A microgrid island detection method and system

Abstract

The present invention relates to the field of power grid detection technology, and in particular to a microgrid island detection method and system. The method comprises the following steps: obtaining microgrid topology data; deploying distributed computing units on the microgrid topology data to generate distributed node deployment data; using sensors to collect microgrid operation data according to the distributed node deployment data to generate initial microgrid operation data; constructing multi-dimensional perturbation signals on the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix; injecting micro-reactive disturbances into the initial microgrid operation data using the multi-dimensional perturbation signal matrix to generate multi-dimensional perturbation power grid change data; performing multi-directional impedance fusion on the multi-dimensional perturbation power grid change data to generate power grid fusion impedance data. The present invention improves the comprehensiveness and reliability of microgrid island detection through multi-dimensional perturbation signals and dynamic timing analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid detection, and in particular to a microgrid island detection method and system. Background Art

[0002] As a distributed energy system that can achieve local energy self-sufficiency, microgrid plays an increasingly important role in the power system. However, the islanding phenomenon of microgrids, that is, the independent operation of microgrids without being connected to the main grid, may lead to problems such as reduced power quality and increased safety hazards. Therefore, effective islanding detection methods are particularly critical. The initial islanding detection methods are mainly based on passive detection technologies, such as changes in voltage, frequency, and phase. Although these methods are simple and easy to use, they may not be able to effectively detect the islanding phenomenon when the load and power generation are well matched. For this reason, active detection technologies have gradually been introduced, including disturbance injection methods and frequency drift methods. These methods detect the islanding state by applying specific disturbances to the microgrid and have higher detection accuracy. In recent years, with the rapid development of smart grid and artificial intelligence technologies, data-driven detection methods have emerged. These methods use technologies such as big data analysis and machine learning to perform deep learning and pattern recognition on power system operation data, thereby achieving more intelligent and accurate islanding detection. However, the current traditional island detection technology has different applicability to different types of distributed energy (such as photovoltaic, wind energy, etc.), and it is difficult to achieve unified detection in microgrids with mixed energy sources. At the same time, there is a lack of effective follow-up processing mechanism after the island is detected, which leads to low comprehensiveness and reliability of island detection. Summary of the invention

[0003] Based on this, it is necessary to provide a microgrid island detection method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, a microgrid island detection method is provided, the method comprising the following steps:

[0005] Step S1: Acquire microgrid topology data; deploy distributed computing units on the microgrid topology data to generate distributed node deployment data; use sensors to collect microgrid operation data based on the distributed node deployment data to generate initial microgrid operation data; construct multi-dimensional perturbation signals on the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix;

[0006] Step S2: using a multi-dimensional disturbance signal matrix to perform a micro-reactive disturbance injection on the initial microgrid operation data to generate multi-dimensional disturbance grid change data; performing multi-directional impedance fusion on the multi-dimensional disturbance grid change data to generate grid fusion impedance data; performing dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data;

[0007] Step S3: construct an island detection framework; use the initial microgrid operation data to make the island detection framework compatible with distributed energy detection, and generate an adaptive island detection framework; use the adaptive island detection framework to perform multi-mode collaborative island detection on the grid time series impedance analysis data, and generate comprehensive grid island detection data; perform grid operation network topology modeling on the distributed node deployment data according to the comprehensive grid island detection data, and generate a grid node operation network topology model;

[0008] Step S4: Perform grid node island detection and discrimination on the grid node operation network topology model to generate abnormal grid nodes; calculate the node abnormality degree of abnormal grid nodes through the grid comprehensive island detection data to obtain the node abnormality coefficient; rotate the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient to generate a grid protection strategy to execute the normal operation of the microgrid.

[0009] The present invention ensures a comprehensive understanding of the entire power grid layout by acquiring the topological structure data of the microgrid, and provides the necessary basic information for subsequent analysis and detection. The deployment of distributed computing units helps to optimize the efficiency of data processing, makes the deployment of distributed nodes more reasonable, and reduces the waste of computing resources. The sensor collects data according to the distributed node deployment data, so that the initial microgrid operation data is more accurate and real-time, covering the key operating parameters of the microgrid. The construction of multi-dimensional perturbation signals can introduce more dimensional information, provide rich data features, and enhance the accuracy and sensitivity of subsequent analysis and detection. The micro-reactive disturbance injection of microgrid operation data using a multi-dimensional perturbation signal matrix effectively simulates the abnormal situation of the power grid and improves the sensitivity of detection. Multi-directional impedance fusion enables the multi-dimensional information of the data to be effectively integrated, improves the perception of changes in the power grid state, and enhances the comprehensiveness of detection. Dynamic time series change trend analysis helps to capture the change trend of the power grid impedance at different time points, further improves the real-time monitoring capability of the power grid state, and provides key data support for island detection. The island detection framework is constructed, and the distributed energy detection is compatible with the initial microgrid operation data, which improves the adaptability of the detection system and enables it to adapt to different microgrid structures and operation modes. The generation of the adaptive island detection framework enables the detection system to have self-adjustment capabilities, optimize the detection process according to the actual operation conditions, and improve the accuracy and efficiency of the detection. Through multi-mode collaborative island detection, combined with the advantages of multiple detection methods, the ability to identify island phenomena under complex power grid structures is improved, and the probability of false alarms and missed alarms is reduced. The final generated grid node operation network topology model provides accurate structural data support for subsequent analysis and processing, which helps to further optimize the management and operation of the power grid. The island detection and discrimination of the grid node operation network topology model can accurately identify abnormal grid nodes, thereby providing specific targets for subsequent processing. The calculation of the node abnormality degree makes the processing of abnormal nodes more targeted, and different countermeasures can be taken according to their severity, which improves the flexibility and effectiveness of dealing with abnormalities. Through the rotation of the grid protection strategy, the protection strategy can be dynamically adjusted according to the actual situation to ensure that the grid can remain stable and safe under different operating conditions, and enhance the risk resistance of the microgrid. Therefore, the present invention improves the comprehensiveness and reliability of microgrid island detection through multi-dimensional disturbance signals and dynamic timing analysis.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: Acquire microgrid topology data;

[0012] Step S12: Identify key nodes of the microgrid topology data to obtain key node data of the microgrid topology;

[0013] Step S13: deploying distributed computing units on the microgrid based on the key node data of the microgrid topology to generate distributed node deployment data;

[0014] Step S14: using sensors to collect microgrid operation data according to distributed node deployment data to generate initial microgrid operation data;

[0015] Step S15: construct a multi-dimensional perturbation signal for the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix.

[0016] By acquiring microgrid topology data and identifying key nodes, the present invention can understand the overall layout and key components of the microgrid in detail, which is helpful for optimizing system design and fault diagnosis. The deployment of distributed computing units based on the key nodes of the microgrid topology can improve computing efficiency and data processing capabilities, and enhance the stability and reliability of the microgrid. By collecting microgrid operation data according to distributed node deployment data through sensors, real-time and accurate operation data can be obtained, which is helpful for real-time monitoring and analysis of the operation status of the microgrid. By constructing a multi-dimensional perturbation signal for the initial microgrid operation data, a multi-dimensional perturbation signal matrix can be generated, which provides basic data support for the dynamic analysis, fault detection and early warning of the system. Through the multi-dimensional perturbation signal matrix, a more in-depth analysis of the microgrid can be performed, including anomaly detection, performance optimization and fault prediction, to improve the intelligence level and adaptive ability of the system.

[0017] Preferably, step S14 comprises the following steps:

[0018] Step S151: extracting multi-dimensional grid feature data from the initial microgrid operation data to obtain multi-dimensional feature data of the microgrid, wherein the multi-dimensional feature data of the microgrid includes microgrid voltage data, microgrid reactive power data, and microgrid harmonic data;

[0019] Step S152: injecting a perturbation signal through the microgrid voltage data, the microgrid reactive power data and the microgrid harmonic data to generate an initial microgrid perturbation signal; performing a signal spectrum characteristic analysis on the initial microgrid perturbation signal to generate perturbation signal spectrum characteristic data;

[0020] Step S153: performing multi-frequency domain superposition on the initial perturbation signal of the microgrid according to the perturbation signal spectrum characteristic data to generate a multi-frequency domain disturbance signal set; performing nonlinear signal superposition on the multi-frequency domain disturbance signal set to generate a nonlinear disturbance signal matrix;

[0021] Step S154: performing perturbation signal spectrum equalization on the nonlinear perturbation signal matrix to generate a multi-dimensional perturbation signal matrix.

[0022] The present invention can accurately obtain the key electrical characteristics of the microgrid by extracting multi-dimensional grid characteristic data (including voltage data, reactive power data and harmonic data) from the initial microgrid operation data, and provide detailed basic data for subsequent signal analysis and optimization. The microgrid voltage, reactive power and harmonic data are injected with perturbation signals, and the initial perturbation signal is analyzed for spectrum characteristics, so that the main frequency components and characteristics of the signal can be identified, which is helpful to understand the response behavior and potential problems of the system under various conditions. According to the perturbation signal spectrum characteristic data, multi-frequency domain superposition is performed, and the generated multi-frequency domain disturbance signal set can more comprehensively simulate various disturbance conditions of the system. This multi-frequency domain analysis can cover more frequency ranges and help reveal potential interference and nonlinear characteristics in the system. By performing nonlinear signal superposition on the multi-frequency domain disturbance signal set, the generated nonlinear disturbance signal matrix can capture the complex nonlinear behavior in the microgrid system, which is helpful to identify and analyze the nonlinear phenomena and their influences in the system. The nonlinear disturbance signal matrix is ​​subjected to perturbation signal spectrum equalization, and the multi-dimensional disturbance signal matrix finally generated can more accurately reflect the actual disturbance condition of the microgrid system. A balanced signal matrix facilitates more accurate system analysis, fault detection, and performance optimization.

[0023] Preferably, step S2 comprises the following steps:

[0024] Step S21: injecting a micro-reactive disturbance into the initial microgrid operation data using a multi-dimensional disturbance signal matrix to generate multi-dimensional disturbance grid change data;

[0025] Step S22: extracting frequency components from the multi-dimensional disturbance power grid change data to obtain a power grid change multi-frequency disturbance component set; calculating independent impedance components from the power grid change multi-frequency disturbance component set to generate power grid independent impedance component data;

[0026] Step S23: performing nonlinear error correction on the independent impedance component data of the power grid through nonlinear effects to generate independent corrected impedance component data of the power grid;

[0027] Step S24: performing multi-directional impedance fusion on the grid independent correction impedance component data to generate grid fusion impedance data; performing dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data.

[0028] The present invention can effectively capture the response of the system under different disturbance conditions by injecting a micro-reactive disturbance into the initial microgrid operation data and generating multi-dimensional disturbance grid change data, which helps to deeply understand the behavior and performance of the microgrid in actual operation. Frequency components are extracted from the multi-dimensional disturbance grid change data to obtain a grid change multi-frequency disturbance component set, which can identify the disturbance components of different frequencies in the grid. This multi-frequency analysis helps to reveal the main interference factors and their frequency characteristics in the system. Independent impedance component calculation is performed on the grid change multi-frequency disturbance component set to generate grid independent impedance component data, which can accurately evaluate the impedance characteristics of each frequency component in the grid. This process helps to understand the impedance response of the system, especially the response to specific frequency disturbances. By performing nonlinear error correction on the grid independent impedance component data and generating grid independent correction impedance component data, the error introduced by nonlinear effects can be reduced to ensure the accuracy and reliability of impedance data. This process improves the accuracy of the data and supports more accurate system analysis and optimization. The independent corrected impedance component data of the power grid is fused in multiple directions to generate power grid fused impedance data, and further dynamic time series change trend analysis is performed to generate power grid time series impedance analysis data. This fusion and dynamic analysis can reveal the overall characteristics of the power grid impedance and its changing trends over time, which is helpful for long-term monitoring and optimization of power grid performance.

[0029] Preferably, step S3 comprises the following steps:

[0030] Step S31: construct an island detection framework; perform grid operation energy analysis on the initial microgrid operation data to generate grid operation energy division data;

[0031] Step S32: Perform distributed energy detection compatibility on the island detection framework through the grid operation energy division data to generate an adaptive island detection framework;

[0032] Step S33: using an adaptive island detection framework to perform multi-mode collaborative island detection on the grid time series impedance analysis data to generate comprehensive grid island detection data;

[0033] Step S34: Perform grid operation network topology modeling on the distributed node deployment data according to the grid comprehensive island detection data to generate a grid node operation network topology model.

[0034] The present invention can systematically identify and analyze the island operation mode by constructing an island detection framework and performing grid operation energy analysis on the initial microgrid operation data, which is helpful to detect the island phenomenon in the microgrid, that is, some areas or nodes in the system are disconnected from the main grid due to faults or other reasons, but can still operate independently. According to the grid operation energy division data, the island detection framework is compatible with distributed energy detection to generate an adaptive island detection framework. This adaptive framework can be adjusted according to the actual energy distribution, improve the accuracy and reliability of island detection, and ensure that it can cope with island phenomena under different operating conditions. The adaptive island detection framework is used to perform multi-mode collaborative island detection on the grid timing impedance analysis data to generate grid comprehensive island detection data. This multi-mode collaborative detection can comprehensively consider different detection modes and data sources, provide more comprehensive island detection results, and improve the comprehensiveness and accuracy of detection. According to the grid comprehensive island detection data, the distributed node deployment data is modeled for the grid operation network topology to generate a grid node operation network topology model. This process helps to visualize and optimize the operation network structure of the grid and support more effective system analysis, troubleshooting and maintenance. Overall, step S3 can significantly improve the robustness and stability of the microgrid system through systematic island detection and network topology modeling. It can timely detect and handle island phenomena, optimize the grid operation status, and ensure the safety and reliability of the grid under various operating conditions.

[0035] Preferably, step S33 includes the following steps:

[0036] Step S331: classifying the power grid energy types for the power grid time series impedance analysis data to generate power grid single energy impedance analysis data and power grid multi-energy impedance analysis data;

[0037] Step S332: using the adaptive island detection framework to perform single energy reactive power offset detection on the single energy impedance analysis data of the power grid, and generating single energy reactive power offset detection data; comparing the single energy reactive power offset detection value with the preset reactive power offset detection threshold, thereby generating single energy microgrid island detection data;

[0038] Step S333: performing energy type identification on the power grid multi-energy impedance analysis data to obtain energy type identification data; performing multi-energy change feature extraction based on the energy type identification data to obtain multi-energy type change feature data;

[0039] Step S334: performing time domain analysis on the power grid multi-energy impedance analysis data to generate a power grid multi-energy impedance time domain curve; performing same peak time domain curve cutting on the power grid multi-energy impedance time domain curve based on a preset peak acquisition range to generate a time domain peak cutting curve;

[0040] Step S335: performing feature change matching on the time domain peak cutting curve and the multi-energy type change feature data to generate a multi-energy feature change matching curve; performing reactive power offset detection on the multi-energy feature change matching curve to generate multi-energy reactive power offset detection data; comparing the multi-energy reactive power offset detection value with a preset reactive power offset detection threshold to generate multi-energy microgrid island detection data;

[0041] Step S336: Perform collaborative island detection on the single-energy microgrid island detection data and the multi-energy microgrid island detection data to generate comprehensive grid island detection data.

[0042] The present invention divides the power grid time-series impedance analysis data into energy types and generates power grid single energy and multi-energy impedance analysis data, which can more carefully understand the impact of different types of energy on power grid impedance and help analyze and compare different energy characteristics. The single energy impedance analysis data is detected by using an adaptive island detection framework to identify and analyze the reactive power offset under a single energy source. This process helps to discover potential problems in a single energy system and generate corresponding island detection data. Energy type identification and multi-energy change feature extraction of power grid multi-energy impedance analysis data can reveal the impedance characteristics and change laws under multiple energy conditions. This feature extraction provides a more comprehensive understanding of complex power grid systems and helps monitor and optimize multi-energy systems. The time domain curve of the power grid multi-energy impedance is analyzed to generate a time domain peak cutting curve, which helps capture and analyze the peak changes in the power grid. This analysis can reveal the changes in the impedance characteristics of the power grid in different time periods and provide more time domain information for island detection. By matching the time-domain peak cutting curve with the characteristic data of multiple energy types, generating a matching curve for multiple energy characteristic changes, and performing reactive power offset detection, the characteristic changes under multiple energy sources can be comprehensively considered, which helps to improve the accuracy and comprehensiveness of island detection and generate multi-energy microgrid island detection data. By collaboratively analyzing the island detection data of single-energy and multi-energy microgrids and generating comprehensive grid island detection data, various types of data and detection results can be integrated to provide a more comprehensive island detection capability, which helps to comprehensively evaluate the island phenomenon of the power grid system and improve the safety and stability of the system.

[0043] Preferably, step S4 comprises the following steps:

[0044] Step S41: performing grid node island detection and discrimination on the grid node operation network topology model. When a grid node island is detected in the grid node operation network topology model, the corresponding grid node is marked as an abnormal grid node.

[0045] Step S42: Performing grid stability analysis on abnormal grid nodes through grid comprehensive island detection data to generate grid node stability analysis data; performing node load performance analysis on abnormal grid nodes to generate node reliability analysis data;

[0046] Step S43: Calculate the node abnormality degree of the abnormal power grid node based on the power grid node stability maintenance calculation formula through the power grid node stability analysis data and the node reliability analysis data to obtain the node abnormality coefficient;

[0047] Step S44: rotating the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient, generating a grid protection strategy to execute normal operation of the microgrid.

[0048] The present invention can accurately identify the grid nodes with islanding phenomenon by performing island detection and discrimination on the network topology model of the grid node operation, and mark them as abnormal grid nodes. This identification ability helps to timely discover potential problems in the grid system and ensure early processing of abnormal situations. Stability analysis and load performance analysis are performed on abnormal grid nodes to generate node stability analysis data and node reliability analysis data. These analyses can evaluate the stability and reliability of abnormal nodes and provide a basis for subsequent treatment measures. Based on the grid node stability calculation formula, the node abnormality degree is calculated by calculating the stability analysis data and the reliability analysis data to obtain the node abnormality coefficient. This calculation quantifies the abnormality degree of the node and helps to more accurately evaluate the severity and impact range of the problem. According to the node abnormality coefficient, the grid protection strategy is rotated for the abnormal grid nodes to generate grid protection strategies. These protection strategies can formulate appropriate countermeasures for different degrees of node abnormalities to ensure the normal operation of the microgrid and system stability. Through systematic abnormal node detection, analysis and protection strategy formulation, the safety and reliability of the grid system can be significantly improved. Timely identification and processing of abnormal nodes and formulation of targeted protection measures can help prevent the expansion of system failures and maintain the normal operation of the grid. The implementation of step S4 can optimize the operating efficiency of the power grid and ensure that the system can respond and adjust quickly when an abnormality occurs through an effective protection strategy, thereby reducing potential system downtime and operating risks.

[0049] Preferably, the grid node stability calculation formula in step S43 is as follows:

[0050]

[0051] In the formula, Represented as a grid node The stability coefficient, It represents the time interval for power grid stability analysis. Represented as a node In time The voltage value, Represented as a node In time The current value, Represented as a node In time The active power, Represented as a node In time The reactive power, Expressed as voltage influence weight coefficient, Expressed as the power impact weight coefficient, Expressed as the impedance influence weight coefficient, Represented as a node The neighboring set of other connected nodes, Represented as a node With Node The impedance value between.

[0052] The present invention analyzes and integrates a collaborative island detection formula. The principle of the formula is to evaluate the stability of the power grid node by comprehensively considering the changes in node voltage, current, active and reactive power, and the voltage difference of neighboring nodes. The formula consists of two main parts: integral term and sum term. The integral term in the formula In Reflects the voltage With current The dynamic relationship between voltage and current. The sensitivity of voltage to current change is the key to measure node stability. The rate of change of current The product with the voltage represents the effect of node current fluctuation on stability. Reflects the node power and the second derivative of reactive power The relationship between the acceleration of reactive power (i.e., the second-order derivative) has an important influence on stability, which takes into account the dynamic characteristics of power changes. The sum term in the formula In Compute Node Its neighboring nodes The voltage difference between the The voltage difference affects the current flow through the impedance, reflecting the node The interaction with its neighboring nodes and its impact on stability. To adjust the sensitivity of voltage to current changes. The value increases the effect of current fluctuation on node stability and helps to identify the effect of current changes on the node. To adjust the weight of the acceleration of power and reactive power change. The value emphasizes the impact of the dynamic characteristics of reactive power changes on stability and helps to identify potential risks of power fluctuations. is the impedance influence weight coefficient, which is used to adjust the impact of the voltage difference of neighboring nodes on node stability. The value increases the impact of neighborhood voltage differences on stability, which helps to identify mutual influences in the network. When using the conventional cooperative island detection formula in the field, the grid node By applying the collaborative island detection formula provided by the present invention, the stability coefficient of the power grid node can be calculated more accurately. The stability coefficient of the network. Through the integral term, the formula comprehensively considers the dynamic characteristics of current and power, which helps to comprehensively evaluate the stability of the node. The sum term considers the interaction between the node and the neighborhood by considering the voltage difference and impedance of the neighboring nodes, and can identify potential unstable factors caused by the network structure. Parameters , and It allows the focus of the formula to be adjusted according to actual conditions, providing flexible analysis tools to meet different grid stability analysis needs.

[0053] Preferably, step S44 includes the following steps:

[0054] Step S441: classifying abnormal power grid nodes into power grid risk levels according to node abnormality coefficients, and generating power grid risk level classification data, wherein the power grid risk level classification data includes low-risk power grid nodes, medium-risk power grid nodes, and high-risk power grid nodes;

[0055] Step S442: Continuously monitor abnormal changes of low-risk power grid nodes to generate protection strategies for low-risk power grid nodes; perform off-grid transfer of medium-risk power grid nodes to generate protection strategies for medium-risk power grid nodes; perform node shutdown processing on high-risk power grid nodes to generate protection strategies for high-risk power grid nodes;

[0056] Step S443: Regulate the grid protection strategy of abnormal grid nodes according to the low-risk grid node protection strategy, the medium-risk grid node protection strategy, and the high-risk grid node protection strategy, and generate a grid protection strategy to execute normal operation of the microgrid.

[0057] The present invention divides the abnormal power grid nodes into low risk, medium risk and high risk by dividing the grid risk level according to the node abnormality coefficient. This classification method can clarify the risk level of different nodes, help to formulate treatment measures in a targeted manner, and improve the accuracy of problem identification and the targetedness of treatment. Continuously monitoring the abnormal changes of low-risk power grid nodes and generating corresponding protection strategies can help to capture potential abnormal changes in time, ensure that the nodes can be effectively protected when small-scale abnormalities occur, and reduce the impact on the system. Off-grid transfer processing is performed on medium-risk power grid nodes to generate protection strategies. This measure can prevent the potential negative impact of medium-risk nodes on the power grid system and protect the stability of the entire system by temporarily isolating them. Node shutdown processing is performed on high-risk power grid nodes to generate protection strategies. This measure can immediately eliminate the threat of high-risk nodes to the power grid, reduce system risks through shutdown processing, and ensure the overall safety of the power grid. According to the protection strategies of different risk levels, the abnormal power grid nodes are regulated by the power grid protection strategy to generate a comprehensive power grid protection strategy. This comprehensive regulation ensures that all abnormal power grid nodes can be properly processed and protected, thereby improving the overall operation safety and stability of the microgrid. By implementing corresponding protection strategies for nodes with different risk levels, the risk of the system can be effectively reduced, ensuring that the power grid can quickly adjust and recover in abnormal situations, and enhancing the stability and reliability of the system. Through classification management and policy regulation, the efficiency of the system's response to abnormal situations can be improved. Ensuring that appropriate protection measures are implemented for nodes with different risk levels helps to quickly respond to various potential power grid problems and reduce the possibility of system failures.

[0058] In this specification, a microgrid island detection system is provided, which is used to execute the above-mentioned microgrid island detection method. The microgrid island detection system includes:

[0059] The perturbation signal generation module is used to obtain microgrid topology data; deploy distributed computing units on the microgrid topology data to generate distributed node deployment data; use sensors to collect microgrid operation data based on the distributed node deployment data to generate initial microgrid operation data; construct multi-dimensional perturbation signals on the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix;

[0060] The impedance analysis module is used to inject a small amount of reactive disturbance into the initial microgrid operation data using a multi-dimensional disturbance signal matrix to generate multi-dimensional disturbance grid change data; to perform multi-directional impedance fusion on the multi-dimensional disturbance grid change data to generate grid fusion impedance data; to perform dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data;

[0061] The island detection module is used to build an island detection framework; the island detection framework is compatible with distributed energy detection through the initial microgrid operation data to generate an adaptive island detection framework; the adaptive island detection framework is used to perform multi-mode collaborative island detection on the grid time series impedance analysis data to generate comprehensive grid island detection data; the distributed node deployment data is modeled for grid operation network topology based on the comprehensive grid island detection data to generate a grid node operation network topology model;

[0062] The grid protection module is used to perform grid node island detection and identification on the grid node operation network topology model to generate abnormal grid nodes; calculate the node abnormality degree of abnormal grid nodes through the grid comprehensive island detection data to obtain the node abnormality coefficient; rotate the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient, and generate a grid protection strategy to execute the normal operation of the microgrid.

[0063] The beneficial effect of the present invention is that by acquiring the microgrid topology data and deploying the distributed computing unit, the data collection and management of the distributed nodes are realized, the comprehensiveness and accuracy of the initial microgrid operation data are ensured, and the multi-dimensional perturbation signal is constructed to lay a solid foundation for subsequent detection. The microgrid is injected with a micro-reactive disturbance by using a multi-dimensional disturbance signal matrix, and combined with the multi-directional impedance fusion technology, the sensitivity and accuracy of the power grid change detection are effectively improved, and the dynamic time series change trend analysis further improves the real-time monitoring and analysis capabilities of the power grid state. By constructing an adaptive island detection framework, the system's compatibility with distributed energy is enhanced, and the multi-mode collaborative island detection improves the reliability and accuracy of the detection. The finally generated network topology model of the power grid node operation provides accurate basic data for further detection and protection measures. The network topology model of the power grid node operation is subjected to island detection and discrimination and the node abnormality coefficient is calculated, so that the abnormal power grid node can be accurately located, and the power grid protection strategy can be dynamically adjusted according to the degree of abnormality to ensure the safe and stable operation of the microgrid. Therefore, the present invention improves the comprehensiveness and reliability of the microgrid island detection through multi-dimensional disturbance signals and dynamic time series analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A schematic diagram of the steps of a microgrid island detection method;

[0065] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0066] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0067] Figure 4 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0068] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0069] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0070] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0071] It should be understood that although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, and the term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0072] To achieve this, please refer to Figures 1 to 4 , a microgrid island detection method, the method comprising the following steps:

[0073] Step S1: Acquire microgrid topology data; deploy distributed computing units on the microgrid topology data to generate distributed node deployment data; use sensors to collect microgrid operation data based on the distributed node deployment data to generate initial microgrid operation data; construct multi-dimensional perturbation signals on the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix;

[0074] Step S2: using a multi-dimensional disturbance signal matrix to perform a micro-reactive disturbance injection on the initial microgrid operation data to generate multi-dimensional disturbance grid change data; performing multi-directional impedance fusion on the multi-dimensional disturbance grid change data to generate grid fusion impedance data; performing dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data;

[0075] Step S3: construct an island detection framework; use the initial microgrid operation data to make the island detection framework compatible with distributed energy detection, and generate an adaptive island detection framework; use the adaptive island detection framework to perform multi-mode collaborative island detection on the grid time series impedance analysis data, and generate comprehensive grid island detection data; perform grid operation network topology modeling on the distributed node deployment data according to the comprehensive grid island detection data, and generate a grid node operation network topology model;

[0076] Step S4: Perform grid node island detection and discrimination on the grid node operation network topology model to generate abnormal grid nodes; calculate the node abnormality degree of abnormal grid nodes through the grid comprehensive island detection data to obtain the node abnormality coefficient; rotate the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient to generate a grid protection strategy to execute the normal operation of the microgrid.

[0077] The present invention ensures a comprehensive understanding of the entire power grid layout by acquiring the topological structure data of the microgrid, and provides the necessary basic information for subsequent analysis and detection. The deployment of distributed computing units helps to optimize the efficiency of data processing, makes the deployment of distributed nodes more reasonable, and reduces the waste of computing resources. The sensor collects data according to the distributed node deployment data, so that the initial microgrid operation data is more accurate and real-time, covering the key operating parameters of the microgrid. The construction of multi-dimensional perturbation signals can introduce more dimensional information, provide rich data features, and enhance the accuracy and sensitivity of subsequent analysis and detection. The micro-reactive disturbance injection of microgrid operation data using a multi-dimensional perturbation signal matrix effectively simulates the abnormal situation of the power grid and improves the sensitivity of detection. Multi-directional impedance fusion enables the multi-dimensional information of the data to be effectively integrated, improves the perception of changes in the power grid state, and enhances the comprehensiveness of detection. Dynamic time series change trend analysis helps to capture the change trend of the power grid impedance at different time points, further improves the real-time monitoring capability of the power grid state, and provides key data support for island detection. The island detection framework is constructed, and the distributed energy detection is compatible with the initial microgrid operation data, which improves the adaptability of the detection system and enables it to adapt to different microgrid structures and operation modes. The generation of the adaptive island detection framework enables the detection system to have self-adjustment capabilities, optimize the detection process according to the actual operation conditions, and improve the accuracy and efficiency of the detection. Through multi-mode collaborative island detection, combined with the advantages of multiple detection methods, the ability to identify island phenomena under complex power grid structures is improved, and the probability of false alarms and missed alarms is reduced. The final generated grid node operation network topology model provides accurate structural data support for subsequent analysis and processing, which helps to further optimize the management and operation of the power grid. The island detection and discrimination of the grid node operation network topology model can accurately identify abnormal grid nodes, thereby providing specific targets for subsequent processing. The calculation of the node abnormality degree makes the processing of abnormal nodes more targeted, and different countermeasures can be taken according to their severity, which improves the flexibility and effectiveness of dealing with abnormalities. Through the rotation of the grid protection strategy, the protection strategy can be dynamically adjusted according to the actual situation to ensure that the grid can remain stable and safe under different operating conditions, and enhance the risk resistance of the microgrid. Therefore, the present invention improves the comprehensiveness and reliability of microgrid island detection through multi-dimensional disturbance signals and dynamic timing analysis.

[0078] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a microgrid island detection method of the present invention. In this example, the microgrid island detection method includes the following steps:

[0079] Step S1: Acquire microgrid topology data; deploy distributed computing units on the microgrid topology data to generate distributed node deployment data; use sensors to collect microgrid operation data based on the distributed node deployment data to generate initial microgrid operation data; construct multi-dimensional perturbation signals on the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix;

[0080] In an embodiment of the present invention, the topological structure data of the microgrid is obtained by accessing the microgrid control center or database. This data includes information such as the connection relationship, cable layout, and switch position of each node in the microgrid. The obtained topological structure data is preprocessed, including removing redundant information, unifying the data format, etc., to ensure the accuracy and consistency of the data. The preprocessed microgrid topological structure data is stored in a distributed database for subsequent calculation and analysis. The microgrid topological structure data is analyzed to identify key nodes in the system, such as generator nodes, transformer nodes, load centers, etc. The identified key nodes are marked as high-priority nodes to provide a basis for the subsequent deployment of distributed computing units. According to the location of the key nodes and the overall structure of the microgrid, the entire microgrid is divided into several distributed computing units, each unit including one or more key nodes and their associated non-key nodes. Corresponding computing resources and data processing capabilities are allocated to each distributed computing unit, and distributed node deployment data is generated, and distributed node deployment data containing node information of each computing unit, computing resource allocation, etc. is generated and stored in a distributed computing platform. Corresponding sensors are deployed at the key nodes and non-key nodes of each distributed computing unit. These sensors are used to monitor the voltage, current, power, frequency and other operating parameters of the nodes in real time. Ensure the calibration accuracy of the sensors after installation, so that they can accurately reflect the actual operating status of the microgrid. Use the deployed sensors to collect the microgrid operating data in real time according to the distributed node deployment data. Transmit the collected operating data to the microgrid monitoring system through the network to ensure the synchronization and real-time nature of the data. Summarize and store the collected operating data to form the initial microgrid operating data, which provides a basis for subsequent analysis and processing. Analyze the initial microgrid operating data to extract key multidimensional feature data, such as voltage fluctuation, reactive power change, harmonic content, etc. The extracted feature data is constructed in multiple dimensions to form an initial data matrix containing the features of each dimension. Based on the actual situation of the power grid operation, design appropriate perturbation signals, which include small-amplitude voltage perturbations, power perturbations, etc., to simulate the dynamic response of the microgrid under different operating conditions. Inject the designed perturbation signals into the initial data matrix to construct a multidimensional signal containing perturbation features. The multidimensional data after the disturbance signal is injected is processed into a matrix to generate a multidimensional disturbance signal matrix. The matrix contains the response data of the microgrid under different disturbance conditions, which can reflect the dynamic characteristics and stability of the microgrid. The generated multidimensional disturbance signal matrix is ​​stored in the data management system for subsequent analysis and decision-making.

[0081] Step S2: using a multi-dimensional disturbance signal matrix to perform a micro-reactive disturbance injection on the initial microgrid operation data to generate multi-dimensional disturbance grid change data; performing multi-directional impedance fusion on the multi-dimensional disturbance grid change data to generate grid fusion impedance data; performing dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data;

[0082] In an embodiment of the present invention, a micro-reactive disturbance model is constructed based on a multidimensional disturbance signal matrix. The model takes into account the operating characteristics and reactive power requirements of the microgrid, and designs appropriate disturbance amplitudes and frequencies to ensure that the disturbance signal is small but detectable. Suitable disturbance signals, including voltage offsets, power fluctuations, etc., are selected as the input source of the reactive disturbance signal. In the key nodes of the microgrid or within the entire network, appropriate nodes are selected for micro-reactive disturbance injection according to the characteristics of the initial microgrid operation data. Using the multidimensional disturbance signal matrix, the designed micro-reactive disturbance signal is injected into the initial microgrid operation data to simulate the dynamic response of the microgrid when subjected to a small reactive disturbance. During the disturbance injection process, the operation data of the microgrid is collected in real time to form multidimensional disturbance grid change data, which reflects the response characteristics of the microgrid under different disturbance conditions. The multidimensional disturbance grid change data is analyzed to calculate the impedance characteristics of each node under different disturbance conditions. The impedance characteristics include parameters such as voltage change and power change, which can reflect the stability and anti-disturbance capability of the power grid. The calculated impedance characteristics are normalized to eliminate the influence between different dimensions and ensure the accuracy of impedance fusion. According to the topological structure and operating characteristics of the power grid, the impedance characteristics of the power grid are divided into different directions, such as north-south direction, east-west direction, etc. The multi-directional impedance fusion algorithm is applied to fuse the impedance characteristics of different directions to generate the fused impedance data of the power grid. The fusion algorithm can consider weighted average, nonlinear superposition and other methods to fully reflect the overall impedance characteristics of the power grid. The fused impedance data is stored as power grid fusion impedance data. This data is used for subsequent time series change trend analysis and can reflect the impedance characteristic changes of the power grid in a global range. The power grid fusion impedance data is arranged in time series to form a dynamic time series data set reflecting the impedance changes of the power grid. The time series data is smoothed to eliminate noise and abnormal fluctuations to ensure that the time series trend is clear and visible. Using the time series analysis algorithm, the dynamic time series change trend analysis of the power grid fusion impedance data is carried out to identify the main trends and laws of the power grid impedance changes over time. The key time series features, such as impedance rising trend, falling trend, periodic changes, etc., are extracted to provide a basis for subsequent stability assessment and risk prediction. The analyzed time series characteristic data is stored as grid time series impedance analysis data. This data is used for further island detection and grid protection strategy formulation to ensure the stable operation of the microgrid.

[0083] Step S3: construct an island detection framework; use the initial microgrid operation data to make the island detection framework compatible with distributed energy detection, and generate an adaptive island detection framework; use the adaptive island detection framework to perform multi-mode collaborative island detection on the grid time series impedance analysis data, and generate comprehensive grid island detection data; perform grid operation network topology modeling on the distributed node deployment data according to the comprehensive grid island detection data, and generate a grid node operation network topology model;

[0084] In the embodiment of the present invention, the basic structure of the island detection framework is designed based on the operating characteristics of the microgrid and the island detection requirements. The structure includes a detection node, a data acquisition unit, an analysis module, and a decision module to form a multi-level and multi-dimensional detection system. The island detection framework adopts a modular design to facilitate expansion and upgrading. Each module is responsible for different functions, such as signal acquisition, data processing, anomaly detection, etc. The initial microgrid operation data is collected and preprocessed to ensure the accuracy and integrity of the data. The preprocessing includes operations such as data cleaning and noise filtering. The initial microgrid operation data is analyzed to determine the adaptability of the island detection framework, identify compatibility issues, and make adjustment suggestions. In the island detection framework, a distributed energy detection module is integrated to enable the framework to adapt to different types of distributed energy (such as solar energy, wind energy, energy storage system, etc.). The distributed energy detection module is tested for compatibility to ensure that it can seamlessly connect with different parts of the initial microgrid operation data to generate an adaptive island detection framework. The introduction of an adaptive adjustment mechanism enables the framework to be adjusted in real time according to different operating environments and energy types, thereby enhancing the accuracy and robustness of island detection. The framework is optimized through data analysis and feedback adjustment to ensure that it can efficiently process the power grid time series impedance analysis data. The power grid time series impedance analysis data is decomposed and classified to form data subsets with different modes, such as low-frequency mode, high-frequency mode, and abnormal mode. Key features, such as impedance change rate and time series fluctuation amplitude, are extracted from each data subset to provide a basis for collaborative detection. According to the characteristics of different modes, multiple detection strategies are designed, and multi-mode collaborative detection is realized through an adaptive island detection framework. The detection results under different modes are fused to generate comprehensive power grid island detection data. The fusion algorithm can use weighted average, Bayesian reasoning and other methods to ensure the comprehensiveness and accuracy of the detection results. The distributed node deployment data is analyzed in detail to identify key nodes and important connection relationships, providing basic data for power grid operation network topology modeling. According to the importance and function of the nodes, they are classified and prioritized to determine the basic structure of the network topology. Based on the distributed node deployment data, a power grid operation network topology model is constructed. The model reflects the connection relationship and operation status between the nodes in the power grid, and can be used for subsequent operation monitoring and anomaly detection. The constructed topology model is optimized to ensure that it can accurately reflect the actual operation of the power grid, especially the network changes under island detection. The generated grid node operation network topology model is stored in the database for subsequent query, analysis and application. The topology model is applied to the real-time monitoring and operation management of the power grid. Through the analysis of the model, potential island risks and operation anomalies can be discovered and handled in a timely manner.

[0085] Step S4: Perform grid node island detection and discrimination on the grid node operation network topology model to generate abnormal grid nodes; calculate the node abnormality degree of abnormal grid nodes through the grid comprehensive island detection data to obtain the node abnormality coefficient; rotate the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient to generate a grid protection strategy to execute the normal operation of the microgrid.

[0086] In an embodiment of the present invention, by taking the network topology model of the power grid node operation as input, the structure is parsed to identify the key nodes and their connection relationships therein. The discrimination conditions for island detection are set, and these conditions may include the isolation of the node, reactive power offset, timing impedance anomaly, etc. The nodes in the topology model are analyzed by the island detection algorithm to identify the nodes that become islands. The algorithm can be based on characteristics such as timing analysis and impedance change trend. The detected island nodes are marked to generate a list of abnormal power grid nodes. Each abnormal node will be accompanied by a detection index and a judgment basis. The calculation formula of the node abnormality coefficient is defined. For example, the abnormality coefficient can be a weighted average of multiple influencing factors, including impedance abnormality degree, reactive power offset, node stability, etc. The comprehensive island detection data of the power grid is taken as input, including impedance abnormality degree data, reactive power offset data, node stability data, etc. According to the above formula, the node abnormality coefficient of each abnormal power grid node is calculated. The higher the abnormality coefficient, the greater the risk of the node. According to the node abnormality coefficient, the abnormal power grid nodes are divided into low-risk, medium-risk and high-risk nodes. The threshold for risk level division is set. For example, nodes below a certain threshold are low risk, those between two thresholds are medium risk, and those above the upper threshold are high risk. Continuously monitor low-risk nodes to prevent abnormal expansion. Develop corresponding protection strategies, such as locally adjusting operating parameters or increasing monitoring frequency. Take more proactive measures for medium-risk nodes, such as temporarily disconnecting nodes from the grid, or limiting their load capacity, and develop off-grid transfer strategies for medium-risk nodes. Take emergency shutdown measures for high-risk nodes to prevent threats to the stability of the entire microgrid. Develop shutdown protection strategies for high-risk nodes. According to the protection strategies for low, medium, and high-risk nodes, implement protection operations on the corresponding abnormal grid nodes. During the execution of the strategy, monitor the node status in real time, and adjust the protection strategy in time according to the operation of the node to ensure the stable operation of the microgrid.

[0087] Preferably, step S1 comprises the following steps:

[0088] Step S11: Acquire microgrid topology data;

[0089] Step S12: Identify key nodes of the microgrid topology data to obtain key node data of the microgrid topology;

[0090] Step S13: deploying distributed computing units on the microgrid based on the key node data of the microgrid topology to generate distributed node deployment data;

[0091] Step S14: using sensors to collect microgrid operation data according to distributed node deployment data to generate initial microgrid operation data;

[0092] Step S15: construct a multi-dimensional perturbation signal for the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix.

[0093] In an embodiment of the present invention, the topological structure information of the microgrid is obtained from the actual physical structure and configuration file of the microgrid, including the connection relationship of components such as power generation units, distribution lines, loads, and energy storage systems. These structural information are extracted through data acquisition equipment or from an existing database to form a microgrid topology diagram containing nodes and connecting lines. The topological structure data is preliminarily verified to ensure data integrity and accuracy, and finally the microgrid topological structure data is generated. The acquired microgrid topological structure data is analyzed to identify key nodes that have a significant impact on power transmission and system stability, such as trunk line connection points, important load nodes, and power generation sources. A graph theory algorithm (such as a PageRank algorithm or a shortest path algorithm) is used to quantify the importance of each node and determine the key nodes. The identified key nodes are classified and marked to generate microgrid topology key node data. According to the distribution of key nodes and the overall structure of the microgrid, a suitable distributed computing unit deployment scheme is designed, taking into account communication delay, load balancing, and fault tolerance between nodes. Distributed computing units are deployed at each key node for real-time monitoring and processing of data of related nodes. The specific deployment information of the distributed computing unit at each key node is recorded and generated, and the distributed node deployment data is finally generated. Connect the corresponding sensors to each deployed distributed computing unit to collect the operation data of each part of the microgrid, such as voltage, current, power, frequency, etc. Ensure the real-time and accuracy of sensor data, and transmit the data to the central control unit through distributed nodes. Perform preliminary processing on the collected microgrid operation data, such as denoising, time synchronization, etc., and finally generate the initial microgrid operation data. Analyze the initial microgrid operation data to identify the dynamic change patterns and potential perturbation signals. Use signal processing techniques (such as wavelet transform or Fourier transform) to decompose and reconstruct the operation data to construct multidimensional perturbation signals. Matrix these perturbation signals to form a multidimensional perturbation signal matrix, which can be used for further analysis and optimization control.

[0094] Preferably, step S14 comprises the following steps:

[0095] Step S151: extracting multi-dimensional grid feature data from the initial microgrid operation data to obtain multi-dimensional feature data of the microgrid, wherein the multi-dimensional feature data of the microgrid includes microgrid voltage data, microgrid reactive power data, and microgrid harmonic data;

[0096] Step S152: injecting a perturbation signal through the microgrid voltage data, the microgrid reactive power data and the microgrid harmonic data to generate an initial microgrid perturbation signal; performing a signal spectrum characteristic analysis on the initial microgrid perturbation signal to generate perturbation signal spectrum characteristic data;

[0097] Step S153: performing multi-frequency domain superposition on the initial perturbation signal of the microgrid according to the perturbation signal spectrum characteristic data to generate a multi-frequency domain disturbance signal set; performing nonlinear signal superposition on the multi-frequency domain disturbance signal set to generate a nonlinear disturbance signal matrix;

[0098] Step S154: performing perturbation signal spectrum equalization on the nonlinear perturbation signal matrix to generate a multi-dimensional perturbation signal matrix.

[0099] In an embodiment of the present invention, the voltage data of each node is extracted from the initial microgrid operation data to analyze the voltage fluctuation, stability and change trend. The initial data is subjected to power factor analysis, reactive power data is extracted, and the reactive power demand and supply of each node are identified. Harmonic analysis methods (such as fast Fourier transform, FFT) are used to extract the harmonic components of each node and identify the amplitude and phase of the main harmonics in the spectrum. The extracted voltage data, reactive power data, and harmonic data are standardized to form unified multi-dimensional feature data of the microgrid, providing a basis for subsequent analysis. Based on the extracted microgrid voltage data, reactive power data, and harmonic data, synthetic perturbation signals are generated, which can be injected into the microgrid system by adding noise, disturbance, or simulating faults. In a simulation or experimental environment, the generated perturbation signal is injected into the microgrid operation data to generate the initial microgrid perturbation signal. The initial microgrid perturbation signal is subjected to spectrum analysis, and tools such as FFT are used to analyze the energy distribution of each frequency component to generate perturbation signal spectrum characteristic data, which will reveal the frequency characteristics of the signal and its potential impact. Using the spectrum characteristic data, the perturbation signals in different frequency domains are superimposed to generate a disturbance signal set containing multiple frequency domain characteristics. These signals can reflect the response characteristics of the microgrid in multiple frequency domains. The nonlinear superposition technology (such as convolution or nonlinear transformation) is applied to superimpose the multi-frequency domain disturbance signal set to generate a nonlinear disturbance signal matrix to simulate the behavior of the microgrid under complex dynamic conditions. The generated signal set is organized in the form of a matrix to form a nonlinear disturbance signal matrix that can reflect the complex dynamic characteristics of the microgrid. By performing spectrum equalization on the nonlinear disturbance signal matrix, the amplitude of different frequency components is adjusted to ensure that the signal distribution in the matrix is ​​more balanced and stable. On the basis of spectrum equalization, the data in the matrix is ​​further processed to eliminate the excessive concentration and sudden abnormalities of the signal, and finally a multidimensional disturbance signal matrix is ​​formed. The generated multidimensional disturbance signal matrix is ​​verified to ensure that it can accurately reflect the dynamic characteristics of the microgrid, and further adjustment and optimization are performed when necessary.

[0100] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0101] Step S21: injecting a micro-reactive disturbance into the initial microgrid operation data using a multi-dimensional disturbance signal matrix to generate multi-dimensional disturbance grid change data;

[0102] Step S22: extracting frequency components from the multi-dimensional disturbance power grid change data to obtain a power grid change multi-frequency disturbance component set; calculating independent impedance components from the power grid change multi-frequency disturbance component set to generate power grid independent impedance component data;

[0103] Step S23: performing nonlinear error correction on the independent impedance component data of the power grid through nonlinear effects to generate independent corrected impedance component data of the power grid;

[0104] Step S24: performing multi-directional impedance fusion on the grid independent correction impedance component data to generate grid fusion impedance data; performing dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data.

[0105] In an embodiment of the present invention, signal components related to reactive power are selected from a multidimensional disturbance signal matrix, and these signals are injected into the initial microgrid operation data in the form of micro-disturbance. The injection process can be performed through a simulator or an actual control device to ensure that the amplitude of the disturbance is small enough and will not affect the normal operation of the microgrid. After the micro-disturbance is injected, the operating state of the power grid is monitored in real time through a distributed sensor network, and the changes in the microgrid after the injection are recorded. The collected operating data is compared with the original data to identify the small changes caused by the disturbance injection, and multi-dimensional disturbance power grid change data is generated as the basis for subsequent analysis. The frequency domain analysis of the multi-dimensional disturbance power grid change data is performed using spectrum analysis technology (such as fast Fourier transform, FFT), and the frequency components therein are extracted to form a multi-frequency disturbance component set. The component set contains the response characteristics of the power grid at different frequencies. For each extracted frequency component, an independent impedance calculation is performed. By measuring the phase difference and amplitude of the voltage and current, the impedance component of the corresponding frequency can be calculated. The impedance Z(f) corresponding to each frequency component is calculated using the formula Z(f)=V(f) / I(f). Normalize the impedance components of all frequencies to generate a grid-independent impedance component data set, which is stored in a matrix format for subsequent nonlinear effect analysis. Perform nonlinear analysis on the grid-independent impedance component data to identify errors caused by nonlinear effects (such as harmonic distortion, equipment saturation, etc.). Use nonlinear regression or machine learning models to predict and quantify these errors. Construct an error correction model to separate the identified nonlinear error components from the original impedance data for correction. The correction formula can be expressed as: ,in The error is caused by the nonlinear effect. The corrected impedance component data is reorganized to generate independent corrected impedance component data of the power grid to ensure the accuracy of the data and to be used for subsequent analysis. According to the spatial and temporal characteristics of the power grid, the independent corrected impedance component data is fused in multiple directions. The fusion process can consider the impedance coupling relationship between power grid nodes in different directions. The fusion method can adopt weighted average or fusion algorithm based on network topology. The fused impedance data should be able to reflect the impedance characteristics of the entire power grid and generate power grid fused impedance data to ensure its comprehensiveness and globality. The fused impedance data is analyzed in time series, and the time series analysis method (such as autoregressive model, ARIMA) is used to identify the trend and periodic characteristics of impedance changes over time. The analyzed time series change trend data is sorted to generate power grid time series impedance analysis data. This data can be used to predict the future operation status of the power grid and provide a basis for optimization control.

[0106] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0107] Step S31: construct an island detection framework; perform grid operation energy analysis on the initial microgrid operation data to generate grid operation energy division data;

[0108] Step S32: Perform distributed energy detection compatibility on the island detection framework through the grid operation energy division data to generate an adaptive island detection framework;

[0109] Step S33: using an adaptive island detection framework to perform multi-mode collaborative island detection on the grid time series impedance analysis data to generate comprehensive grid island detection data;

[0110] Step S34: Perform grid operation network topology modeling on the distributed node deployment data according to the grid comprehensive island detection data to generate a grid node operation network topology model.

[0111] In an embodiment of the present invention, an extensible island detection framework is designed based on the island effect characteristics of the power system. The framework should include a monitoring module, a data processing module and a detection algorithm module of the grid node to ensure that the island state of the grid can be monitored and analyzed in real time. The framework should define an interface with the microgrid operation data to ensure that the microgrid operation data can be obtained in real time for island detection. The initial microgrid operation data is analyzed to identify different types of energy supply and their distribution in the grid. Common energy types include wind energy, solar energy, energy storage devices, etc. The balance relationship between the power generation and load demand of each energy supply is analyzed to identify the operation mode that causes the island phenomenon. According to the analysis results, the grid operation energy division data is generated, which represents the energy classification information and its distribution in the grid in the form of a matrix or a chart. According to the grid operation energy division data, the island detection framework is adjusted to make it compatible with different types of distributed energy. The extended detection module should be able to process characteristic data of different energy types, such as intermittency, energy storage characteristics, etc. According to the characteristics of different energy sources, the detection parameters and algorithms in the framework are adjusted to enable it to adaptively detect the island state of various types of distributed energy. Based on the extended detection module, the detection algorithm of the framework is optimized to enable it to maintain efficient island detection capabilities under dynamically changing energy distribution. All extended modules and optimized algorithms are integrated into the island detection framework to generate an island detection framework with adaptive capabilities. Multiple island detection modes are defined in the adaptive island detection framework, such as detection modes based on voltage fluctuations, frequency changes, impedance mutations, etc. Multiple island detection modes are applied simultaneously to analyze the grid time series impedance analysis data to ensure the accuracy and comprehensiveness of the detection results. The island status data obtained by each mode detection are integrated to eliminate redundant information and generate a unified grid island status data set. The integrated data is further comprehensively analyzed to identify the island area in the grid and its dynamic changes, and generate comprehensive grid island detection data. The comprehensive grid island detection data is used to analyze the island area in the grid and its impact on the overall grid operation, and determine the node deployment strategy that needs to be adjusted. According to the island detection results, the operation importance of each node is evaluated, and the nodes that need to be monitored or adjusted are identified. Based on the distributed node deployment data and combined with the island detection data, a network topology model for grid operation is constructed. The model should be able to reflect the connection relationship between nodes, data flow and its changes in the island state. According to the actual detection results and operation requirements, the node connection and data path in the topology model are dynamically adjusted to ensure that the power grid can still maintain stable operation in the island state. The constructed power grid node operation network topology model is simulated and verified to ensure that the model can accurately reflect the actual operation of the power grid. According to the simulation results, the node deployment and connection strategy in the model are further optimized to finally generate a stable power grid node operation network topology model.

[0112] Preferably, step S33 includes the following steps:

[0113] Step S331: classifying the power grid energy types for the power grid time series impedance analysis data to generate power grid single energy impedance analysis data and power grid multi-energy impedance analysis data;

[0114] Step S332: using the adaptive island detection framework to perform single energy reactive power offset detection on the single energy impedance analysis data of the power grid, and generating single energy reactive power offset detection data; comparing the single energy reactive power offset detection value with the preset reactive power offset detection threshold, thereby generating single energy microgrid island detection data;

[0115] Step S333: performing energy type identification on the power grid multi-energy impedance analysis data to obtain energy type identification data; performing multi-energy change feature extraction based on the energy type identification data to obtain multi-energy type change feature data;

[0116] Step S334: performing time domain analysis on the power grid multi-energy impedance analysis data to generate a power grid multi-energy impedance time domain curve; performing same peak time domain curve cutting on the power grid multi-energy impedance time domain curve based on a preset peak acquisition range to generate a time domain peak cutting curve;

[0117] Step S335: performing feature change matching on the time domain peak cutting curve and the multi-energy type change feature data to generate a multi-energy feature change matching curve; performing reactive power offset detection on the multi-energy feature change matching curve to generate multi-energy reactive power offset detection data; comparing the multi-energy reactive power offset detection value with a preset reactive power offset detection threshold to generate multi-energy microgrid island detection data;

[0118] Step S336: Perform collaborative island detection on the single-energy microgrid island detection data and the multi-energy microgrid island detection data to generate comprehensive grid island detection data.

[0119] In an embodiment of the present invention, the grid time series impedance analysis data is preprocessed to identify different energy types therein, such as solar energy, wind energy, energy storage system, etc. The classification algorithm is used to divide the time series impedance data to distinguish single energy impedance data and impedance data of multiple energy combinations. The grid single energy impedance analysis data (including time series impedance data of a single energy type) and the grid multi-energy impedance analysis data (including time series impedance data of multiple energy types) are generated. The grid single energy impedance analysis data is input into the adaptive island detection framework to perform reactive power offset detection, calculate the reactive power offset of a single energy, and obtain detection data. A preset reactive power offset detection threshold is used as a benchmark to determine whether the detection value exceeds the threshold. If the detection value exceeds the threshold, a single energy microgrid island detection data is generated, and potential island risk areas are marked. The grid multi-energy impedance analysis data is processed using an energy identification algorithm to identify the different energy types contained therein, and the identified energy types are recorded as energy type identification data. Based on the energy type identification data, the change characteristics between multiple energy sources, such as frequency changes, power fluctuations, etc., are analyzed, these characteristics are extracted, and multi-energy type change characteristic data are generated. Perform time domain analysis on the multi-energy impedance analysis data to generate a time domain curve reflecting the change of multi-energy impedance. According to the preset peak acquisition range, process the time domain curve, identify and cut out the same peak time domain segment, and save the cut curve as the time domain peak cutting curve for subsequent analysis. Perform matching analysis on the time domain peak cutting curve and the multi-energy type change characteristic data to identify the characteristic change trend of different energy types, and generate a multi-energy characteristic change matching curve based on the matching results. Perform reactive power offset detection on the multi-energy characteristic change matching curve to obtain multi-energy reactive power offset detection data. Compare the detection data with the preset reactive power offset threshold to determine whether there is an abnormality. If the detection data exceeds the threshold range, generate multi-energy microgrid island detection data to mark potential island risks. Integrate the single energy and multi-energy island detection data to form a complete island detection data set. Use the collaborative analysis algorithm to further analyze the integrated data to identify more complex island phenomena or potential risks. Based on the collaborative detection results, generate comprehensive grid island detection data for further grid topology modeling and optimization.

[0120] As an example of the present invention, refer to Figure 4 As shown, in this example, step S4 includes:

[0121] Step S41: performing grid node island detection and discrimination on the grid node operation network topology model. When a grid node island is detected in the grid node operation network topology model, the corresponding grid node is marked as an abnormal grid node.

[0122] Step S42: Performing grid stability analysis on abnormal grid nodes through grid comprehensive island detection data to generate grid node stability analysis data; performing node load performance analysis on abnormal grid nodes to generate node reliability analysis data;

[0123] Step S43: Calculate the node abnormality degree of the abnormal power grid node based on the power grid node stability maintenance calculation formula through the power grid node stability analysis data and the node reliability analysis data to obtain the node abnormality coefficient;

[0124] Step S44: rotating the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient, generating a grid protection strategy to execute normal operation of the microgrid.

[0125] In an embodiment of the present invention, by analyzing the network topology model of the grid node operation, each grid node in the model and its connection relationship are identified. The grid nodes in the model are detected for islanding using an island detection algorithm, including disconnected nodes or isolated areas. When an islanding phenomenon is detected in a grid node or area, the corresponding node is marked as an abnormal grid node. All grid nodes marked as abnormal are recorded, and a list of abnormal nodes is generated for subsequent analysis. The grid comprehensive island detection data is input into the grid stability analysis system to evaluate the stability of the marked abnormal grid nodes. The stability indicators of the node, such as voltage fluctuation, frequency change, and load fluctuation, are analyzed to generate grid node stability analysis data. The load analysis model is applied to the abnormal grid nodes to evaluate the load bearing capacity, power transmission efficiency, load distribution and other performance of the nodes. Node reliability analysis data is generated based on the analysis results, and the reliability indicators and potential risks of the nodes are recorded. The grid node stability calculation formula is applied, and the grid node stability analysis data and the node reliability analysis data are combined to perform a comprehensive analysis on the abnormal nodes. The grid node stability analysis data and the node reliability analysis data are input into the formula to calculate the node abnormality coefficient of each abnormal grid node. According to the calculation results, the abnormal coefficient of each abnormal node is recorded for subsequent protection strategy formulation. According to the calculated node abnormal coefficient, the appropriate protection strategy is selected for the abnormal grid node to ensure the safe operation of the grid. For nodes with high abnormal coefficients, more stringent protection measures are implemented first, such as load transfer, line switching, isolation protection, etc. A complete set of grid protection strategies is generated, including specific operation steps and emergency measures to ensure the normal operation of the microgrid. The formulated grid protection strategy is applied to the actual operation of the microgrid to ensure that the microgrid can still operate stably under islanding or abnormal conditions.

[0126] Preferably, the grid node stability calculation formula in step S43 is as follows:

[0127]

[0128] In the formula, Represented as a grid node The stability coefficient, It represents the time interval for power grid stability analysis. Represented as a node In time The voltage value, Represented as a node In time The current value, Represented as a node In time The active power, Represented as a node In time The reactive power, Expressed as voltage influence weight coefficient, Expressed as the power impact weight coefficient, Expressed as the impedance influence weight coefficient, Represented as a node The neighboring set of other connected nodes, Represented as a node With Node The impedance value between.

[0129] The present invention analyzes and integrates a collaborative island detection formula. The principle of the formula is to evaluate the stability of the power grid node by comprehensively considering the changes in node voltage, current, active and reactive power, and the voltage difference of neighboring nodes. The formula consists of two main parts: integral term and sum term. The integral term in the formula In Reflects the voltage With current The dynamic relationship between voltage and current. The sensitivity of voltage to current change is the key to measure node stability. The rate of change of current The product with the voltage represents the effect of node current fluctuation on stability. Reflects the node power and the second derivative of reactive power The relationship between the acceleration of reactive power (i.e., the second-order derivative) has an important influence on stability, which takes into account the dynamic characteristics of power changes. The sum term in the formula In Compute Node Its neighboring nodes The voltage difference between the The voltage difference affects the current flow through the impedance, reflecting the node The interaction with its neighboring nodes and its impact on stability. To adjust the sensitivity of voltage to current changes. The value increases the effect of current fluctuation on node stability and helps to identify the effect of current changes on the node. To adjust the weight of the acceleration of power and reactive power change. The value emphasizes the impact of the dynamic characteristics of reactive power changes on stability and helps to identify potential risks of power fluctuations. is the impedance influence weight coefficient, which is used to adjust the impact of the voltage difference of neighboring nodes on node stability. The value increases the impact of neighborhood voltage differences on stability, which helps to identify mutual influences in the network. When using the conventional cooperative island detection formula in the field, the grid node By applying the collaborative island detection formula provided by the present invention, the stability coefficient of the power grid node can be calculated more accurately. The stability coefficient of the network. Through the integral term, the formula comprehensively considers the dynamic characteristics of current and power, which helps to comprehensively evaluate the stability of the node. The sum term considers the interaction between the node and the neighborhood by considering the voltage difference and impedance of the neighboring nodes, and can identify potential unstable factors caused by the network structure. Parameters , and It allows the focus of the formula to be adjusted according to actual conditions, providing flexible analysis tools to meet different grid stability analysis needs.

[0130] Preferably, step S44 includes the following steps:

[0131] Step S441: classifying abnormal power grid nodes into power grid risk levels according to node abnormality coefficients, and generating power grid risk level classification data, wherein the power grid risk level classification data includes low-risk power grid nodes, medium-risk power grid nodes, and high-risk power grid nodes;

[0132] Step S442: Continuously monitor abnormal changes of low-risk power grid nodes to generate protection strategies for low-risk power grid nodes; perform off-grid transfer of medium-risk power grid nodes to generate protection strategies for medium-risk power grid nodes; perform node shutdown processing on high-risk power grid nodes to generate protection strategies for high-risk power grid nodes;

[0133] Step S443: Regulate the grid protection strategy of abnormal grid nodes according to the low-risk grid node protection strategy, the medium-risk grid node protection strategy, and the high-risk grid node protection strategy, and generate a grid protection strategy to execute normal operation of the microgrid.

[0134] In the embodiment of the present invention, risk assessment is performed on each abnormal power grid node based on the node abnormality coefficient calculated in the previous step. If the node abnormality coefficient is lower than the preset low risk threshold, the abnormal power grid node is marked as a low risk power grid node. If the node abnormality coefficient is between the low risk threshold and the high risk threshold, the abnormal power grid node is marked as a medium risk power grid node. If the node abnormality coefficient is higher than the preset high risk threshold, the abnormal power grid node is marked as a high risk power grid node. Each abnormal node is classified as low risk, medium risk or high risk, and power grid risk level classification data is generated, and the risk level of each abnormal node is recorded as the basis for subsequent strategy formulation. The power grid nodes marked as low risk are continuously monitored to monitor their abnormal changes. According to the monitoring results, a low risk power grid node protection strategy is formulated to ensure that measures can be taken in time when the abnormal changes exceed the preset range. The off-grid transfer operation is implemented for the medium risk power grid node, that is, the node is isolated from the power grid to reduce the potential impact on the overall power grid. A protection strategy after off-grid is generated for the medium risk node to ensure that the node can operate stably during the transfer process. Emergency shutdown processing is performed on the high risk power grid node to prevent it from causing serious impact on the power grid system. Develop detailed shutdown protection strategies for high-risk nodes, including emergency response measures and recovery plans after shutdown. Integrate the protection strategies for low-risk, medium-risk, and high-risk nodes to form a comprehensive grid protection strategy. Based on the integrated protection strategy, dynamically regulate abnormal grid nodes to ensure that risk nodes at all levels are properly handled. Apply the generated grid protection strategy to the actual microgrid operation process to ensure the safety, stability, and normal operation of the grid. During the execution process, monitor the effectiveness of the protection strategy in real time and make necessary adjustments based on actual conditions.

[0135] In this specification, a microgrid island detection system is provided, which is used to execute the above-mentioned microgrid island detection method. The microgrid island detection system includes:

[0136] The perturbation signal generation module is used to obtain microgrid topology data; deploy distributed computing units on the microgrid topology data to generate distributed node deployment data; use sensors to collect microgrid operation data based on the distributed node deployment data to generate initial microgrid operation data; construct multi-dimensional perturbation signals on the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix;

[0137] The impedance analysis module is used to inject a small amount of reactive disturbance into the initial microgrid operation data using a multi-dimensional disturbance signal matrix to generate multi-dimensional disturbance grid change data; to perform multi-directional impedance fusion on the multi-dimensional disturbance grid change data to generate grid fusion impedance data; to perform dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data;

[0138] The island detection module is used to build an island detection framework; the island detection framework is compatible with distributed energy detection through the initial microgrid operation data to generate an adaptive island detection framework; the adaptive island detection framework is used to perform multi-mode collaborative island detection on the grid time series impedance analysis data to generate comprehensive grid island detection data; the distributed node deployment data is modeled for grid operation network topology based on the comprehensive grid island detection data to generate a grid node operation network topology model;

[0139] The grid protection module is used to perform grid node island detection and identification on the grid node operation network topology model to generate abnormal grid nodes; calculate the node abnormality degree of abnormal grid nodes through the grid comprehensive island detection data to obtain the node abnormality coefficient; rotate the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient, and generate a grid protection strategy to execute the normal operation of the microgrid.

[0140] The present invention realizes data collection and management of distributed nodes by acquiring microgrid topology data and deploying distributed computing units, ensures the comprehensiveness and accuracy of initial microgrid operation data, and lays a solid foundation for subsequent detection through multi-dimensional perturbation signal construction. The microgrid is injected with micro-reactive disturbance by using a multi-dimensional disturbance signal matrix, and combined with multi-directional impedance fusion technology, the sensitivity and accuracy of power grid change detection are effectively improved, and the dynamic time series change trend analysis further improves the real-time monitoring and analysis capabilities of the power grid state. By constructing an adaptive island detection framework, the system's compatibility with distributed energy is enhanced, and multi-mode collaborative island detection improves the reliability and accuracy of detection. The resulting network topology model of the power grid node operation provides accurate basic data for further detection and protection measures. The network topology model of the power grid node operation is subjected to island detection and discrimination and the node abnormality coefficient is calculated, so that abnormal power grid nodes can be accurately located, and the power grid protection strategy can be dynamically adjusted according to the degree of abnormality to ensure the safe and stable operation of the microgrid. Therefore, the present invention improves the comprehensiveness and reliability of microgrid island detection through multi-dimensional disturbance signals and dynamic time series analysis.

[0141] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A microgrid island detection method, characterized in that: The following steps are involved: Step S1: Acquire microgrid topology data; Perform distributed computing unit deployment on microgrid topology data to generate distributed node deployment data; Using sensors to collect microgrid operation data based on distributed node deployment data to generate initial microgrid operation data; constructing multi-dimensional perturbation signals for the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix; Step S2: injecting a small amount of reactive disturbance into the initial microgrid operation data using a multidimensional disturbance signal matrix to generate multidimensional disturbance grid change data; Perform multi-directional impedance fusion on multi-dimensional disturbance power grid change data to generate power grid fusion impedance data; perform dynamic time series change trend analysis on power grid fusion impedance data to generate power grid time series impedance analysis data; Step S3: construct an island detection framework; use the initial microgrid operation data to make the island detection framework compatible with distributed energy detection to generate an adaptive island detection framework; use the adaptive island detection framework to perform multi-mode collaborative island detection on the grid time series impedance analysis data to generate comprehensive grid island detection data; According to the power grid integrated island detection data, the distributed node deployment data is modeled for the power grid operation network topology to generate a power grid node operation network topology model; step S3 includes the following steps: Step S31: construct an island detection framework; perform grid operation energy analysis on the initial microgrid operation data to generate grid operation energy division data; Step S32: Perform distributed energy detection compatibility on the island detection framework through the grid operation energy division data to generate an adaptive island detection framework; Step S33: using an adaptive island detection framework to perform multi-mode collaborative island detection on the grid time series impedance analysis data to generate grid comprehensive island detection data; Step S33 includes the following steps: Step S331: classifying the power grid time-series impedance analysis data by power grid energy type, generating power grid single energy impedance analysis data and power grid multi-energy impedance analysis data; Step S332: using the adaptive island detection framework to perform single energy reactive power offset detection on the single energy impedance analysis data of the power grid, and generating single energy reactive power offset detection data; comparing the single energy reactive power offset detection value with the preset reactive power offset detection threshold, thereby generating single energy microgrid island detection data; Step S333: performing energy type identification on the power grid multi-energy impedance analysis data to obtain energy type identification data; performing multi-energy change feature extraction based on the energy type identification data to obtain multi-energy type change feature data; Step S334: performing time domain analysis on the power grid multi-energy impedance analysis data to generate a power grid multi-energy impedance time domain curve; performing same peak time domain curve cutting on the power grid multi-energy impedance time domain curve based on a preset peak acquisition range to generate a time domain peak cutting curve; Step S335: performing feature change matching on the time domain peak cutting curve and the multi-energy type change feature data to generate a multi-energy feature change matching curve; performing reactive power offset detection on the multi-energy feature change matching curve to generate multi-energy reactive power offset detection data; comparing the multi-energy reactive power offset detection value with a preset reactive power offset detection threshold to generate multi-energy microgrid island detection data; Step S336: Perform collaborative island detection on the single-energy microgrid island detection data and the multi-energy microgrid island detection data to generate comprehensive grid island detection data; Step S34: performing grid operation network topology modeling on the distributed node deployment data according to the grid comprehensive island detection data, and generating a grid node operation network topology model; Step S4: Perform grid node island detection and discrimination on the grid node operation network topology model to generate abnormal grid nodes; calculate the node abnormality degree of abnormal grid nodes through the grid comprehensive island detection data to obtain the node abnormality coefficient; rotate the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient to generate a grid protection strategy to execute the normal operation of the microgrid.

2. The microgrid island detection method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Acquire microgrid topology data; Step S12: Identify key nodes of the microgrid topology data to obtain key node data of the microgrid topology; Step S13: deploying distributed computing units on the microgrid based on the key node data of the microgrid topology to generate distributed node deployment data; Step S14: using sensors to collect microgrid operation data according to distributed node deployment data to generate initial microgrid operation data; Step S15: construct a multi-dimensional perturbation signal for the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix.

3. The microgrid island detection method according to claim 2, characterized in that: Step S14 includes the following steps: Step S151: extracting multi-dimensional grid feature data from the initial microgrid operation data to obtain multi-dimensional feature data of the microgrid, wherein the multi-dimensional feature data of the microgrid includes microgrid voltage data, microgrid reactive power data, and microgrid harmonic data; Step S152: injecting a perturbation signal through the microgrid voltage data, the microgrid reactive power data and the microgrid harmonic data to generate an initial microgrid perturbation signal; performing a signal spectrum characteristic analysis on the initial microgrid perturbation signal to generate perturbation signal spectrum characteristic data; Step S153: performing multi-frequency domain superposition on the initial perturbation signal of the microgrid according to the perturbation signal spectrum characteristic data to generate a multi-frequency domain disturbance signal set; performing nonlinear signal superposition on the multi-frequency domain disturbance signal set to generate a nonlinear disturbance signal matrix; Step S154: performing perturbation signal spectrum equalization on the nonlinear perturbation signal matrix to generate a multi-dimensional perturbation signal matrix.

4. The microgrid island detection method according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: injecting a micro-reactive disturbance into the initial microgrid operation data using a multi-dimensional disturbance signal matrix to generate multi-dimensional disturbance grid change data; Step S22: extracting frequency components from the multi-dimensional disturbance power grid change data to obtain a power grid change multi-frequency disturbance component set; calculating independent impedance components from the power grid change multi-frequency disturbance component set to generate power grid independent impedance component data; Step S23: performing nonlinear error correction on the independent impedance component data of the power grid through nonlinear effects to generate independent corrected impedance component data of the power grid; Step S24: performing multi-directional impedance fusion on the grid independent correction impedance component data to generate grid fusion impedance data; performing dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data.

5. The microgrid island detection method according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: performing grid node island detection and discrimination on the grid node operation network topology model. When a grid node island is detected in the grid node operation network topology model, the corresponding grid node is marked as an abnormal grid node. Step S42: Performing grid stability analysis on abnormal grid nodes through grid comprehensive island detection data to generate grid node stability analysis data; performing node load performance analysis on abnormal grid nodes to generate node reliability analysis data; Step S43: Calculate the node abnormality degree of the abnormal power grid node based on the power grid node stability maintenance calculation formula through the power grid node stability analysis data and the node reliability analysis data to obtain the node abnormality coefficient; Step S44: rotating the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient, generating a grid protection strategy to execute normal operation of the microgrid.

6. The microgrid island detection method according to claim 5, characterized in that: The grid node stability calculation formula in step S43 is as follows: In the formula, Represented as a grid node The stability coefficient, It represents the time interval for power grid stability analysis. Represented as a node In time The voltage value, Represented as a node In time The current value, Represented as a node In time The active power, Represented as a node In time The reactive power, Expressed as voltage influence weight coefficient, Expressed as the power impact weight coefficient, Expressed as the impedance influence weight coefficient, Represented as a node The neighboring set of other connected nodes, Represented as a node With Node The impedance value between.

7. The microgrid island detection method according to claim 5, characterized in that: Step S44 includes the following steps: Step S441: classifying abnormal power grid nodes into power grid risk levels according to node abnormality coefficients, and generating power grid risk level classification data, wherein the power grid risk level classification data includes low-risk power grid nodes, medium-risk power grid nodes, and high-risk power grid nodes; Step S442: Continuously monitor abnormal changes of low-risk power grid nodes to generate protection strategies for low-risk power grid nodes; perform off-grid transfer of medium-risk power grid nodes to generate protection strategies for medium-risk power grid nodes; perform node shutdown processing on high-risk power grid nodes to generate protection strategies for high-risk power grid nodes; Step S443: Regulate the grid protection strategy of abnormal grid nodes according to the low-risk grid node protection strategy, the medium-risk grid node protection strategy, and the high-risk grid node protection strategy, and generate a grid protection strategy to execute normal operation of the microgrid.

8. A microgrid island detection system, characterized in that: Used to execute the microgrid island detection method according to claim 1, the microgrid island detection system comprises: The perturbation signal generation module is used to obtain microgrid topology data; deploy distributed computing units on the microgrid topology data to generate distributed node deployment data; use sensors to collect microgrid operation data based on the distributed node deployment data to generate initial microgrid operation data; construct multi-dimensional perturbation signals on the initial microgrid operation data to generate a multi-dimensional perturbation signal matrix; The impedance analysis module is used to inject a small amount of reactive disturbance into the initial microgrid operation data using a multi-dimensional disturbance signal matrix to generate multi-dimensional disturbance grid change data; to perform multi-directional impedance fusion on the multi-dimensional disturbance grid change data to generate grid fusion impedance data; to perform dynamic time series change trend analysis on the grid fusion impedance data to generate grid time series impedance analysis data; The island detection module is used to build an island detection framework; the island detection framework is compatible with distributed energy detection through the initial microgrid operation data to generate an adaptive island detection framework; the adaptive island detection framework is used to perform multi-mode collaborative island detection on the grid time series impedance analysis data to generate comprehensive grid island detection data; the distributed node deployment data is modeled for grid operation network topology based on the comprehensive grid island detection data to generate a grid node operation network topology model; The grid protection module is used to perform grid node island detection and identification on the grid node operation network topology model to generate abnormal grid nodes; calculate the node abnormality degree of abnormal grid nodes through the grid comprehensive island detection data to obtain the node abnormality coefficient; rotate the grid protection strategy for the abnormal grid nodes according to the node abnormality coefficient, and generate a grid protection strategy to execute the normal operation of the microgrid.

Images