Adaptive topological structure optimization method, system and equipment for power distribution network and medium

Abstract

The invention discloses a power distribution network adaptive topological structure optimization method, system and device and a medium, and the method comprises the steps: obtaining related parameters of a power distribution network system, dividing the power distribution network system into different layers and different regions based on the related parameters of the power distribution network system, and constructing a layering and partitioning model; coordinating and solving the layering and partitioning model through an intelligent algorithm, and obtaining control variables of different regions; constructing an objective function based on the control variable, and solving the objective function by adopting an improved particle swarm optimization method to obtain a topological structure; and optimizing the topological structure by adopting a double-layer optimization strategy to obtain an optimization result, evaluating the optimization result, and feeding back the evaluation result to the target function. According to the method, the randomness and volatility challenges caused by multi-source resource access can be effectively dealt with, and the limitation of the traditional separation design is broken through the collaborative optimization of topology and control, so that the operation efficiency, the stability and the anti-interference capability of the power distribution network are improved, and the collaborative evolution of a topological structure and a control strategy is realized.

Description

Technical Field

[0001] This invention relates to the field of distribution network topology technology, and in particular to a method, system, device and medium for adaptive topology optimization of distribution networks. Background Technology

[0002] With the large-scale integration of distributed energy sources, electric vehicles, energy storage systems, and other multi-source resources into distribution networks, the operating environment of traditional distribution networks has undergone significant changes. These multi-source resources exhibit characteristics such as randomness and volatility, making the operating state of distribution networks more complex and variable, and placing higher demands on the optimization of distribution network topology. Traditional distribution network topology optimization methods are no longer adequate to adapt to these changes, and a new method is urgently needed to improve the adaptability of distribution networks to multi-source resources and achieve efficient and stable operation.

[0003] However, existing technologies have significant shortcomings in addressing the challenges posed by the access of multiple resources. Current distribution network topology optimization methods are primarily based on static network reconfiguration, which cannot effectively handle the randomness and volatility of multiple resources. Traditional network reconfiguration methods are typically based on deterministic load models, making it difficult to adapt to the fluctuating characteristics of distributed resources, resulting in optimization results that fail to achieve the expected effects in actual operation. Furthermore, existing topology optimization and control strategies are often independent, lacking a system-level collaborative optimization framework, and failing to fully leverage the system's resilience potential. This makes it difficult for distribution networks to achieve coordinated optimization of topology and control strategies when facing the access of multiple resources, impacting the operational efficiency and stability of the distribution network. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an adaptive topology optimization method for distribution networks to solve the problems that traditional distribution network topology optimization methods cannot adapt to the randomness and volatility of multi-source resources and the lack of coordination between topology and control strategies.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for adaptive topology optimization of a distribution network, comprising:

[0008] Obtain relevant parameters of the power distribution network system, and based on these parameters, divide the power distribution network system into different levels and regions to construct a hierarchical partitioning model;

[0009] The hierarchical partitioning model is solved in a coordinated manner using intelligent algorithms to obtain control variables for different regions.

[0010] Based on the control variables, an objective function is constructed, and the objective function is solved using an improved particle swarm optimization method to obtain the topological structure.

[0011] A two-layer optimization strategy is used to optimize the topology, the optimization results are obtained, the optimization results are evaluated, and the evaluation results are fed back to the objective function.

[0012] As a preferred embodiment of the adaptive topology optimization method for distribution networks described in this invention, the distribution network system is divided into different levels and regions, and a hierarchical partitioning model is constructed, including:

[0013] Based on relevant parameters of the distribution network system, the distribution network system is divided into physical layer, control layer and decision layer;

[0014] Each floor is divided into different zones based on geographical distribution, electrical characteristics, and control requirements.

[0015] The beneficial effects of this preferred technical solution are as follows: dividing the power distribution network into physical, control, and decision layers, and further dividing the region according to geographical distribution, can clearly define the hierarchical relationship between physical equipment, control signals, and decision logic through a three-layer architecture, transforming the system from a traditional single physical network into a multi-dimensional collaborative framework; the regional division, combined with parameters such as electrical distance and control radius, can realize a control mode of local autonomy plus global coordination, which reduces communication overhead and computational load within the region, and ensures the effective transmission of global optimization instructions through inter-layer mapping matrices, providing a structured foundation for hierarchical management of multi-source resources and regional adaptive control, and solving the problem of low optimization efficiency caused by fuzzy system architecture and unreasonable regional division in traditional methods.

[0016] As a preferred embodiment of the adaptive topology optimization method for distribution networks described in this invention, the method involves: coordinating the solution of the hierarchical partitioning model using intelligent algorithms to obtain control variables for different regions, including:

[0017] Based on a hierarchical and partitioned coordination mechanism of a multi-agent system, each region is equipped with a region controller for self-optimization control of resources within the region; each layer is equipped with a hierarchical coordinator for coordination between different regions within the same layer; and a global coordinator is set up for coordination decision-making between different layers.

[0018] The distributed optimization problem in the hierarchical partitioning coordination mechanism is solved by the alternating direction multiplier method. The coordination of the regional controller, hierarchical coordinator and global coordinator is iteratively coordinated until the constraint deviation between regions is less than a preset threshold. The control variables output by each regional controller are used to form the control variable set of different regions.

[0019] The beneficial effects of this preferred technical solution are as follows: by constructing a collaborative architecture of regional controller, hierarchical coordinator and global coordinator through a multi-agent system, and combining it with the alternating direction multiplier method to achieve distributed optimization, the complex global optimization problem can be decomposed into parallel solution tasks for each region, thereby reducing computational complexity.

[0020] As a preferred embodiment of the adaptive topology optimization method for distribution networks described in this invention, the method includes: constructing an objective function based on the control variables, solving the objective function using an improved particle swarm optimization method, and obtaining the topology, including:

[0021] Based on the control variables, the distribution network topology state space and topology transformation operation set are defined, and the load status, distributed generation status and energy storage status are monitored in real time through the distribution network system state perception function.

[0022] Based on the state monitored by the state-aware function, an objective function is constructed, and the objective function is solved using an improved particle swarm optimization method. The weights of the objective function are dynamically adjusted according to the state of the distribution network system to obtain a topology structure that adapts to fluctuations in multi-source resources.

[0023] As a preferred embodiment of the adaptive topology optimization method for distribution networks described in this invention, the method employs a two-layer optimization strategy to optimize the topology and obtain optimization results, including:

[0024] Obtain the state variables, topology variables, and control variables of the power distribution system, and construct the dynamic equations of the power distribution network system;

[0025] Based on the dynamic equations of the distribution network system, a two-layer optimization strategy is adopted for solving the problem. The outer layer uses an improved genetic algorithm to determine the optimal topology, while the inner layer uses a model predictive control method to solve for the optimal control sequence under the optimal topology. The topology and control strategy are synergistically optimized to obtain the optimization result that includes the optimal topology and the corresponding control strategy.

[0026] The beneficial effects of this preferred technical solution are as follows: Based on the dynamic equations of the distribution network system, a two-layer optimization strategy of outer-layer genetic algorithm and inner-layer model predictive control is adopted, which can break the limitations of the traditional design of separate topology optimization and control strategy: the outer-layer genetic algorithm takes the topology structure as the optimization object and selects the optimal topology through chromosome encoding and fitness function; the inner layer solves the continuous control variables such as energy storage charging and discharging power under a fixed topology and handles real-time fluctuations through rolling time domain optimization.

[0027] As a preferred embodiment of the adaptive topology optimization method for distribution networks described in this invention, the method includes: evaluating the optimization results and feeding the evaluation results back to the objective function, comprising:

[0028] Establish optimization indicators, which include topological characteristics, functional characteristics, and control characteristics;

[0029] The optimization results are quantitatively evaluated based on the optimization indicators, and a comprehensive resilience index is calculated.

[0030] A resilience enhancement strategy is formulated based on the comprehensive resilience index, and the comprehensive resilience index is fed back to the weight adjustment mechanism of the objective function.

[0031] As a preferred embodiment of the adaptive topology optimization method for distribution networks described in this invention, the objective function is expressed as:

[0032]

[0033] Where f1 represents network loss, f2 represents voltage deviation, f3 represents load balancing, and w1, w2, and w3 are weighting coefficients, with w1 + w2 + w3 = 1, a k T represents a topology transformation operation. i Let S represent a feasible topology of the distribution network, S represent the distribution network system state awareness function, and A represent the topology transformation operation set.

[0034] In a second aspect, the present invention provides a distribution network adaptive topology optimization system, comprising: a data acquisition module for acquiring relevant parameters of the distribution network system;

[0035] The model building module is used to divide the power distribution network system into different levels and regions based on the relevant parameters of the power distribution network system, and build a hierarchical and partitioned model.

[0036] The model solving module is used to solve the hierarchical partitioning model in a coordinated manner through intelligent algorithms to obtain control variables for different regions.

[0037] The topology acquisition module is used to construct an objective function based on the control variables, solve the objective function using an improved particle swarm optimization method, and obtain the topology structure.

[0038] The comprehensive optimization module is used to optimize the topology using a two-layer optimization strategy, obtain the optimization results, evaluate the optimization results, and feed the evaluation results back to the objective function.

[0039] Thirdly, the present invention provides an electronic device, comprising:

[0040] Memory and processor;

[0041] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the adaptive topology optimization method for the power distribution network.

[0042] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the power distribution network adaptive topology optimization method.

[0043] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention can effectively address the challenges of randomness and volatility brought about by the access of multiple resources. It achieves a balance between local autonomy and global coordination through hierarchical and partitioned modeling, improves the optimization efficiency between regions and layers by utilizing a multi-source collaborative coordination mechanism, enhances the system's adaptability to resource fluctuations by using an adaptive topology optimization algorithm, breaks through the limitations of traditional separate design by co-optimizing topology and control, and forms an optimization closed loop by combining system resilience assessment. This improves the operating efficiency, stability, and anti-interference capability of the distribution network, realizes the co-evolution of topology structure and control strategy, and better adapts to the complex scenario of high proportion of distributed energy access. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of the overall process of the adaptive topology optimization method for power distribution networks according to an embodiment of the present invention. Detailed Implementation

[0046] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0047] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for adaptive topology optimization of a distribution network is provided, comprising:

[0048] S100: Obtain relevant parameters of the distribution network system, and based on these parameters, divide the distribution network system into different levels and regions to construct a hierarchical and partitioned model;

[0049] S101: Use intelligent algorithms to coordinate the solution of hierarchical partitioning models and obtain control variables for different regions;

[0050] S102: Based on control variables, construct the objective function, solve the objective function using an improved particle swarm optimization method, and obtain the topological structure;

[0051] S103: A two-layer optimization strategy is adopted to optimize the topology, obtain the optimization results, evaluate the optimization results, and feed the evaluation results back to the objective function.

[0052] It should be noted that by acquiring distribution network system parameters to construct a hierarchical and partitioned model, the system can be divided into a physical layer, a control layer, a decision layer, and multiple regions based on electrical characteristics and control requirements. This provides a clear architecture for hierarchical management and regional autonomous control of multi-source resources. The model is solved using intelligent algorithms, leveraging a multi-agent mechanism involving regional controllers, hierarchical coordinators, and a global coordinator. This enables resource self-optimization control within each region and coordinates constraints between regions and layers, thereby obtaining control variables for each region to guide topology adjustments. An objective function containing indicators such as network loss and voltage deviation is constructed based on the control variables and solved using an improved particle swarm optimization method. This allows for dynamic optimization of the topology structure according to the system state, enhancing adaptability to distributed resource fluctuations. A two-layer optimization strategy further optimizes the topology structure. The outer layer uses a genetic algorithm to determine the topology, while the inner layer uses model predictive control to solve the control strategy. The optimization results are evaluated and fed back to the objective function, achieving collaborative optimization of topology and control strategies. This invention effectively solves the problems of traditional distribution network topology optimization methods being unable to adapt to the random fluctuations of multi-source resources and lacking coordination between independent topology and control design, thus improving the operating efficiency of the distribution network and its adaptability to complex scenarios.

[0053] Example 2, refer to Figure 1 As an embodiment of the present invention, based on the above embodiment, a method for adaptive topology optimization of power distribution networks is provided.

[0054] In this embodiment of the invention, step S100 involves obtaining relevant parameters of the distribution network system, dividing the distribution network system into different levels and regions based on these parameters, and constructing a hierarchical partitioning model. It also includes sub-steps A1-A2:

[0055] A1: Based on relevant parameters of the distribution network system, the distribution network system is divided into physical layer, control layer and decision layer;

[0056] A2: Each layer is divided into different areas based on geographical distribution, electrical characteristics, and control requirements.

[0057] In this embodiment of the invention, a hierarchical and partitioned model of the power distribution network is first constructed based on complex network theory. The power distribution network system is divided into three layers: the physical layer, the control layer, and the decision layer. Each layer is further divided into multiple functional areas according to geographical distribution, electrical characteristics, and control requirements.

[0058] For example, the physical layer model can be represented as a weighted directed graph:

[0059] G p =V p E p W p

[0060] Among them, V p =v1,v2,...,v n This represents a set of physical nodes, including substations, distribution transformers, switching equipment, load nodes, and distributed resource access points; E p =e ij |v i ,v j ∈V p Indicates physical connection relationship; W p =w ij |e ij ∈E p The edge weights represent the impedance characteristics of the line.

[0061] For example, the control layer model is represented as follows:

[0062] G c =V c E c W c

[0063] Among them, V c =c1,c2,...,c m This represents a set of control nodes, including various controllers and intelligent terminals; E c Indicates the control signal transmission path; W c This indicates the control weight, reflecting the control priority and degree of influence.

[0064] For example, the decision-making layer model is represented as follows:

[0065] G d =V d E d W d

[0066] Among them, V d =d1,d2,...,d k E represents the set of decision nodes, including optimization decision units at all levels; d Indicates the path of decision-making information transmission; W d It indicates the decision weight, reflecting the scope and priority of the decision's impact.

[0067] Inter-layer relationships are mapped through matrix M pc and M cd express:

[0068] M pc =[m ij ]n×m

[0069] Mcd = [m ij ]m×k

[0070] Where, m ij This represents the degree of association between node i and node j. When nodes are fully associated, m represents the degree of association between them. ij =1, m when there is no correlation ij =0, 0 < m when partially correlated ij <1.

[0071] In this embodiment of the invention, the power distribution network partitioning can be represented as follows, based on electrical characteristics and control requirements:

[0072] Z = Z1, Z2, ..., Z r

[0073] Among them, each region Z i It includes a set of physical nodes, control nodes, and decision nodes:

[0074]

[0075] The zoning principle takes into account the following factors: electrical distance, control radius, communication delay, resource distribution, and load characteristics.

[0076] The region partitioning employs an improved spectral clustering algorithm, defining the electrical distance matrix D = [d ij ]n×n, where:

[0077]

[0078] Among them, R ij and X ij These are the equivalent resistance and reactance between node i and node j, respectively.

[0079] Based on the electrical distance matrix, construct a similarity matrix S = [s ij n×n:

[0080]

[0081] Where σ is the kernel parameter, determined through cross-validation.

[0082] The Laplacian matrix L = DS is calculated using the spectral clustering algorithm, where D is the degree matrix and the diagonal elements are...

[0083] Solve the characteristic equation: Lv = λv, and select the eigenvectors corresponding to the r smallest non-zero eigenvalues ​​to form the characteristic matrix. The K-means algorithm is used to cluster the row vectors of the feature matrix to obtain the final partitioning result.

[0084] In one alternative implementation, a hierarchical division can be performed using a function-based positioning method. Specifically, based on the functional positioning of nodes in the distribution network, nodes such as substations and distribution transformers with energy conversion and transmission functions are divided into the physical layer, various controllers and intelligent terminals that realize real-time control functions are divided into the control layer, and decision-making units at all levels that undertake system optimization decision-making functions are divided into the decision-making layer, so as to clarify the functional positioning and boundaries of different levels.

[0085] In one alternative implementation, a fuzzy C-means clustering method can be used for region division. Specifically, by comprehensively considering parameters such as electrical distance, load characteristics, and resource distribution, a multi-dimensional feature vector is constructed. The membership degree of each node to different regions is calculated using the fuzzy C-means clustering algorithm. Based on the membership degree, the region affiliation of the node is determined, thereby realizing the region division of each layer of the distribution network and improving the flexibility and adaptability of the partitioning.

[0086] In this embodiment of the invention, step S101 involves using an intelligent algorithm to coordinate the solution of a hierarchical partitioning model and obtain control variables for different regions; it also includes sub-steps B1-B2:

[0087] B1: A hierarchical and partitioned coordination mechanism based on a multi-agent system, which equips each region with a region controller for self-optimization control of resources within the region; each layer is equipped with a hierarchical coordinator for coordination between different regions within the same layer; and a global coordinator is set up for coordination decision-making between different layers.

[0088] B2: Solve the distributed optimization problem in the hierarchical partition coordination mechanism by using the alternating direction multiplier method. Iterate through coordination, regional controllers, hierarchical coordinators and global coordinators until the constraint deviation between regions is less than a preset threshold. Obtain the control variables output by each regional controller to form the control variable set of different regions.

[0089] In this embodiment of the invention, to achieve coordinated operation of each layer and region, a hierarchical partitioning coordination mechanism based on a multi-agent system is designed. Each region is equipped with a Region Controller (RC), responsible for the self-optimization control of resources within the region; each layer is equipped with a Layer Coordinator (LC), responsible for coordination between different regions within the same layer; and a Global Coordinator (GC) is set up to be responsible for coordination decisions between different layers.

[0090] For example, the area controller RC i Decision model:

[0091]

[0092] stg i (x i ,u i )≤0

[0093] h i (x i ,u i ) = 0

[0094] x i ∈X i ,u i ∈U i

[0095] Where, x i Let u be the state variable of region i. i To control the variable, f i Let g be the objective function. i and h i These are inequality and equality constraints, respectively, X i and U i These are the feasible regions for the state and control variables, respectively.

[0096] For example, the hierarchical coordinator LC j Coordination model:

[0097]

[0098] stc j (x,u)=0(16)

[0099] Where Ij is the region index set of the j-th layer, λ j For Lagrange multipliers, c j (x,u) is a coupling constraint, representing the mutual influence between different regions.

[0100] An example of the decision-making model of the Global Coordinator (GC):

[0101]

[0102] std(x,u)=0(18)

[0103] Where J is the total number of levels, μ is the global coordination multiplier, and d(x,u) is the global coupling constraint.

[0104] In this embodiment of the invention, to improve coordination efficiency, a distributed optimization algorithm based on the Alternating Direction Multiplier Method (ADMM) is adopted:

[0105] Step 1: RC controllers for each area iParallel solution of local optimization problems;

[0106] Step 2: Hierarchical Coordinator (LC) j Update the Lagrange multiplier λ based on the solution of each region controller j ;

[0107] Step 3: The Global Coordinator (GC) updates the global coordinator multiplier μ;

[0108] Step 4: Feed back the updated multipliers to each region controller, and repeat steps 1-3 until convergence.

[0109] In one alternative implementation, an improved ADMM method incorporating an adaptive relaxation factor can enhance coordination efficiency. This involves introducing a relaxation factor that dynamically adjusts with the number of iterations into the multiplier update step of the traditional ADMM algorithm. When constraint deviations are large in the early stages of iteration, the relaxation factor is increased to accelerate multiplier updates; as iterations progress, the relaxation factor is gradually decreased to ensure convergence accuracy. Simultaneously, during the parallel solution phase for the region controller, parallel computing acceleration techniques are employed, such as utilizing GPU heterogeneous computing or distributed memory computing frameworks, to distribute the local optimization problems of each region across multiple computing nodes for parallel solution.

[0110] In another optional implementation, a historical data-driven coordination optimization method can be employed. A historical coordination database is constructed to store the optimization results of each regional controller under different operating scenarios, the multiplier update trajectories of hierarchical coordinators, and the constraint handling schemes of the global coordinator. When the system enters a new operating scenario, a similarity matching algorithm retrieves historical coordination parameters from the database for similar scenarios, using these parameters as the initial values ​​for the current iteration, avoiding iteration from scratch. Combined with an online learning mechanism, real-time optimization data from the current scenario is used to correct the historical parameters.

[0111] The convergence criterion is:

[0112] |c j (x,u)|∞≤ε1,|d(x,u)|∞≤ε2 (19)

[0113] Wherein, ε1 and ε2 are preset thresholds.

[0114] In one alternative implementation, a distributed coordination method based on communication topology optimization can be chosen to improve coordination efficiency. Specifically, by analyzing the communication traffic and interaction frequency between each area controller, hierarchical coordinator, and global coordinator, a weighted communication topology graph is constructed. The minimum spanning tree algorithm or spectral clustering algorithm from graph theory is then used to optimize the communication links, reducing redundant communication paths and shortening the transmission delay of critical coordination information. Simultaneously, a dynamic bandwidth allocation mechanism is set for the communication links; when a large deviation in inter-area coupling constraints is detected, higher bandwidth is automatically allocated to the communication links between the relevant coordinators.

[0115] In another alternative implementation, a hierarchical coordination method based on model reduction can be chosen to improve coordination efficiency. Specifically, for the high-dimensional state space in the hierarchical and zonal model of the distribution network, Proper Orthogonal Decomposition (POD) or local linearization reduction methods are used to reduce the dimensionality of each regional model, thereby reducing the computational load of the regional controller in solving local optimization problems. In the hierarchical coordinator and the global coordinator, an error compensation mechanism between the reduced model and the original model is established, and the accuracy loss caused by the reduced model is dynamically corrected through iterative learning.

[0116] In this embodiment of the invention, step S102, which involves constructing an objective function based on control variables, solving the objective function using an improved particle swarm optimization method, and obtaining the topological structure, further includes sub-steps C1-C2:

[0117] C1: Based on control variables, define the distribution network topology state space and topology transformation operation set, and monitor the load status, distributed generation status and energy storage status in real time through the distribution network system state awareness function;

[0118] C2: Based on the state monitored by the state-aware function, an objective function is constructed, and an improved particle swarm optimization method is used to solve the objective function. The weights of the objective function are dynamically adjusted according to the state of the distribution network system to obtain a topology that adapts to the fluctuations of multi-source resources.

[0119] In this embodiment of the invention, the distribution network topology state space is defined as follows:

[0120] T = T1, T2, ..., T N

[0121] Among them, T i This represents a feasible topology of the distribution network, where N is the total number of feasible topologies.

[0122] Topology transformation operation set:

[0123] A = a1, a2, ..., a K

[0124] Among them, a k This represents a topology transformation operation, such as switch switching, line connection or disconnection, etc., where K is the total number of operations.

[0125] Topology transformation function:

[0126] T i+1 =τ(T) i ,a k ,θ)

[0127] Where τ is the transformation function and θ are system parameters, including load status and distributed resource output. To adapt to the fluctuating characteristics of distributed resources, a system state awareness function is defined:

[0128] S=φ(L,G,E,Ω)

[0129] Where L = L1, L2, ..., L n Let G represent the set of load states, G = G1, G2, ..., G m Let E represent the set of distributed generation states, E = E1, E2, ..., E p Let Ω represent the set of energy storage states, and let Ω represent the set of network operation constraints.

[0130] In this embodiment of the invention, an adaptive topology optimization objective function is designed based on state awareness:

[0131]

[0132] Where f1 represents network loss, f2 represents voltage deviation, f3 represents load balancing degree, w1, w2, and w3 are weighting coefficients, and w1 + w2 + w3 = 1.

[0133] Network loss calculation:

[0134]

[0135] Among them, R mn Let P be the resistance of the line (m,n). mn and Q mn These are the active and reactive power of the line, respectively.

[0136] V m This represents the node voltage.

[0137] Voltage deviation calculation:

[0138]

[0139] Among them, V j Let V be the voltage at node j. ref This is the reference voltage.

[0140] Load balancing calculation:

[0141]

[0142] Among them, S j For the load of feeder j, b represents the maximum allowable load, and b represents the total number of feeders.

[0143] The weighting coefficients employ an adaptive adjustment strategy:

[0144] w1(t+1)=w1(t)+Δw1(t)

[0145] w2(t+1)=w2(t)+Δw2(t)

[0146] w3(t+1) = 1 - w1(t+1) - w2(t+1)

[0147] in,

[0148]

[0149] In the formula, α1 and α2 are the learning rates, δ1(t) and δ2(t) are the adaptive factors, and J is the system performance evaluation index.

[0150] The adaptive factor is dynamically adjusted according to the system state:

[0151]

[0152] Where, β i η is the sensitivity coefficient. i (t) represents the system state change index.

[0153] To improve algorithm efficiency, an improved particle swarm optimization method is used to solve the topology optimization problem. Position update of particle i:

[0154] x i (t+1)=x i (t)+v i (t+1)

[0155] Speed ​​updates:

[0156] v i (t+1)=w·v i (t)+c1r1[p i (t)-x i (t)]+c2r2[p g (t)-x i (t)]

[0157] Where w is the inertia weight, c1 and c2 are acceleration constants, r1 and r2 are random numbers between [0,1], and p i (t) represents the optimal position of particle i, p g (t) represents the globally optimal position.

[0158] In one alternative implementation, an improved genetic algorithm can be used to solve the objective function. Specifically, the algorithm performance is optimized by designing adaptive crossover and mutation probabilities. The crossover probability is dynamically adjusted based on the individual's fitness value. For individuals with fitness above average, a smaller crossover probability is used to retain superior genes; for individuals with low fitness, a larger crossover probability is used to enhance search capability. The mutation probability is related to the number of iterations and gradually decreases as iterations proceed, thereby balancing global search and local optimization capabilities. This allows the genetic algorithm to search the distribution network topology state space more efficiently and avoid getting trapped in local optima.

[0159] In another alternative implementation, simulated annealing can be used to solve the objective function. Specifically, a dynamic temperature update strategy and an adaptive neighborhood search mechanism are introduced. Temperature updates employ a non-linear decay method, cooling down rapidly in the early stages of iteration to expand the search range, and then slowly cooling down in later stages for finer searching. The neighborhood search range is dynamically adjusted based on the quality of the current solution. When the fitness of a solution is better than the historical average, the neighborhood search range is reduced to improve local optimization accuracy; conversely, it is expanded to enhance global exploration capabilities. Through the coordinated adaptive adjustment of temperature and neighborhood, the convergence speed and global optimization capability of the algorithm in distribution network topology optimization are improved.

[0160] In this embodiment of the invention, step S103 employs a two-layer optimization strategy to optimize the topology, obtain optimization results, evaluate the optimization results, and feed the evaluation results back to the objective function. It also includes sub-steps D1-D4:

[0161] D1: Obtain the state variables, topology variables, and control variables of the power distribution system, and construct the dynamic equations of the power distribution network system;

[0162] D2: Based on the dynamic equations of the distribution network system, a two-layer optimization strategy is adopted for solving the problem. The outer layer uses an improved genetic algorithm to determine the optimal topology, while the inner layer uses a model predictive control method to solve for the optimal control sequence under the optimal topology. The topology and control strategy are optimized in a coordinated manner to obtain the optimization result that includes the optimal topology and the corresponding control strategy.

[0163] D3: Establish optimization indicators, including topological characteristics, functional characteristics, and control characteristics; quantify the optimization results based on the optimization indicators, and calculate the comprehensive resilience index.

[0164] D4: A resilience enhancement strategy is formulated based on the comprehensive resilience index, and a weight adjustment mechanism is used to feed the comprehensive resilience index back to the objective function.

[0165] In this invention, traditional methods often separate topology optimization from control strategy design. This invention proposes a collaborative optimization method for topology and control strategy, establishing a unified optimization framework. The system state variable x(t), topology variable y(t), and control variable u(t) are defined, and the system dynamic equation is:

[0166]

[0167] The goal of collaborative optimization is:

[0168]

[0169] g(x(t),y(t),u(t),t)≤0

[0170] h(x(t),y(t),u(t),t)=0

[0171] y(t)∈Y, u(t)∈U

[0172] Where L is the instantaneous cost function, g and h are the inequality and equality constraints, respectively, and Y and U are the feasible regions of the topological and control variables, respectively.

[0173] Since the topological variable y(t) is typically a discrete variable, while the control variable u(t) is a continuous variable, this problem belongs to mixed-integer nonlinear programming. This invention employs a two-level optimization strategy:

[0174] Outer layer optimization:

[0175] min y(t) J outer (y(t))=J (y( t))

[0176] Inner layer optimization:

[0177] J (y( t))=min u(t) J inner (y(t),u(t))

[0178] st

[0179] The outer layer optimization uses an improved genetic algorithm, while the inner layer optimization uses the model predictive control (MPC) method.

[0180] For example, the chromosome encoding in a genetic algorithm represents the topological structure, and the fitness function is:

[0181]

[0182] Here, ε is a small positive number to prevent the denominator from being zero.

[0183] The MPC controller, under a fixed topology y(t), solves for the optimal control sequence through rolling time-domain optimization:

[0184]

[0185] stx(t+k+1|t)=f(x(t+k|t),y(t),u(t+k|t),t+k)

[0186] g(x(t+k|t),y(t),u(t+k|t),t+k)≤0

[0187] h(x(t+k|t),y(t),u(t+k|t),t+k)=0

[0188] u(t+k|t)∈U

[0189] Where, N p To predict the time domain, x ref For the reference state, Q and R are the weight matrices for the state error and control cost, respectively.

[0190] To improve computational efficiency, a distributed computing framework is adopted, which distributes optimization tasks to regional controllers and achieves result coordination and convergence through hierarchical coordinators and a global coordinator.

[0191] In this embodiment of the invention, a method for assessing the resilience of distribution networks based on complex network theory is proposed, which quantifies system resilience from three dimensions: topological characteristics, functional characteristics, and control characteristics.

[0192] For example, topology resilience metrics:

[0193] R T =α1·R con +α2·R clu +α3·R het (50)

[0194] Among them, R con R is a network connectivity metric. clu R is the clustering coefficient. het α1, α2, and α3 are the heterogeneity indicators, and α1, α2, and α3 are the weighting coefficients.

[0195] Connectivity index calculation:

[0196]

[0197] Where, d ij Let be the shortest path length between node i and node j.

[0198] Clustering coefficient calculation:

[0199]

[0200] Among them, E i Let k be the number of edges between the neighbors of node i. i Let be the degree of node i.

[0201] Heterogeneity index calculation:

[0202]

[0203] in, <k>and <k 2 > These are the first and second moments of the nodal degree, respectively.

[0204] Functional resilience indicators:

[0205] R F =β1·R load +β2·R volt +β3·R loss

[0206] Among them, R load For load supply capacity, R volt For voltage recovery capability, R loss For loss control capability, β1, β2, and β3 are weighting coefficients.

[0207] Load supply capacity calculation:

[0208]

[0209] Where, n L P represents the number of load nodes. i sup and P i dem These represent the actual power supply and the required power supply for node i, respectively.

[0210] Voltage recovery capability calculation:

[0211]

[0212] Among them, V i and V ref These are the actual voltage and reference voltage of node i, respectively.

[0213] Loss control capability calculation:

[0214]

[0215] Among them, P loss P represents the total network loss. total This represents the total transmission power.

[0216] Controlling resilience indicators:

[0217] R C =γ1·R ada +γ2·R coor +γ3·R resp

[0218] Among them, R ada To control adaptive capability, R coor For coordination ability, R resp γ1, γ2, and γ3 are weighting coefficients for the response speed.

[0219] Calculation of adaptive control capability:

[0220]

[0221] Where u(t) is the actual control input, u opt (t) represents the ideal optimal control input.

[0222] Coordination ability calculation:

[0223]

[0224] Where, n Z J represents the number of regions. i For the local optimization objective of region i, J global The goal is to optimize the overall system.

[0225] Response speed calculation:

[0226]

[0227] Among them, T ref For reference response time, T act This represents the actual response time.

[0228] Overall resilience index:

[0229] R = w T ·R T +w F ·R F +w C ·R C

[0230] Among them, w T w F w C Here are the weighting coefficients, and w T +w F +w C =1.

[0231] Based on the resilience assessment results, this invention proposes a resilience enhancement strategy: improving the system's ability to withstand failures by increasing redundant paths and optimizing the network structure; improving the system's ability to maintain functions by introducing flexible resources and optimizing resource allocation; and improving the system's response and self-recovery capabilities by adopting improved control algorithms and optimized coordination mechanisms.

[0232] Example 3, referring to Tables 1-3, is an embodiment of the present invention. Based on the above embodiments, an adaptive topology optimization method for distribution networks is provided. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0233] The algorithm was simulated and verified using MATLAB R2024a, with an improved IEEE 123-node distribution system as an example. This system includes 25 distributed photovoltaic (PV) power generation points, 10 energy storage systems, and 15 electric vehicle (EV) charging stations. The simulation considered typical daily load variations, PV power generation fluctuations, and EV charging characteristics.

[0234] Experimental results show that, compared with traditional methods, the method of this invention significantly improves system performance and resilience:

[0235] Table 1. Performance comparison of different methods on the IEEE 123-node system.

[0236]

[0237] Table 2 Comparison of resilience indices of different methods in fault recovery scenarios

[0238]

[0239]

[0240] Table 3 Comparison of system performance of different hierarchical partitioning strategies

[0241]

[0242] Experimental results show that the proposed hierarchical partitioning multi-source collaborative adaptive topology optimization method outperforms traditional methods in all performance indicators. Network loss reduction reaches 18.7%, voltage deviation improvement reaches 23.5%, load balancing improvement reaches 17.9%, and the computation time is only 32.6 seconds, significantly lower than centralized optimization methods. In fault recovery scenarios, the comprehensive resilience index of this method reaches 0.837, and the system recovery time is only 8.6 minutes, nearly twice as fast as traditional methods. Through the optimized hierarchical partitioning strategy, the system coordination convergence time is significantly shortened, communication overhead is reduced, and overall efficiency is improved by 16.8%.

[0243] By analyzing system performance under different scenarios, it was found that the method of this invention performs particularly well in dealing with scenarios with high volatility of distributed resources. When the photovoltaic output fluctuation coefficient exceeds 30%, the voltage deviation of traditional methods increases sharply, while this method effectively maintains system voltage stability through adaptive topology adjustment and multi-source coordinated control. In addition, in terms of multi-source resource coordination, this method achieves optimized coordination of photovoltaic power generation, energy storage system and controllable load, effectively smoothing power fluctuations and improving the economy and reliability of system operation.

[0244] Simulation results fully demonstrate the significant advantages of the method of the present invention in improving the operating efficiency of distribution networks, enhancing system resilience, and coping with distributed resource fluctuations, and it has good theoretical value and application prospects.

[0245] Example 4 illustrates a schematic scheme for a distribution network adaptive topology optimization method. It should be noted that the technical solution of this distribution network adaptive topology optimization system belongs to the same concept as the technical solution of the aforementioned distribution network adaptive topology optimization method. Details not described in detail in this example can be found in the description of the technical solution of the aforementioned distribution network adaptive topology optimization method.

[0246] This embodiment also provides a distribution network adaptive topology optimization system, including:

[0247] The data acquisition module is used to acquire relevant parameters of the power distribution network system;

[0248] The model building module is used to divide the distribution network system into different levels and regions based on relevant parameters of the distribution network system, and build a hierarchical and partitioned model.

[0249] The model solving module is used to coordinate the solution of the hierarchical and partitioned model through intelligent algorithms to obtain the control variables of different regions.

[0250] The topology acquisition module is used to construct an objective function based on control variables, solve the objective function using an improved particle swarm optimization method, and obtain the topology structure.

[0251] The comprehensive optimization module is used to optimize the topology using a two-layer optimization strategy, obtain the optimization results, evaluate the optimization results, and feed the evaluation results back to the objective function.

[0252] This embodiment also provides an electronic device suitable for adaptive topology optimization of distribution networks, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the adaptive topology optimization method for distribution networks proposed in the above embodiments.

[0253] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the method for implementing adaptive topology optimization of the distribution network as proposed in the above embodiments.

[0254] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for implementing adaptive topology optimization of distribution networks proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0255] Based on the above description of the implementation methods, those skilled in the art will clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0256] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.< / k>

Claims

1. A method for adaptive topology optimization of a distribution network, characterized in that, include: Obtain relevant parameters of the power distribution network system, and based on these parameters, divide the power distribution network system into different levels and regions to construct a hierarchical partitioning model; The hierarchical partitioning model is solved in a coordinated manner using intelligent algorithms to obtain control variables for different regions. Based on the control variables, an objective function is constructed, and the objective function is solved using an improved particle swarm optimization method to obtain the topological structure. A two-layer optimization strategy is used to optimize the topology, the optimization results are obtained, the optimization results are evaluated, and the evaluation results are fed back to the objective function.

2. The adaptive topology optimization method for distribution networks as described in claim 1, characterized in that, The power distribution network system is divided into different levels and regions, and a hierarchical zoning model is constructed, including: Based on relevant parameters of the distribution network system, the distribution network system is divided into physical layer, control layer and decision layer; Each floor is divided into different zones based on geographical distribution, electrical characteristics, and control requirements.

3. The adaptive topology optimization method for distribution networks as described in claim 2, characterized in that, The hierarchical partitioning model is solved using intelligent algorithms to obtain control variables for different regions, including: Based on a hierarchical and partitioned coordination mechanism of a multi-agent system, each region is equipped with a region controller for self-optimization control of resources within the region; each layer is equipped with a hierarchical coordinator for coordination between different regions within the same layer; and a global coordinator is set up for coordination decision-making between different layers. The distributed optimization problem in the hierarchical partitioning coordination mechanism is solved by alternating direction multiplier method. The coordination of the regional controller, hierarchical coordinator and global coordinator is iteratively coordinated until the constraint deviation between regions is less than a preset threshold. The control variables output by each regional controller are used to form the control variable set of different regions.

4. The adaptive topology optimization method for distribution networks as described in claim 3, characterized in that, Based on the control variables, an objective function is constructed, and the objective function is solved using an improved particle swarm optimization method to obtain the topological structure, including: Based on the control variables, the distribution network topology state space and topology transformation operation set are defined, and the load status, distributed generation status and energy storage status are monitored in real time through the distribution network system state perception function. Based on the state monitored by the state-aware function, an objective function is constructed, and an improved particle swarm optimization method is used to solve the objective function. The weights of the objective function are dynamically adjusted according to the state of the distribution network system to obtain a topology structure that adapts to the fluctuations of multi-source resources.

5. The adaptive topology optimization method for distribution networks as described in claim 4, characterized in that, A two-layer optimization strategy is used to optimize the topology, and the optimization results are obtained, including: Obtain the state variables, topology variables, and control variables of the power distribution system, and construct the dynamic equations of the power distribution network system; Based on the dynamic equations of the distribution network system, a two-layer optimization strategy is adopted for solving the problem. The outer layer uses an improved genetic algorithm to determine the optimal topology, while the inner layer uses a model predictive control method to solve for the optimal control sequence under the optimal topology. The topology and control strategy are synergistically optimized to obtain the optimization result that includes the optimal topology and the corresponding control strategy.

6. The adaptive topology optimization method for distribution networks as described in claim 5, characterized in that, The optimization results are evaluated, and the evaluation results are fed back to the objective function, including: Establish optimization indicators, which include topological characteristics, functional characteristics, and control characteristics; The optimization results are quantitatively evaluated based on the optimization indicators, and a comprehensive resilience index is calculated. A resilience enhancement strategy is formulated based on the comprehensive resilience index, and the comprehensive resilience index is fed back to the weight adjustment mechanism of the objective function.

7. The adaptive topology optimization method for distribution networks as described in claim 4, characterized in that, The objective function is expressed as: Where f1 represents network loss, f2 represents voltage deviation, f3 represents load balancing, and w1, w2, and w3 are weighting coefficients, with w1 + w2 + w3 = 1, a k T represents a topology transformation operation. i Let S represent a feasible topology of the distribution network, S represent the distribution network system state awareness function, and A represent the topology transformation operation set.

8. A distribution network adaptive topology optimization system, using the method described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire relevant parameters of the power distribution network system; The model building module is used to divide the power distribution network system into different levels and regions based on the relevant parameters of the power distribution network system, and build a hierarchical and partitioned model. The model solving module is used to solve the hierarchical partitioning model in a coordinated manner through intelligent algorithms to obtain control variables for different regions. The topology acquisition module is used to construct an objective function based on the control variables, solve the objective function using an improved particle swarm optimization method, and obtain the topology structure. The comprehensive optimization module is used to optimize the topology using a two-layer optimization strategy, obtain the optimization results, evaluate the optimization results, and feed the evaluation results back to the objective function.

9. An electronic device, characterized in that, include: A memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the distribution network adaptive topology optimization method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores computer-executable instructions that, when executed by a processor, implement the steps of the power distribution network adaptive topology optimization method according to any one of claims 1 to 7.

Images