A microgrid dynamic coordination control method and system for distributed energy

Abstract

The present invention discloses a method and system for dynamic coordinated control of microgrids for distributed energy, and relates to the field of microgrid energy management technology. Step one, real-time collection of multi-source microgrid data, edge computing and data preprocessing operations, and early judgment of island mode; Step two, aggregation of the structure and status of each microgrid, dynamic construction and update of the microgrid topology, analysis of the microgrid topology structural changes and health status, and assessment of the comprehensive risk of microgrid nodes; Step three, calculation of the actual distributable power of each type of load, and autonomous control and elastic mode switching; Step four, based on the actual distributable power of each type of load, real-time assessment of the microgrid collaborative energy dispatching capability, role identification and implementation of optimization strategies. This solves the problem that as the scale of access to distributed energy and microgrids continues to expand, some extreme events cause sudden failures in the main network, resulting in delayed island emergency dispatch response.

Description

Technical Field

[0001] The present invention relates to the technical field of microgrid energy management, and in particular to a microgrid dynamic coordination control method and system for distributed energy. Background Art

[0002] With the widespread adoption of renewable energy sources such as photovoltaics and wind power, as well as energy storage technologies, traditional centralized power systems are gradually transforming into distributed, intelligent new power systems. Microgrids, as key carriers of distributed energy, are being widely deployed in industrial parks, buildings, factories, and urban and rural power grids. They enable autonomous production, consumption, regulation, and optimization of local energy, improving energy efficiency and security.

[0003] For example, the invention patent with announcement number CN119205336A discloses a distributed energy multi-microgrid optimization control method for electric energy trading, constructs a distributed energy multi-microgrid optimization control constraint model for electric energy trading; constructs a distributed energy multi-microgrid optimization control two-stage model for electric energy trading; uses the ADMM algorithm to solve the distributed energy multi-microgrid optimization control constraint model for electric energy trading and the distributed energy multi-microgrid optimization control two-stage model for electric energy trading, and proposes a multi-microgrid cooperation cost minimization and benefit distribution plan. This invention solves the existing problems of distributed energy consumption and Nash cooperation benefit distribution, maximizes the overall benefits of the alliance, and allows each microgrid to reduce operating costs caused by uncertainty through cooperation.

[0004] For example, the invention patent with publication number CN118469154B discloses an intelligent microgrid evaluation system and method based on distributed energy consumption, which relates to the field of smart grids for power systems, and in particular to an intelligent microgrid evaluation system and method based on distributed energy consumption. The system includes collecting and preprocessing the energy production, consumption, and storage of each node in the microgrid, as well as key data information during the operation of the microgrid, through intelligent sensing devices and sensors; classifying the microgrid according to its scale, defining and calculating a first energy consumption efficiency; obtaining an adjustment parameter for the energy consumption efficiency based on the load demand and energy loss; and making appropriate adjustments to the energy cost. The energy consumption efficiency adjustment parameter is compared with a preset energy consumption efficiency adjustment parameter to obtain the adjusted energy consumption efficiency; and then, a microgrid evaluation model is established based on the relevant data information. The present invention realizes real-time evaluation and optimized regulation of the microgrid, thereby improving the energy consumption efficiency and reliability of the microgrid.

[0005] However, in the process of implementing the technical solutions of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0006] As the number of microgrids increases, their types become more diverse, and their load structures become increasingly complex, the random fluctuations of distributed power sources and loads make balancing energy supply and demand more difficult. Events such as localized equipment failures and extreme weather conditions can easily lead to isolated operations and energy shortages. Furthermore, the coexistence of multiple microgrids in campuses and regions places higher demands on energy collaboration, flexible load response, emergency coordination, and intelligent scheduling.

[0007] Therefore, in order to solve the above problems, there is an urgent need for a microgrid dynamic coordination control method and system for distributed energy. Summary of the Invention

[0008] Technical problems solved

[0009] In response to the shortcomings of the existing technology, the present invention provides a dynamic coordinated control method and system for distributed energy microgrids, which solves the problem of delayed emergency dispatch response of isolated islands due to sudden failures of the main grid caused by some extreme events as the access scale of distributed energy and microgrids continues to expand.

[0010] Technical Solution

[0011] To achieve the above objectives, the present invention is implemented through the following technical solutions: a dynamic coordinated control method for distributed energy microgrids, including: step one, real-time collection of multi-source microgrid data, edge computing and data preprocessing operations, real-time monitoring of anomalies, and early judgment of island mode; step two, based on the multi-source microgrid data, aggregating the structure and status of each microgrid, dynamically constructing and updating the microgrid topology map, and analyzing the structural changes and health status of the microgrid topology map in real time, and evaluating the comprehensive risk of the microgrid nodes based on the early judgment results of the island mode and the microgrid topology map; step three, calculating the actual distributable power of each type of load based on the comprehensive risk of the microgrid nodes and the health status of the microgrid, and performing autonomous control and elastic mode switching based on the actual distributable power of each type of load; step four, based on the microgrid topology map, on the basis of the actual distributable power of each type of load, real-time evaluation of the microgrid collaborative energy scheduling capability, and performing role identification and implementation of optimization strategies based on the evaluation results of the collaborative energy scheduling capability of each microgrid.

[0012] Furthermore, multi-source microgrid data is collected in real time, and edge computing and data preprocessing operations are performed. The specific process of real-time monitoring of anomalies is as follows: Real-time collection of multi-source microgrid data: Multi-source microgrid data is collected in real time through IoT terminal devices and smart sensors, as well as local energy storage monitoring systems and distributed power generation monitoring systems. Multi-source microgrid data includes: voltage, frequency, phase, energy storage SOC, energy storage capacity, distributed power supply output, load data, and active power and reactive power; the sliding average method is used to remove invalid, abnormal, and noisy multi-source microgrid data in the collection process through local filtering, and the multi-source microgrid data is compressed. Through sliding window detection and logical rule screening methods, abnormal outliers caused by equipment failures and signal interference are automatically identified and eliminated, and the multi-source microgrid data is standardized and normalized; at the same time, based on the long short-term memory network method, the multi-source microgrid data of each node is dynamically monitored, and electrical anomalies and communication anomalies are identified in real time; and the multi-source microgrid data is stored in a multi-source microgrid database.

[0013] Furthermore, the specific process of early judgment of the island mode is as follows: the voltage of the last ten minutes is obtained through the multi-source microgrid database and the average value is calculated as the historical steady-state average value, the current real-time voltage collected is subtracted from the historical steady-state average value and the average value is taken to obtain the voltage deviation; at the same time, the frequency of the last ten minutes is obtained through the multi-source microgrid database and the average value is calculated as the frequency reference, the current real-time frequency collected is subtracted from the frequency reference and the average value is taken to obtain the frequency deviation; two voltages and two frequencies are continuously obtained based on the sliding time window, the difference between the two voltages is divided by the sliding time window length and the absolute value is taken to obtain the voltage change rate, and the difference between the two frequencies is divided by the sliding time window length and the absolute value is taken to obtain the frequency change rate; the voltage deviation is multiplied by the frequency deviation, and then divided by the sum of the voltage change rate, the frequency change rate and the constant 1 to obtain the voltage-frequency joint fluctuation intensity; the voltage deviation is subtracted from the product of the frequency deviation and the response weight factor to obtain the voltage-frequency response deviation value; the voltage-frequency joint fluctuation intensity and the voltage-frequency response deviation value are added to obtain the island judgment value; the island judgment value is compared in real time When the islanding threshold is less than the first-level threshold, the system is deemed to be in a normal state, maintaining normal grid-connected operation. Multi-source microgrid data is regularly archived and monitored for ongoing monitoring. When the islanding threshold is greater than or equal to the first-level threshold and less than the second-level threshold, the system is deemed to be in a mild islanding state, initiating a local rapid self-check process at the edge to verify the health of the voltage and frequency acquisition channels and communication links. Repeating this process multiple times within a specified timeframe, if the system remains in the mild islanding state, the building's public lighting load is automatically reduced, and the energy storage system enters standby mode. Early warning information is automatically sent to maintenance personnel. When the islanding threshold is greater than or equal to the second-level threshold, the system is deemed to be in a high-risk islanding state, automatically switching to islanding mode, disconnecting from the main grid and activating local distributed power sources and energy storage. Critical loads are prioritized, while non-critical loads are automatically offlined. The energy storage system dynamically discharges according to pre-set priorities. Emergency alarms are triggered locally and on the cloud platform, recording multi-source microgrid data and islanding thresholds throughout the entire islanding event process. Regional microgrids are then coordinated for mutual assistance and archiving of the islanding event.

[0014] Furthermore, based on the multi-source microgrid data, the structure and status of each microgrid are aggregated, the microgrid topology is dynamically constructed and updated, and the specific process of real-time analysis of the changes in the microgrid topology structure and health status is as follows: through the edge gateway, smart terminal and main station communication interface, the topology structure, node online status and the opening and closing status of circuit breakers and interconnection line key equipment of all microgrids are regularly collected, and the microgrid operation data is obtained in real time. The microgrid operation data includes voltage, frequency, load data, energy storage SOC and energy storage capacity; all topology structures, node online status, opening and closing status data of key equipment and microgrid operation data are encoded in a standardized format and normalized, and uploaded to the scheduling and analysis platform periodically through a secure data channel; each microgrid unit is abstracted as a node in the microgrid topology. If two microgrid units are directly connected through a Nodes whose interconnecting lines are connected are considered neighbors. The interconnecting lines between nodes are abstracted as edges in the microgrid topology map. The topological structure, node online status, on / off status data of key equipment, and microgrid operation data collected in real time are used to dynamically construct and update the microgrid topology map. Based on the latest microgrid topology map and the microgrid operation data of each node, the node features and edge features in the microgrid topology map are extracted, the topological relationship is converted into an adjacency matrix, and the node features and edge features are combined into a feature matrix as the input of the graph convolutional neural network algorithm. The graph convolutional neural network algorithm automatically integrates the characteristics of each node itself and the characteristics of neighboring nodes, and considers the topological relationship between nodes. It performs deep learning analysis on the physical and state relationships between nodes, and automatically extracts the health status of nodes in the microgrid topology map, including abnormal connections, weakly connected areas, and node health degradation trend information.

[0015] Furthermore, the specific process of evaluating the comprehensive risk of microgrid nodes based on the early judgment results of the island mode and the microgrid topology map is as follows: by extracting the health status of the nodes in the microgrid topology map, automatically summarizing and normalizing them to obtain the health of node i and the health of neighboring node j; obtaining real-time load data, and filtering out the maximum value as the maximum design capacity based on the historical load data obtained in the last twenty days from the multi-source microgrid database, and dividing the current real-time load data by the maximum design capacity to obtain the load ratio of node i; obtaining the neighbor node set of node i based on the microgrid topology map, and obtaining the length of the neighbor node set The number of neighbor nodes is obtained by calculating the health degree; the opening and closing status of the key equipment of the circuit breaker and the tie line between nodes i and j is obtained to obtain the connectivity strength of nodes i and j; the island determination value is calculated to obtain the island signal strength of node i; the health degree of node i is subtracted from the health degree of neighbor node j and the absolute value is taken, and then divided by the sum of the connectivity strength of nodes i and j and the constant 1 to obtain the health degree between neighbors; based on the set of neighbor nodes, the health degrees of all neighbors are accumulated and then divided by the number of neighbor nodes to obtain the average health degree between neighbors; the average health degree between neighbors is multiplied by the health weight factor and then multiplied by the health degree weight factor of node i and the load ratio of node i. , the island signal strength of node i is added to obtain the comprehensive risk assessment value of node i; the comprehensive risk assessment value is compared with the first and second level risk thresholds in real time. When the comprehensive risk assessment value is less than the first level risk threshold, the microgrid node is considered to be risk-free, and the node maintains normal monitoring and daily operation. The comprehensive risk assessment value and load change trend of each node are automatically recorded and archived regularly; when the comprehensive risk assessment value is greater than or equal to the first level risk threshold and less than the second level risk threshold, the microgrid node is considered to have a moderate risk, and a risk attention reminder is automatically pushed to the operation and maintenance personnel. The node is highlighted in yellow on the platform interface; local equipment self-check , check the circuit breakers, energy storage status, and key points of the communication link; reduce non-critical interruptible loads and schedule energy storage charging in advance; when the comprehensive risk assessment value is greater than or equal to the secondary risk threshold, the microgrid node is considered to be at high risk, and a red alert is automatically triggered, with a pop-up window and SMS notification to the relevant responsible persons; only the core load is retained, and all non-critical loads are forced offline; energy storage and distributed power sources enter high-priority power supply mode; automatically prepare and switch to island operation to quickly isolate high-risk nodes; automatically trigger the energy mutual assistance mechanism with adjacent microgrids; self-start special emergency inspections and fault location; and archive all risk events.

[0016] Furthermore, the specific process of calculating the actual distributable power for each type of load based on the comprehensive risk of the microgrid node and the health status of the microgrid is as follows: based on the historical active power in the multi-source microgrid database, the maximum active power in the past twenty days is selected to obtain the current maximum allowable power of the i-th load; the real-time active power and the current maximum allowable power are obtained and automatically summed to obtain the local total available power supply; the instantaneous active power of all devices under the i-th load is collected in real time and summed to obtain the real-time required power of the i-th load; all online load categories in the current park that need to participate in the allocation are automatically traversed and the total number is calculated to obtain the number of all load categories; based on the current number of all load categories, the real-time required power of all load categories is accumulated to obtain the total required power of all load categories; the real-time required power of the i-th load is divided by the total required power of all load categories, and then multiplied by the local total available power supply to obtain the theoretical allocated power value of the i-th load; the theoretical allocated power value of the i-th load is compared with the current maximum allowable power of the i-th load, and the smaller of the two is taken as the actual allocated power value of the i-th load.

[0017] Furthermore, the specific process of autonomous control and flexible mode switching based on the actual distributable power of various loads is as follows: based on the actual distributed power value of various loads, the core load is prioritized, and the remaining loads are operated in staggered manner: if the actual distributed power value is equal to the real-time demand power, it is considered that the load can maintain normal operation and is fully guaranteed; if the actual distributed power value is less than the real-time demand power, it is considered that the load needs to be operated in staggered manner and is partially guaranteed; priority is given to ensuring that all demands of the core load are met; power supply to important but adjustable loads is automatically reduced in batches and time periods; for interruptible loads, some charging equipment is automatically suspended, non-emergency production line power consumption is postponed, and interruptible loads are transferred to off-peak periods and restarted; load staggered operation instructions are issued to each terminal device, and detailed records are kept of each load reduction, staggered and guarantee process; local energy and load data are continuously monitored, and the grid-connected mode is smoothly switched back when the main network is restored.

[0018] Furthermore, based on the microgrid topology, on the basis of the actual distributable power of each type of load, the specific process of real-time evaluation of the microgrid collaborative energy scheduling capability is as follows: obtain multi-source microgrid data, multiply the energy storage SOC with the energy storage capacity and add the distributed power output to obtain the local total available energy of microgrid i and the local total available energy of neighboring microgrid j; obtain the active power of all types of loads in real time, and automatically accumulate the current total energy demand of microgrid i and the current total energy demand of neighboring microgrid j based on the length of the scheduling time window; obtain the set of neighboring microgrids of microgrid i according to the microgrid topology, and use breadth-first search to automatically calculate the shortest path from microgrid node i to microgrid node j. The energy transmission loss percentage between the two microgrids is calculated based on the multi-source microgrid data and normalized to obtain the collaborative dispatch distance between microgrids i and j. The local total available energy of neighbor microgrid j is subtracted from the current total demand energy of neighbor microgrid j, and then divided by the sum of the collaborative dispatch distance between microgrids i and j and a constant 1 to obtain the energy surplus and shortage value. Based on the set of neighbor microgrids, the energy surplus and shortage values ​​of all microgrids are accumulated to obtain the total energy surplus and shortage value. The total energy surplus and shortage value is multiplied by the neighborhood support item weight factor to obtain the domain support item. The local total available energy of microgrid i is subtracted from the current total demand energy of microgrid i, and then added to the domain support item to obtain the collaborative energy dispatch value of microgrid i.

[0019] Furthermore, the specific process of role identification and implementation of optimization strategy based on the collaborative energy dispatch capability evaluation results of each microgrid is as follows: automatic identification is performed based on the collaborative energy dispatch value of each microgrid. If the collaborative energy dispatch value of the microgrid is greater than 0, the microgrid is determined to be an energy supplier and is given priority as an energy output and mutual support node; if the collaborative energy dispatch value of the microgrid is less than 0, the microgrid is determined to be an energy demander and is given priority to obtain regional collaborative energy support; all microgrids are sorted in descending order of priority according to the collaborative energy dispatch value, with large positive values ​​having high priority for energy supply and large negative values ​​having high priority for gaps; the collaborative energy dispatch value and the collaborative dispatch distance are used to automatically match the supply and demand parties with the nearest energy path with the least loss; an energy flow table is generated, and instructions to release excess energy are automatically issued to the supplier, and the demand party receives instructions to adjust energy; if resources are limited, the gap is sorted according to the gap Priority is adjusted in sequence until all resource allocations and demands are met; elastic load and interruptible load nodes automatically adjust local load distribution according to the collaborative energy scheduling value to perform peak-shifting operation and load reduction; at the same time, the supply-side microgrid automatically adjusts the output of local energy storage and distributed power sources to release available energy; the demand-side microgrid updates its own load configuration in real time, gives priority to protecting the core load, and automatically enters a partial load reduction state; each round of energy adjustment decisions, execution details, energy flow and load adjustment based on the collaborative energy scheduling value are automatically recorded, and the collaborative energy scheduling value, microgrid role, actual adjustment amount, response time and energy loss of each microgrid before and after the adjustment are archived; for abnormal scheduling events where the adjustment is not fully met, the execution fails, and the key load is affected, the abnormal causes and response results are automatically marked and summarized; the knowledge base is continuously updated using all archived scheduling events.

[0020] The second aspect of the present invention provides a microgrid dynamic coordination control system for distributed energy, including: a multi-source perception and intelligent analysis module, a regional topology and state fusion module, an autonomous control and elastic mode switching module and a regional collaborative scheduling and self-learning optimization module; wherein the multi-source perception and intelligent analysis module is used to collect multi-source microgrid data in real time, perform edge computing and data preprocessing operations, monitor anomalies in real time, and make early judgments on island modes; the regional topology and state fusion module is used to aggregate the structure and state of each microgrid based on multi-source microgrid data, dynamically construct and update the microgrid topology map, and analyze the changes in the microgrid topology map structure in real time. The module is used to evaluate the comprehensive risk of microgrid nodes based on the early judgment results of the island mode and the microgrid topology map; the autonomous control and elastic mode switching module is used to calculate the actual distributable power of each type of load according to the comprehensive risk of the microgrid nodes and the health status of the microgrid, and perform autonomous control and elastic mode switching according to the actual distributable power of each type of load; the regional collaborative scheduling and self-learning optimization module is used to evaluate the collaborative energy scheduling capability of the microgrid in real time based on the actual distributable power of each type of load based on the microgrid topology map, and perform role identification and implement optimization strategies according to the collaborative energy scheduling capability evaluation results of each microgrid.

[0021] Beneficial effects

[0022] The present invention has the following beneficial effects:

[0023] (1) The present invention collects multi-source microgrid data in real time, combines edge computing with data preprocessing operations, efficiently filters and compresses various data anomalies and noises, and dynamically monitors multi-source microgrid data of each node based on the long short-term memory network method, thereby achieving real-time monitoring of electrical anomalies and communication anomalies, and being able to judge and respond to islanding modes in an early stage, thereby effectively improving the active security defense and early warning capabilities of the microgrid system.

[0024] (2) The present invention aggregates the structure and status of each microgrid through multi-source microgrid data, dynamically constructs and updates the microgrid topology map, automatically analyzes the structural changes and health status of the microgrid topology map through the graph convolutional neural network algorithm, and combines the early judgment results of the island mode to conduct a multi-dimensional and dynamic assessment of the comprehensive risk of the microgrid nodes, thereby achieving accurate identification and risk judgment of weakly connected areas, abnormal connections and node health degradation trends, and improving the overall resilience and risk prevention and control capabilities of the system.

[0025] (3) The present invention dynamically calculates the actual distributable power of each type of load, implements autonomous control and flexible mode switching for the actual distributable power of each type of load, gives priority to core loads, automatically performs load peak shifting, batch reduction and interruptible load management, realizes refined energy allocation and flexible optimization, and improves the energy reliability and energy efficiency of the system in multiple scenarios.

[0026] (4) The present invention, through real-time evaluation of the microgrid collaborative energy dispatching capability, realizes efficient matching of energy suppliers and demanders, optimal allocation of energy flows and archiving of abnormal adjustment events through role identification and optimization strategy implementation, supports dynamic mutual assistance and energy regulation between regional microgrids, and can continuously update the knowledge base and self-learning optimization dispatching strategy, which greatly enhances the intelligence, coordination and self-adaptation capabilities of the regional energy system.

[0027] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A flow chart of a dynamic coordinated control method for distributed energy microgrids;

[0029] Figure 2 This is a structural diagram of a microgrid dynamic coordination control system for distributed energy;

[0030] Figure 3 This is the change diagram of the collaborative energy scheduling value of the microgrid nodes. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. As those skilled in the art will understand, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0032] See also Figure 1-Figure 3, an embodiment of the present invention provides a technical solution: a microgrid dynamic coordinated control method and system for distributed energy, comprising: step one, real-time collection of multi-source microgrid data, edge computing and data preprocessing operations, real-time monitoring of anomalies, and early judgment of island mode; step two, based on the multi-source microgrid data, aggregating the structure and status of each microgrid, dynamically constructing and updating the microgrid topology map, and analyzing the structural changes and health status of the microgrid topology map in real time, and evaluating the comprehensive risk of the microgrid node based on the early judgment result of the island mode and the microgrid topology map; step three, calculating the actual distributable power of each type of load according to the comprehensive risk of the microgrid node and the health status of the microgrid, and performing autonomous control and elastic mode switching according to the actual distributable power of each type of load; step four, based on the microgrid topology map, on the basis of the actual distributable power of each type of load, real-time evaluation of the microgrid collaborative energy dispatching capability, and performing role identification and implementing optimization strategy according to the evaluation result of the collaborative energy dispatching capability of each microgrid.

[0033] Specifically, multi-source microgrid data is collected in real time, edge computing and data preprocessing operations are performed, and the specific process of real-time monitoring of anomalies is as follows: Real-time collection of multi-source microgrid data: Multi-source microgrid data is collected in real time through IoT terminal devices and smart sensors as well as local energy storage monitoring systems and distributed power generation monitoring systems. Multi-source microgrid data includes: voltage, frequency, phase, energy storage SOC, energy storage capacity, distributed power supply output, load data, and active power and reactive power; the comprehensive collection of multiple types of data can provide rich real-time basis for the operating status of the microgrid, and effectively support the subsequent application of various intelligent algorithms and the improvement of system risk prevention and control capabilities. The sliding average method is used to remove invalid, abnormal and noisy multi-source microgrid data during the collection process through local filtering, and the multi-source microgrid data is compressed. Through sliding window detection and logical rule screening methods, abnormal outliers caused by equipment failures and signal interference are automatically identified and eliminated, and the multi-source microgrid data is standardized and normalized to improve data quality and reduce the error rate, providing a solid foundation for microgrid anomaly detection and trend analysis; at the same time, based on the long short-term memory network method, the multi-source microgrid data of each node is dynamically monitored, and electrical anomalies and communication anomalies are identified in real time, which effectively enhances the perception sensitivity and early warning capabilities of abnormal changes; and the multi-source microgrid data is stored in the multi-source microgrid database. By establishing a complete data storage and traceability system, data guarantee can be provided for subsequent fault tracing, operation and maintenance decision-making and system self-learning optimization.

[0034] In this implementation plan, by collecting multi-source microgrid data in real time, performing edge computing and data preprocessing operations, and monitoring anomalies in real time, it is possible to improve data quality, enhance anomaly identification capabilities, and provide efficient and accurate data support for the multi-source microgrid database, thereby improving the system's intelligence level and operational safety.

[0035] Specifically, the specific process of early judgment of the islanding mode is as follows: the voltage of the last ten minutes is obtained through the multi-source microgrid database and the average value is calculated as the historical steady-state average value, the voltage deviation is obtained by subtracting the currently collected real-time voltage from the historical steady-state average value and taking the average value; at the same time, the frequency of the last ten minutes is obtained through the multi-source microgrid database and the average value is calculated as the frequency reference, the frequency deviation is obtained by subtracting the currently collected real-time frequency from the frequency reference and taking the average value; it is used to sensitively reflect the degree of deviation of the current electrical state from the normal level, providing a reliable basis for subsequent judgment of islanding risks. Based on the sliding time window, two voltages and two frequencies are continuously obtained. The voltage change rate is obtained by dividing the difference between the two voltages by the sliding time window length and taking the absolute value. The frequency change rate is obtained by dividing the difference between the two frequencies by the sliding time window length and taking the absolute value. The voltage deviation is multiplied by the frequency deviation, and then divided by the sum of the voltage change rate, the frequency change rate and the constant 1 to obtain the voltage-frequency joint fluctuation intensity. The voltage-frequency response deviation value is obtained by subtracting the product of the frequency deviation and the response weight factor from the voltage deviation. The voltage-frequency joint fluctuation intensity is added to the voltage-frequency response deviation value to obtain the islanding judgment value. The islanding judgment value is compared with the first and second level judgment thresholds in real time. The hierarchical judgment can minimize false alarms and maintain system stability. When the islanding judgment value is less than the first level judgment threshold, it is judged to be normal and the normal grid-connected operation state is maintained. The multi-source microgrid data monitored is archived regularly and continuously observed. When When the islanding judgment value is greater than or equal to the first-level judgment threshold and less than the second-level judgment threshold, it is judged as a mild islanding state, and the edge local rapid self-inspection process is started to verify the health status of the voltage, frequency acquisition channels and communication links; the judgment is repeated multiple times within a limited time. If it continues to be in the mild islanding state range, the building public lighting load is automatically reduced, and the energy storage system enters standby mode; and early warning information is automatically pushed to the operation and maintenance personnel; when the islanding judgment value is greater than or equal to the second-level judgment threshold, it is judged as a high-risk islanding state, and it automatically switches to the islanding operation mode, disconnects from the main grid, and enables local distributed power supply and energy storage; critical loads are prioritized, and non-critical loads are automatically offline; the energy storage system dynamically discharges according to the preset priority; triggers emergency alarms on the local and cloud platforms, records multi-source microgrid data and islanding judgment values ​​of the entire process of the islanding event, and links regional microgrids for mutual assistance to archive the islanding event.

[0036] The specific formula for the island determination value is:

[0037] ;

[0038] Where, Represents the islanding judgment value, which is used to determine whether the microgrid is in an islanding state in the early stage; Indicates voltage deviation, reflecting the fluctuation amplitude of the current voltage relative to the normal state, and is a sensitive indicator of abnormal changes; Indicates frequency deviation, which is used to reflect the degree of system frequency fluctuation and is also an important basis for abnormality identification; Indicates the voltage change rate, which indicates the instantaneous change speed of voltage over time. The larger the value, the more severe the voltage fluctuation. It is used to correct the sensitivity of abnormality judgment. Indicates the frequency change rate, which indicates the speed at which the frequency changes over time, reflects the dynamic volatility of the system, and is used to correct and suppress false alarms caused by normal fluctuations; Represents the response weight factor, which is obtained by least squares straight-line fitting of multiple sets of voltage deviation and frequency deviation data sets. It is used to measure the linear empirical response relationship between voltage and frequency deviation, and its value range is between 0.5 and 2.0.

[0039] In this implementation plan, voltage and frequency deviations and their change rates are obtained in real time through a multi-source microgrid database, the voltage-frequency joint fluctuation intensity and the voltage-frequency response deviation value are calculated, and the islanding judgment value is obtained. The first-level and second-level judgment thresholds are compared in real time, and a graded response is implemented. Monitoring data is archived regularly, early warning information is automatically pushed, the islanding operation mode is switched, emergency alarms and regional microgrid mutual assistance are triggered, and the entire process of islanding events is archived.

[0040] Specifically, based on multi-source microgrid data, the structure and status of each microgrid are aggregated, the microgrid topology is dynamically constructed and updated, and the specific process of real-time analysis of the changes in the microgrid topology structure and health status is as follows: through the edge gateway, smart terminal and main station communication interface, the topology structure, node online status and the opening and closing status of circuit breakers and interconnection line key equipment of all microgrids are regularly collected, and the microgrid operation data is obtained in real time. The microgrid operation data includes voltage, frequency, load data, energy storage SOC and energy storage capacity; by comprehensively collecting structure and operation information, a dynamic panoramic perception of the entire microgrid system status can be achieved; all topology structures, node online status, opening and closing status data of key equipment and microgrid operation data are encoded in a standardized format and normalized, and periodically uploaded to the scheduling and analysis platform through a secure data channel, which ensures the security and efficiency of data transmission, and also provides a basis for data consistency and comparability for subsequent intelligent analysis; each microgrid unit is abstracted as a node in the microgrid topology. If two microgrid units are directly connected by a tie line, they are considered neighbor nodes. The tie lines between nodes are abstracted as edges in the microgrid topology graph. The real-time collected topology structure, node online status, key equipment on / off status data, and microgrid operation data are used to dynamically construct and update the microgrid topology graph to dynamically reflect changes in the network structure. Based on the latest microgrid topology graph and the microgrid operation data of each node, the node features and edge features in the microgrid topology graph are extracted, the topological relationship is converted into an adjacency matrix, and the node features and edge features are combined into a feature matrix as the input of the graph convolutional neural network algorithm. The graph convolutional neural network algorithm automatically integrates the characteristics of each node itself and the characteristics of neighboring nodes, and considers the topological relationship between nodes. It performs deep learning analysis on the physical and state relationships between nodes, and automatically extracts the health status of nodes in the microgrid topology graph, including abnormal connections, weakly connected areas, and node health degradation trend information, realizing automatic identification and trend analysis of the health status and operation risks of the microgrid.

[0041] In this implementation plan, the structure and status of each microgrid are aggregated based on multi-source microgrid data, the microgrid topology map is dynamically constructed and updated, and the structural changes and health status of the microgrid topology map are analyzed in real time, realizing dynamic panoramic perception and risk trend analysis of the microgrid system status, and improving the system's intelligence level and proactive operation and maintenance capabilities.

[0042] Specifically, the specific process of evaluating the comprehensive risk of microgrid nodes based on the early judgment results of the island mode and the microgrid topology diagram is as follows: by extracting the health status of the nodes in the microgrid topology diagram, automatically summarizing and normalizing them to obtain the health status of node i and the health status of neighboring node j, which is conducive to the dynamic comparison and standardized evaluation of the health status between nodes; obtaining real-time load data, and filtering out the maximum value as the maximum design capacity based on the historical load data obtained in the last twenty days from the multi-source microgrid database, and dividing the current real-time load data by the maximum design capacity to obtain the load ratio of node i, which effectively reflects the current load pressure of the node and provides a quantitative basis for risk assessment; according to the microgrid The network topology diagram is used to obtain the neighbor node set of node i, and the length of the neighbor node set is obtained to obtain the number of neighbor nodes; the opening and closing status of the circuit breaker and the key equipment of the tie line between nodes i and j are obtained to obtain the connectivity strength of nodes i and j, which further refines the physical connection and interoperability between the node and the neighbor; the island judgment value is calculated to obtain the island signal strength of node i, which timely reflects the risk level of the node affected by the island; the health of node i is subtracted from the health of neighbor node j and the absolute value is taken, and then divided by the sum of the connectivity strength of nodes i and j and the constant 1 to obtain the health between neighbors; based on the neighbor node set, the health of all neighbors is accumulated and then divided by the number of neighbor nodes to obtain the health between neighbors. The average health value of the neighbors is multiplied by the health weight factor, and then added with the health of node i, the load ratio of node i, and the island signal strength of node i to obtain the comprehensive risk assessment value of node i; the comprehensive risk assessment value is compared with the first and second level risk thresholds in real time. When the comprehensive risk assessment value is less than the first level risk threshold, the microgrid node is considered to have no risk, and the node maintains normal monitoring and daily operation. The comprehensive risk assessment value and load change trend of each node are automatically recorded and archived regularly; when the comprehensive risk assessment value is greater than or equal to the first level risk threshold and less than the second level risk threshold, the microgrid node is considered to have moderate risk, and a risk attention reminder is automatically pushed to the operation and maintenance personnel, saving time. The point is highlighted in yellow on the platform interface; local equipment self-inspection is performed to check the circuit breaker, energy storage status, and key points of the communication link; non-critical interruptible loads are reduced, and energy storage charging is scheduled in advance; when the comprehensive risk assessment value is greater than or equal to the secondary risk threshold, the microgrid node is considered to be at high risk, and a red alert is automatically triggered, with a pop-up window and SMS notification to the relevant responsible persons; only the core load is retained, and all non-critical loads are forced offline; energy storage and distributed power sources enter high-priority power supply mode; automatically prepare and switch to island operation to quickly isolate high-risk nodes; automatically trigger the energy mutual assistance mechanism with adjacent microgrids; self-start special emergency inspections and fault location; and archive all risk events.

[0043] The specific formula for the comprehensive risk assessment value is:

[0044] ;

[0045] Where, Represents the comprehensive risk assessment value of node i, which comprehensively quantifies the comprehensive operation risk of each node in the entire microgrid network and is used for intelligent monitoring, island warning and dispatch assistance; Indicates the health of node i; Indicates the health of neighbor node j, reflecting the health status of the node itself. The larger the value, the higher the node health. The value decreases when it is closer to abnormality and degradation. Indicates the load ratio of node i, which is used to measure the ratio of the current load of node i to the maximum design capacity, reflecting the load pressure of the node. The higher the load, the greater the system pressure; represents the set of neighbor nodes of node i; Indicates the number of neighbor nodes; Indicates the connectivity strength between nodes i and j. A higher value indicates stronger connectivity. For example, the value is high when the circuit breaker is closed and the line is healthy, and vice versa. Indicates the island signal strength. If the island determination value is less than the determination threshold, the island signal strength is 0. If the island determination value is greater than or equal to the determination threshold, the island signal strength is 1. It represents the health weight factor. The comprehensive risk assessment value of each node under different events is compared with the actual occurrence of island events. The least squares method is used for fitting, and the health weight factor that minimizes the prediction error is selected as the optimal health weight factor. The value range is between 0.1 and 2.0.

[0046] In this implementation plan, the comprehensive risk assessment value is dynamically calculated through multi-dimensional indicators such as node health, load ratio, number of neighbor nodes, connectivity strength and island signal strength, and the first and second level risk thresholds are compared in real time. Risk attention reminders are automatically pushed, red alerts are triggered, and island operation and energy mutual assistance are switched, effectively improving the risk identification, graded response and security assurance capabilities of the microgrid system.

[0047] Specifically, the specific process of calculating the actual distributable power of each type of load based on the comprehensive risk of the microgrid node and the health status of the microgrid is as follows: based on the historical active power in the multi-source microgrid database, the maximum active power in the last twenty days is selected to obtain the current maximum allowable power of the i-th type of load, fully considering the historical extreme value changes on the load side to improve the safety margin of the allocation strategy; obtaining the real-time active power and the current maximum allowable power, automatically adding them up to obtain the local total available power supply, which can dynamically reflect the overall level of the current deployable energy resources; collecting the instantaneous active power of all devices under the i-th type of load in real time and summing them up to obtain the i-th type of load Real-time power demand; automatically traverse all load categories that are online and need to participate in the current park and count the total number to obtain the number of all load categories; based on the current number of all load categories, accumulate the real-time power demand of all load categories to obtain the total power demand of all load categories; divide the real-time power demand of the i-th load by the total power demand of all load categories, and then multiply it by the local total available power supply to obtain the theoretical distribution power value of the i-th load; compare the theoretical distribution power value of the i-th load with the current maximum allowable power of the i-th load, and take the smaller one as the actual distribution power value of the i-th load.

[0048] The specific formula for the actual allocated power value is:

[0049] ;

[0050] Where, It represents the actual distributed power value of the i-th type of load, and is used to fairly, efficiently and intelligently distribute all available power to each type of load according to various factors during island emergency operation; Indicates the current maximum allowable power of the i-th category load, representing the maximum power supply that this category of load can withstand under safe operation, and is used to limit the allocated value to not exceed the safety boundary of the equipment; Indicates the total available local power supply, representing the total amount of energy that can be allocated by the system; It represents the real-time power demand of the i-th type of load, reflecting the instantaneous power demand of this type of load; Indicates the number of all current load categories, indicating the total number of load types participating in the current scheduling and distribution, such as lighting, power, air conditioning, electric vehicle charging, etc. Represents the total power demand of all load categories, used to ensure proportional distribution; Represents the minimum function, that is, the smaller value between the current maximum allowed power and the available power obtained by allocating the power ratio according to the real-time demand is selected.

[0051] In this implementation plan, the actual distributable power for each type of load is calculated based on the comprehensive risk of the microgrid nodes and the health status of the microgrid. This can dynamically reflect the local total available power supply and the real-time demand of each load. Combined with the historical active power extremes and current load demand, it ensures that critical loads are prioritized, thereby improving energy utilization efficiency and power supply security.

[0052] Specifically, the specific process of autonomous control and flexible mode switching based on the actual distributable power of various loads is as follows: according to the actual distributed power value of various loads, the core load is prioritized, and the remaining loads are staggered: if the actual distributed power value is equal to the real-time demand power, it is considered that the load can maintain normal operation and is fully guaranteed; if the actual distributed power value is less than the real-time demand power, it is considered that the load needs to be staggered and partially guaranteed; this method ensures the continuity of key businesses and system reliability through dynamic allocation; gives priority to ensuring that all demands of core loads are met; automatically reduces power supply in batches and time periods for important but adjustable loads, which helps to smooth the load curve and alleviate the risk of local overload; automatically suspends some charging equipment for interruptible loads, postpones power consumption of non-emergency production lines, and transfers interruptible loads to off-peak periods before restarting; sends load staggered operation instructions to each terminal device, and keeps detailed records of each load reduction, staggered and guarantee process; continuously monitors local energy and load data, and smoothly switches back to grid-connected mode when the main network is restored.

[0053] In this implementation plan, autonomous control and flexible mode switching are carried out according to the actual distributable power of various loads, which can prioritize core loads, improve energy utilization efficiency, reduce peak pressure, smooth the load curve, optimize energy scheduling strategies, and ensure system flexibility and traceability of operation data.

[0054] Specifically, based on the microgrid topology, and on the basis of the actual distributable power of each type of load, the specific process of real-time evaluation of the microgrid collaborative energy scheduling capability is as follows: obtain multi-source microgrid data, multiply the energy storage SOC by the energy storage capacity and add the distributed power output to obtain the local total available energy of microgrid i and the local total available energy of neighboring microgrid j; it not only reflects the energy reserve and power generation capacity of each microgrid and its surrounding nodes, but also provides an accurate basis for subsequent collaborative scheduling. The active power of all types of loads is obtained in real time, and the current total energy demand of microgrid i and the current total energy demand of neighboring microgrid j are automatically accumulated based on the length of the scheduling time window, dynamically presenting the immediate energy pressure of each node; according to the microgrid topology, the set of neighboring microgrids of microgrid i is obtained, and the number of nodes passed by the shortest path from microgrid node i to microgrid node j is automatically calculated using breadth-first search; the percentage of energy transmission loss between the two microgrids is calculated based on the multi-source microgrid data and normalized to obtain the collaborative scheduling distance between microgrids i and j, effectively combining The physical structure of the network and the actual operation efficiency; the energy surplus and shortage value is obtained by subtracting the current total energy demand of the neighbor microgrid j from the local total available energy of the neighbor microgrid j, and then dividing it by the sum of the collaborative scheduling distance between microgrids i and j and the constant one. Based on the set of neighbor microgrids, the energy surplus and shortage values ​​of all microgrids are accumulated to obtain the total energy surplus and shortage value. The total energy surplus and shortage value is multiplied by the neighborhood support item weight factor to obtain the domain support item. The local total available energy of microgrid i is subtracted from the current total energy demand of microgrid i, and then added to the domain support item to obtain the collaborative energy scheduling value of microgrid i.

[0055] The specific formula of the collaborative energy scheduling value is:

[0056] ;

[0057] Where, It represents the collaborative energy dispatch value of microgrid i, which is used to measure the current collaborative dispatch support capability of the i-th microgrid and reflects the current overall energy surplus or shortage capability of microgrid i. It is an important basis for determining its role and priority in regional collaborative dispatch. If it is a positive value, it means that the microgrid node has energy surplus. If it is a negative value, it means that the microgrid node has an energy gap. If it is 0, it means that the node is currently in supply and demand balance; represents the total local available energy of microgrid i, which represents the total energy that can be dispatched and released at present; represents the local total available energy of neighboring microgrid j, reflecting the dispatchability of neighboring nodes; It represents the current total energy demand of microgrid i, summarizing the total energy consumed by all load categories in the current scheduling period, and is the local real-time demand for electricity; represents the current total energy demand of the neighboring microgrid j, reflecting the local load pressure of the neighboring node; represents the set of neighboring microgrids of microgrid i; represents the coordinated dispatch distance between microgrids i and j, indicating the transmission difficulty and loss degree when dispatching energy; It represents the weight factor of the neighborhood support item, which is used to adjust the influence of neighboring node energy surplus and shortage on the scheduling value of this node. It is obtained by fitting the historical collaborative energy scheduling values ​​of the regional microgrid using the least squares method, and its value range is between 0.1 and 2.0.

[0058] The neighborhood support weight factor is set to 0.5. Based on the differences in local available total energy and current total energy demand of each microgrid node, as well as the presence of different neighboring nodes, the collaborative energy dispatch values ​​of different microgrid nodes are calculated based on the collaborative dispatch distance between different microgrids. This is shown in Table 1.

[0059] Table 1 Collaborative energy scheduling value data table

[0060]

[0061] like Figure 3 As shown in Table 1, it is a diagram of the collaborative energy scheduling value of the microgrid node provided in the embodiment of the present application. Figure 3 It can be seen that the energy surplus and shortage distribution of each node, different microgrid nodes have different neighboring nodes, and according to the local available total energy and current total demand energy of each microgrid node, the obtained collaborative energy scheduling value has certain differences, where positive value nodes represent supportable energy, and negative value nodes represent urgent need for support.

[0062] In this implementation plan, based on the microgrid topology map and the actual distributable power of each type of load, the microgrid collaborative energy scheduling capability is evaluated in real time, which can reflect the energy reserve and power generation capacity, dynamically present the energy pressure, effectively combine the physical structure and operation efficiency, achieve a dynamic balance of energy surplus and shortage, promote multi-node collaboration, and enhance regional energy self-healing and efficient mutual assistance capabilities.

[0063] Specifically, the process for role identification and implementation of optimization strategies based on the collaborative energy dispatch capability assessment results of each microgrid is as follows: Each microgrid's collaborative energy dispatch value is automatically identified. If the microgrid's collaborative energy dispatch value is greater than 0, the microgrid is considered an energy supplier and is prioritized as an energy output and mutual support node. If the microgrid's collaborative energy dispatch value is less than 0, the microgrid is considered an energy demander and is prioritized for regional collaborative energy support. This process uses an intelligent algorithm to automatically identify and dynamically adapt energy suppliers and demanders. All microgrids are sorted in descending order of priority based on their collaborative energy dispatch value, with those with large positive values ​​giving higher priority to energy supplies and those with large negative absolute values ​​giving higher priority to energy shortages. This ensures efficient and orderly energy allocation. The collaborative energy dispatch value and the collaborative dispatch distance are used to automatically match suppliers and demanders with the nearest energy pathway with the lowest energy loss, further improving energy flow efficiency. An energy flow table is generated, automatically issuing instructions to suppliers to release excess energy, and to demanders to adjust energy. If resources are limited, adjustments are made based on the priority of the shortage until all resources are allocated and demand is met, demonstrating flexible scheduling and refined management. Elastic load and interruptible load nodes automatically adjust local load distribution based on the collaborative energy dispatch value, performing staggered operation and load reduction. At the same time, the supply-side microgrid automatically adjusts the output of local energy storage and distributed power sources to release available energy. The demand-side microgrid updates its own load configuration in real time, prioritizes core loads, and automatically enters a partial load reduction state. Each round of energy adjustment decisions, execution details, energy flows, and load adjustments based on the collaborative energy dispatch value are automatically recorded. The collaborative energy dispatch value, microgrid role, actual adjustment amount, response time, and energy loss of each microgrid before and after the adjustment are archived, ensuring the transparency and traceability of the decision-making process. Abnormal dispatch events where adjustments are not fully met, execution fails, or critical loads are affected are automatically marked and summarized as to the cause of the abnormality and the effectiveness of the response. The knowledge base is continuously updated using all archived dispatch events.

[0064] In this implementation plan, role identification is performed and optimization strategies are implemented based on the collaborative energy dispatch capability assessment results of each microgrid, which enables automatic identification and dynamic adaptation of energy supply and demand sides, improves energy flow efficiency, embodies flexible dispatch and refined management, ensures transparency and traceability of the decision-making process, and provides rich data support for subsequent optimization strategies and intelligent dispatch model training.

[0065] Reference Figure 2As shown, the second aspect of the present invention provides a microgrid dynamic coordination control system for distributed energy, which is applied to the above-mentioned microgrid dynamic coordination control method for distributed energy, including: a multi-source perception and intelligent analysis module, a regional topology and state fusion module, an autonomous control and elastic mode switching module and a regional collaborative scheduling and self-learning optimization module; wherein the multi-source perception and intelligent analysis module is used to collect multi-source microgrid data in real time, perform edge computing and data preprocessing operations, monitor anomalies in real time, and perform early judgment of island mode; the regional topology and state fusion module is used to aggregate the structure and state of each microgrid based on multi-source microgrid data, and dynamically construct and update the microgrid topology map , and analyze the structural changes and health status of the microgrid topology in real time, and evaluate the comprehensive risk of the microgrid nodes based on the early judgment results of the island mode and the microgrid topology; the autonomous control and elastic mode switching module is used to calculate the actual distributable power of each type of load according to the comprehensive risk of the microgrid nodes and the health status of the microgrid, and perform autonomous control and elastic mode switching according to the actual distributable power of each type of load; the regional collaborative scheduling and self-learning optimization module is used to evaluate the microgrid collaborative energy scheduling capability in real time based on the microgrid topology and the actual distributable power of each type of load, and perform role identification and implement optimization strategies according to the collaborative energy scheduling capability evaluation results of each microgrid.

[0066] In this implementation plan, based on the multi-module fusion strategy, it is possible to achieve real-time monitoring of anomalies, early judgment of island modes, dynamic assessment of comprehensive risks, optimization of load distribution, improvement of elastic control capabilities, strengthening of regional energy coordination and promotion of self-learning optimization of intelligent scheduling strategies.

[0067] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0068] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. The preferred embodiments do not describe all details in detail, nor do they limit the invention to the specific embodiments described. As those skilled in the art will appreciate, numerous modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A microgrid dynamic coordination control method for distributed energy, characterized in that: include: Step 1: Real-time collection of multi-source microgrid data, edge computing and data preprocessing, real-time monitoring of anomalies, and early identification of islanding modes. Step 2: Based on multi-source microgrid data, the structure and status of each microgrid are aggregated, and the microgrid topology is dynamically constructed and updated. The structural changes and health status of the microgrid topology are analyzed in real time. The comprehensive risk of microgrid nodes is assessed based on the early judgment results of the island mode and the microgrid topology. Step 3: Calculate the actual distributable power for each type of load based on the comprehensive risk of the microgrid nodes and the health status of the microgrid, and perform autonomous control and flexible mode switching based on the actual distributable power for each type of load; Step 4: Based on the microgrid topology, the collaborative energy dispatch capability of the microgrid is evaluated in real time based on the actual distributable power of each load type. The role is identified and the optimization strategy is implemented based on the collaborative energy dispatch capability evaluation results of each microgrid. The specific process of early judgment of the island mode is as follows: The voltage of the last ten minutes is obtained through the multi-source microgrid database and the average value is calculated as the historical steady-state average value. The current real-time voltage collected is subtracted from the historical steady-state average value and the average value is taken to obtain the voltage deviation. At the same time, the frequency of the last ten minutes is obtained through the multi-source microgrid database and the average value is calculated as the frequency reference. The current real-time frequency collected is subtracted from the frequency reference and the average value is taken to obtain the frequency deviation. Two voltages and two frequencies are continuously obtained based on the sliding time window. The difference between the two voltages is divided by the length of the sliding time window and the absolute value is taken to obtain the voltage change rate. The difference between the two frequencies is divided by the length of the sliding time window and the absolute value is taken to obtain the frequency change rate. The voltage deviation is multiplied by the frequency deviation, and then divided by the sum of the voltage change rate, the frequency change rate and a constant of one to obtain the voltage-frequency joint fluctuation intensity. The voltage deviation is subtracted from the product of the frequency deviation and the response weight factor to obtain the voltage-frequency response deviation value. The voltage-frequency joint fluctuation intensity and the voltage-frequency response deviation value are added to obtain the island judgment value. Compare the islanding determination value with the first and second level determination thresholds in real time. When the islanding determination value is less than the first level determination threshold, it is determined to be normal, and the normal grid-connected operation state is maintained. The multi-source microgrid data monitored is regularly archived and continuously observed. When the islanding threshold is greater than or equal to the first-level threshold and less than the second-level threshold, a mild islanding state is detected. A local rapid self-check process is initiated at the edge to verify the health of the voltage and frequency acquisition channels and communication links. Repeating this process multiple times within a specified timeframe, if the mild islanding state persists, the building's public lighting load is automatically reduced, and the energy storage system enters standby mode. And automatically push early warning information to operation and maintenance personnel; When the islanding judgment value is greater than or equal to the secondary judgment threshold, it is judged as a high-risk islanding state, and automatically switches to the islanding operation mode, disconnects from the main grid, and enables local distributed power supply and energy storage; critical loads are prioritized, and non-critical loads are automatically offline; the energy storage system dynamically discharges according to the preset priority; triggers local and cloud platform emergency alarms, records multi-source microgrid data and islanding judgment values ​​for the entire process of the islanding event, and links regional microgrids for mutual assistance to archive the islanding event.

2. A method for dynamic coordinated control of distributed energy microgrids according to claim 1, characterized in that: The specific process of real-time collection of multi-source microgrid data, edge computing and data preprocessing, and real-time monitoring of anomalies is as follows: Real-time collection of multi-source microgrid data: IoT terminal devices and smart sensors, as well as local energy storage monitoring systems and distributed generation monitoring systems, collect multi-source microgrid data in real time. Multi-source microgrid data includes: voltage, frequency, phase, energy storage SOC, energy storage capacity, distributed power output, load data, and active and reactive power. The sliding average method is used to remove invalid, abnormal and noisy multi-source microgrid data during the collection process through local filtering, and the multi-source microgrid data is compressed. The sliding window detection and logical rule screening methods are used to automatically identify and eliminate abnormal outliers caused by equipment failures and signal interference, and the multi-source microgrid data is standardized and normalized. At the same time, based on the long short-term memory network method, the multi-source microgrid data of each node is dynamically monitored, and electrical anomalies and communication anomalies are identified in real time. The multi-source microgrid data are stored in a multi-source microgrid database.

3. A method for dynamic coordinated control of distributed energy microgrids according to claim 1, characterized in that: The specific process of aggregating the structure and status of each microgrid based on multi-source microgrid data, dynamically constructing and updating the microgrid topology map, and analyzing the changes in the microgrid topology map structure and health status in real time is as follows: Through the edge gateway, smart terminal and master station communication interface, the topology structure of all microgrids, the online status of nodes, and the on / off status of key equipment such as circuit breakers and tie lines are regularly collected, and microgrid operation data is obtained in real time. Microgrid operation data includes voltage, frequency, load data, energy storage SOC and energy storage capacity. All topology structure, node online status, on / off status data of key equipment and microgrid operation data are encoded in a standardized format and normalized, and periodically uploaded to the scheduling and analysis platform through a secure data channel. Each microgrid unit is abstracted as a node in the microgrid topology. If two microgrid units are directly connected by a tie line, they are considered neighbors. The tie lines between nodes are abstracted as edges in the microgrid topology. The microgrid topology is dynamically constructed and updated using the real-time collected topology structure, node online status, key equipment on / off status data, and microgrid operation data. Based on the latest microgrid topology and the microgrid operation data of each node, the node features and edge features in the microgrid topology are extracted, the topological relationship is converted into an adjacency matrix, and the node features and edge features are combined into a feature matrix as the input of the graph convolutional neural network algorithm. The graph convolutional neural network algorithm automatically integrates the characteristics of each node itself and the characteristics of neighboring nodes, and considers the topological relationship between nodes. It performs deep learning analysis on the physical and state relationships between nodes, and automatically extracts the health status of nodes in the microgrid topology, including abnormal connections, weakly connected areas, and node health degradation trend information.

4. A method for dynamic coordinated control of distributed energy microgrids according to claim 1, characterized in that: The specific process of evaluating the comprehensive risk of microgrid nodes based on the early judgment results of the island mode and the microgrid topology diagram is as follows: By extracting the health status of the nodes in the microgrid topology, the health of node i and the health of neighboring node j are automatically summarized and normalized. Real-time load data is obtained, and the maximum value of the historical load data within the past 20 days is filtered out from the multi-source microgrid database as the maximum design capacity. The load ratio of node i is obtained by dividing the current real-time load data by the maximum design capacity. Obtain the neighbor node set of node i according to the microgrid topology map, and obtain the length of the neighbor node set to obtain the number of neighbor nodes; Obtain the on / off status of the key equipment of the circuit breaker and tie line between nodes i and j to obtain the connectivity strength between nodes i and j; calculate the islanding judgment value to obtain the islanding signal strength of node i; subtract the health of neighbor node j from the health of node i and take the absolute value, and then divide it by the sum of the connectivity strength of nodes i and j and a constant of one to obtain the inter-neighbor health; based on the set of neighboring nodes, accumulate all inter-neighbor healths and divide it by the number of neighboring nodes to obtain the average inter-neighbor health; multiply the average inter-neighbor health by the health weight factor, and then add it to the health of node i, the load ratio of node i, and the islanding signal strength of node i to obtain the comprehensive risk assessment value of node i; The comprehensive risk assessment value is compared with the first and second level risk thresholds in real time. When the comprehensive risk assessment value is less than the first level risk threshold, the microgrid node is considered risk-free and the node maintains normal monitoring and daily operation. The comprehensive risk assessment value and load change trend of each node are automatically recorded and archived regularly. When the comprehensive risk assessment value is greater than or equal to the first-level risk threshold and less than the second-level risk threshold, the microgrid node is considered to have a moderate risk. A risk attention reminder is automatically sent to the operation and maintenance personnel, and the node is highlighted in yellow on the platform interface. Local equipment self-inspections are performed to verify circuit breakers, energy storage status, and key communication link points. Non-critical interruptible loads are reduced, and energy storage charging is scheduled in advance. When the comprehensive risk assessment value is greater than or equal to the secondary risk threshold, the microgrid node is considered to be at high risk and a red alert is automatically triggered, with a pop-up window and SMS notification to the relevant responsible persons; Only core loads are retained, and all non-critical loads are forced offline; energy storage and distributed power sources enter high-priority energy supply mode; island operation is automatically prepared and switched to quickly isolate high-risk nodes; energy mutual assistance mechanisms with adjacent microgrids are automatically triggered; special emergency inspections and fault location are automatically started; and all risk events are archived.

5. A method for dynamic coordinated control of distributed energy microgrids according to claim 1, characterized in that: The specific process of calculating the actual distributable power for each type of load based on the comprehensive risk of the microgrid node and the health status of the microgrid is as follows: Based on the historical active power in the multi-source microgrid database, the maximum active power in the last twenty days is selected to obtain the current maximum allowable power of the i-th type load; Obtain real-time active power and current maximum allowable power, and automatically add them up to obtain the local total available power supply; Collect the instantaneous active power of all devices under the i-th load in real time and sum them up to obtain the real-time required power of the i-th load; Automatically traverse all load categories that are online and need to participate in the current park and count the total number to obtain the number of all load categories; based on the current number of all load categories, accumulate the real-time demand power of all load categories to obtain the total demand power of all load categories; divide the real-time demand power of the i-th load by the total demand power of all load categories, and then multiply it by the local total available power supply to obtain the theoretical distribution power value of the i-th load; compare the theoretical distribution power value of the i-th load with the current maximum allowable power of the i-th load, and take the smaller one as the actual distribution power value of the i-th load.

6. A method for dynamic coordinated control of distributed energy microgrids according to claim 1, characterized in that: The specific process of autonomous control and flexible mode switching based on the actual distributable power of various loads is as follows: Based on the actual allocated power values ​​of various loads, the core load is prioritized, and the remaining loads are operated in staggered peaks: if the actual allocated power value is equal to the real-time required power, the load is considered to be able to maintain normal operation and is fully guaranteed; if the actual allocated power value is less than the real-time required power, the load is considered to need to operate in staggered peaks and is partially guaranteed; prioritizing ensuring that all demands of the core load are met; Automatically reduce power supply in batches and time periods for important but adjustable loads; For interruptible loads, some charging equipment will be automatically suspended, power consumption of non-emergency production lines will be postponed, and interruptible loads will be transferred to off-peak periods before being restarted; load peak-shifting operation instructions will be issued to each terminal device, and detailed records will be kept of each load reduction, peak-shifting and guarantee process; local energy and load data will be continuously monitored, and the system will smoothly switch back to grid-connected mode when the main grid is restored.

7. A method for dynamic coordinated control of distributed energy microgrids according to claim 1, characterized in that: The specific process of real-time evaluation of the microgrid collaborative energy dispatching capability based on the microgrid topology diagram and the actual distributable power of each type of load is as follows: Obtain multi-source microgrid data, multiply the energy storage SOC by the energy storage capacity and add the distributed power output to obtain the local total available energy of microgrid i and the local total available energy of neighboring microgrid j; The active power of all load categories is obtained in real time. Based on the length of the scheduling time window, the current total energy demand of microgrid i and the current total energy demand of neighboring microgrid j are automatically accumulated. The set of neighboring microgrids of microgrid i is obtained based on the microgrid topology map, and the number of nodes passed by the shortest path from microgrid node i to microgrid node j is automatically calculated using breadth-first search. The energy transmission loss percentage between the two microgrids is calculated based on multi-source microgrid data and normalized to obtain the collaborative scheduling distance between microgrids i and j. The energy surplus or shortage value is obtained by subtracting the current total energy demand of neighbor microgrid j from the local total available energy of neighbor microgrid j, and then dividing it by the sum of the collaborative scheduling distance between microgrids i and j and a constant of one. Based on the set of neighbor microgrids, the energy surplus or shortage values ​​of all microgrids are accumulated to obtain the total energy surplus or shortage value. The total energy surplus or shortage value is multiplied by the neighborhood support item weight factor to obtain the domain support item. The current total energy demand of microgrid i is subtracted from the local total available energy of microgrid i, and then added to the domain support item to obtain the collaborative energy scheduling value of microgrid i.

8. A method for dynamic coordinated control of distributed energy microgrids according to claim 1, characterized in that: The specific process of identifying roles and implementing optimization strategies based on the collaborative energy dispatch capability evaluation results of each microgrid is as follows: Automatically identify each microgrid based on its collaborative energy dispatch value. If the value is greater than 0, the microgrid is considered an energy supplier and is given priority as an energy output and mutual support node. If the value is less than 0, the microgrid is considered an energy demander and is given priority in receiving regional collaborative energy support. All microgrids are sorted in descending order of priority according to the collaborative energy dispatch value. The energy supply with a large positive value has a high priority, and the gap with a large negative absolute value has a high priority. Automatically match suppliers and demanders with the nearest energy pathway with minimal loss using collaborative energy dispatch value and collaborative dispatch distance; Generate an energy flow table, automatically issue instructions to suppliers to release excess energy, and demanders receive energy adjustment instructions. If resources are limited, adjustments are made in order of gap priority until all resources are allocated and demand is met. Flexible load and interruptible load nodes automatically adjust local load distribution based on the coordinated energy scheduling value, implementing peak-shifting operation and load reduction. At the same time, the supply-side microgrid automatically adjusts the output of local energy storage and distributed power sources to release available energy; the demand-side microgrid updates its own load configuration in real time, giving priority to core loads and automatically entering a partial load reduction state; Automatically record each round of energy adjustment decisions, execution details, energy flow and load adjustment based on the collaborative energy scheduling value, and archive the collaborative energy scheduling value, microgrid role, actual adjustment amount, response time and energy loss of each microgrid before and after the adjustment; automatically mark and summarize the abnormal causes and response results for abnormal scheduling events such as incomplete adjustment, execution failure and impact on key loads; continuously update the knowledge base using all archived scheduling events.

9. A microgrid dynamic coordination control system for distributed energy, using the microgrid dynamic coordination control method for distributed energy as claimed in claim 1, characterized in that: include: Multi-source perception and intelligent analysis module, regional topology and state fusion module, autonomous control and elastic mode switching module, and regional collaborative scheduling and self-learning optimization module; The multi-source perception and intelligent analysis module is used to collect multi-source microgrid data in real time, perform edge computing and data preprocessing operations, monitor anomalies in real time, and make early judgments on islanding modes. The regional topology and status fusion module is used to aggregate the structure and status of each microgrid based on multi-source microgrid data, dynamically construct and update the microgrid topology map, and analyze the structural changes and health status of the microgrid topology map in real time. It also evaluates the comprehensive risk of microgrid nodes based on the early judgment results of the island mode and the microgrid topology map; The autonomous control and elastic mode switching module is used to calculate the actual distributable power of each type of load based on the comprehensive risk of the microgrid nodes and the health status of the microgrid, and perform autonomous control and elastic mode switching according to the actual distributable power of each type of load; The regional collaborative scheduling and self-learning optimization module is used to evaluate the collaborative energy scheduling capability of the microgrid in real time based on the actual distributable power of each type of load based on the microgrid topology map, and to identify roles and implement optimization strategies based on the collaborative energy scheduling capability evaluation results of each microgrid.

Images