Hierarchical coordination control system and method for AC / DC hybrid microgrid group

Abstract

The invention discloses a hierarchical coordination control system and method for an AC / DC hybrid micro-grid group, and relates to the technical field of smart grids and micro-grids, the system comprises a dynamic partition aggregation unit, a flexible interconnection control unit and a wide-area coordination control unit. The method is advantaged in that firstly, operation efficiency and reliability of the micro-grid group are substantially improved, stable supply of electric energy is guaranteed, secondly, the system has high flexibility and adaptivity, dynamic partition aggregation and flexible interconnection control can be carried out according to different source load scene characteristic parameters, various complex conditions can be effectively handled, response speed and stability of the system are improved, and the method is suitable for large-scale popularization and application. And finally, the implementation of the system also contributes to reducing the operation cost, realizes double improvement of economy and reliability by optimizing energy scheduling and energy storage management, and lays a solid foundation for sustainable development of the microgrid group.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grids and microgrids, and in particular to a hierarchical coordinated control system and method for an AC / DC hybrid microgrid group. Background Art

[0002] With the rapid development of distributed energy, AC / DC hybrid microgrids have received extensive attention and application as an efficient and flexible energy supply method. However, in actual operation, there are still some problems that need to be solved in AC / DC hybrid microgrids.

[0003] First, the control method of traditional microgrid groups is relatively simple and lacks a hierarchical coordination mechanism, which leads to irrational resource allocation, serious energy waste, and low operating efficiency. Secondly, the microgrid group lacks adaptability and is difficult to cope with complex changes in different source and load scenarios, resulting in reduced system stability and even safety accidents. In addition, the existing microgrid group lacks the dual guarantee of economy and reliability during operation, resulting in high operating costs and low user satisfaction.

[0004] Therefore, there is an urgent need for a hierarchical coordinated control system for an AC / DC hybrid microgrid group that can solve the above problems in order to improve the operating efficiency, stability and economy of the microgrid group. Summary of the invention

[0005] The purpose of the present invention is to make up for the shortcomings of the prior art and to provide a hierarchical coordinated control system and method for an AC / DC hybrid microgrid group, which can significantly improve the operating efficiency and reliability of the microgrid group through refined hierarchical management and coordinated control, and ensure the stable supply of electric energy. Secondly, the system is highly flexible and adaptable, and can perform dynamic partition aggregation and flexible interconnection control according to different source-load scenario characteristic parameters, effectively cope with various complex situations, and improve the response speed and stability of the system. Finally, the implementation of the system also helps to reduce operating costs, and achieves a dual improvement in economy and reliability by optimizing energy scheduling and energy storage management, laying a solid foundation for the sustainable development of the microgrid group.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a hierarchical coordination control system of an AC / DC hybrid microgrid group includes a dynamic partition aggregation unit, a flexible interconnection control unit and a wide area coordination control unit;

[0007] The dynamic partition aggregation unit is based on a multi-dimensional scene evaluation function Dynamically partition and aggregate the AC / DC hybrid microgrid in the substation area, where F i Represents the characteristic parameters of different source-load scenarios, α iis the weight coefficient determined by the reinforcement learning algorithm based on a large amount of historical data training, G represents the environmental factor parameter, β represents its corresponding weight, H represents the time factor parameter, γ represents its corresponding weight, I represents the microgrid structure stability parameter, δ represents its corresponding weight;

[0008] The flexible interconnection control unit is based on an adaptive interaction strategy through an interactive control function P out =f(P in ,ω,θ) to determine the output power, where P in is the input power, ω is the interaction weight parameter dynamically adjusted according to the real-time source-load change, and θ is the interaction control parameter vector;

[0009] The wide-area coordination control unit uses a distributed collaborative evolutionary algorithm to perform multi-level coordinated control of the wide-area distribution network source, network, load and storage, where the strategy set of each participant is S = {S1, S2, ..., S n}, the profit function is R={R1,R2,…,R n}, by simulating the inheritance, variation and selection operations in the biological evolution process, the strategy combination of each participant is continuously iterated and optimized.

[0010] Furthermore, the dynamic partition aggregation unit includes a scene feature collection subunit, a partition aggregation algorithm subunit and a dynamic adjustment subunit. The scene feature collection subunit collects multi-dimensional feature parameters of power supply, load, environment and microgrid structure in the substation in real time. The partition aggregation algorithm subunit adopts a clustering partition algorithm based on density peak. First, the substation microgrid node set is defined as N = {n1, n2, ..., n m}, for each node n j , calculate its comprehensive eigenvalue C nj =λ 1j P s,j +λ 2j L j +λ 3j T j +λ 4j S j +λ 5j J j , where P s,j is the power characteristic parameter of node j, L j is the load characteristic parameter of node j, T j is the environmental parameter of node j, S j is the microgrid structure parameter of node j, J j is the time-varying characteristic parameter of node j;

[0011] λ 1j , 2j , 3j , 4j ,5j It is dynamically determined by the particle swarm optimization algorithm. The particle swarm optimization algorithm is as follows: Assume that the particle swarm consists of multiple particles, each particle represents a set of weight coefficients, and the position of the particle is represented by X = {λ1,λ2,λ3,λ4,λ5}. The fitness function is defined as Where K is the number of sample nodes, C target,k is the target comprehensive eigenvalue, and the particle velocity update formula is V i (t+1)=wV i (t)+c1r1(P best,i -X i (t))+c2r2(G best -X i (t)), the position update formula is X i (t+1)=X i (t0+V i (t+10, where V i (t) is the velocity of particle i at time t, w is the inertia weight, and its initial value is set to 0.8. It gradually decreases with the number of iterations to balance the global search and local search capabilities. Where t is the current iteration number, T max is the maximum number of iterations, c1 and c2 are learning factors, set to 1.5 and 1.7 respectively, r1 and r2 are random numbers in the interval [0, 1], P best,i is the best historical position of particle i, G best For the best historical position of the entire particle swarm, the position and velocity of the particles are updated iteratively to find the weight coefficient combination that minimizes the fitness function;

[0012] Then cluster partitioning is performed according to the density peak of the comprehensive eigenvalue. Let node n j The local density is where d jk For node n j With node n k The distance between c is the cutoff distance, and the appropriate value is determined through experiments. Node n j The distance index is For the node with the largest local density, its distance index is δ j =max k d jk, select nodes with larger local density and larger distance index as cluster centers, and assign other nodes to the partitions to which the cluster centers with the closest distance and larger local density belong, so that the comprehensive characteristic values ​​of nodes in the partitions have higher similarity, and the comprehensive characteristic values ​​of nodes between partitions are quite different. The dynamic adjustment subunit realizes dynamic adjustment of the partition aggregation results by dynamically adjusting the weight coefficient and recalculating the comprehensive characteristic value according to the real-time changing characteristic parameters.

[0013] Furthermore, the flexible interconnection control unit includes a power interaction control subunit, a mode switching control subunit and a fault processing subunit, and the power interaction control subunit controls the power interaction of the low-voltage flexible interconnection device according to an adaptive interaction strategy;

[0014] The mode switching control subunit is based on the mode switching criterion function M switch =g(P in ,P out ,ω,η determine whether to switch the interactive mode, where P in is the input power, P out is the output power, ω is the interaction weight parameter, and its value is determined by the fuzzy logic algorithm combined with the current source-load state. The steps of the fuzzy logic algorithm are as follows: define the input variables as the source-load power difference E and the power change rate EC, and the output variable as the interaction weight parameter ω. Divide the input variables and the output variables into several fuzzy sets. The input variable E can be divided into fuzzy sets of positive large (PB), positive small (PS), zero (ZE), negative small (NS), and negative large (NB). EC is also divided similarly. The fuzzy rule is determined according to expert experience, that is, if E is PB and EC is PB, then ω is large. The centroid method is used for fuzzy reasoning and defuzzification to obtain a specific interaction weight parameter value. η is the mode switching threshold parameter vector. When the mode switching criterion function meets specific conditions, switch to different interaction modes, including but not limited to constant power mode and constant voltage mode.

[0015] When a fault occurs in the flexible interconnection device, the fault processing subunit takes measures such as quickly isolating the faulty part, starting the backup device and adjusting the interaction strategy to ensure the safe and stable operation of the microgrid.

[0016] Furthermore, the wide area coordination control unit includes a co-evolutionary computing subunit, a multi-level coordination subunit and an information interaction subunit;

[0017] The co-evolutionary computing subunit uses a distributed co-evolutionary algorithm to calculate the optimal strategy of each participant. In the algorithm, each participant is regarded as an individual, and the strategy of the individual is represented by a gene sequence. The gene sequence consists of multiple gene positions, and each gene position represents a decision variable. New individuals are generated through genetic operations (crossover and mutation). The crossover operation adopts a single-point crossover method, randomly selects a crossover point, and exchanges part of the genes of two parent individuals at the crossover point to generate two offspring individuals. The mutation operation adopts a random mutation method, and randomly changes the gene positions of the individual with a certain mutation probability. The fitness of the individual is evaluated according to the benefit function, and the benefit function is Where K is the number of different profit indicators, r k (S) is the profit value of individual strategy S under the kth profit index. Individuals with high fitness are selected for the next generation of evolution. The elite retention strategy is adopted to directly retain several individuals with the highest fitness in each generation to the next generation;

[0018] The multi-level coordination subunit coordinates participants at different levels by establishing a hierarchical coordination function H coord =h(S l1 ,S l2 ,…,S ln ), where S li represents the set of strategies of the participants at the i-th level, h is the coordination function, and the coordination strategy is determined by solving a multi-objective optimization problem. The objective function includes power balance and cost minimization among the levels;

[0019] The information interaction subunit establishes an efficient information transmission channel, uses encryption and verification technology to ensure the accuracy and security of information, and realizes real-time information interaction and sharing among all participants.

[0020] Furthermore, the control system further comprises a monitoring management unit for monitoring and managing the operating status of the AC / DC hybrid microgrid group, wherein the monitoring management unit comprises a data acquisition subunit, a status monitoring subunit and a control decision subunit;

[0021] The data acquisition subunit uses a distributed sensor network to collect various operating data of the microgrid in real time, including but not limited to voltage, current and power;

[0022] The state monitoring subunit analyzes and processes the collected data using the state evaluation function Evaluate the operating status of the microgrid, where X j Indicates key operating parameters, including but not limited to voltage deviation and frequency deviation, Y jrepresents abnormal index parameters, including but not limited to overcurrent and overvoltage, k and l are weights determined by data analysis, and the weights are determined by principal component analysis. After the data is standardized, the covariance matrix is ​​calculated, the eigenvalues ​​and eigenvectors are solved, and the weights of each parameter are determined according to the size of the eigenvalues;

[0023] The control decision subunit formulates corresponding control decisions based on the operating status evaluation results and the preset control strategy, and transmits the decision information to each control unit for execution.

[0024] Furthermore, the units in the system exchange information with each other through a high-speed communication network, the communication network adopts a self-organizing mesh topology, and the communication protocol includes data format, transmission mode and error checking content;

[0025] The data format adopts a customized data packet structure, including fields for source address, destination address, data type and data content. The transmission method adopts a hybrid transmission method based on the combination of time division multiple access and code division multiple access to improve the efficiency and reliability of data transmission. In the time division multiple access part, time is divided into multiple time slots, and different nodes send data in different time slots. In the code division multiple access part, spread spectrum code is used to modulate data, and different nodes use different spread spectrum codes to distinguish different data sources. Error checking adopts a combination of cyclic redundancy check and hash check to ensure accurate transmission of information. The cyclic redundancy check is used to detect errors in data transmission and generate a check code. The receiver determines whether the data is correct based on the check code. The hash check is used to verify the integrity of the data, and a hash operation is performed on the data to generate a hash value. The receiver performs the same hash operation on the received data and compares whether the two hash values ​​are consistent.

[0026] Furthermore, the system also includes an energy management unit, including an energy prediction subunit, an optimization scheduling subunit and an energy storage management subunit. The energy prediction subunit uses an innovative prediction algorithm that combines deep learning with time series analysis to predict power output and load demand. The prediction model is P pred =f pred (t,ω1,ω2,…,ω n ), where t is the time variable, ω is the model parameter determined by training, and a deep neural network is used for training. The network structure includes an input layer, several hidden layers and an output layer. The input layer receives historical data, including but not limited to power output power, load demand and environmental factors over a period of time in the past. The hidden layer extracts the features of the data through nonlinear transformation, and the output layer outputs the predicted power value. The model parameters are adjusted through the back propagation algorithm to minimize the error between the predicted value and the actual value.

[0027] The optimization dispatch subunit takes the minimum operation cost and the highest reliability of the microgrid as the objective function. The objective function is F obj =μC+νR, where C represents the operating cost, including the power generation cost and the equipment maintenance cost, R represents the reliability index, including the voltage stability and the power supply reliability, μ and ν are weight coefficients, and the optimal energy scheduling scheme is solved by the multi-objective optimization algorithm. The genetic algorithm is used for solving, and the individual represents an energy scheduling scheme, which is continuously evolved through crossover, mutation and selection operations to find the scheduling scheme that makes the objective function optimal;

[0028] The energy storage management subunit controls the charging and discharging of the energy storage device according to the energy scheduling plan, and uses the energy storage control function E ctrl =g ctrl (S es ,ω es ,θ es ) controls the state of the energy storage device, where S es is a set of state parameters of the energy storage device, including but not limited to power and charge / discharge power, ω es is the energy storage control weight parameter, which is determined according to the type and performance of the energy storage device, θ es is the control parameter vector.

[0029] Furthermore, each control unit of the system adopts a distributed control architecture and is distributed on different nodes of the microgrid. The control unit of each node has a certain degree of autonomous decision-making ability, and makes control decisions through local information and information interaction with adjacent nodes. The nodes exchange information and coordinate control through a high-speed communication network, and a distributed consistency algorithm is used to ensure that the control decisions of each node are coordinated and consistent;

[0030] The distributed consensus algorithm iteratively updates the state variables of each node so that the state variables of all nodes gradually converge to a consistent value. The state variable update function is X i (t+1)=f cons (X i (t),N i ,ω cons ), where X i (t) represents the state variable of node i at time t, N i is the set of adjacent nodes of node i, ω cons is the consistency weight parameter, which is determined according to the connection strength and communication quality between nodes. The stronger the connection strength and the better the communication quality, the larger the weight parameter. Initially, the state variables of each node are set randomly, and then gradually tend to a consistent value through continuous iterative updates. When the difference between the state variables of all nodes is less than a preset threshold τ, consistency is considered to be achieved.

[0031] Furthermore, the system also includes a user interaction unit, including a user interface subunit and a demand response subunit;

[0032] The user interface subunit provides a user-friendly operation interface, which graphically displays the operating status and control parameters of the microgrid;

[0033] The demand response subunit adjusts the operation strategy of the microgrid according to the user's electricity demand and feedback. The demand response function is R dr =h dr (D user ,ω dr ,θ dr ), where D user Represents user demand parameters, including power consumption and power consumption time, ω dr is the demand response weight parameter, which is determined according to the user's electricity priority and urgency, θ dr is the demand response control parameter vector. When the user demand changes, the demand response subunit meets the user's electricity demand by adjusting the partition aggregation, flexible interconnection control and wide-area coordinated control strategy of the microgrid.

[0034] On the other hand, a method for using a hierarchical coordinated control system of an AC / DC hybrid microgrid group is provided, and the specific steps of the method are as follows:

[0035] Step 1: System startup and initialization

[0036] Start each unit of the hierarchical coordination control system of the AC / DC hybrid microgrid group;

[0037] Each subunit is initialized, including random initialization of particle positions in the particle swarm optimization algorithm and random setting of node state variables in the distributed consensus algorithm;

[0038] Step 2: Data collection and feature extraction

[0039] The scene feature acquisition subunit collects multi-dimensional feature parameters of power supply, load, environment and microgrid structure in the substation in real time;

[0040] Process the collected data and extract key features, including various parameters required to calculate the comprehensive characteristic values ​​of the nodes;

[0041] Step 3: Dynamic Partition Aggregation

[0042] The partition aggregation algorithm subunit uses the density peak-based clustering partition algorithm to perform dynamic partition aggregation based on the collected feature parameters;

[0043] The weight coefficient is determined by the particle swarm optimization algorithm, the comprehensive characteristic value of the node is calculated, and the partition range is determined according to the density peak;

[0044] Step 4: Flexible interconnection control

[0045] The flexible interconnection control unit determines the interaction weight parameter ω through the fuzzy logic algorithm according to the current source-load status and partition situation;

[0046] According to the interactive control function P out =f(P in ,ω,θ) control the low voltage flexible interconnection device to perform power interaction;

[0047] The mode switching control subunit is based on the mode switching criterion function M switch =g(P in ,P out ,ω,η) determine whether to switch the interaction mode;

[0048] Step 5: Wide-area coordinated control

[0049] The co-evolution computing subunit of the wide-area coordination control unit uses a distributed co-evolution algorithm to calculate the optimal strategy of each participant;

[0050] The multi-level coordination subunits establish a hierarchical coordination function H coord =h(S l1 ,S l2 ,…,S ln ) Coordinate with stakeholders at different levels;

[0051] The information interaction subunit ensures real-time information interaction and sharing among all participants;

[0052] Step 6: Monitoring and Management

[0053] The data acquisition subunit of the monitoring and management unit continuously collects microgrid operation data;

[0054] The condition monitoring subunit uses the condition evaluation function Evaluate the microgrid operating status;

[0055] The control decision subunit formulates corresponding control decisions according to the operating status and transmits them to each control unit for execution;

[0056] Step 7: Energy Management

[0057] The energy prediction subunit uses a prediction algorithm that combines deep learning with time series analysis to predict power output and load demand. The prediction model is P pred =f pred (t,ω1,ω2,…,ω n );

[0058] The optimization dispatch subunit takes the minimization of the operation cost and the maximization of the reliability of the microgrid as the objective function F obj=μC+νR, and the optimal energy scheduling solution is solved through a multi-objective optimization algorithm;

[0059] The energy storage management subunit controls the energy storage control function E according to the energy scheduling plan. ctrl =g ctrl (S es ,ω es ,θ es ) Control the charge and discharge status of energy storage equipment;

[0060] Step 8: User Interaction

[0061] The user interface subunit provides users with a friendly operation interface, which graphically displays the operating status and control parameters of the microgrid;

[0062] The demand response subunit adjusts the operation strategy of the microgrid according to the user's electricity demand and feedback. When the user's demand changes, the user's electricity demand is met by adjusting the strategies of partition aggregation, flexible interconnection control and wide-area coordinated control;

[0063] Step 9: Dynamic Adjustment and Coordination

[0064] When parameters in the system change, including changes in source-load status and user demand, the dynamic partition aggregation unit, flexible interconnection control unit, and wide-area coordination control unit are dynamically adjusted according to the changes;

[0065] Each unit exchanges information and coordinates control through a high-speed communication network to ensure stable and efficient operation of the system;

[0066] Step 10: Continue to run and optimize

[0067] The system operates continuously, constantly monitoring and analyzing the operating status and data of the microgrid;

[0068] According to the operating conditions and new data, the algorithms and parameters in the system are optimized and adjusted to improve the performance and adaptability of the system.

[0069] Compared with the prior art, the hierarchical coordinated control system and method of the AC / DC hybrid microgrid group has the following beneficial effects:

[0070] 1. The present invention adopts a hierarchical coordinated control system, which can realize refined management and control of AC / DC hybrid microgrid groups. The dynamic partition aggregation unit dynamically partitions and aggregates the AC / DC hybrid microgrid in the substation area according to the multi-dimensional scenario evaluation function to ensure that the source-load characteristics, environmental factors, time factors and microgrid structure stability in each partition are fully considered. This dynamic partitioning method helps to optimize resource allocation, reduce energy waste and improve the operating efficiency of the system. At the same time, the collaborative work of the flexible interconnection control unit and the wide-area coordination control unit can achieve the goals of power balance, cost minimization and reliability maximization for each participant in the microgrid, further improving the overall reliability and stability of the system.

[0071] 2. The present invention can adjust the partitioning strategy in real time according to the actual operating conditions through the dynamic partitioning aggregation unit to adapt to different source-load scenarios and changes in environmental factors. The flexible interconnection control unit can determine the interaction weight parameters through the fuzzy logic algorithm according to the current source-load status and partition situation, and control the low-voltage flexible interconnection device to perform power interaction according to the interactive control function to achieve flexible interaction between sources and loads. In addition, the wide-area coordination control unit can coordinate participants at different levels to ensure power balance and cost minimization between levels, thereby enhancing the overall adaptability and flexibility of the system.

[0072] 3. The present invention monitors and manages the operating status of the microgrid in real time through the monitoring management unit, which can promptly discover and handle abnormal situations to ensure the safe and stable operation of the system. At the same time, the energy management unit can formulate the optimal energy scheduling plan based on the prediction results of the power output power and load demand to achieve the economic operation of the microgrid. In addition, the user interaction unit provides users with a friendly operation interface and a convenient interaction method, so that users can intuitively understand the operating status and control parameters of the microgrid and adjust the operating strategy of the microgrid according to their own needs. This user interaction method not only improves user participation and satisfaction, but also helps to promote the sustainable development of the microgrid.

[0073] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0075] Figure 1 The present invention is a flow chart of a hierarchical coordinated control system and method for an AC / DC hybrid microgrid group. DETAILED DESCRIPTION

[0076] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0077] Example: AC / DC Hybrid Microgrid in Industrial Park

[0078] 1. Implementation Background

[0079] In modern industrial parks, the electricity demands of many enterprises are highly diverse and dynamic. The electricity load of different types of factories varies significantly over time during the production process. At the same time, some enterprises in the park have also installed distributed power sources, such as solar power generation systems and small wind power generation devices. In order to achieve efficient use of energy, improve power supply reliability and reduce energy costs, a hierarchical coordinated control system of AC / DC hybrid microgrid groups is introduced in the industrial park.

[0080] 2. System deployment and parameter setting

[0081] (1) System startup and initialization:

[0082] Turn on each unit of the system. In the particle swarm optimization algorithm, the initial position of the particle is represented by X = {0.2, 0.3, 0.4, 0.1, 0.5}, which means that at the beginning of the algorithm, a set of weight coefficients represented by each particle is randomly set to these values. In the distributed consensus algorithm, the node state variables are randomly set in the interval [0, 1] to ensure that the system has a certain degree of randomness and diversity in the initial state.

[0083] (2) Data collection and feature extraction:

[0084] The scene feature collection subunit collects the power supply type, load characteristics, environmental parameters and microgrid structure parameters of each factory in real time through sensors installed at various locations in the park.

[0085] For example, for a factory mainly engaged in manufacturing, its power supply mainly includes a 500kW solar power generation system and a 300kW traditional thermal power backup power supply. During a specific time period, the factory's load demand is an average of 400kW, and can reach 450kW during peak hours. In terms of environmental parameters, the current temperature is collected to be 25°C and the humidity is 60%. The microgrid structure parameters include information such as the length of the connection line between the factory and other factories and the line impedance.

[0086] (3) Dynamic partition aggregation:

[0087] Evaluate functions based on multi-dimensional scenarios Partition aggregation is performed, where the weight coefficient α is determined by the reinforcement learning algorithm i are {0.2, 0.15, 0.25, 0.1, 0.3} respectively, where F i The weight β corresponding to the environmental factor parameter G (such as temperature, humidity, etc.) is set to 0.15. Under the current environment, changes in temperature and humidity will have a certain impact on the efficiency of the factory's electrical equipment, thereby indirectly affecting the operation status of the microgrid. The weight γ corresponding to the time factor parameter H is 0.1. Considering that the production activities of the factory are different in different time periods, the electricity demand will also change. The weight δ corresponding to the microgrid structure stability parameter I is 0.2. Ensuring the structural stability of the microgrid is crucial for the reliable operation of the system.

[0088] For each node (factory) in the park, calculate its comprehensive characteristic value C n =λ1P s +λ2L+λ3T+λ4S+λ5J, where, P s are power characteristic parameters. For the above-mentioned manufacturing factories, their power characteristic parameters may include the output power stability of the solar power generation system, the reliability of the traditional thermal power backup power supply, etc. L is the load characteristic parameter, such as the power load curve of different production equipment in the factory. T is the environmental parameter. Factors such as the current temperature and humidity will affect the heat dissipation and insulation performance of the equipment, thereby affecting the power demand. S is the microgrid structure parameter, such as the connection method and line capacity with other factories. J is the time-varying characteristic parameter, which reflects the changing trend of the factory's power demand over time.

[0089] The weight coefficients were determined by the particle swarm optimization algorithm. After several iterations, the weight coefficients were determined to be {0.3, 0.2, 0.15, 0.2, 0.15} for a specific plant node, which means that the importance of each parameter was further optimized in the calculation of the comprehensive eigenvalue of this node.

[0090] Cluster partitioning is performed according to the density peak of the comprehensive eigenvalue, and the local density of the node is ρ = ∑ k≠j χ(d jk -d c ), where d jk is the distance between nodes, and the cutoff distance d is determined by experiment c For example, after multiple experiments and data analysis, it is determined that when d c = 100 meters, different areas can be divided better, and the distance index of the node is (For the node with the largest local density, its distance index is δ = max k d jk ), select the node with larger local density and larger distance index as the cluster center, and assign other nodes to the partition to which the cluster center with the closest distance and larger local density belongs. In this way, factories with similar power consumption characteristics and close geographical locations can be divided into the same partition, which is convenient for centralized management and coordinated control.

[0091] (4) Flexible interconnection control:

[0092] The flexible interconnection control unit determines the interaction weight parameter ω through the fuzzy logic algorithm according to the current source-load status and partition situation. In the fuzzy logic algorithm, the input variables are defined as the source-load power difference E and the power change rate EC. E is divided into five fuzzy sets: positive large (PB), positive small (PS), zero (ZE), negative small (NS), and negative large (NB). For example, when the source-load power difference is greater than 50kW, it belongs to the positive large (PB) set; when the source-load power difference is between 10kW and 50kW, it belongs to the positive small (PS) set. The power change rate EC is also divided similarly. The fuzzy rules are determined according to expert experience, such as "if E is PS and EC is ZE, then ω is medium", etc. The interaction weight parameter is determined to be 0.6 through fuzzy reasoning and defuzzification through the center of gravity method.

[0093] According to the interactive control function P out =f(P in ,ω,θ) controls the low voltage flexible interconnection device to perform power interaction, where the input power P in The output power is 50kW, and the output power is 45kW through the function calculation. The function takes into account the input power, interaction weight parameters and other control parameter vectors θ to ensure the reasonable distribution and transmission of power. The mode switching control subunit is based on the mode switching criterion function M switch =g(P in ,P out,ω,η) determines whether to switch the interactive mode. The current mode switching threshold parameter vector η is set to {40,50,0.5,0.3}, which does not meet the switching condition temporarily. This means that under the current input power, output power, interactive weight parameters and threshold parameters, the system does not need to switch the interactive mode.

[0094] (5) Wide-area coordinated control:

[0095] The co-evolutionary computing subunit uses a distributed co-evolutionary algorithm to calculate the optimal strategy of each participant. In the algorithm, each participant is regarded as an individual, and the individual's strategy is represented by a gene sequence. The gene sequence consists of multiple gene bits, and each gene bit represents a decision variable. For example, one gene bit may represent the power distribution ratio of a factory, and another gene bit may represent the control parameters of a flexible interconnection device.

[0096] New individuals are generated through genetic operations (crossover and mutation). The crossover operation adopts a single-point crossover method, randomly selects a crossover point, and exchanges some genes of two parent individuals at the crossover point to generate two offspring individuals. For example, for two individuals representing different factory strategies, the gene bits of the power distribution ratio are exchanged at the crossover point to generate a new strategy combination. The mutation operation adopts a random mutation method to randomly change the gene bits of individuals with a certain mutation probability. For example, the power distribution ratio gene bits of a factory are randomly adjusted with a probability of 5% to increase the diversity of strategies.

[0097] The fitness of an individual is evaluated according to the profit function, which is Suppose there are three benefit indicators, corresponding to different benefit values ​​in terms of cost, reliability and efficiency. For example, one benefit indicator may be the operating cost of the microgrid, including power generation cost, equipment maintenance cost, etc.; another benefit indicator may be power supply reliability, such as voltage stability, power supply interruption time, etc.; the third benefit indicator may be energy utilization efficiency, selecting individuals with high fitness for the next generation of evolution, and adopting an elite retention strategy to retain the top 10% individuals with the highest fitness to the next generation, so as to ensure that excellent strategies are retained and inherited during the evolution process.

[0098] The multi-level coordination subunits establish a hierarchical coordination function H coord =h9S l1 ,S l2 ,…,S ln) Coordinate participants at different levels to ensure power balance and cost minimization between levels. For example, large factories and small factories in the park can be used as participants at different levels respectively, and the power allocation and interaction strategies between them can be adjusted through coordination functions to optimize the overall system. The information interaction subunit ensures real-time information interaction and sharing among all participants. Through the high-speed communication network, each factory and control unit can exchange information such as power demand, power generation, equipment status, etc. in time to make rapid decisions and coordinated control.

[0099] (6) Monitoring and management:

[0100] The data acquisition subunit continuously collects microgrid operation data, such as voltage, current, power, etc., and monitors the operation status of the microgrid in real time through sensors installed on various nodes and lines. For example, the power consumption and voltage data of each factory are collected every 5 minutes.

[0101] The condition monitoring subunit uses the condition evaluation function To evaluate the operation status of the microgrid, the weights k and l are determined to be 0.6 and 0.4 by principal component analysis. j Indicates key operating parameters, such as voltage deviation, frequency deviation, etc. j Indicates abnormal indicator parameters, such as overcurrent, overvoltage, etc. For example, if the voltage deviation of a factory exceeds 5%, it is considered that there are certain problems with the operation status of the factory and adjustment and optimization are needed.

[0102] (7) Energy management:

[0103] The energy prediction subunit adopts a prediction algorithm that combines deep learning with time series analysis. The prediction model is P pred =f pred (t,ω1,ω2,…,ω n ), trained through a deep neural network, input historical data, such as power output power, load demand, environmental factors, etc. in the past period of time. For example, the solar power generation power, factory power load, and environmental parameters such as temperature and humidity at different time periods every day in the past week are input. The model parameter ω is determined according to the historical data training. By continuously adjusting the model parameters, the error between the predicted value and the actual value is minimized.

[0104] The optimization dispatch subunit takes the minimization of the operation cost and the maximization of the reliability of the microgrid as the objective function F obj=μC+νR, the weight coefficient μ is set to 0.7, ν is set to 0.3, and the optimal energy scheduling scheme is solved by a multi-objective optimization algorithm, where C represents the operating cost, including power generation cost, equipment maintenance cost, etc., and R represents the reliability index, such as voltage stability, power supply reliability, etc. For example, the power generation distribution of each factory, the charging and discharging strategy of the energy storage equipment, and the control parameters of the flexible interconnection device are adjusted through the optimization algorithm to achieve the goals of minimizing cost and maximizing reliability.

[0105] The energy storage management subunit controls the energy storage control function E according to the energy scheduling plan. ctrl =g ctrl (S es ,ω es ,θ es ) controls the charge and discharge state of the energy storage device, where S es is a set of state parameters of energy storage equipment, such as power, charge and discharge power, etc., ω es is the energy storage control weight parameter, which is set to 0.8 according to the type and performance of the energy storage device. es To control the parameter vector, for example, when the power load of the microgrid is low, the excess power is stored in the energy storage device, and when the power load is peak, the power in the energy storage device is released to balance the supply and demand.

[0106] (8) User Interaction:

[0107] The user interface subunit provides a friendly operating interface for park managers, displaying the operating status and control parameters of the microgrid. Through the graphical interface, managers can intuitively understand the power consumption, power generation, energy storage equipment status and other information of each factory. For example, the interface can display the real-time power consumption curve of each factory, the output power of the solar power generation system and the power level of the energy storage equipment.

[0108] The demand response subunit adjusts the operation strategy of the microgrid according to the changes in the electricity demand of the factories in the park. For example, when a factory temporarily increases its electricity consumption, the demand response subunit can adjust the partition aggregation and flexible interconnection control strategy to meet its electricity demand. Specifically, the partition where the factory is located can temporarily interact with other partitions, or adjust the output power of the flexible interconnection device to ensure that the factory's electricity demand is met.

[0109] (9) Dynamic adjustment and coordination:

[0110] When the factory's production plan adjustment causes the source and load status to change, the dynamic partition aggregation unit, flexible interconnection control unit and wide-area coordination control unit will be dynamically adjusted according to the changes. For example, if a factory adds new production equipment, resulting in a substantial increase in power load, the system will automatically adjust the partition aggregation algorithm, recalculate the node comprehensive characteristic values, and make partition adjustments. At the same time, the flexible interconnection control unit will adjust the interaction weight parameters and power interaction strategy according to the new source and load status, and the wide-area coordination control unit will also recalculate the optimal strategy of each participant to ensure the stable operation of the microgrid.

[0111] Each unit exchanges information and coordinates control through a high-speed communication network to ensure stable and efficient operation of the system. The communication network adopts a self-organizing mesh topology with high reliability and scalability. The data format adopts a customized data packet structure, including fields such as source address, destination address, data type, and data content. The transmission method adopts a hybrid transmission method based on time division multiple access and code division multiple access to improve the efficiency and reliability of data transmission. Error checking adopts a combination of cyclic redundancy check and hash check to ensure accurate transmission of information.

[0112] (10) Continuous operation and optimization:

[0113] The system runs continuously, constantly monitoring and analyzing the operating status and data of the microgrid. Through real-time data collection and analysis algorithms, the system can promptly identify potential problems and optimization opportunities. For example, by monitoring parameters such as voltage deviation and frequency deviation, the stability of the microgrid can be determined. By analyzing the power load curve and the power generation curve, the energy scheduling strategy can be optimized.

[0114] According to the operating conditions and new data, the algorithms and parameters in the system are optimized and adjusted to improve the performance and adaptability of the system. For example, according to the data feedback from actual operation, the parameters of the particle swarm optimization algorithm are adjusted to improve the accuracy of partition aggregation, the fuzzy rules and weight parameters of the fuzzy logic algorithm are adjusted, the flexible interconnection control strategy is optimized, and the genetic operation parameters and benefit function of the distributed collaborative evolution algorithm are adjusted to improve the effect of wide-area coordinated control. At the same time, as the factories in the park develop and change, the system can continuously adapt to new needs and challenges to achieve continuous optimization and improvement.

[0115] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A hierarchical coordinated control system for an AC / DC hybrid microgrid group, characterized in that: The system includes a dynamic partition aggregation unit, a flexible interconnection control unit and a wide area coordination control unit; The dynamic partition aggregation unit is based on a multi-dimensional scene evaluation function Dynamically partition and aggregate the AC / DC hybrid microgrid in the substation area, where F i Represents the characteristic parameters of different source-load scenarios, α i is the weight coefficient determined by the reinforcement learning algorithm based on a large amount of historical data training, G represents the environmental factor parameter, β represents its corresponding weight, H represents the time factor parameter, γ represents its corresponding weight, I represents the microgrid structure stability parameter, δ represents its corresponding weight; The flexible interconnection control unit is based on an adaptive interaction strategy through an interactive control function P out =f(P in ,ω,θ) to determine the output power, where P in is the input power, ω is the interaction weight parameter dynamically adjusted according to the real-time source-load change, and θ is the interaction control parameter vector; The wide-area coordination control unit uses a distributed collaborative evolutionary algorithm to perform multi-level coordinated control of the wide-area distribution network source, network, load and storage, where the strategy set of each participant is S = {S1, S2, ..., S n }, the profit function is R={R1,R2,…,R n }, by simulating the inheritance, variation and selection operations in the biological evolution process, the strategy combination of each participant is continuously iterated and optimized.

2. A hierarchical coordinated control system for an AC / DC hybrid microgrid group according to claim 1, characterized in that: The dynamic partition aggregation unit includes a scene feature acquisition subunit, a partition aggregation algorithm subunit and a dynamic adjustment subunit. The scene feature acquisition subunit collects multi-dimensional feature parameters of power supply, load, environment and microgrid structure in the substation in real time. The partition aggregation algorithm subunit adopts a clustering partition algorithm based on density peak. First, the substation microgrid node set is defined as N = {n1, n2, ..., n m }, for each node n j , calculate its comprehensive eigenvalue Where P s,j is the power characteristic parameter of node j, L j is the load characteristic parameter of node j, T j is the environmental parameter of node j, S j is the microgrid structure parameter of node j, J j is the time-varying characteristic parameter of node j; λ 1j , 2j , 3j , 4j , 5j It is dynamically determined by the particle swarm optimization algorithm. The particle swarm optimization algorithm is as follows: Assume that the particle swarm consists of multiple particles, each particle represents a set of weight coefficients, and the position of the particle is represented by X = {λ1,λ2,λ3,λ4,λ5}. The fitness function is defined as Where K is the number of sample nodes, C target,k is the target comprehensive eigenvalue, and the particle velocity update formula is V i (t+1)=wV i (t)+c1r1(P best,i -X i (t))+c2r2(G best -X i (t)), the position update formula is X i (t+1)=X i (t)+V i (t+1), where V i (t) is the velocity of particle i at time t, w is the inertia weight, and its initial value is set to 0.

8. It gradually decreases with the number of iterations to balance the global search and local search capabilities. Where t is the current iteration number, T max is the maximum number of iterations, c1 and c2 are learning factors, set to 1.5 and 1.7 respectively, r1 and r2 are random numbers in the interval [0, 1], P best,i is the best historical position of particle i, G best For the best historical position of the entire particle swarm, the position and velocity of the particles are updated iteratively to find the weight coefficient combination that minimizes the fitness function; Then cluster partitioning is performed according to the density peak of the comprehensive eigenvalue. Let node n j The local density is ρ j =∑ k≠j χ ( d jk -d c ) , where d jk For node n j With node n k The distance between c is the cutoff distance, and the appropriate value is determined through experiments. Node n j The distance index is For the node with the largest local density, its distance index is δ j =max k d jk , select nodes with larger local density and larger distance index as cluster centers, and assign other nodes to the partitions to which the cluster centers with the closest distance and larger local density belong, so that the comprehensive characteristic values ​​of nodes in the partitions have higher similarity, and the comprehensive characteristic values ​​of nodes between partitions are quite different. The dynamic adjustment subunit realizes dynamic adjustment of the partition aggregation results by dynamically adjusting the weight coefficient and recalculating the comprehensive characteristic value according to the real-time changing characteristic parameters.

3. The hierarchical coordination control system of an AC / DC hybrid microgrid group according to claim 1 is characterized in that: The flexible interconnection control unit includes a power interaction control subunit, a mode switching control subunit and a fault processing subunit, and the power interaction control subunit controls the power interaction of the low-voltage flexible interconnection device according to an adaptive interaction strategy; The mode switching control subunit is based on the mode switching criterion function M switch =g(P in ,P out ,ω,η) determines whether to switch the interaction mode, where P in is the input power, P out is the output power, ω is the interaction weight parameter, and its value is determined by the fuzzy logic algorithm combined with the current source-load state. The steps of the fuzzy logic algorithm are as follows: define the input variables as the source-load power difference E and the power change rate EC, and the output variable as the interaction weight parameter ω. Divide the input variables and the output variables into several fuzzy sets. The input variable E can be divided into fuzzy sets of positive large (PB), positive small (PS), zero (ZE), negative small (NS), and negative large (NB). EC is also divided similarly. The fuzzy rule is determined according to expert experience, that is, if E is PB and EC is PB, then ω is large. The centroid method is used for fuzzy reasoning and defuzzification to obtain a specific interaction weight parameter value. η is the mode switching threshold parameter vector. When the mode switching criterion function meets specific conditions, switch to different interaction modes, including but not limited to constant power mode and constant voltage mode. When a fault occurs in the flexible interconnection device, the fault processing subunit takes measures such as quickly isolating the faulty part, starting the backup device and adjusting the interaction strategy to ensure the safe and stable operation of the microgrid.

4. The hierarchical coordination control system of an AC / DC hybrid microgrid group according to claim 1, characterized in that: The wide area coordination control unit includes a co-evolutionary computing subunit, a multi-level coordination subunit and an information interaction subunit; The co-evolutionary computing subunit uses a distributed co-evolutionary algorithm to calculate the optimal strategy of each participant. In the algorithm, each participant is regarded as an individual, and the strategy of the individual is represented by a gene sequence. The gene sequence consists of multiple gene positions, and each gene position represents a decision variable. New individuals are generated through genetic operations (crossover and mutation). The crossover operation adopts a single-point crossover method, randomly selects a crossover point, and exchanges part of the genes of two parent individuals at the crossover point to generate two offspring individuals. The mutation operation adopts a random mutation method, and randomly changes the gene positions of the individual with a certain mutation probability. The fitness of the individual is evaluated according to the benefit function, and the benefit function is Where K is the number of different profit indicators, r k (S) is the profit value of individual strategy S under the kth profit index. Individuals with high fitness are selected for the next generation of evolution. The elite retention strategy is adopted to directly retain several individuals with the highest fitness in each generation to the next generation; The multi-level coordination subunit coordinates participants at different levels by establishing a hierarchical coordination function H coord =h(S l1 ,S l2 ,…,S ln ), where S li represents the set of strategies of the participants at the i-th level, h is the coordination function, and the coordination strategy is determined by solving a multi-objective optimization problem. The objective function includes power balance and cost minimization among the levels; The information interaction subunit establishes an efficient information transmission channel, uses encryption and verification technology to ensure the accuracy and security of information, and realizes real-time information interaction and sharing among all participants.

5. The hierarchical coordination control system of an AC / DC hybrid microgrid group according to claim 1, characterized in that: The control system also includes a monitoring and management unit for monitoring and managing the operating status of the AC / DC hybrid microgrid group, wherein the monitoring and management unit includes a data acquisition subunit, a status monitoring subunit and a control decision subunit; The data acquisition subunit uses a distributed sensor network to collect various operating data of the microgrid in real time, including but not limited to voltage, current and power; The state monitoring subunit analyzes and processes the collected data using the state evaluation function Evaluate the operating status of the microgrid, where X j Indicates key operating parameters, including but not limited to voltage deviation and frequency deviation, Y j represents abnormal index parameters, including but not limited to overcurrent and overvoltage, k and l are weights determined by data analysis, and the weights are determined by principal component analysis. After the data is standardized, the covariance matrix is ​​calculated, the eigenvalues ​​and eigenvectors are solved, and the weights of each parameter are determined according to the size of the eigenvalues; The control decision subunit formulates corresponding control decisions based on the operating status evaluation results and the preset control strategy, and transmits the decision information to each control unit for execution.

6. The hierarchical coordination control system of an AC / DC hybrid microgrid group according to claim 1, characterized in that: The units in the system exchange information through a high-speed communication network, which adopts a self-organizing mesh topology. The communication protocol includes data format, transmission mode and error checking content; The data format adopts a customized data packet structure, including fields for source address, destination address, data type and data content. The transmission method adopts a hybrid transmission method based on the combination of time division multiple access and code division multiple access to improve the efficiency and reliability of data transmission. In the time division multiple access part, time is divided into multiple time slots, and different nodes send data in different time slots. In the code division multiple access part, spread spectrum code is used to modulate data, and different nodes use different spread spectrum codes to distinguish different data sources. Error checking adopts a combination of cyclic redundancy check and hash check to ensure accurate transmission of information. The cyclic redundancy check is used to detect errors in data transmission and generate a check code. The receiver determines whether the data is correct based on the check code. The hash check is used to verify the integrity of the data, and a hash operation is performed on the data to generate a hash value. The receiver performs the same hash operation on the received data and compares whether the two hash values ​​are consistent.

7. The hierarchical coordination control system of an AC / DC hybrid microgrid group according to claim 1, characterized in that: The system also includes an energy management unit, including an energy prediction subunit, an optimization scheduling subunit and an energy storage management subunit. The energy prediction subunit uses an innovative prediction algorithm that combines deep learning with time series analysis to predict power output and load demand. The prediction model is P pred =f pred (t,ω1,ω2,…,ω n ), where t is the time variable, ω is the model parameter determined by training, and a deep neural network is used for training. The network structure includes an input layer, several hidden layers and an output layer. The input layer receives historical data, including but not limited to power output power, load demand and environmental factors over a period of time in the past. The hidden layer extracts the features of the data through nonlinear transformation, and the output layer outputs the predicted power value. The model parameters are adjusted through the back propagation algorithm to minimize the error between the predicted value and the actual value. The optimization dispatch subunit takes the minimum operation cost and the highest reliability of the microgrid as the objective function. The objective function is F obj =μC+νR, where C represents the operating cost, including the power generation cost and the equipment maintenance cost, R represents the reliability index, including the voltage stability and the power supply reliability, μ and ν are weight coefficients, and the optimal energy scheduling scheme is solved by the multi-objective optimization algorithm. The genetic algorithm is used for solving, and the individual represents an energy scheduling scheme, which is continuously evolved through crossover, mutation and selection operations to find the scheduling scheme that makes the objective function optimal; The energy storage management subunit controls the charging and discharging of the energy storage device according to the energy scheduling plan, and uses the energy storage control function E ctrl =g ctrl (S es ,ω es ,θ es ) controls the state of the energy storage device, where S es is a set of state parameters of the energy storage device, including but not limited to power and charge / discharge power, ω es is the energy storage control weight parameter, which is determined according to the type and performance of the energy storage device, θ es is the control parameter vector.

8. The hierarchical coordination control system of an AC / DC hybrid microgrid group according to claim 1, characterized in that: The control units of the system adopt a distributed control architecture and are distributed on different nodes of the microgrid. The control unit of each node has a certain degree of autonomous decision-making ability, and makes control decisions through local information and information interaction with adjacent nodes. The nodes exchange information and coordinate control through a high-speed communication network, and a distributed consistency algorithm is used to ensure that the control decisions of each node are coordinated and consistent. The distributed consensus algorithm iteratively updates the state variables of each node so that the state variables of all nodes gradually converge to a consistent value. The state variable update function is X i (t+1)=f cons ( X i (t),N i ,ω cons ), where X i( t ) represents the state variable of node i at time t, N i is the set of adjacent nodes of node i, ω cons is the consistency weight parameter, which is determined according to the connection strength and communication quality between nodes. The stronger the connection strength and the better the communication quality, the larger the weight parameter. Initially, the state variables of each node are set randomly, and then gradually tend to a consistent value through continuous iterative updates. When the difference between the state variables of all nodes is less than a preset threshold τ, consistency is considered to be achieved.

9. The hierarchical coordination control system of an AC / DC hybrid microgrid group according to claim 1, characterized in that: The system also includes a user interaction unit, including a user interface subunit and a demand response subunit; The user interface subunit provides a user-friendly operation interface, which graphically displays the operating status and control parameters of the microgrid; The demand response subunit adjusts the operation strategy of the microgrid according to the user's electricity demand and feedback. The demand response function is R dr =h dr (D user ,ω dr ,θ dr ), where D user Represents user demand parameters, including power consumption and power consumption time, ω dr is the demand response weight parameter, which is determined according to the user's electricity priority and urgency, θ dr is the demand response control parameter vector. When the user demand changes, the demand response subunit meets the user's electricity demand by adjusting the partition aggregation, flexible interconnection control and wide-area coordinated control strategy of the microgrid.

10. A method for using a hierarchical coordinated control system of an AC / DC hybrid microgrid group according to any one of claims 1 to 9, characterized in that: The specific steps of this method are as follows: Step 1: System startup and initialization Start each unit of the hierarchical coordination control system of the AC / DC hybrid microgrid group; Each subunit is initialized, including random initialization of particle positions in the particle swarm optimization algorithm and random setting of node state variables in the distributed consensus algorithm; Step 2: Data collection and feature extraction The scene feature acquisition subunit collects multi-dimensional feature parameters of power supply, load, environment and microgrid structure in the substation in real time; Process the collected data and extract key features, including various parameters required to calculate the comprehensive characteristic values ​​of the nodes; Step 3: Dynamic Partition Aggregation The partition aggregation algorithm subunit uses the density peak-based clustering partition algorithm to perform dynamic partition aggregation based on the collected feature parameters; The weight coefficient is determined by the particle swarm optimization algorithm, the comprehensive characteristic value of the node is calculated, and the partition range is determined according to the density peak; Step 4: Flexible interconnection control The flexible interconnection control unit determines the interaction weight parameter ω through the fuzzy logic algorithm according to the current source-load status and partition situation; According to the interactive control function P out =f(P in ,ω,θ) control the low voltage flexible interconnection device to perform power interaction; The mode switching control subunit is based on the mode switching criterion function M switch =g(P in ,P out ,ω,η) determine whether to switch the interaction mode; Step 5: Wide-area coordinated control The co-evolution computing subunit of the wide-area coordination control unit uses a distributed co-evolution algorithm to calculate the optimal strategy of each participant; The multi-level coordination subunits establish a hierarchical coordination function H coord =h(S l1 ,S l2 ,…,S ln ) Coordinate with stakeholders at different levels; The information interaction subunit ensures real-time information interaction and sharing among all participants; Step 6: Monitoring and Management The data acquisition subunit of the monitoring and management unit continuously collects microgrid operation data; The condition monitoring subunit uses the condition evaluation function Evaluate the microgrid operating status; The control decision subunit formulates corresponding control decisions according to the operating status and transmits them to each control unit for execution; Step 7: Energy Management The energy prediction subunit uses a prediction algorithm that combines deep learning with time series analysis to predict power output and load demand. The prediction model is P pred =f pred (t,ω1,ω2,…,ω n ); The optimization dispatch subunit takes the minimization of the operation cost and the maximization of the reliability of the microgrid as the objective function F obj =μC+νR, and the optimal energy scheduling solution is solved through a multi-objective optimization algorithm; The energy storage management subunit controls the energy storage control function E according to the energy scheduling plan. ctrl =g ctrl (S es ,ω es ,θ es ) Control the charge and discharge status of energy storage equipment; Step 8: User Interaction The user interface subunit provides users with a friendly operation interface, which graphically displays the operating status and control parameters of the microgrid; The demand response subunit adjusts the operation strategy of the microgrid according to the user's electricity demand and feedback. When the user's demand changes, the user's electricity demand is met by adjusting the strategies of partition aggregation, flexible interconnection control and wide-area coordinated control; Step 9: Dynamic Adjustment and Coordination When parameters in the system change, including changes in source-load status and user demand, the dynamic partition aggregation unit, flexible interconnection control unit, and wide-area coordination control unit are dynamically adjusted according to the changes; Each unit exchanges information and coordinates control through a high-speed communication network to ensure stable and efficient operation of the system; Step 10: Continue to run and optimize The system operates continuously, constantly monitoring and analyzing the operating status and data of the microgrid; According to the operating conditions and new data, the algorithms and parameters in the system are optimized and adjusted to improve the performance and adaptability of the system.

Images