Substituting the values, we get: number of paths = 5, total capacity = 75kW, total capacity ratio = 2.5, weighted latency sum = 26ms, its reciprocal = 0.03846, coupling degree = 7.54. If the preset threshold is 6, then the coupling degree of this node meets the standard and can be used as a cooperative node.

[0032] Furthermore, in the feature extraction stage, let the power change morphology feature values ​​extracted by node K within a 10s time window be step intensity 0.82 and ramp slope 0.45, the voltage fluctuation fingerprint feature values ​​be high-frequency ripple density 15 and short-term drop slope -80V / s, and the control response curve feature values ​​be the average interval between command arrival and execution 20ms, overshoot 5%, and steady-state error 1%, forming the original feature vector Fk=[0.82,0.45,15,-80,20,5,1]. The feature encoding network after deep metric learning is mapped to the metric space to obtain the vector Fk'=[0.61,-0.33,0.78]; the original feature vector Fm of another cooperative node M is encoded in the same way to obtain Fm'=[0.59,-0.35,0.81].

[0033] Furthermore, the similarity discrimination network calculates the distance between the two entities using the Euclidean distance approach: Δx = 0.61 - 0.59 = 0.02, Δy = -0.33 - (-0.35) = 0.02, Δz = 0.78 - 0.81 = -0.03. If the similarity score is defined as S = 1 - D (normalized to the [0,1] interval), then S ≈ 0.9588. Setting a similarity threshold of 0.90, since S is higher than the threshold, events K and M within the current time window are paired as candidate associated event pairs. The source nodes K and M, the time window of 10 seconds, and the similarity of 0.9588 are recorded for subsequent spatiotemporal coupling verification. This numerical example demonstrates that coupling calculation can effectively quantify the linkage strength between nodes. The feature vector, after encoding and distance measurement, can objectively reflect the similarity of responses of different nodes under the same disturbance or control action, thereby achieving reliable screening of candidate associated events.

[0034] Step S30: Perform double verification on the candidate associated event pairs to obtain valid associated event pairs, and connect them in chronological order to form a disturbance propagation event chain that reflects the transmission of disturbances or control processes across nodes; In this step, the candidate related event pair set generated in step S20 undergoes dual verification. This verification process is based on a spatiotemporal coupling verification model, aiming to determine whether the event pair has a real physical or logical connection from two dimensions: temporal continuity and spatial topological causal reachability. This filters out false connections and event combinations that co-occur accidentally. Specifically, temporal continuity verification and spatial topological causal reachability verification are performed on the candidate related event pair in sequence. The candidate related event pair is confirmed as a valid related event pair only if both verifications pass. All valid related event pairs are then sorted in ascending order by the timestamp of their first event, and the connection between adjacent event pairs is determined sequentially. The connectivity test merges adjacent event pairs that satisfy temporal continuity and spatial topological reachability to form one or more disturbance propagation event chains. Specifically, based on the occurrence time of the first event in a candidate associated event pair, a temporal neighborhood width is set to determine whether the occurrence time of the second event falls within this temporal neighborhood. Simultaneously, it verifies whether the two events conform to preset temporal dependency rules; if both are satisfied, the temporal continuity check passes. In the topology model, the two nodes belonging to the candidate associated event pair are located, and an exploration is performed from the source node to the target node. If at least one path exists that satisfies causal directionality and is allowed by the transmission medium and permission configuration, the spatial topological causal reachability check passes.

[0035] In some embodiments, combined with Figure 2 As shown, for each event pair to be verified, their timestamp information and spatial location identifiers are extracted. The time axis continuity verification part adopts the sliding time window comparison method, that is, taking the occurrence time of the first event as the benchmark, setting a reasonable time neighborhood width, checking whether the occurrence time of the second event falls within the neighborhood, and examining whether there is a time sequence dependency between the two events as specified by business rules or physical mechanisms, such as control commands must respond before the controlled object, and fault occurrence must occur before protection action triggering, etc. If the timing logic and time interval both meet the predefined conditions, the time continuity is determined to be valid.

[0036] Furthermore, the spatial topological causal reachability verification part relies on the system's preset spatial connection graph. This graph uses nodes to represent physical devices or logical units and directed or undirected edges to represent possible causal action paths. The processing logic is as follows: locate the nodes to which two events belong in the connection graph, run a path search algorithm, which supports a hybrid strategy of breadth-first search and cost heuristics. Starting from the source node, it explores adjacent nodes layer by layer until reaching the target node, and accumulates the results of transmission feasibility and direction constraint verification on the path. If there is at least one path that satisfies causal directionality and is allowed by the transmission medium and permission configuration, then the spatial topological causal reachability is determined to be valid. The candidate event pair is confirmed as a valid associated event pair if and only if both temporal continuity and spatial causal reachability are passed. The confirmed valid associated event pairs are then concatenated and integrated in chronological order.

[0037] Furthermore, all valid event pairs are sorted in ascending order based on the timestamp of the first event. Adjacent event pairs are then checked for temporal and spatial connectivity; that is, the second event of one event pair and the first event of the next event pair must satisfy the continuity and reachability requirements in both time and space. If they are connectable, they are merged into a longer event sequence; otherwise, they are preserved as independent sub-chains as breakpoints. This connectivity check and merging process is repeated until no new connectable event pairs are generated, thus forming one or more cross-node transmission trajectories of disturbance or control processes, i.e., disturbance propagation event chains. This event chain faithfully reflects the dynamic process of disturbance or control signals being sequentially transmitted through multiple nodes within the system.

[0038] For example, a method for determining the time and space dimensions during the construction of a dual-verification and disturbance propagation event chain, as well as the data processing logic for event cascading, are provided. Assume the candidate associated event pair set contains event pairs (E1, E2) and (E3, E4). E1 occurs at t1=12:00:01.200, belonging to the grid-connected interface G; E2 occurs at t2=12:00:01.450, belonging to the energy storage main controller B; E3 occurs at t3=12:00:01.480, belonging to the load feeder start L; and E4 occurs at t4=12:00:01.720, belonging to the protection relay R. The time axis continuity verification uses a sliding time window comparison method. That is, taking the time of the first event t1 as the benchmark, a time neighborhood width Δt=0.5s is set, and the subsequent event times are checked to see if |t2-t1|≤Δt and conform to the business timing rules. The calculated value |t2-t1| = 0.250s, which is less than 0.5s. Since the rule requires that control or disturbance events occur before responses, if E1 is a disturbance trigger and E2 is an energy storage response, the timing logic is valid, and the time continuity judgment passes. Similarly, |t3-t1| = 0.280s ≤ 0.5s, |t4-t3| = 0.240s ≤ 0.5s, and E3 is a load-side anomaly and E4 is a relay protection action, also satisfying the timing rule, and the time continuity is valid in both cases.

[0039] Furthermore, the spatial topological causal reachability verification is based on a pre-defined directed spatial connection graph. It is assumed that there is an allowed energy / command transmission path P1 from G to B, a path P2 from B to L, and a path P3 from L to R. The path search algorithm employs a hybrid strategy of breadth-first search and cost-based heuristics, exploring adjacent nodes layer by layer from the source node G. At P1, the cumulative transmission feasibility verification is passed (the medium is a power cable, and permission configuration allows it), and the direction from G to B conforms to causal directionality. At P2, the path from B to L is also feasible and the direction is correct. At P3, the path from L to R also satisfies the direction and medium conditions. Since each path satisfies both transmission feasibility and direction constraints, the spatial topological causal reachability is determined to be valid. Therefore, event pairs (E1, E2) and (E3, E4) both pass both temporal and spatial verifications, confirming them as valid associated event pairs.

[0040] Furthermore, the valid event pairs are arranged in ascending order of the time of the first event as [(E1,E2),(E3,E4)]. The connectivity of adjacent event pairs is checked: E2 time t2=12:00:01.450, node B; E3 time t3=12:00:01.480, node L, with a time difference of |t3-t2|=0.030s. If the time neighborhood is still 0.5s, then continuity is satisfied. Spatially, B and L have a direct path P2 in the graph, with the direction from B to L, satisfying reachability, and therefore can be connected. (E1,E2) and (E3,E4) are merged into the sequence E1→E2→E3→E4, forming a disturbance propagation event chain, reflecting the cross-node transmission trajectory of the disturbance from the grid connection interface, through the energy storage main control response, load feeder anomaly, to the relay protection action. This numerical example shows that temporal continuity can be determined by |t2-t1|≤Δt and conforming to the business sequence rules, spatial reachability can be confirmed by path search and direction verification, and event connection is recursively merged according to time difference and path existence, ultimately generating a complete chain description of system disturbance propagation.

[0041] Step S40: Segmentally analyze the internal response signals of each node in the disturbance propagation event chain, identify the energy transfer characteristics using external excitation fluctuations and internal response fluctuation signals, obtain the real part of the transfer characteristics through phase unwinding and equivalent time delay mean correction, convert it into the linear response growth rate, and extract the steady-state amplitude information in the disturbance propagation event chain to determine the nonlinear saturation intensity parameters. In this step, the time-series response signals of the acquisition node before and after the event are collected. Using the abrupt change points of the signal amplitude change rate and statistical variance as segment boundaries, an adaptive threshold segmentation method is employed to divide the response signals into initial transient, transition, and steady-state segments. Each segment's signal is aligned in the time domain with the corresponding external excitation fluctuation signal. Energy transfer characteristics are obtained through cross-correlation peak detection and sliding window covariance analysis. The real part of the transfer characteristics is obtained after phase unwinding and equivalent time delay mean correction. Specifically, this includes: continuously correcting phase jumps caused by sampling or transmission; when a phase change exceeds a preset reasonable range, accumulating or decreasing compensation based on the trend of phase changes before and after, restoring the phase curve to smoothness and monotonicity; statistically analyzing the difference distribution between the excitation and response peak times within each segment, removing outliers, and taking a weighted average. This average is used as the equivalent transmission time delay of the node and is used as the horizontal coordinate reference for correcting the transfer characteristic curve, yielding the corrected real part of the transfer characteristics reflecting system inertia and transmission delay.

[0042] In some embodiments, for each node in the chain, its time-series response signal before and after the event is collected, and the signal interval is divided by an adaptive threshold segmentation method, that is, the point of change of signal amplitude and statistical variance is used as the segment boundary, and the overall response is divided into an initial transient segment, a transition segment and a steady-state segment. For each segment of signal, time-domain alignment comparison is performed with the corresponding external excitation fluctuation signal, and the energy transfer characteristics between the two are identified by cross-correlation peak detection and sliding window covariance analysis, including the concentrated period of energy input, response hysteresis profile and amplitude amplification or attenuation degree. Then, phase unwinding processing is performed: continuous correction is applied to phase jumps caused by sampling or transmission. That is, when a phase change is detected to exceed a preset reasonable range, the compensation amount is accumulated or decreased according to the trend of phase change before and after, so that the phase curve is restored to smoothness and monotonicity. On this basis, the equivalent time delay mean correction is calculated. The processing logic is as follows: statistically analyze the difference distribution of the peak times of excitation and response in each segment, remove outliers and take the weighted average. This mean is used as the equivalent transmission time delay of the node and is used as the horizontal coordinate reference for correcting the transmission characteristic curve. Finally, the real part of the corrected transmission characteristic reflecting the system inertia and transmission delay is obtained.

[0043] Furthermore, the linear response growth rate is calculated: under the assumption of small disturbances, the rate of change of response amplitude at the beginning and end of the transient segment is taken, and the growth exponent per unit time is calculated by combining the equivalent time delay. Positive values ​​indicate a natural divergence trend, negative values ​​indicate a natural convergence trend, and zero values ​​approximate critical stability. After completing the transfer characteristic analysis, steady-state amplitude information in the disturbance propagation event chain is extracted to evaluate the system's resource-constrained performance under continuous or large disturbances. The amplitude plateau interval of each node's response signal after entering the steady-state segment is located. The maximum and average values ​​within this interval are statistically analyzed and compared with the rated or tolerable limits of the corresponding resources for that node. These resources include the capacity margin of energy storage devices, the power limit of inverter units, and the thermal stability limit of transmission lines. When the steady-state amplitude approaches or exceeds the limit, the exceedance amplitude and duration are recorded. The nonlinear saturation intensity parameters are quantified using piecewise fitting and extreme value clustering methods. For example, suppose the sampling period of the time series response signal of a node before and after a disturbance event is 1ms. The amplitude sequence is stable at 100 units before the trigger, rises rapidly to 180 after the trigger, and then slowly falls back to around 130. When using the adaptive threshold segmentation method, the rate of change of the signal amplitude is first calculated. At 10ms after the trigger, the rate of change suddenly increases from 0 to 80 units / ms, and the statistical variance suddenly increases from 5 in the steady segment to 120. This moment is defined as the starting point of the initial transient segment. At 60ms, the rate of change drops below 5 and the variance falls back to 10. This is defined as the end of the transient segment and the starting point of the transition segment. After 150ms, the amplitude fluctuation is maintained within ±2, and the variance is about 3. This is defined as the starting point of the steady-state segment. The transient segment interval is denoted as [10ms, 60ms], the transition segment as [60ms, 150ms], and the steady-state segment as [150ms, 300ms].

[0044] Aligning each segment with the external excitation fluctuation signal in the time domain, cross-correlation peak detection revealed that the excitation peak occurred at 20ms, the response peak at 55ms, and the lag was 35ms. Sliding window covariance analysis showed that the energy input was concentrated in the 20~40ms range, and the response amplitude was amplified by about 1.3 times compared to the excitation. During phase unwinding, a phase jump from 358° to 5° was detected at 65ms, exceeding the preset reasonable range (a jump of >180° is considered abnormal). Based on the trend before and after, 360° was added to the 5° base to obtain 365°, which was then normalized to 5° to make the curve continuous. Subsequently, the distribution of the time difference between the peak excitation and response in the transient segment was statistically analyzed as {34,35,36,80,37}ms. After removing the outlier of 80ms, the average of the remaining values ​​was (34+35+36+37) / 4=35.5ms, which is the equivalent transmission delay τ. After correcting the x-axis of the transfer characteristic using τ, the transient segment starts with a response amplitude of 100, ends with an amplitude of 130, and has a time span of 50ms. Under the assumption of small-amplitude perturbation, the linear response growth rate is calculated as follows: growth exponent ≈ (end amplitude / start amplitude)^(1 / equivalent duration) - 1. Substituting this into the unit time scale, the growth exponent ≈ (130 / 100)^(1 / (50ms)) - 1 = (1.3)^(20) - 1 ≈ 18.79 - 1 ≈ 17.79 (positive values ​​indicate a natural divergence trend). The steady-state amplitude plateau range is 130 ± 2, with a maximum value of 132 and an average value of 130.5. This node is an inverter unit with a power limit of 125. The over-limit amplitude is 7, and the duration ranges from 160ms to 300ms, totaling 140ms. When constructing the amplitude-response inhibition rate curve, it was found that when the amplitude is below 110, the inhibition rate increases linearly with the amplitude, and the slope is stable. After exceeding 115, the increase in inhibition rate slows down, and tends to flatten out around 125. The inflection point was identified as amplitude 115 and inhibition rate 20%. The saturation intensity measure was taken as the ratio of amplitude to inhibition rate at the inflection point, which was 5.75, to describe the upper limit of regulation capability under large signals and the ease with which saturation is reached. This numerical example shows that the segment boundary can be determined by the rate of change and variance mutation points, the time delay mean is obtained by weighted averaging after outlier removal, the linear response growth rate is estimated from the amplitude at the beginning and end of the transient segment and the equivalent time delay, and the nonlinear saturation intensity is quantified by the ratio of the amplitude-response inhibition rate curve inflection point, thus fully realizing the accurate extraction of node dynamic characteristics and resource-constrained performance.

[0045] Step S50: Input the linear response growth rate and nonlinear saturation intensity parameters into the state trend discrimination model. Based on the discrimination results output by the model, divide the microgrid into several collaborative processing units according to function and control level, and apply fine-grained coordinated control authority to the collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands.

[0046] In some embodiments, combined with Figure 3As shown, the rate of change of key performance indicators of the system is continuously monitored, and the frequency of control command issuance and the trigger threshold of protection actions are dynamically adjusted according to the rate of change. Key performance indicators include frequency offset rate, voltage sag amplitude, and communication delay fluctuation. Specifically, when the rate of change of key performance indicators enters the preset first fluctuation range, the frequency of control command issuance is increased, while the trigger threshold of protection actions is decreased; when the rate of change of key performance indicators is in the second fluctuation range, the frequency of control command issuance is decreased, while the trigger threshold of protection actions is increased.

[0047] It should be noted that the first fluctuation range is the high fluctuation range, and the second fluctuation range is the low fluctuation range. The state trend discrimination model has multiple pre-set operating state classification standards, including stable mode, limit cycle oscillation mode, and transition mode between the two, and sets a judgment range based on the index range for each mode. When the model is executed, it first performs sign and amplitude analysis on the linear response growth rate to distinguish three cases: positive growth, negative decay, and near-zero change; then, combined with the magnitude of the nonlinear saturation intensity parameter, it assesses the possibility that the system will enter an amplitude-limited state under a large signal excitation.

[0048] The specific judgment logic is as follows: if the growth rate is positive and the saturation intensity parameter is higher than the preset high saturation threshold, it is determined that the system may form a continuous periodic energy exchange under the influence of external disturbances or internal control, that is, tend to the limit cycle oscillation mode; if the growth rate is negative and the saturation intensity parameter is in the low to medium range, it is determined that the system can gradually dissipate the disturbance energy and return to the stable equilibrium point, that is, tend to the stable mode; if the growth rate is close to zero and the saturation intensity parameter is high, it is determined that the system is in a critical state and is prone to oscillation or instability due to small additional disturbances.

[0049] During the discrimination process, the model employs a hierarchical judgment and cross-validation strategy. First, a coarse classification is performed on a single indicator dimension, followed by a fine classification through region mapping in a two-dimensional indicator space, avoiding misjudgments caused by extreme values ​​of a single indicator. When the discrimination result indicates that the current operating point may evolve into a limit cycle oscillation mode, the operating point is marked as a high-risk operating point, and its coordinate information, indicator value, discrimination basis, and timestamp are recorded in the operating status database, achieving a forward-looking quantitative prediction of potential risk states. The marking results can be used for priority scheduling of subsequent control and protection strategies.

[0050] For example, a hierarchical judgment and cross-validation logic for a state trend discrimination model is provided. Suppose the linear response growth rate extracted from a certain operating point in stage S4 is λ = +16.2 (positive and with a large amplitude), and the nonlinear saturation intensity parameter is σ = 5.75. The state trend discrimination model has pre-defined judgment intervals: when the linear response growth rate is positive and its amplitude is greater than the set positive threshold λ_thr = 5, it is classified as positive growth; when it is negative and less than the negative threshold λ_thr = -5, it is classified as negative decay; and values ​​between these two are considered near-zero changes. The nonlinear saturation intensity parameter σ is divided into a low interval σ ≤ 3, a middle interval 3 < σ ≤ 6, and a high interval σ > 6, where the high saturation threshold σ_high = 6. The model first performs a coarse classification of single indicators: λ = +16.2 satisfies λ > λ_thr and has a positive sign, and is classified as positive growth; σ = 5.75 falls into the middle interval. The model then proceeds to a two-dimensional index space region mapping and fine classification. Based on preset mode partitioning rules, if the region is in the "positive growth and high saturation range," it is classified as a limit cycle oscillation mode; if it is in the negative decay and medium saturation range, it is classified as a stable mode; and if it is in the near-zero change and high saturation range, it is classified as a critical state. In this example, λ is positively growing, but σ only reaches the upper limit of the medium range. According to the rules, supplementary judgment is needed in the cross-validation stage, taking into account the magnitude of the growth rate and the degree to which the saturation intensity approaches the high threshold. The model is set to treat it as a critical partial oscillation mode when λ is significantly higher than λ_thr and the difference between σ and σ_high is less than δ=0.5. Here, σ_high-σ=6-5.75=0.25<δ, therefore, although this operating point does not fully meet the high saturation condition, it has a high probability of entering the limit cycle oscillation mode and is considered a high-risk operating point.

[0051] In the labeling stage, the system writes the coordinate information of the running point (such as node number and time point), the index value (λ=16.2, σ=5.75), the criteria for positive growth and σ approaching the high saturation threshold, and the timestamp into the running status database to achieve forward-looking quantitative prediction. This numerical example shows that the sign and amplitude analysis of the linear response growth rate can distinguish between growth, decay, and critical trends; the nonlinear saturation intensity parameter provides an assessment of the possibility of large signal limitation; and the two-dimensional region mapping and threshold proximity supplementary judgment can reduce misjudgments caused by extreme single indexes. Thus, when the growth rate is significantly positive and the saturation intensity is close to the high limit, potential oscillation risks can be accurately identified and control strategies can be prioritized.

[0052] In some embodiments, based on the electrical topology of the microgrid and considering the control authority and business function roles of each node, the system is divided into functional units such as generation aggregation units, load management units, energy storage regulation units, and grid connection interface units. Above this functional division, a hierarchical structure of regional control layers, local control layers, and equipment control layers is overlaid according to a control hierarchy model, ensuring that each collaborative handling unit possesses both functional integrity and controllability. After the division is completed, the system retrieves the status trend judgment results and extracts all collaborative handling units containing high-risk operating points as priority handling targets. For these collaborative handling units involving high-risk operating points, the system applies fine-grained coordination control. The specific processing logic is as follows: a multi-objective optimization mechanism is introduced in the control command generation stage. This mechanism takes into account three requirements: suppressing oscillations, maintaining power quality, and avoiding resource overload. Adjustable weight coefficients are set for the control variables of different units. In the command issuance stage, a consistency and timeliness guarantee protocol is adopted. That is, each unit's local controller first generates a local command draft based on the weight coefficient, and then the coordination controller collects all drafts for global consistency verification to eliminate conflicting commands or commands with excessive superposition effects. The issuance time point is set according to the timing requirements of the event chain to ensure that the control and protection commands are synchronized in timing and logic.

[0053] During execution, the system continuously monitors the rate of change of key performance indicators, including frequency offset rate, voltage sag, and communication delay fluctuation. Real-time values ​​of each indicator are acquired at a fixed sampling period, and the difference in change between adjacent periods is calculated and divided by the sampling period to obtain the rate of change. This rate of change is compared with a preset rate warning range. If the rate of change of a certain indicator enters a high-fluctuation range, the system is determined to be in a state of rapid degradation or increased disturbance. Based on the rate of change level, the system dynamically adjusts the frequency of control command issuance and the trigger threshold of protective actions. The processing logic is as follows: when the rate of change is high, the command issuance frequency is increased to accelerate the response, while the protective action trigger threshold is appropriately lowered to ensure timely intervention; when the rate of change is low, the command issuance frequency is reduced to avoid frequent adjustments that could cause equipment wear or communication congestion, and the protective trigger threshold is appropriately increased to prevent malfunctions. This dynamic adjustment mechanism effectively prevents excessive control from causing secondary disturbances or the entrapment of risks due to slow response.

[0054] For example, a dynamic adjustment logic example based on the rate of change of key performance indicators is provided: Assume that a microgrid electrical topology includes rooftop photovoltaic arrays (power generation), commercial building loads (loads), lithium battery energy storage cabinets (energy storage), and common coupling points (grid connection). Based on functional attributes and control permissions, it is divided into a power generation aggregation unit U1 (including a group of photovoltaic inverters), a load management unit U2 (including a load aggregation controller), an energy storage regulation unit U3 (including an energy storage cabinet and a local BMS), and a grid connection interface unit U4 (including a grid-connected converter and a protection terminal). Control layers are superimposed on this functional division. U1 to U4 each have a regional control layer (responsible for cross-device coordination), a local control layer (responsible for local station optimization), and an equipment control layer (responsible for individual execution), forming a collaborative handling unit that combines functional integrity and control operability. The state trend judgment result shows that a certain node within U3 is a high-risk operating point (λ=+16.2, σ=5.75), therefore U3 is extracted as a priority handling target. A multi-objective optimization mechanism is introduced in the control command generation stage, setting weight coefficients for three objectives: oscillation suppression w1=0.5, maintaining power quality w2=0.3, and avoiding resource overload w3=0.2 (weights can be adjusted online according to operating conditions). The local controller generates drafts based on these weights for variables such as energy storage charging and discharging rates and reactive power support strength. For example, the draft suggests reducing charging power by 20% (significantly contributing to oscillation suppression) and fine-tuning the voltage reference value (secondarily contributing to power quality). The coordinating controller collects drafts from U3 and other units, performs global consistency checks, and eliminates the superimposed effects of conflicts such as increased power generation by U1 and reduced charging by U3. Commands are then uniformly issued at t=12:00:02.000 according to the event chain timing requirements to ensure logical synchronization.

[0055] During execution, the rate of change of key performance indicators is continuously monitored. Assuming a sampling period T = 1 second, the frequency offset rate within a certain period is Δf1 = 0.05 Hz, and in the next period Δf2 = 0.12 Hz. Therefore, the rate of change of frequency offset is r_f = (0.12 - 0.05) / 1 = 0.07 Hz / s; the rate of change of voltage sag is r_v = 0.04 pu / s; and the rate of change of communication delay fluctuation is r_d = 2 ms / s. Within the preset rate warning range, the high fluctuation range is r ≥ 0.06 (taking frequency as an example). At this point, r_f = 0.07 enters the high fluctuation range, indicating that the system is in a state of rapid degradation or increased disturbance. Dynamic adjustments are made based on the rate of change level: when the rate of change is high, the command issuance frequency is increased from the default once every 5 seconds to once every 1 second, while the protection action trigger threshold is lowered from a frequency deviation of 0.2 Hz to 0.15 Hz to ensure timely intervention; when the rate of change is low, the opposite adjustment is made to reduce the risk of equipment wear and communication congestion. This numerical example demonstrates that functional and control hierarchy division can identify high-risk units, multi-objective weight optimization and consistency verification ensure instruction coordination, and the formula for calculating the rate of change of key performance indicators is r=(X K1 -Xk The ) / T can reflect the rate of degradation in real time and dynamically adjust the release frequency and protection threshold accordingly, effectively preventing excessive regulation from causing secondary disturbances or slow response from solidifying risks.

[0056] In the above method steps, by constructing a space-energy coupling topology model covering the entire network, unified modeling and real-time sensing of multi-source heterogeneous information such as distributed power sources, energy storage, loads, and grid connection interfaces are achieved. This overcomes the limitation of traditional single-point monitoring in capturing distributed interactive effects, making the analysis of disturbance propagation paths and causal relationships more refined and reliable. Through in-depth segmented analysis and equivalent time delay correction of signals within the event chain, not only are linear response characteristics reflecting system inertia and transmission delay obtained, but also the degree of resource constraint and nonlinear saturation intensity under large signal conditions are quantified, enhancing the model's universality and identification accuracy for strong and weak disturbance scenarios. By fusing linear growth rate and nonlinear saturation intensity... The state trend discrimination model can predict in advance whether the operating point will slide into the limit cycle oscillation or instability, and accurately locate high-risk operating points, realizing the transformation from passive response to active early warning, and reducing the probability of cascading failures caused by sudden instability. In addition, based on the division of functional and control-level collaborative processing units and fine-grained coordinated control strategies, combined with dynamic commands and threshold adjustment mechanisms based on the rate of change of key performance indicators, the balance between the timeliness and consistency of control and protection actions is ensured. This avoids secondary disturbances caused by overly aggressive regulation and prevents the risk solidification caused by slow response, thereby improving the overall dynamic stability, disturbance resistance robustness and intelligent coordination level of security protection of the microgrid.

[0057] Please see Figure 4 , Figure 4 This embodiment illustrates a microgrid monitoring system based on coordinated control and security protection linkage. In this embodiment, the microgrid monitoring system based on coordinated control and security protection linkage is used to perform the above-mentioned... Figure 1 The steps in the corresponding embodiments. Please refer to the details. Figure 1 as well as Figure 1 The relevant descriptions in the corresponding embodiments are shown below. For ease of explanation, only the parts relevant to this embodiment are shown. See also... Figure 4 The microgrid monitoring system 200 based on coordinated control and security protection includes: Topology module 210 is configured to collect spatial layout parameters and control parameters of equipment in various areas of the microgrid, and combine them with the time-series operation data streams of multiple monitoring nodes to establish a topology model covering the entire network. The associated event filtering module 220 is configured to select collaborative nodes that are strongly coupled with key monitoring points based on the topology model, extract state feature vectors representing system behavior from the original data streams of each node and the fused data streams of the collaborative nodes, and perform similarity calculations on the state feature vectors to determine candidate associated event pairs originating from the same disturbance or control process. The verification module 230 is configured to perform double verification on the candidate associated event pairs to obtain valid associated event pairs, and to connect them in chronological order to form a disturbance propagation event chain that reflects the transmission of disturbances or control processes across nodes. The correction module 240 is configured to perform segmented analysis of the internal response signals of each node in the disturbance propagation event chain, identify the energy transfer characteristics by using external excitation fluctuations and internal response fluctuation signals, obtain the real part of the transfer characteristics by phase unwinding and equivalent time delay mean correction, convert it into a linear response growth rate, and extract steady-state amplitude information in the disturbance propagation event chain to determine the nonlinear saturation intensity parameters. The dynamic adjustment module 250 is configured to input the linear response growth rate and nonlinear saturation intensity parameters into the state trend discrimination model. Based on the discrimination results output by the model, the microgrid is divided into several collaborative processing units according to function and control level. Fine-grained coordinated control is applied to the collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands.

[0058] The microgrid monitoring system based on coordinated control and security protection provided in this embodiment achieves unified modeling and real-time sensing of multi-source heterogeneous information such as distributed power sources, energy storage, loads, and grid-connected interfaces by constructing a space-energy coupling topology model covering the entire network. This overcomes the limitations of traditional single-point monitoring in capturing distributed interactive effects, making the analysis of disturbance propagation paths and causal relationships more refined and reliable. Through in-depth segmented analysis and equivalent time delay correction of signals within the event chain, it not only obtains linear response characteristics reflecting system inertia and transmission delay but also quantifies the degree of resource constraint and nonlinear saturation intensity under large signal conditions, enhancing the model's universality and identification accuracy for strong and weak disturbance scenarios. By using a linear growth rate... By integrating nonlinear saturation intensity into the state trend discrimination model, it can predict in advance whether the operating point will slide towards the limit cycle oscillation or instability, and accurately locate high-risk operating points, realizing the transformation from passive response to active early warning, and reducing the probability of cascading failures caused by sudden instability. Furthermore, relying on the division of functional and control-level collaborative processing units and fine-grained coordinated control strategies, combined with dynamic commands and threshold adjustment mechanisms based on the rate of change of key performance indicators, it ensures a balance between the timeliness and consistency of control and protection actions, avoiding secondary disturbances caused by overly aggressive regulation, and preventing the risk solidification caused by slow response. Thus, it improves the overall dynamic stability, disturbance rejection robustness, and intelligent coordination level of security protection of the microgrid.

[0059] It should be understood that the modules in the system provided in this embodiment are used to execute... Figure 1 The steps in the corresponding embodiments, and for Figure 1 The steps in the corresponding embodiments have been explained in detail in the above embodiments. Please refer to them for details. Figure 1 as well as Figure 1 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0060] In some embodiments, an electronic device is also provided, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the microgrid monitoring method based on coordinated control and security protection linkage described in the above embodiments.

[0061] Please see Figure 5 , Figure 5 This diagram illustrates a structural block diagram of a server in an electronic device according to this embodiment. The server 500 includes a processor 501, a memory 502, and a computer program 503 stored in the memory 502 and executable on the processor 501. For example, this program may be a microgrid monitoring method based on coordinated control and security protection linkage. When the processor 501 executes the computer program 503, it implements the steps of the microgrid monitoring method based on coordinated control and security protection linkage described in the above embodiments, for example... Figure 1 Steps S10 to S50 correspond to the embodiments. Alternatively, the processor 501 executes the computer program 503 to implement the above. Figure 4 The functions of each module in the corresponding embodiments, for example, Figure 4 For details on the functions of the modules shown (e.g., topology module 210), please refer to [link / reference]. Figure 4 The relevant descriptions in the corresponding embodiments are not repeated here.

[0062] For example, computer program 503 can be divided into one or more units, one or more units are stored in memory 502 and executed by processor 501 to complete the technical solution provided in the above embodiments. One or more units can be a series of computer program instruction segments capable of performing a specific function, which are used to describe the execution process of computer program 503 in server 500.

[0063] The electronic device may include, but is not limited to, a processor 501 and a memory 502. Those skilled in the art will understand that... Figure 5This is merely an example of server 500 in an electronic device and does not constitute a limitation on server 500. It may include more or fewer components than shown, or combine certain components, or different components. For example, a turntable terminal device may also include input / output terminal devices, network access terminal devices, buses, etc.

[0064] The processor 501 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0065] The memory 502 can be an internal storage unit of the server 500, such as the server 500's hard drive or memory. The memory 502 can also be an external storage terminal device of the server 500, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the server 500. Furthermore, the memory 502 can include both internal storage units and external storage terminal devices of the server 500. The memory 502 is used to store computer programs and other programs and data required by the turntable terminal device. The memory 502 can also be used to temporarily store data that has been output or will be output.

[0066] In one embodiment, a computer-readable storage medium is also provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the microgrid monitoring method based on coordinated control and security protection linkage as described in the above embodiments.

[0067] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0068] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be non-volatile or volatile. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0069] The terms "first," "second," "third," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects and not to describe a particular order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, it may include a series of steps or units, or optionally, steps or units not listed, or other steps or units inherent to these processes, methods, products, or devices.

[0070] The accompanying drawings show only the portions relevant to this application, not all of them. Before discussing exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations may be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations may be rearranged. The process may be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process may correspond to a method, function, procedure, subroutine, subprogram, etc.

[0071] The terms “component,” “module,” “system,” “unit,” etc., used in this specification are used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. For example, a unit can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, a thread of execution, a program, and / or distributed between two or more computers. Furthermore, these units can be executed from various computer-readable media on which various data structures are stored. Units can communicate, for example, via local and / or remote processes based on signals having one or more data packets (e.g., data from a second unit interacting with another unit between a local system, a distributed system, and / or a network; for example, the Internet interacting with other systems via signals).

[0072] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0073] Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The reference to "embodiment" herein means that a specific feature, structure, or characteristic described in connection with an embodiment can be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily indicate the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

Claims

1. A microgrid monitoring method based on coordinated control and security protection linkage, characterized in that, include: Collect spatial layout parameters and control parameters of equipment in various areas of the microgrid, and combine them with the time-series operation data streams of multiple monitoring nodes to establish a topology model covering the entire network; Based on the aforementioned topology model, collaborative nodes strongly coupled with key monitoring points are selected. State feature vectors representing system behavior are extracted from the original data streams of each node and the fused data streams of the collaborative nodes. Similarity calculations are performed on the state feature vectors to determine candidate related event pairs originating from the same disturbance or control process. The candidate associated event pairs are double-checked to obtain valid associated event pairs, and then connected in chronological order to form a disturbance propagation event chain that reflects the transmission of disturbances or control processes across nodes. The internal response signals of each node in the disturbance propagation event chain are analyzed in segments. The energy transfer characteristics are identified by using external excitation fluctuations and internal response fluctuation signals. The real part of the transfer characteristics is obtained by phase unwinding and equivalent time delay mean correction, and converted into the linear response growth rate. The steady-state amplitude information in the disturbance propagation event chain is extracted to determine the nonlinear saturation intensity parameters. The linear response growth rate and nonlinear saturation intensity parameters are input into the state trend discrimination model. Based on the discrimination results output by the model, the microgrid is divided into several collaborative processing units according to function and control level. Fine-grained coordinated control is applied to the collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands.

2. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 1, characterized in that, The establishment of a topology model covering the entire network includes: A basic graph structure is constructed using devices in each area of ​​the microgrid as nodes and physical wiring and logical control links between devices as edges. The spatial layout parameters and control parameters are embedded in the node attributes, and the energy transfer direction, capacity limit, command transmission protocol and communication delay index are embedded in the edge attributes. The time-series running data stream is mapped to corresponding nodes and edges to form a dynamic attribute layer, resulting in the topology model that simultaneously describes the physical location of the device, the energy transfer path, and the command transfer path.

3. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 2, characterized in that, The selection of collaborative nodes strongly coupled with key monitoring points based on the topology model includes: In the topology model, identify key monitoring points that have a significant impact on the overall network operation status. These key monitoring points include at least one of the following: power exchange hubs, the beginning of important load feeders, grid connection interfaces, and the main control location of energy storage systems. For each key monitoring point, the coupling degree between it and other nodes in the network on the energy transmission path and command transmission path is calculated based on three indicators: the number of paths, the total transmission capacity of the paths, and the weighted sum of command response delay. Nodes with a coupling degree higher than a preset threshold are selected as cooperative nodes.

4. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 3, characterized in that, The extraction of state feature vectors representing system behavior from the original data streams of each node and the fused data streams of collaborating nodes includes: For a single node or a collaborative node, the running data segment is extracted according to the time window, and the window length is set according to the duration of the typical disturbance or control process to be captured. Multidimensional features are extracted within the window, including power change morphology features, voltage fluctuation fingerprint features, and control response curve features. In the extraction process, configurable feature templates and sliding calculations are used to obtain state feature vectors with consistent structures under the same operating conditions.

5. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 1, characterized in that, The process of performing double verification on the candidate associated event pairs to obtain valid associated event pairs, and then concatenating them in chronological order to form a disturbance propagation event chain reflecting the cross-node transmission of disturbances or control processes, includes: For each candidate associated event pair, perform temporal continuity verification and spatial topological causal reachability verification in sequence. If both verifications pass, the candidate associated event pair is confirmed as a valid associated event pair. All valid related event pairs are arranged in ascending order according to the timestamp of the first event. The connectivity of adjacent event pairs is determined in turn. Adjacent event pairs that satisfy temporal continuity and spatial topological reachability are merged to form one or more disturbance propagation event chains.

6. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 5, characterized in that, The sequential execution of temporal continuity verification and spatial topological causal reachability verification on candidate associated event pairs includes: Based on the occurrence time of the first event in the candidate related event pair, a time neighborhood width is set to determine whether the occurrence time of the second event falls within the time neighborhood. At the same time, it is verified whether the two events meet the preset time sequence dependency rules. If both are met, the time continuity check passes. In the topology model, locate the two nodes to which the candidate associated event pair belongs, and explore from the source node to the target node. If there is at least one path that satisfies causal directionality and is allowed by the transmission medium and permission configuration, then the spatial topology causal reachability verification passes.

7. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 1, characterized in that, The segmented analysis of the internal response signals of each node in the disturbance propagation event chain, using external excitation fluctuations and internal response fluctuation signals to identify energy transfer characteristics, includes: The time-series response signals of the acquisition node before and after the event are collected. The response signals are divided into initial transient segment, transition segment and steady-state segment by the segmentation boundary with the abrupt change point of the signal amplitude change rate and statistical variance. By aligning each signal segment with the corresponding external excitation fluctuation signal in the time domain, the energy transfer characteristics are obtained through cross-correlation peak detection and sliding window covariance analysis.

8. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 1, characterized in that, The real part of the transfer characteristic obtained after phase unwinding and equivalent time delay mean correction includes: Continuous correction is performed for phase jumps caused by sampling or transmission. When a phase change is detected to exceed a preset reasonable range, the compensation amount is accumulated or decreased according to the trend of phase change before and after, so that the phase curve is restored to a smooth and monotonous state. The difference distribution of the peak times of excitation and response within each segment is statistically analyzed. After removing outliers, a weighted average is taken. This average is used as the equivalent transmission delay of the node and is used as the horizontal axis reference for correcting the transfer characteristic curve, thus obtaining the real part of the corrected transfer characteristic that reflects the system inertia and transmission delay.

9. The microgrid monitoring method based on coordinated control and security protection linkage according to claim 1, characterized in that, The process of dividing the microgrid into several collaborative processing units according to function and control levels, and applying fine-grained coordinated control to collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands, includes: The system continuously monitors the rate of change of key performance indicators and dynamically adjusts the frequency of control command issuance and the trigger threshold of protection actions based on the rate of change. The key performance indicators include frequency offset rate, voltage sag amplitude, and communication delay fluctuation. Specifically, when the rate of change of key performance indicators enters the preset first fluctuation range, the frequency of control command issuance is increased while the threshold for triggering protective actions is lowered; when the rate of change of key performance indicators is in the second fluctuation range, the frequency of control command issuance is reduced while the threshold for triggering protective actions is increased.

10. A microgrid monitoring system based on coordinated control and security protection linkage, characterized in that, include: The topology module is configured to collect spatial layout parameters and control parameters of equipment in various areas of the microgrid, and combine them with the time-series operation data streams of multiple monitoring nodes to establish a topology model covering the entire network. The associated event filtering module is configured to select collaborative nodes that are strongly coupled with key monitoring points based on the topology model, extract state feature vectors representing system behavior from the original data streams of each node and the fused data streams of the collaborative nodes, and perform similarity calculations on the state feature vectors to determine candidate associated event pairs originating from the same disturbance or control process. The verification module is configured to perform double verification on the candidate associated event pairs to obtain valid associated event pairs, and to connect them in chronological order to form a disturbance propagation event chain that reflects the transmission of disturbances or control processes across nodes. The correction module is configured to perform segmented analysis of the internal response signals of each node in the disturbance propagation event chain, identify energy transfer characteristics using external excitation fluctuations and internal response fluctuation signals, obtain the real part of the transfer characteristics through phase unwinding and equivalent time delay mean correction, convert it into a linear response growth rate, and extract steady-state amplitude information in the disturbance propagation event chain to determine nonlinear saturation intensity parameters. The dynamic adjustment module is configured to input the linear response growth rate and nonlinear saturation intensity parameters into the state trend discrimination model. Based on the discrimination results output by the model, the microgrid is divided into several collaborative processing units according to function and control level. Fine-grained coordination control is applied to the collaborative processing units involving risky operating points to ensure the consistency and timeliness of control and protection commands.

Images