Coordinated planning method and device for flexible power distribution network

Abstract

The application provides a kind of flexible power distribution network collaborative planning method and device, method includes: obtaining the grid frame parameter of target power distribution network, planning parameter and source load output typical scene data;With preset partition condition as constraint, construct the partition constraint of distribution network expressed by analytical formula;With the sum of investment cost and operating cost of target power distribution network minimum as target, construct flexible power distribution network constraint set, flexible power distribution network constraint set at least includes distribution network partition constraint, energy storage system integrated intelligent soft switch planning scheme generation constraint, energy storage system integrated intelligent soft switch planning constraint, energy storage system integrated intelligent soft switch operation constraint and flexible power distribution network operation constraint;Flexible power distribution network constraint set is solved, and the topology and capacity configuration scheme of energy storage system integrated intelligent soft switch are obtained, to guide the physical grid frame structure upgrading of target power distribution network. Improve the utilization of intelligent soft switch.

Description

Technical Field

[0001] This application relates to the field of power distribution network planning and operation control technology, and in particular to a collaborative planning method for flexible power distribution networks, a collaborative planning device for flexible power distribution networks, computer equipment, computer-readable storage media, and computer program products. Background Technology

[0002] Flexible distribution networks refer to a new type of distribution network that uses intelligent soft switches and other technologies to achieve flexible interconnection between feeders and actively adjust power flow distribution. With the large-scale integration of distributed power sources and new loads, the uncertainty and complexity of distribution network operation have increased significantly. Traditional rigid network structures are difficult to adapt to the drastic fluctuations in source and load. Therefore, reasonable flexible distribution network planning is needed to improve the economy and operational safety of the distribution network.

[0003] Current flexible distribution network planning primarily employs a phased approach. First, planners, based on the distribution network's topology and electrical parameters, use methods such as clustering algorithms, sensitivity analysis, or community detection to divide the network into several relatively independent zones. Then, based on these zones, planners utilize heuristic algorithms (such as genetic algorithms or particle swarm optimization) or mathematical programming methods (such as mixed-integer linear programming) to configure intelligent soft switches and energy storage systems to address power imbalances within each zone, taking into account the load and power characteristics of each zone. Finally, planners use the determined zone layout and device configuration scheme as input, and simulate and verify the planning scheme through AC power flow calculations or optimal power flow models. This involves calculating electrical quantities such as node voltage amplitude, line transmission power, network losses, and the operating status of intelligent soft switches under various scenarios. The plan evaluates whether it meets safety constraints such as voltage over-limit, line overload, and power backflow, and adjusts the device installation locations and capacities based on the evaluation results. The final output is a smart soft switch deployment scheme and its corresponding planned capacity that meets both economic and safety requirements.

[0004] However, because the actual control capabilities of the smart soft switches to be configured are unknown during the partitioning process, the partition boundaries are often based solely on static topology or electrical distance, making it difficult to predict how flexible interconnections will change power flow distribution. Therefore, the resulting partitions may not be globally optimal. Furthermore, device configuration can only be optimized within a fixed partitioning framework, artificially limiting the cross-regional power allocation capabilities of smart soft switches. This planning approach ultimately reduces the utilization efficiency of integrated smart soft switches in energy storage systems and the grid planning's adaptability to source-load uncertainties. Summary of the Invention

[0005] The purpose of this application is to provide a collaborative planning method, a collaborative planning device, a computer device, a computer-readable storage medium, and a computer program product for flexible distribution networks, so as to improve the utilization efficiency of integrated intelligent soft switches in energy storage systems and the adaptability of power grid planning to source-load uncertainties.

[0006] To address the aforementioned technical problems, this application provides the following technical solutions: The first aspect of this application provides a collaborative planning method for flexible distribution networks. The method includes: acquiring the network parameters, planning parameters, and typical source-load output scenario data of the target distribution network; constructing a distribution network zoning constraint expressed analytically based on the network parameters and with preset zoning conditions as constraints; constructing a flexible distribution network constraint set based on the planning parameters and the typical source-load output scenario data, with the objective of minimizing the sum of the investment cost and operating cost of the target distribution network; the flexible distribution network constraint set includes at least the distribution network zoning constraint, the intelligent soft switch planning scheme generation constraint for integrated energy storage system, the intelligent soft switch planning constraint for integrated energy storage system, the intelligent soft switch operation constraint for integrated energy storage system, and the flexible distribution network operation constraint; solving the flexible distribution network constraint set to obtain the topology and capacity configuration scheme of the intelligent soft switch for integrated energy storage system, so as to guide the physical network structure upgrade of the target distribution network.

[0007] Compared to existing technologies, the collaborative planning method for flexible distribution networks provided in the first aspect of this application constructs analytically expressed distribution network zoning constraints based on preset zoning conditions. Based on the goal of minimizing investment and operating costs, it jointly constructs a flexible distribution network constraint set by combining distribution network zoning constraints, intelligent soft-switching planning scheme generation constraints, planning constraints, operating constraints, and flexible distribution network operating constraints. This transforms the traditional sequential decoupling mode of zoning boundary delineation and intelligent soft-switching resource allocation with energy storage into an integrated joint optimization mode. This avoids the artificial separation caused by the unpredictable flexible control capabilities of zoning, enabling the zoning results to adaptively match the power flow regulation and energy time-shifting capabilities of the intelligent soft-switching with energy storage. Ultimately, the intelligent soft-switching with integrated energy storage can achieve cross-regional power allocation and cross-time period energy shifting based on the globally optimal zoning boundaries, improving the utilization efficiency of the intelligent soft-switching with integrated energy storage and the distribution network's adaptability to source-load uncertainties.

[0008] In other embodiments provided in this application, based on network parameters and with preset partitioning conditions as constraints, an analytical expression of distribution network partitioning constraints is constructed, including: defining binary variables to characterize node partitioning affiliation, line partitioning affiliation, and whether a line is a partitioning boundary line, based at least on the node set, line set, and topological connection relationship in the network parameters; and using the binary variables as carriers, constructing an analytical expression of distribution network partitioning constraints with constraints to ensure the unique affiliation of nodes and lines, maintain the balance of partitioning scale, and ensure the radial topological structure within the partition.

[0009] By constructing analytically expressed distribution network partition constraints based on node sets, line sets, and topological connections, ensuring unique ownership of nodes and lines, maintaining balanced partition size, and guaranteeing radial topology within partitions, traditional heuristic or clustering-based partitioning rules can be transformed into strictly mathematically analytical optimization constraints, thereby improving the standardization, balance, and solvability of distribution network partitioning.

[0010] In other embodiments provided in this application, binary variables are used as carriers to construct analytically expressed distribution network zoning constraints, ensuring the unique affiliation of nodes and lines, maintaining zoning size balance, and guaranteeing the radial topology within a zoning zone. This includes: based on the line set and line connectivity parameters, using binary variables to represent whether each line is a zoning boundary line, setting that the zoning boundary line does not belong to any zoning zone, and maintaining the zoning boundary line as a physical line used to transmit zoning interconnection power, thus constructing zoning boundary constraints; based on the node set, line set, and topology connection relationship, using binary variables to represent whether each node belongs to a certain zoning zone and whether each line belongs to a certain zoning zone, setting that any node belongs to only one zoning zone, and setting that any... Each non-boundary line belongs to only one partition, and the non-boundary line and its two endpoints are assigned to the same partition, thus constructing a unique attribution constraint for nodes and lines. Based on the node set and the preset number of partitions, the number of nodes in each partition is set to fall within a preset equilibrium range to avoid generating single clusters, super-large clusters, or isolated node clusters, thus constructing a partition size equilibrium constraint. Based on the node set, line set, and the partition attribution relationship between each node and line, the difference between the number of nodes and the number of lines in each partition is set to be greater than or equal to 0 and less than or equal to 1, thus constructing a radial topology constraint within the partition. The partition boundary constraint, the unique attribution constraint for nodes and lines, the partition size equilibrium constraint, and the radial topology constraint within the partition are determined as the distribution network partition constraints.

[0011] By using binary variables to construct partition boundary constraints, unique node and line ownership constraints, partition size balance constraints, and radial topology constraints within partitions, the distribution network partitioning rules can be fully embedded into the optimization model in a mathematical analytical form. Under the combined effect of clearly distinguishing between boundary lines and internal lines, ensuring consistency of nodes and lines across constraints, controlling partition size balance and avoiding isolated nodes, and ensuring that the necessary conditions for radial topology are met within each partition, the clarity of partition boundaries, the logical rigor of partition ownership, and the practical reliability of the planning scheme are improved.

[0012] In other embodiments provided in this application, with the goal of minimizing the sum of investment cost and operating cost of the target distribution network, a flexible distribution network constraint set is constructed based on network parameters, planning parameters, typical source and load output scenario data, and distribution network zoning constraints. This includes: constructing an objective function with the goal of minimizing the sum of investment cost over the entire life cycle of the target distribution network and annualized operating cost calculated based on typical source and load output scenario data; and constructing a flexible distribution network constraint set by limiting the feasible domain of the objective function based on planning parameters, typical source and load output scenario data, and distribution network zoning constraints.

[0013] By constructing an objective function with the goal of minimizing the sum of the investment cost over the entire life cycle of the target distribution network and the annualized operating cost calculated based on typical source and load output scenario data, the optimization framework can be integrated with the one-time construction investment and the long-term operating economy across multiple scenarios and time periods. This improves the accuracy of the economic assessment of the planning scheme over the entire life cycle and enhances the comprehensive economic adaptability to source and load uncertainties.

[0014] In other embodiments provided in this application, an objective function is constructed with the goal of minimizing the sum of the investment cost over the entire life cycle of the target distribution network and the annualized operating cost calculated based on typical source-load output scenario data. This includes: constructing the investment cost based on the sum of site development costs, equipment purchase costs, and line construction costs of intelligent soft switches; constructing the annualized operating cost by weighted summing of the total operating costs under each scenario based on the probability of occurrence of each scenario in typical source-load output scenarios; and uniformly converting the investment cost and operating cost of each future year to the current base year based on the present value coefficient, and equally allocating the one-time construction investment to each year within the investment recovery period based on the capital recovery coefficient, thereby processing the investment cost and annualized operating cost into equal-annual values ​​to obtain the objective function expressed in the form of equal-annual values.

[0015] By incorporating the site development costs, equipment purchase costs, and line construction costs of smart soft switches into the investment cost, and by applying equal-annual values ​​to the investment cost and annualized operating cost based on the present value factor and capital recovery factor, it is possible to achieve full-caliber investment modeling for smart soft switches with energy storage and to scientifically compare and convert costs between different years, thereby improving the comprehensiveness of the total investment and operating cost assessment and the comparability of the time dimension.

[0016] In other embodiments provided in this application, the method further includes: constructing equipment purchase costs based on the sum of the products of the installed capacity of the AC-DC converter, the DC-DC converter, and the energy storage battery and the corresponding unit capacity investment cost; and constructing the operating cost for the corresponding scenario by multiplying the sum of the line loss cost, voltage over-limit penalty cost, line current over-limit penalty cost, and self-balancing deviation penalty cost at each moment in any scenario by a time interval and performing time-series accumulation.

[0017] By independently including the capacity investment costs of AC-DC converters, DC-DC converters, and energy storage batteries in the equipment purchase cost, and by accumulating the line loss cost, voltage over-limit penalty cost, line current over-limit penalty cost, and self-balancing deviation penalty cost in time series to construct the total operating cost, it is possible to accurately characterize the internal equipment cost structure of smart soft switches containing energy storage and the various operating costs of flexible distribution networks in the spatiotemporal dimensions, thereby improving the accuracy of equipment investment modeling and the comprehensiveness of operating cost assessment.

[0018] In other embodiments provided in this application, based on planning parameters, typical source-load output scenario data, and distribution network zoning constraints, a set of constraints for flexible distribution networks is constructed to limit the feasible domain of the objective function. This includes: based on the candidate installation locations and candidate scheme sets in the planning parameters, limiting each candidate location to at most one planning scheme, and associating the binary variable representing whether a certain planning scheme is selected with the required line length of the scheme, thus constructing a planning scheme generation constraint for an integrated smart soft switch for energy storage systems; based on the candidate installation locations, equipment capacity upper and lower limits, and investment cost parameters in the planning parameters, limiting the installation capacity range of AC-DC converters, DC-DC converters, and energy storage batteries, and associating the installation capacity of each component with the binary variable representing whether to deploy a smart soft switch, thus constructing planning constraints for an integrated smart soft switch for energy storage systems; and based on the time-series characteristics in the typical source-load output scenario data, describing the power balance of each port of the smart soft switch at each time point in each scenario. The temporal evolution of the relationship between the energy storage system and the state of charge of the energy storage battery is determined, and the power range of each port, the lower limit of converter loss, and the boundary of the state of charge of the energy storage battery are defined to construct the integrated intelligent soft-switching operation constraints of the energy storage system. Based on the node affiliation relationship and line topology of each distributed balance unit defined by the distribution network zoning constraints, the power balance relationship of each node at each time under each scenario and the analytical relationship between voltage and power at both ends of the line are described. The square variables of line current and node voltage are introduced, and the nonlinear AC power flow equation is replaced by the second-order cone form inequality constraints. The square of each node voltage and the square of each line current are limited to the preset safety upper and lower limits, respectively, to construct the flexible distribution network operation constraints. The distribution network zoning constraints, the intelligent soft-switching planning scheme generation constraints of the integrated energy storage system, the intelligent soft-switching planning constraints of the integrated energy storage system, the intelligent soft-switching operation constraints of the integrated energy storage system, and the flexible distribution network operation constraints are determined as the flexible distribution network constraint set.

[0019] By generating constraints, planning constraints, and operational constraints for the integrated smart soft switch of the energy storage system, and constructing flexible distribution network operational constraints based on the distribution network zoning constraints using a second-order cone form of the square of the node voltage and the square of the line current, the system can collaboratively model equipment deployment decisions, energy storage timing operation characteristics, and radial AC power flow within the zoning boundaries in a single model. This improves the coupling consistency between planning and operation of smart soft switches with energy storage and the accuracy of AC power flow modeling.

[0020] In other embodiments provided in this application, solving the constraint set of the flexible distribution network includes: decomposing the non-convex constraints in the constraint set of the flexible distribution network into the difference between a first convex function and a second convex function; and performing successive linearization iterative approximation on the concave terms in the difference at the solution of the previous iteration to solve the constraint set of the flexible distribution network.

[0021] By transforming the original non-convex constraints into a sequence of iteratively solvable convex subproblems and introducing a penalty term to tighten the cone relaxation bias, the physical characteristics of the nonlinear electrical constraints can be accurately restored while ensuring global convergence, thereby obtaining an engineering-feasible solution with strict optimality guarantee.

[0022] In other embodiments provided in this application, the non-convex constraints in the flexible distribution network constraint set are decomposed into the difference between a first convex function and a second convex function, including: based on the second-order cone constraints in the flexible distribution network constraint set, an initial solution is obtained by using convex relaxation techniques to obtain an initial solution for the distribution network state variables; based on the initial solution, a first cone deviation value characterizing the distribution network power flow constraint relaxation deviation and a second cone deviation value characterizing the intelligent soft-switching converter loss constraint relaxation deviation are calculated; when the first cone deviation value or the second cone deviation value is greater than a preset convergence threshold, the original nonlinear cone constraints are transformed into the difference between the first convex function and the second convex function. The difference of convex functions, where the second convex function is a linear approximation of the first convex function at the solution of the previous iteration; the concave terms in the difference are successively linearized iteratively approximated at the solution of the previous iteration, and the constraint set of the flexible distribution network is solved, including: adding a penalty term to the optimization model with the objective of minimizing the sum of investment cost and operating cost, the penalty term being associated with the slack variable of the difference between the first and second convex functions, constructing a modified optimization model and solving iteratively; updating the coefficient of the penalty term and repeating the iteration until the first cone deviation value and the second cone deviation value both satisfy the convergence condition, and stopping the iteration.

[0023] After obtaining the initial solution based on second-order cone relaxation, the cone deviation values ​​of power flow constraints and converter loss constraints are calculated respectively. When the deviation exceeds the threshold, the original non-convex cone constraint is transformed into the form of the difference between the convex function and its first-order Taylor expansion linear approximation. A penalty term associated with the relaxation variable is added to the objective function for iterative solution. The cone relaxation gap can be gradually tightened without introducing additional nonlinear complexity, thereby improving the accuracy and convergence reliability of the flexible distribution network collaborative planning model.

[0024] The second aspect of this application provides a collaborative planning device for flexible distribution networks. The device includes: an acquisition module for acquiring network parameters, planning parameters, and typical source-load output scenario data of a target distribution network; a first construction module for constructing an analytically expressed distribution network partition constraint based on the network parameters and preset partition conditions; a second construction module for constructing a flexible distribution network constraint set based on the planning parameters and the typical source-load output scenario data, with the objective of minimizing the sum of investment cost and operating cost of the target distribution network. The flexible distribution network constraint set includes at least the distribution network partition constraint, a planning scheme generation constraint for an integrated smart soft switch of an energy storage system, a planning constraint for an integrated smart soft switch of an energy storage system, an operating constraint for an integrated smart soft switch of an energy storage system, and an operating constraint for the flexible distribution network; and a planning module for solving the flexible distribution network constraint set to obtain the topology and capacity configuration scheme of the integrated smart soft switch of the energy storage system, so as to guide the physical network structure upgrade of the target distribution network.

[0025] A third aspect of this application provides a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the method of the first aspect.

[0026] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method of the first aspect.

[0027] The fifth aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the method of the first aspect.

[0028] The collaborative planning device for flexible distribution networks provided in the second aspect of this application, the computer equipment provided in the third aspect, the computer-readable storage medium provided in the fourth aspect, and the computer program product provided in the fifth aspect have the same or similar beneficial effects as the collaborative planning method for flexible distribution networks provided in the first aspect. Attached Figure Description

[0029] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of this application are illustrated by way of example and not limitation, with the same or corresponding reference numerals denoteing the same or corresponding parts, wherein: Figure 1 This is a flowchart illustrating the collaborative planning method for flexible distribution networks in the embodiments of this application. Figure 1 ; Figure 2 This is a flowchart illustrating the collaborative planning method for flexible distribution networks in the embodiments of this application. Figure 2 ; Figure 3 This is a schematic diagram of the structure of the collaborative planning device for flexible power distribution networks in the embodiments of this application. Figure 1 ; Figure 4 This is a schematic diagram of the structure of the collaborative planning device for flexible power distribution networks in the embodiments of this application. Figure 2 ; Figure 5 This is a schematic diagram of the structure of the computer device in the embodiments of this application. Detailed Implementation

[0030] Exemplary embodiments of this application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of this application are shown in the drawings, it should be understood that this application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0031] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains.

[0032] It should be noted that the user data, data acquisition, and / or use involved in the embodiments of this application strictly comply with the laws, regulations, and industry standards of relevant countries and regions. The collection and acquisition of data involved in the embodiments of this application are all done in advance by actively prompting or prominently displaying information to inform users and obtaining authorization, or by obtaining full authorization from all parties. The processing, manipulation, forwarding, and use of data involved in the embodiments of this application are all carried out on the premise that the user or relevant party is fully informed and authorized. In implementing the various embodiments of this application, the types of data or information, scope of use, and usage scenarios that may be involved are informed to users or relevant parties and authorization is obtained through appropriate means. The specific methods of notification and authorization may vary according to the actual situation, and this application is not limited in this regard. The processing of personal information involved in the embodiments of this application is carried out under the premise of having a legal basis (such as obtaining the consent of the personal information subject or being necessary for the performance of a contract), and is only processed within the prescribed or agreed scope. Sensitive personal information such as biometric information, medical and health information, financial account information, and precise location information involved in the embodiments of this application are all processed under the premise of having a specific purpose and sufficient necessity, and with the separate authorization and consent of the user or relevant party. In some embodiments of this application, if the user or related party refuses to process personal information other than the information necessary for the basic functions, it will not affect the use of the basic functions of the embodiments of this application.

[0033] The current phased planning approach, which first divides the system into zones based on static topology or electrical distance and then configures integrated smart soft switches for energy storage systems within a fixed zoning framework, is problematic because the actual control capability of the integrated smart soft switches for energy storage systems cannot be predicted during the zoning phase. This can lead to zoning results that are not globally optimal. Furthermore, the device configuration phase is constrained by fixed artificial zoning boundaries, limiting cross-zone power allocation capabilities. Consequently, the utilization efficiency of integrated smart soft switches for energy storage systems and the adaptability of the distribution network to uncertainties in source and load are reduced.

[0034] To address the problems of low equipment utilization efficiency and insufficient adaptability to source load uncertainties caused by the decoupling of zoning and equipment configuration in current phased planning, the inventors recognized that the root cause lies in the informational disconnect between zoning decisions and equipment control capabilities. Therefore, it is essential to construct a unified optimization framework that can simultaneously accommodate zoning boundary variables and equipment configuration variables, so that zoning boundaries are no longer fixed, artificially preset parameters, but rather decision variables that can adaptively adjust based on the flexible control potential of the equipment. Simultaneously, considering that this joint optimization will introduce complex nonlinear electrical constraints (such as power flow equations and converter losses), direct solutions will face significant mathematical difficulties. This further compels the inventors to find an efficient algorithm that can preserve nonlinear accuracy while ensuring solution convergence. This leads to a complete technical breakthrough path: from the pain point of efficiency loss due to decoupling, to the necessity of collaborative modeling, and finally to the search for a high-precision, solvable algorithm.

[0035] In view of this, embodiments of this application provide a collaborative planning method, a collaborative planning device, a computer device, a computer-readable storage medium, and a computer program product for flexible distribution networks. By acquiring the grid parameters, planning parameters, and typical source-load output scenario data of the distribution network, an analytically expressed distribution network zoning constraint is constructed. With the goal of minimizing the sum of investment cost and operating cost, the distribution network zoning constraint, the intelligent soft-switching planning scheme generation constraint of the integrated energy storage system, the planning constraint, the operating constraint, and the flexible distribution network operating constraint are jointly constructed into a flexible distribution network constraint set. This transforms the delineation of zoning boundaries and the allocation of intelligent soft-switching resources with energy storage from the traditional decoupling mode to an integrated joint optimization mode, ultimately obtaining the collaboratively optimal equipment topology and capacity.

[0036] First, the collaborative planning method for flexible distribution networks provided in the embodiments of this application will be described.

[0037] The collaborative planning method for flexible distribution networks is applied to distribution network physical systems that include distributed generation sources and new types of loads. Specifically, the output volatility of distributed generation sources exceeds a preset volatility threshold, and the temporal randomness of new types of loads exceeds a preset randomness threshold.

[0038] Distributed power sources refer to power sources connected to the distribution network that are mainly renewable energy sources (such as photovoltaic and wind power). Their output volatility refers to the characteristic that the power generation capacity changes drastically in a short period of time due to changes in meteorological conditions (such as sunlight and wind speed). When their statistical indicators (such as power change rate and variance) exceed the preset fluctuation threshold, they will significantly affect the power balance of the power grid.

[0039] New types of loads mainly refer to electrical equipment with high time uncertainty, such as electric vehicle charging piles. Their temporal randomness refers to the unpredictability of load access time, charging duration, and power demand in the time dimension. When their probability distribution characteristics (such as peak-valley difference rate and randomness entropy value) exceed the preset randomness threshold, it will exacerbate the risk of power grid supply and demand imbalance.

[0040] The output volatility of distributed power sources refers to the degree to which the power generation of renewable energy sources such as photovoltaics and wind power changes drastically in a short period of time due to meteorological conditions (such as cloud cover and sudden changes in wind speed). It is usually quantified by statistical indicators such as power change rate, variance, or ramp rate. The preset fluctuation threshold is an upper limit of allowable fluctuation set to ensure the safe operation of the power grid. When the actual fluctuation exceeds this threshold, it must be smoothed out through flexible control measures (such as smart soft switching).

[0041] The temporal randomness of new loads refers to the unpredictability of the access time, charging duration, and power demand of electrical equipment such as electric vehicle charging piles. It is usually measured by probability distribution characteristics (such as peak-valley difference rate and randomness entropy value). The preset randomness threshold is the maximum random deviation boundary allowed to maintain the supply and demand balance of the system. When it exceeds this threshold, it is necessary to rely on the energy time-shifting capability of the energy storage system for adjustment.

[0042] The distribution network physical system is the physical power network that carries the aforementioned source loads. It consists of primary equipment such as nodes, lines, and transformers. The planning method in this application improves the economic efficiency and safety of the system in a highly uncertain environment through the coordinated optimization of partitioning and flexible interconnection equipment.

[0043] Figure 1 This is a flowchart illustrating the collaborative planning method for flexible distribution networks in the embodiments of this application. Figure 1 See Figure 1 As shown, the method may include: S11: Obtain the network structure parameters, planning parameters, and typical source load output data of the target distribution network.

[0044] The target distribution network refers to an active distribution network that includes distributed power sources, new loads, and intelligent soft switches that can be integrated through energy storage systems to achieve flexible interconnection between feeders. Examples include urban distribution networks or industrial park distribution networks that are connected to a high proportion of distributed photovoltaic power and electric vehicle charging piles.

[0045] Network parameters are the basic data that describe the physical topology and electrical characteristics of a distribution network, such as node set, line set, line resistance and reactance, line current carrying capacity, and feeder connection relationship.

[0046] Planning parameters are data that characterize the technical and economic characteristics of the equipment to be planned and the information on candidate sites, such as candidate installation locations for smart soft switches, investment cost per unit capacity of AC-DC converters, investment cost per unit capacity of energy storage batteries, discount rate, and investment payback period.

[0047] Typical source and load output scenario data are used to characterize the uncertainty of distributed power sources and loads in the time dimension. These are multi-scenario time-series data, such as photovoltaic daily output curves, wind power output curves, and electric vehicle daily charging load curves. Each scenario corresponds to an occurrence probability.

[0048] Specifically, network parameters can be obtained by reading Geographic Information System (GIS) data, Common Information Model (CIM) files, or the power grid topology database in the dispatch automation system of the target distribution network. These parameters include node numbers, line connection relationships, line resistance and reactance, line thermal stability limit capacity, and the rated capacity of loads and distributed power sources connected to each node.

[0049] Planning parameters can be set based on the actual project and the technical and economic manuals provided by the equipment manufacturers, including the candidate installation locations of intelligent soft switches and energy storage systems, the unit capacity investment cost of AC-DC converters, DC-DC converters and energy storage batteries, site development costs, unit length line construction costs, annual interest rates and investment payback periods.

[0050] Typical power generation scenario data can be generated by statistically clustering photovoltaic and wind power output data and various load electricity consumption data over a historical year. Several representative daily power output curves and load curves are selected as typical scenarios, and a corresponding probability of occurrence is assigned to each scenario to characterize the uncertainty of distributed power generation output and load demand.

[0051] After obtaining the network parameters, planning parameters, and typical source-load output scenario data of the target distribution network, the node voltage reference values, power reference values, and line impedance parameters in the network parameters can be normalized to unify all electrical quantities to the same reference system. The photovoltaic output curves, wind power output curves, and load curves in the typical source-load output scenario data are normalized to map their numerical range to the [0,1] interval. A synchronous back-substitution reduction algorithm based on probabilistic distance is used to cluster and reduce the normalized source-load output scenarios, retaining a finite number of representative typical scenarios and their corresponding occurrence probabilities. The scenario data used for power flow calculation needs to be restored to the original dimensions after reduction, or normalization processing (mean of 0, variance of 1) can be used instead of normalization processing.

[0052] By standardizing the grid parameters, the dimensional differences caused by different voltage and power levels can be eliminated, improving the stability and convergence of numerical calculations in the optimization model. Normalizing and reducing the source-load output scenarios can significantly reduce the number of scenarios and the scale of time-series variables in the model while preserving the statistical characteristics of the original data, thereby reducing the computational complexity of the mixed-integer programming model and improving solution efficiency.

[0053] S12: Based on the grid parameters and with preset partitioning conditions as constraints, construct analytical expressions for distribution network partitioning constraints.

[0054] The aim here is to divide the distribution network into several power supply zones with clear electrical boundaries, which do not require 100% real-time autonomy but can achieve cross-regional power distribution through intelligent soft switches integrated with energy storage systems, thereby improving the distribution network's ability to accommodate high proportions of distributed sources and loads and its operational flexibility.

[0055] Based on the network parameters and with preset partitioning conditions as constraints, an analytical expression of distribution network partitioning constraints is constructed. This transforms the topological connection relationship and partitioning rules of the distribution network into mathematical expressions that can be embedded in the optimization model. This provides a strict analytical constraint basis for the adaptive decision-making of partitioning boundaries, making partitioning decisions no longer independent heuristic steps, but endogenous variables that can participate in subsequent collaborative optimization.

[0056] Using pre-defined zoning conditions as constraints refers to using a series of pre-set distribution network zoning rules as restrictions when constructing analytically expressed distribution network zoning constraints. These rules are used to define the boundaries of feasible zoning schemes. Pre-defined zoning conditions may include: whether a line belongs to a zoning boundary line; whether a boundary line does not belong to any zoning but remains a physical line for transmitting interconnection power; unique node and line affiliation conditions; zoning scale balancing conditions; and radial topology conditions within a zoning. Furthermore, pre-defined zoning conditions can be flexibly expanded according to specific planning needs. For example, they may also include scale balancing conditions based on load capacity (using the total load capacity within a zoning instead of the number of nodes as a balancing indicator); implicit boundary identification conditions based on power transmission direction (automatically identifying boundaries through differences in node affiliation at both ends of a line); affiliation consistency conditions based on connectivity path transmission (ensuring that nodes within the same connectivity domain belong to the same zoning); cross-zone isolation conditions based on feeder affiliation (nodes and lines on different feeders belong to different zoning); and radial topology equivalence conditions based on root node count (each zoning has one and only one root node, and the number of lines within a zoning equals the number of nodes minus one).

[0057] In the specific construction, binary variables can be introduced to represent the partition affiliation of nodes and lines, whether a line is a partition boundary line, and constraints such as consistent affiliation of nodes and lines, balanced partition size, and radial topology within a partition can be established. The partitioning rules are then embedded into the optimization model in an analytical mathematical form.

[0058] Alternatively, based on the node electrical distance and power transmission sensitivity in the network parameters, a continuous variable is introduced to characterize the electrical coupling strength between nodes. The goal is to minimize the sum of the average electrical distances between nodes in each partition. Each node is limited to exactly one partition, the total number of nodes in each partition is limited to not exceeding a preset upper limit, and the power transmission of the tie lines between partitions is limited to not exceeding a given threshold. This constructs an analytical distribution network partition constraint based on the degree of electrical coupling.

[0059] By using electrical coupling degree and power transmission sensitivity as the basis for division, the partitioning results can be made to better fit the electrical characteristics of the distribution network, reduce the demand for high-power transmission across partitions, thereby reducing the power exchange amplitude between partitions and improving the self-balancing capability of each distributed balancing unit.

[0060] S13: With the goal of minimizing the sum of investment and operating costs of the target distribution network, construct a set of constraints for a flexible distribution network based on planning parameters and typical source-load output scenario data. The set of constraints for a flexible distribution network includes at least distribution network zoning constraints, constraints for generating intelligent soft-switching planning schemes for integrated energy storage systems, constraints for planning intelligent soft-switching for integrated energy storage systems, constraints for operating intelligent soft-switching for integrated energy storage systems, and constraints for operating flexible distribution networks.

[0061] The objective of minimizing the sum of investment and operating costs of the target distribution network refers to incorporating the one-time investment in distribution network construction (such as equipment purchase, site development, and line construction costs) and the long-term economic costs of operation in multiple scenarios and time periods (such as network loss costs, over-limit penalties, and imbalance deviation costs) into the optimization objective within the same optimization framework in order to seek the most economically efficient planning scheme.

[0062] Based on grid parameters, planning parameters, typical source and load output scenario data, and distribution network zoning constraints, a flexible distribution network constraint set is constructed. This set integrates the device capacity decision of the integrated intelligent soft switch of the energy storage system, the time-series operation characteristics of multiple scenarios, and the network physical constraints after distribution network zoning into the optimization model, thereby providing a complete feasible domain constraint framework for collaborative optimization of zoning boundaries and device configuration.

[0063] In specific construction, the investment cost can be calculated by using the sum of the site development cost, equipment purchase cost, and line construction cost of the smart soft switch as the objective function, and the weighted sum of network loss cost, voltage over-limit penalty, line overload penalty, and imbalance deviation penalty under multiple scenarios as the objective function for the operating cost. Planning constraints, time-series operation constraints, and power flow constraints of the flexible distribution network based on second-order cone relaxation can be established separately, thereby jointly constraining and minimizing the investment cost and operating cost in the same optimization model.

[0064] Alternatively, based on the candidate nodes and discrete capacity levels of distributed energy storage in the planning parameters, integer variables can be used to characterize the selection of the energy storage level for each node, and the total energy storage capacity of each node can be limited to a certain proportion of the capacity of the upstream transformer, thus constructing the planning constraints of the energy storage system. Based on the time-series characteristics of typical source-load output scenario data, the charging and discharging power balance and state of charge recursive relationship of the energy storage system at each time under each scenario can be described, and the charging and discharging power can be limited to the rated power of the energy storage and the state of charge can be maintained between the upper and lower safety limits, thus constructing the operation constraints of the energy storage system. Based on the boundaries of each distributed balance unit defined by the distribution network zoning constraints, a DC power flow model that ignores network losses is used to describe the relationship between the voltage phase angle and injected power of each node with linearized equation constraints, and the voltage phase angle difference of each node can be limited to a preset safety upper limit, thus constructing simplified flexible distribution network operation constraints. The planning constraints of the energy storage system, the operation constraints of the energy storage system, and the simplified flexible distribution network operation constraints are determined as the flexible distribution network constraint set.

[0065] By employing discrete capacity tiers and linearized DC power flow equations, the size of integer variables and the degree of non-convexity and nonlinearity in the model can be significantly reduced. This greatly improves the solution efficiency and reduces the consumption of computing resources while ensuring the basic planning accuracy, making it suitable for rapid planning scenarios in large-scale distribution networks.

[0066] S14: Solve the constraint set of the flexible distribution network to obtain the topology and capacity configuration scheme of the integrated intelligent soft switch of the energy storage system, so as to guide the physical network structure upgrade of the target distribution network.

[0067] In the specific solution, the convex difference programming algorithm can be used to solve the constraint set of the flexible distribution network. By decomposing the non-convex constraints into the difference of convex functions and successively linearizing and iterating the concave terms, the second-order cone relaxation deviation can be tightened.

[0068] Alternatively, a model-free solution method based on reinforcement learning can be employed. Specifically, the partitioning allocation variables, smart soft switch deployment variables, and energy storage capacity variables in the flexible distribution network constraint set are collectively defined as the agent's action space. Network parameters and typical source-load output scenario data are defined as the state space. The negative sum of investment and operating costs, along with the penalty terms for each constraint, are used to construct the reward function. Offline training and online decision-making are performed using algorithms such as deep Q-networks or near-end policy optimization, directly outputting a near-optimal topology and capacity configuration scheme. This approach eliminates the need for convex relaxation or linearization of the original problem, naturally handling planning problems with mixed non-convex, discrete, and continuous variables. Furthermore, after training, it can achieve rapid generalization solutions for different network parameters.

[0069] After solving the constraint set of the flexible distribution network, the following results can be output to guide the upgrading of the physical network structure of the target distribution network: the partition boundaries of the distributed balance unit (i.e., the electrical boundary locations between each partition and the set of lines marked as boundary lines), the topology and capacity planning scheme of the integrated smart soft switch of the energy storage system (including the number of ports of the smart soft switch, the location of each port access node, the capacity of the AC-DC converter, the capacity of the DC-DC converter, and the capacity of the energy storage battery), and the total life cycle investment and operation cost of the target distribution network (including the one-time construction investment cost and the annualized operation cost based on multi-scenario time-series simulation).

[0070] The partition boundaries of the target distribution network refer to the electrical boundary locations between distributed balancing units determined through optimization solutions. Specifically, this includes the set of lines marked as partition boundary lines and the information of the nodes at both ends. For example, in a distribution network containing 10 feeders, the solution results mark lines L5, L8, and L12 as boundary lines, indicating that the two nodes connected by these lines belong to different balancing units, and the lines themselves do not belong to any partition, but are used to transmit interconnection power between partitions.

[0071] The topology and capacity configuration scheme for an integrated smart soft switch in an energy storage system refers to the device connection configuration, installation location, and capacity specifications of each component device of the integrated smart soft switch, determined through optimization solutions. The topology and capacity configuration scheme can include the topology structure, access location, DC-DC converter capacity, AC-DC converter capacity, and battery capacity. The topology structure refers to the number of ports and their connections in the smart soft switch; the access location refers to the distribution network nodes or feeders connected to each port of the smart soft switch; the DC-DC converter capacity refers to the power specifications of the converter connecting the energy storage battery to the DC bus; the AC-DC converter capacity refers to the apparent power specifications of the converter connecting the AC distribution network to the DC bus; and the battery capacity refers to the energy storage specifications of the energy storage system. For example, the solution suggests installing a three-terminal smart soft switch between nodes N7, N12, and N15, where port 1 connects to node N7 on feeder A, port 2 connects to node N12 on feeder B, and port 3 connects to node N15 on feeder C. The energy storage system is integrated into the DC bus of the smart soft switch, with an AC-DC converter capacity of 500kVA, a DC-DC converter capacity of 200kW, and an energy storage battery capacity of 800kWh.

[0072] The investment and operating cost of the target distribution network refers to the comprehensive economic cost of the target distribution network over its entire life cycle, obtained through optimization solutions. This includes the one-time construction investment cost and the annualized operating cost based on multi-scenario time-series simulations. The investment cost covers site development costs for intelligent soft switches, equipment purchase costs (AC-DC converters, DC-DC converters, energy storage batteries), and line construction costs. The operating cost covers line loss costs, voltage exceedance penalty costs, line current exceedance penalty costs, and self-balancing deviation penalty costs. For example, the solution results show that the total investment cost of this planning scheme is 12 million yuan (800,000 yuan per year after equal annual value amortization), the annualized operating cost is 450,000 yuan, and the comprehensive equal annual value cost is 1.25 million yuan per year.

[0073] After obtaining the zoning boundaries of the target distribution network, the topology and capacity configuration scheme of the integrated smart soft switch for energy storage systems, and the investment and operating costs of the target distribution network, the following steps are taken: First, based on the solved zoning boundaries of the distributed balance unit, the electrical boundary positions between each zone are marked on the existing distribution network geographical wiring diagram. The corresponding boundary lines are identified as inter-zone tie lines, and a modified geographical wiring diagram is generated to guide on-site construction personnel in implementing zone isolation or adding power measurement devices. Then, based on the solved topology and capacity configuration scheme of the integrated smart soft switch for energy storage systems, the port connection relationships of the smart soft switch, the location of each port access node, the capacity of the AC-DC converter, the capacity of the DC-DC converter, and the capacity of the energy storage battery are marked on the primary system electrical wiring diagram, forming the primary system electrical wiring diagram to guide equipment selection and procurement, installation location determination, and on-site wiring construction. Finally, the investment and operating costs are included as an economic evaluation conclusion in the planning output document, providing a quantitative decision-making basis for project approval and budgeting, thus completing the entire process guidance from planning data to engineering transformation.

[0074] The above guidance process can be carried out manually by power grid planning engineers based on the output planning data to modify the geographical wiring diagram and select equipment, or it can be executed automatically in a computer system.

[0075] During automatic execution, the system outputs the solved partition boundary data, intelligent soft switch topology and capacity configuration scheme, and investment and operating costs in a structured data format (such as JSON or XML) and connects it to the power distribution network computer-aided design system or geographic information system platform. The system has a built-in automatic mapping module that can automatically identify and mark interval tie lines and generate modified layers of the geographic wiring diagram based on the partition boundary data. Simultaneously, the system automatically calls the corresponding equipment models from the standard equipment library based on the topology and capacity configuration scheme, automatically annotating port connection relationships and capacity parameters on the primary system electrical wiring diagram, and generating construction guidance documents containing equipment bills of materials, installation location coordinates, and terminal block correspondences. This achieves automatic conversion from optimization results to engineering drawings and construction instructions without manual intervention.

[0076] As can be seen from the above, the collaborative planning method for flexible distribution networks provided in this application constructs analytically expressed distribution network zoning constraints based on preset zoning conditions. Based on the goal of minimizing investment and operating costs, it jointly constructs a flexible distribution network constraint set by combining distribution network zoning constraints, intelligent soft-switching planning scheme generation constraints, planning constraints, operating constraints, and flexible distribution network operating constraints. This transforms the traditional sequential decoupling mode of zoning boundary delineation and intelligent soft-switching resource allocation with energy storage into an integrated joint optimization mode. This avoids the artificial separation caused by the unpredictable flexible control capabilities of zoning, enabling the zoning results to adaptively match the power flow regulation and energy time-shifting capabilities of intelligent soft-switching with energy storage. Ultimately, this allows the intelligent soft-switching integrated with energy storage to achieve cross-regional power allocation and cross-time period energy shifting based on the globally optimal zoning boundaries, improving the utilization efficiency of the intelligent soft-switching integrated with energy storage and the distribution network's adaptability to source-load uncertainties.

[0077] Furthermore, as a response to Figure 1 In a refinement and extension of the method shown, this application also provides a collaborative planning method for flexible power distribution networks.

[0078] Figure 2 This is a flowchart illustrating the collaborative planning method for flexible distribution networks in the embodiments of this application. Figure 2 See Figure 2 As shown, the method may include: S21: Obtain the network structure parameters, planning parameters, and typical scenario data of source load output of the target distribution network.

[0079] Step S21 here is implemented in the same way as step S11 in the previous embodiment. Please refer to the relevant description in the previous embodiment. It will not be repeated here.

[0080] S22: Based at least on the node set, line set, and topology connection relationship in the network parameters, define binary variables to characterize the node partition affiliation, line partition affiliation, and whether a line is a partition boundary line; using binary variables as carriers, construct analytical expressions of distribution network partition constraints to ensure the unique affiliation of nodes and lines, maintain the balance of partition size, and ensure the radial topology structure within the partition.

[0081] Based on the node set, line set, and topological connection relationship, a binary variable z is introduced. i,p Indicate whether node i belongs to partition p, z ij,p Characterize whether line ij belongs to partition p, γ ij This characterizes whether line ij is a partition boundary line, thus transforming the physical topological connection relationship into a mathematical variable that can be analytically expressed.

[0082] Ensuring the unique affiliation of nodes and lines means that in the partitioning results, any node belongs to one and only one distributed balance unit, any non-boundary line belongs to one and only one distributed balance unit, and the non-boundary line must belong to the same partition as the nodes at both ends.

[0083] Maintaining a balanced distribution size means setting upper and lower limits on the number of nodes (or load capacity) in each distribution area so that the size of each distributed balancing unit is roughly the same. This avoids situations where some distribution areas have too many nodes, forming an oversized cluster, or some distribution areas have too few nodes, forming an isolated cluster, or the entire distribution network is divided into only a single cluster.

[0084] Ensuring a radial topology within a partition means that within each distributed balance unit, the difference between the number of nodes and the number of lines must be greater than or equal to 0 and less than or equal to 1, which satisfies the basic necessary condition of a tree structure (connectivity and no cycles).

[0085] In addition, the distribution network zoning constraints can be constructed analytically based on the set of nodes, set of lines and topological connections in the network parameters, with the goal of minimizing the expected power exchange amplitude on the tie lines between each zone.

[0086] Minimizing the expected power exchange amplitude on the tie lines between zones means minimizing the absolute value of active power (or reactive power) transmitted between zones through boundary lines when dividing distributed balance units.

[0087] In this way, it is hoped that each partition can achieve internal source-load power self-balancing as much as possible under typical operating scenarios, reducing power dependence on adjacent partitions. When constructing partition constraints based on this objective, it is necessary to associate the node affiliation variable with the power variable on the boundary line, so that the partition optimization results tend to classify nodes with a high degree of power matching into the same partition, thereby reducing the power exchange level between partitions and improving the self-consistent operation capability of each distributed balancing unit.

[0088] To fully embed the distribution network zoning rules into the optimization model in an analytical mathematical form driven by binary variables, thereby simultaneously achieving explicit boundary identification, unique attribution guarantee, scale balance control, and radial topology verification within a unified framework, the following approach can be adopted: separately constructing zoning boundary constraints, unique attribution constraints for nodes and lines, zoning scale balance constraints, and radial topology constraints within the zoning, and then jointly determining them as distribution network zoning constraints.

[0089] Specifically, step S22 above may include: S221: Based on the set of lines and the connectivity parameters of the lines, binary variables are used to represent whether each line is a partition boundary line. The partition boundary line is set to not belong to any partition, and the partition boundary line is kept as a physical line and used to transmit partition communication power, thus constructing partition boundary constraints.

[0090] This involves mathematically modeling the decision of which lines in the distribution network serve as boundaries between zones: introducing a binary variable to explicitly mark whether each line is a boundary line, and using an equality constraint to link the line's connectivity, zone affiliation, and boundary attributes. This ensures that lines marked as boundaries do not belong to any zone, but still exist as physical lines and perform power transmission functions between zones, thereby eliminating the ambiguity of zone affiliation and laying the topological foundation for subsequent calculations of power exchange between zones.

[0091] Right now:

[0092] In the formula, Ω p Ω represents the set of distribution network zones. b z represents the set of distribution network lines. ij,p Indicates whether line ij belongs to partition p, α ij Indicates whether line ij is in a connected state, γ ij This indicates whether line ij belongs to the partition boundary line.

[0093] S222: Based on the node set, the line set, and the topology connection relationship, binary variables are used to represent whether each node belongs to a certain partition and whether each line belongs to a certain partition. It is set that any node belongs to only one partition, any non-boundary line belongs to only one partition, and non-boundary lines and the nodes at both ends belong to the same partition, thus constructing a unique ownership constraint for nodes and lines.

[0094] This involves applying consistency constraints to the partition affiliation of nodes and lines to avoid affiliation conflicts or logical contradictions. A binary variable represents whether each node belongs to a particular partition, and an equality constraint ensures that each node belongs to exactly one partition. Simultaneously, for non-boundary lines (i.e., lines that are connected but not marked as boundaries), two inequality constraints force the line and its two endpoints to belong to the same partition, thus guaranteeing topological connectivity and affiliation consistency within each partition. This provides an accurate basis for subsequent calculations of power balance and power flow distribution within each partition.

[0095] Right now:

[0096] In the formula, Ω nLet z represent the set of nodes in the distribution network. i,p Indicate whether node i belongs to partition p, z j,p Indicates whether node j belongs to partition p.

[0097]

[0098] In the formula, Ω n,h and Ω n,h' Let Ω represent the set of nodes on feeders h and h', respectively. b,h and Ω b,h' Let h and h' represent the sets of lines on feeders h and h' respectively, and kl represent another line. kl,p This indicates whether line kl belongs to partition p. This constraint means that, without considering network reconfiguration, nodes and lines on different feeders belong to different partitions.

[0099] S223: Based on the node set and the preset number of partitions, the number of nodes contained in each partition is set to fall within the preset balance range in order to avoid the generation of single clusters, super-large clusters or isolated node clusters, and to build partition size balance constraints.

[0100] This means preventing severe imbalances in the scale of the partitioning results and ensuring a relatively uniform number of nodes in each distributed balancing unit. Based on the preset total number of partitions and the total number of nodes in the distribution network, a node number fluctuation range parameter is introduced to calculate the minimum and maximum number of nodes allowed for each partition. Then, an inequality constraint forces the actual number of nodes in each partition to fall within this equilibrium range. This constraint effectively avoids unreasonable situations such as a single partition covering the entire distribution network, some partitions having too many nodes forming super-large clusters, or some partitions having too few nodes forming isolated node clusters. This ensures a relatively balanced control capability for each partition and provides a similarly scaled optimization basis for the capacity configuration of subsequent integrated smart soft switching of energy storage systems.

[0101] Right now:

[0102] In the formula, Ω n N represents the set of nodes in the distribution network. n,p min and N n,p max N represents the minimum and maximum number of nodes in a distribution network zone, respectively. n N represents the number of nodes. p This indicates the preset number of partitions, and σ represents the percentage of the floating range of the number of nodes.

[0103] S224: Based on the node set, the line set, and the partition affiliation of each node and line, set the difference between the number of nodes and the number of lines in each partition to be greater than or equal to 0 and less than or equal to 1, and construct radial topological constraints within the partition.

[0104] This means ensuring that the topology within each distributed balancing unit satisfies the basic requirements of a radial (tree) network. Based on the attribution variables of nodes and lines within each partition, the total number of nodes and lines in each partition are counted. Then, an inequality constraint forces the difference between the number of nodes and lines in each partition to be between 0 and 1. In a connected, loop-free radial network, the difference between the number of nodes and lines is exactly 1. When only isolated nodes exist within a partition (nodes but no lines), the difference is 0. Therefore, this inequality allows both single-node partitions (difference of 0) and normal radial subnets (difference of 1), while the upper limit of 1 eliminates the possibility of loops within a partition (loops would cause the difference to be less than or equal to 0), thus guaranteeing the radial characteristics of the topology within each partition and providing a prerequisite for subsequent power flow calculation methods based on the radial assumption.

[0105] Right now:

[0106] In the formula, Ω b This represents a set of distribution network lines.

[0107] In addition, based on the set of lines, continuous variables can be introduced to represent the direction of active power transmission on each line. When the nodes at both ends of a line belong to different partitions, the line is implicitly identified as a partition boundary line, and the absolute value of active power transmitted on the boundary line is limited to not exceeding the preset upper limit of the boundary transmission capacity, thus constructing an implicit partition boundary identification constraint based on power flow direction.

[0108] Furthermore, based on the set of nodes and the set of lines, auxiliary binary variables can be introduced to represent the subordinate relationship between nodes. If there is a connected path between two nodes and all intermediate nodes on the path belong to the same partition as the two endpoints, then the two nodes belong to the same partition. Logical constraints (such as linearized inequalities in the form of if-then) are used to describe the above transitive attribution rules, and attribution consistency constraints based on the transitive nature of connected paths are constructed.

[0109] Alternatively, based on the node set and the sum of the rated power of the loads connected to each node, the total load capacity contained in each partition can be set to fall within a preset balance range, replacing the method of using the number of nodes as the balance index, and constructing a partition-scale balance constraint based on load capacity.

[0110] Furthermore, based on the node set, line set, and the partition affiliation relationship of each node and line, the total number of lines in each partition can be set to equal the total number of nodes in that partition minus the number of root nodes contained in that partition, where each partition has one and only one root node (such as the feeder outlet bus), and a radial topology equivalence constraint based on the root node count can be constructed.

[0111] By employing implicit boundary identification through power flow, the explicit introduction of binary boundary variables can be avoided, reducing the number of integer variables in the model. Using transitive connectivity constraints to ensure consistent attribution allows for the same attribution logic without explicitly associating node affiliations for each line. Using load capacity instead of node count as a scalability metric makes the partition size more closely reflect actual electrical carrying capacity rather than simply node counting. Using root node counting to constrain radial topology provides radial decision conditions that are equivalent to but formally different from node-line difference constraints. This allows for different mathematical expressions for model solving while preserving the integrity of partition constraints, adapting to the numerical stability requirements of different solvers or problem structures.

[0112] S225: Define the partition boundary constraints, the unique ownership constraints of nodes and lines, the partition scale balance constraints, and the radial topology constraints within the partition as the distribution network partition constraints.

[0113] The above four types of constraints are combined and formally confirmed as distribution network zoning constraints, making them a unified set of constraints used in subsequent optimization models to limit the feasible domain of zoning schemes.

[0114] S23: Construct an objective function with the goal of minimizing the sum of the investment cost over the entire life cycle of the target distribution network and the annualized operating cost calculated based on typical source and load output scenario data; construct a set of constraints for the flexible distribution network based on planning parameters and typical source and load output scenario data to limit the feasible domain of the objective function.

[0115] By incorporating the investment cost throughout the entire lifecycle of the distribution network and the annualized operating cost based on multi-scenario probability weighting into the same minimization expression, the advantages and disadvantages of different planning schemes can be uniformly measured from the perspective of lifecycle economics. This avoids the short-sighted decision-making caused by focusing only on investment costs or typical daily operating costs in traditional methods. As a result, the optimization model is guided to seek the best balance between investment and operation between the delineation of zoning boundaries and the configuration of intelligent soft switches that integrate energy storage systems, ultimately reducing the overall cost of the distribution network.

[0116] In order to compare investment costs and operating costs incurred in different years with a uniform dimension throughout the entire life cycle, the present value factor can be used to discount the costs of each future year to the current base year, and the capital recovery factor can be used to equally distribute the one-time construction investment to each year of the investment recovery period, thereby obtaining the objective function expressed in the form of equal annual values.

[0117] Specifically, step S23 above may include: S231: The sum of site development costs, equipment purchase costs, and line construction costs for an integrated intelligent soft switch based on an energy storage system constitutes the investment cost.

[0118] The investment cost of an integrated smart soft switch for energy storage systems is broken down into three parts: site development costs, equipment purchase costs, and line construction costs. Site development costs are represented by multiplying the unit site development cost by the investment decision variable. Equipment purchase costs are represented by the sum of the products of the capacity of the AC-DC converter, DC-DC converter, and energy storage battery and the corresponding unit capacity investment cost. Line construction costs are represented by multiplying the unit length line construction price by the length of the new line to be built. This comprehensively describes the one-time capital expenditure required to invest in and build a smart soft switch.

[0119] Right now:

[0120] In the formula, CO This indicates the investment cost of intelligent soft switches. ST , DE and BR The costs include site development, equipment purchase, and line construction for intelligent soft switches. ST and c BR These represent the site development cost and the unit length line construction cost for intelligent soft switches, respectively. CT c CH and c BT S represents the unit capacity investment cost of the AC-DC converter, DC-DC converter, and energy storage battery, respectively. r CT S r CH and S r BT χ represents the capacity of the AC-DC converter, DC-DC converter, and energy storage battery deployed using scheme r with intelligent soft switching. r It is a binary variable representing whether the intelligent soft switch for planning scheme r is enabled, D. r R represents the length of the line that needs to be constructed for the intelligent soft switch using scheme r, and R represents the set of planning schemes.

[0121] The above is the sum of the products of the installed capacity of the AC-DC converter, the DC-DC converter, and the energy storage battery, and the corresponding unit capacity investment cost, to construct the equipment purchase cost.

[0122] In addition, the equipment purchase cost can be calculated by multiplying the installed capacity of the AC-DC converter, the DC-DC converter, and the energy storage battery in the smart soft switch by the corresponding unit capacity investment cost and summing them. Alternatively, the equipment purchase cost can be calculated by multiplying the total apparent power capacity of the converters at each port of the smart soft switch by the comprehensive unit capacity investment cost, plus the fixed configuration cost of the energy storage battery.

[0123] By adopting a method that combines the investment cost per unit capacity with the fixed configuration cost of energy storage, the expression for equipment purchase cost can be simplified, the number of investment cost variables in the model can be reduced, the nonlinearity of the mixed integer programming problem can be reduced, and thus the solution efficiency can be improved.

[0124] S232: Based on the probability of occurrence of each scenario in typical source load output scenarios, the total operating cost under each scenario is weighted and summed to construct the annualized operating cost.

[0125] By multiplying the total operating cost of the distribution network (including network loss cost, over-limit penalty cost, and self-balance deviation penalty cost) calculated under each typical scenario by the probability of occurrence corresponding to that scenario, and summing the results for all scenarios, an annualized expected value of operating cost considering source-load uncertainty is obtained. This allows the assessment of operating costs to reflect the probability of different scenarios occurring, avoiding the bias caused by relying solely on a single deterministic scenario for decision-making.

[0126] Right now:

[0127] In the formula, s LS , s VL and s MA Let Ω represent the total distribution network loss cost, over-limit penalty cost, and self-balancing deviation penalty cost in scenario s, respectively. s Represents a set of typical scenarios, ψ s This represents the actual probability of scenario s.

[0128]

[0129] In the formula, s,t LN,LS and s,t EP,LS Let c represent the line loss cost and the smart soft-switching loss cost at time t in scenario s, respectively. LS This represents the cost of energy loss per unit kilowatt-hour, calculated using the unit electricity price. Δt represents the time interval in the scenario, and R... ij P represents the resistance of line ij. r,s,t EP,LS This represents the active power loss of the intelligent soft switch using scheme r, in Ω. T This represents a set of time points (or time periods) in a typical scenario.

[0130]

[0131] In the formula, s,t VT,VL and s,t LN,VL c represents the penalty cost caused by the node voltage and line current exceeding the limit at time t in scenario s, respectively. VL This represents the unit load failure cost caused by exceeding limits, calculated using the unit electricity price, δ. i,s,t VT and δ ij,s,t LN Let Ω represent the voltage and current over-limit penalty coefficients at time t in scenario s, respectively. n Ω represents the set of nodes. n,ij Let represent the set of downstream nodes of line ij, let represent the set of lines in partition p, and let represent the set of nodes in partition p.

[0132]

[0133] In the formula, s,t MA,BP This represents the self-balancing deviation penalty cost caused by the active power transmitted along the interval boundary line at time t in scenario s. s,t MA,SP This represents the penalty cost incurred by the balancing node of feeder h exchanging power with the outside world at time t in scenario s. s MA ,SN c represents the penalty cost incurred by the balancing node in scenario s due to exchanging power with external systems. MA P represents the penalty price coefficient caused by partition or system imbalance deviation. ij,s,t Ω represents the active power transmitted by line ij at time t in scenario s. h Represents the set of distribution network feeders, where h is the feeder index and P is the feeder index. h,s,tThis indicates that the balancing node of feeder h exchanges power to the outside.

[0134] The above is the sum of line loss cost, voltage over-limit penalty cost, line current over-limit penalty cost and self-balancing deviation penalty cost at each moment in any scenario, multiplied by the time interval and accumulated in time sequence to construct the operating cost for the corresponding scenario.

[0135] In addition, the operating cost for any given scenario can be constructed by multiplying the sum of the distributed power curtailment penalty cost, energy storage cycle aging cost, and carbon emission penalty cost at any given moment by a time interval and then summing them sequentially. Specifically, the distributed power curtailment penalty cost is calculated based on the product of the curtailed power and the unit curtailment penalty coefficient; the energy storage cycle aging cost is calculated based on the conversion relationship between the depth of charge / discharge and the unit cycle aging cost; and the carbon emission penalty cost is calculated based on the equivalent carbon emissions corresponding to electricity purchased from the upper-level grid and the unit carbon price.

[0136] By introducing curtailment penalty costs, models can be incentivized to prioritize the utilization of distributed renewable energy, thereby improving the efficiency of new energy sources. Introducing energy storage cycle aging costs can prevent overcharging and discharging of energy storage systems, extending their actual lifespan. Introducing carbon emission penalty costs can guide planning schemes to reduce reliance on high-carbon-emission upstream grid power, improving the low-carbon operation level of the distribution network.

[0137] S233: Based on the present value coefficient, the investment cost and operating cost of each future year are uniformly converted to the current base year, and based on the capital recovery coefficient, the one-time construction investment is equally allocated to each year within the investment recovery period. The investment cost and annualized operating cost are processed into equal-annual values ​​to obtain the objective function expressed in the form of equal-annual values.

[0138] By using a present value factor, the investment and operating costs incurred in the y-th year are uniformly discounted to the current base year, thus eliminating the impact of the time value of money on cross-year cost comparisons. Furthermore, by using a capital recovery factor, the one-time construction investment is equally distributed across each year of the investment recovery period, allowing the investment cost and annualized operating cost to be directly added under an equal-annual value basis, thereby constructing a unified objective function reflecting the economic efficiency throughout the entire life cycle.

[0139] Right now:

[0140] In the formula, Ω Y Let y represent the set of years, d represent the annual interest rate, L represent the investment payback period, and λ represent the investment payback period. y ε represents the present value factor, used to allocate costs incurred in different years to the current base year; ε represents the capital recovery factor, used to evenly distribute construction costs across the payback period years. COIndicates investment cost, OP This represents the annualized operating cost.

[0141] To facilitate the solution, the absolute value term in the operating cost is linearized as follows:

[0142] In the formula, K ij,s,t and H ij,s,t P is represented as an auxiliary variable introduced by linearization. max,uba and P min,uba These represent the upper and lower limits of the allowable power difference for the balancing unit, respectively, with M representing a very large positive value.

[0143]

[0144] In the formula, J h,s,t P is represented as an auxiliary variable introduced by linearization. i LD,N ω represents the rated active power of the load at node i. max,ep and ω min,ep These represent the upper and lower limits of the allowed external switching power, respectively, c MA This represents the penalty price factor (unit electricity price or penalty factor) resulting from zoning or system imbalance deviations, Ω. h Ω represents the set of feeders in a distribution network. n,h Let P represent the set of nodes on feeder h. h,s,t This represents the external power exchanged by the balancing node of feeder h at time t in scenario s (positive value indicates power received from the upstream grid, negative value indicates power sent to the upstream grid).

[0145]

[0146] In the formula, W h,s and w h,s,t It is an auxiliary variable introduced for linearization, ω max,en This represents the maximum allowable amount of electricity exchanged, Δt represents the time interval in the scenario, and Ω. T This represents a set of time points (or time periods) in a typical scenario.

[0147] Based on the objective function of minimizing the annual value, as another economic evaluation method, the objective function can also be constructed with the goal of maximizing the net present value within the planning period of the target distribution network. The net present value is defined as the present value of the sum of the reduced power outage loss cost, reduced network loss cost, and delayed grid upgrade investment due to the commissioning of the integrated smart soft switch of the energy storage system in each year of the planning period, minus the present value of the investment cost. Then, based on the planning parameters, typical source and load output scenario data, and distribution network zoning constraints, the feasible domain of the objective function is limited, and a set of constraints for flexible distribution networks is constructed.

[0148] By shifting the objective from cost minimization to net present value maximization, the comprehensive benefits of integrated smart soft-switching energy storage system investments can be quantified more intuitively. These benefits include reduced power outage losses, lower grid losses, and investment savings from delaying traditional grid upgrades. This allows for direct evaluation of the marginal returns of each investment at the planning level, avoiding over- or under-allocation. This approach is particularly suitable for scenarios where investment priorities need to be ranked among multiple competing projects, providing decision-makers with an optimization guide that aligns more closely with economic benefit evaluation practices.

[0149] After constructing the objective function, in order to limit the feasible domain of the objective function, it is necessary to further construct a set of constraints for the flexible distribution network.

[0150] Specifically, step S23 above may include S23a: Based on the candidate installation locations and candidate scheme set in the planning parameters, each candidate location is limited to selecting at most one planning scheme, and the binary variable representing whether a certain planning scheme is selected is associated with the line length required by the scheme, thus constructing a smart soft-switching planning scheme generation constraint for the integrated energy storage system.

[0151] This constraint restricts the selection of candidate construction schemes for integrated smart soft switches for energy storage systems. In actual planning, multiple different construction schemes for integrated smart soft switches for energy storage systems may exist at the same installation location, such as different numbers of ports, different capacity levels, or different topology connection methods. The core function of this constraint is to ensure that each candidate location can only be selected for a maximum of one construction scheme, avoiding duplicate investment or scheme conflicts.

[0152] In the specific construction, a binary variable is assigned to each planning scheme for each candidate location. A value of 1 indicates that the scheme is selected, and 0 indicates that it is not selected. A constraint is then imposed to ensure that the sum of the binary variables for all schemes at the same location is less than or equal to 1, thus ensuring that at most one scheme can be selected for each location. Simultaneously, this binary variable is multiplied and associated with the required line construction length for the corresponding scheme. When a scheme is selected, the line construction cost is included in the total investment; when a scheme is not selected, the line construction cost is zero. Furthermore, this constraint can be extended to limit the total number of selected schemes across all candidate locations to no more than a preset upper limit, thereby controlling the total construction scale of the integrated smart soft switch for energy storage systems.

[0153] S23b: Based on the candidate installation locations, equipment capacity upper and lower limits, and investment cost parameters in the planning parameters, the installation capacity range of AC-DC converters, DC-DC converters, and energy storage batteries is limited. The installation capacity of each component device is associated with the binary variable representing whether to deploy a smart soft switch, thus constructing the planning constraints for an integrated smart soft switch for the energy storage system.

[0154] This section presents a mathematical model for the deployment decision of an integrated smart soft switch for energy storage systems. First, based on the candidate installation locations and the upper and lower capacity limits of various converters and energy storage batteries given in the planning parameters, the installed capacity of AC-DC converters, DC-DC converters, and energy storage batteries is limited to their maximum allowable values. Simultaneously, if deployment is chosen, the capacity cannot be lower than the set minimum value. Second, the installed capacity of these three components is associated with a binary variable representing whether to deploy the smart soft switch. When the binary variable is 1, the capacity of each device can freely take values ​​within its allowable range; when the binary variable is 0, the capacity of each device must be zero, thus achieving a joint decision on whether to deploy and the resulting capacity.

[0155] Right now:

[0156] In the formula, S r,i CT S represents the capacity of the intelligent soft-switching port i converter in the planning scheme r. r CT S represents the total capacity of the intelligent soft-switching converter in the planning scheme r. CT,min and S CT,max S represents the maximum installed capacity of the intelligent soft-switching converter. r CH and S r BT These represent the installed capacity of the DC-DC converter and the energy storage battery, respectively, in Ω. a χ represents the set of candidate installation locations for intelligent soft switches.r It is a binary variable representing whether to use the intelligent soft switch of the planning scheme r, Ω r This represents the set of nodes connected to the intelligent soft switch planning scheme r.

[0157] S23c: Based on the time-series characteristics in typical source-load output scenario data, this paper describes the power balance relationship of each port of the intelligent soft switch and the time-series evolution relationship of the state of charge of the energy storage battery at each moment under each scenario. It also defines the power range of each port, the lower limit of converter loss and the boundary of the state of charge of the energy storage battery, and constructs the operation constraints of the integrated intelligent soft switch of the energy storage system.

[0158] This section describes the technical characteristics of the integrated smart soft switch in the energy storage system at the operational level. First, based on the time-series characteristics of typical source-load output scenario data, a balance relationship is established between the sum of active power injected into each port of the smart soft switch and the active power on the DC-DC converter side. Simultaneously, the time-series recursive change law of the energy storage battery's state of charge with charging and discharging power is described. Second, the active and reactive power injected into each port are limited to not exceeding the corresponding converter's installed capacity. The converter's active power loss is limited to not less than a certain value determined by its transmission power (i.e., the lower limit of loss), and the energy storage battery's state of charge is always maintained between the allowable minimum and maximum values. This ensures that the smart soft switch's operational state meets physical feasibility requirements in any scenario and at any time.

[0159] Right now:

[0160] In the formula, N T P represents the number of points in a typical scenario. r,i,s,t CT and Q r,i,s,t CT P represents the active power and reactive power injected into port i of the intelligent soft switch with scheme r at time t in scenario s, respectively. r,s,t CH P represents the active power injected by the DC-DC converter with scheme r at time t in scenario s. r,i,s,t CT,LS and P r,s,t CH,LS Let P represent the losses at port i of the AC-DC converter and the DC-DC converter of scheme r at time t in scenario s, respectively. r,s,t EP,LS κ represents the total loss of the intelligent soft switch with scheme r at time t in scenario s. ac and κ dc These represent the loss coefficients of the AC-DC converter and the DC-DC converter, respectively. int , min and max These represent the initial, minimum, and maximum states of charge of the energy storage battery, respectively, Ω. n,r Let g represent the set of nodes connected to the intelligent soft switch in the planning scheme r, and g represent the index of the current time in the time-series accumulation, with a value ranging from 0 to N. T-1 S r BT This represents the installed capacity of the energy storage battery in the planning scheme r.

[0161] S23d: Based on the node affiliation and line topology within each distributed balance unit constrained by the distribution network zoning constraints, it describes the power balance relationship of each node at each time under each scenario and the analytical relationship between voltage and power at both ends of the line. It introduces the square variables of line current and node voltage, and uses second-order cone form inequality constraints to replace the nonlinear AC power flow equations. It also limits the square of each node voltage and the square of each line current to be maintained within the preset safety upper and lower limits, respectively, to construct flexible distribution network operation constraints.

[0162] This section describes the AC power flow characteristics and safety boundaries of a flexible distribution network at the operational level. First, based on the node affiliation relationships and line topology within each distributed balance unit constrained by the distribution network zoning constraints, the active and reactive power balance relationships of each node at each time point under different scenarios are established (i.e., the sum of injected power equals the sum of outflow power). Simultaneously, the analytical relationship between the node voltages at both ends of the line and the line transmission power is described. Second, to avoid the difficulty of solving nonlinear AC power flow equations directly, line current square variables and node voltage square variables are introduced, transforming the original quadratic power flow equations into second-order cone-shaped inequality constraints. These constraints are mathematically equivalent to the original power flow equations but exhibit convexity. Finally, the square values ​​of each node voltage are constrained to remain between preset safe upper and lower voltage limits, while the square values ​​of each line current are constrained to not exceed the upper limit of the line's thermal stability capacity, thereby ensuring that the planning scheme can meet the safe operation requirements of the distribution network under any scenario.

[0163] Right now:

[0164] In the formula, P ij,s,t and Q ij,s,t Q represents the active power and reactive power transmitted by line ij at time t in scenario s, respectively. jk,s,t Let X represent the reactive power transmitted by line jk at time t in scenario s, where jk represents the set of all downstream lines originating from node j and flowing to adjacent node k. ij It is the reactance of line ij, l ij,s,t =I ij,s,t 2V represents the square of the current in line ij at time t in scenario s. i,s,t =V i,s,t 2 P represents the square of the voltage at node i at time t in scene s. j,s,t and Q j,s,t Ω represents the active power and reactive power injected at node j, respectively. b This represents a set of distribution network lines.

[0165] S23e: The constraints of distribution network zoning, intelligent soft-switching planning scheme of integrated energy storage system, intelligent soft-switching planning constraints of integrated energy storage system, intelligent soft-switching operation constraints of integrated energy storage system, and flexible distribution network operation constraints are determined as the flexible distribution network constraint set.

[0166] That is, the above three types of constraints are combined and formally confirmed as a set of constraints that limit the feasible region of the objective function.

[0167] The constraint set of a flexible distribution network is a mathematical expression system composed of equality constraints and inequality constraints. The known quantities include network parameters (node ​​set, line set, line resistance and reactance, line thermal stability capacity), planning parameters (candidate installation location, unit capacity investment cost of equipment, discount rate, investment payback period, upper and lower limits of equipment capacity), and typical source and load output scenario data (photovoltaic output curve, load curve, and scenario occurrence probability at each time under each scenario). The unknown quantities to be solved include binary variables representing partition affiliation (node ​​affiliation variable, line affiliation variable, whether the line is a boundary line variable), binary variables representing whether to install smart soft switches, continuous variables representing the installed capacity of AC-DC converters, DC-DC converters, and energy storage batteries, and distribution network operation state variables at each time under each scenario (node ​​voltage square, line current square, line active and reactive power, power injected into each port of smart soft switches, converter losses, energy storage charging and discharging power, and state of charge, etc.).

[0168] Furthermore, based on the candidate installation locations and discrete capacity levels in the planning parameters, integer variables can be used to characterize the selection of AC-DC converters, DC-DC converters, and energy storage batteries installed at each candidate location, while limiting the installed capacity of each device to not exceed its corresponding maximum capacity level, thus constructing intelligent soft-switching planning constraints based on discrete capacity selection. Based on the time-series characteristics in typical source-load output scenario data, a linearized energy storage loss model is used to approximate the charging and discharging efficiency of the energy storage system, and a piecewise linearization method is used to approximate the relationship between converter losses and transmission power, while limiting the state of charge of the energy storage battery at the beginning and end of the scheduling cycle. The system constructs intelligent soft-switching operation constraints based on approximate linearization; based on the node affiliation and line topology within each distributed balance unit defined by the distribution network zoning constraints, it adopts a DC power flow model that ignores line resistance and reactance to describe the active power balance relationship of each node, uses linear inequality constraints to limit the voltage phase angle difference of each node to not exceed the preset safety upper limit, and ignores line current constraints and reactive power constraints to construct simplified flexible distribution network operation constraints; the intelligent soft-switching planning constraints based on discrete capacity selection, the intelligent soft-switching operation constraints based on approximate linearization, and the simplified flexible distribution network operation constraints are determined as the flexible distribution network constraint set.

[0169] By replacing continuous capacity variables with discrete capacity levels, replacing nonlinear loss cone constraints with piecewise linearization approximation, and replacing AC power flow second-order cone constraints with DC power flow, the original mixed-integer second-order cone programming problem can be transformed into a mixed-integer linear programming problem. This significantly improves the solution speed and reduces the solver memory usage while sacrificing some modeling accuracy. It is suitable for engineering planning scenarios that require large-scale candidate location traversal or rapid acquisition of approximate optimal solutions.

[0170] Based on the above simplified linearization method to quickly obtain approximate solutions, for scenarios that require high-precision solutions and can accept a certain amount of computational overhead, the following iterative solution method based on convexity programming can also be used to ensure solution accuracy.

[0171] S24: Decompose the non-convex constraints in the flexible distribution network constraint set into the difference between the first convex function and the second convex function; perform successive linearization iterative approximation on the concave terms in the difference at the solution of the previous iteration, solve the flexible distribution network constraint set, and obtain the topology and capacity configuration scheme of the integrated smart soft switch of the energy storage system to guide the physical network structure upgrade of the target distribution network.

[0172] Specifically, the above-mentioned solution process, which decomposes non-convex constraints into the difference of convex functions and approximates them by successive linearization, is carried out in the following manner.

[0173] Based on the second-order cone constraint in the flexible distribution network constraint set, a convex relaxation technique is used for initial solution to obtain the initial solution of the distribution network state variables. Based on the initial solution, the first cone deviation value representing the relaxation deviation of the distribution network power flow constraint and the second cone deviation value representing the relaxation deviation of the intelligent soft-switching converter loss constraint are calculated. When the first cone deviation value or the second cone deviation value is greater than a preset convergence threshold, the original nonlinear cone constraint is transformed into the form of the difference between the first convex function and the second convex function, where the second convex function is a linear approximation of the first convex function at the solution of the previous iteration. A penalty term is added to the optimization model with the objective of minimizing the sum of investment cost and operating cost. The penalty term is associated with the relaxation variable of the difference between the first convex function and the second convex function, and a modified optimization model is constructed for iterative solution. The coefficient of the penalty term is updated and the iteration is repeated until both the first cone deviation value and the second cone deviation value satisfy the convergence condition, at which point the iteration stops.

[0174] Based on the second-order cone constraint in the constraint set of flexible distribution network, the convex relaxation technique is used for initial solution to obtain the initial solution of the distribution network state variables.

[0175] Here, we first ignore the strictly nonlinear characteristics of some non-convex constraints in the constraint set of the flexible distribution network, and directly relax the rotated second-order cone constraints (such as the cone constraints in the AC power flow equations and the converter loss cone constraints) to the standard second-order cone form, thus transforming the original mixed-integer non-convex programming problem into a mixed-integer second-order cone programming problem. We then use commercial solvers (such as MOSEK, CPLEX, GUROBI, etc.) to solve this relaxed convex optimization problem, obtaining a set of initial numerical solutions for the distribution network operating state variables (including the square of node voltage, the square of line current, line active and reactive power, converter injected power, and losses). This initial solution is the starting point for subsequent iterative solutions, and the magnitude of its cone relaxation deviation determines whether further convexity iterative correction is needed.

[0176] Based on the initial solution, the first cone deviation value characterizing the power flow constraint relaxation deviation of the distribution network and the second cone deviation value characterizing the loss constraint relaxation deviation of the smart soft-switching converter are calculated.

[0177] This quantifies the extent to which the initial solution deviates from the original non-convex constraints. Specifically, the squared values ​​of the currents of each line, the squared values ​​of the node voltages, and the transmitted active and reactive power from the initial solution are substituted into the original expression for the power flow constraints of the distribution network. The maximum difference between the left and right sides of the equation is calculated as the first cone deviation value. Simultaneously, the injected active and reactive power and the corresponding loss values ​​at each converter port from the initial solution are substituted into the original expression for the converter loss constraints. The maximum difference between the left and right sides of the equation is calculated as the second cone deviation value. These two deviation values ​​reflect the degree of deviation between the approximate solution after convex relaxation and the original non-convex constraints, and serve as the basis for determining whether subsequent iterative corrections are needed.

[0178] Right now:

[0179] In the formula, σ PF The cone deviation σ represents the power flow constraint of the distribution network. LS The cone deviation, x, represents the loss constraint of a smart soft-switching AC-DC converter. * This represents the initial solution of the distribution network state variables obtained by solving a second-order cone programming problem. ij,s,t * Let v represent the initial solution of the square of the current in line ij at time t in scenario s. i,s,t * P represents the initial solution of the squared voltage at node i at time t in scenario s. ij,s,t * and Q ij,s,t * Let P represent the initial solutions for the active and reactive power transmitted by line ij at time t in scenario s, respectively. r,i,s,t CT,LS,* P represents the initial solution of the loss at port i of the AC-DC converter with scheme r at time t in scenario s. r,i,s,t CT,* and Q r,i,s,t CT,* Let κ represent the initial solutions for active and reactive power injected into port i of the intelligent soft switch with scheme r at time t in scenario s, respectively. ac This represents the loss factor of the AC-DC converter.

[0180] When the first cone deviation value or the second cone deviation value is greater than the preset convergence threshold, the original nonlinear cone constraint is transformed into the form of the difference between the first convex function and the second convex function, where the second convex function is a linear approximation of the first convex function at the solution of the previous iteration.

[0181] This is used to reconstruct the original non-convex constraints when the cone deviation does not meet accuracy requirements. Specifically, the original nonlinear cone constraints (such as the rotated second-order cone constraints or converter loss cone constraints in the power flow equations of a distribution network) are rewritten as the difference between two convex functions. The first convex function maintains the original nonlinear structure, while the second convex function is a linear approximation of the first convex function at the solution of the previous iteration, based on its first-order Taylor expansion. Through this transformation of subtracting its linear approximation from the convex function, the original non-convex constraint is transformed into a convex constraint condition, allowing convex optimization techniques to be used to solve it in each iteration. Simultaneously, the true boundary of the original non-convex constraint is approximated by continuously updating the linearization points.

[0182] Right now:

[0183] In the formula, w PF and u PF This represents a quadratic convex function defined according to the power flow constraints of the distribution network.

[0184]

[0185] In the formula, x m Let u represent the solution of the distribution network state variables in the m-th iteration. PF In solving x m Approximate linearization at point m, where m represents the iteration number and l ij,s,t m Let v represent the m-th iteration solution of the square of the current in line ij at time t in scenario s. i,s,t m P represents the m-th iteration solution of the square voltage at node i at time t in scenario s. ij,s,t m and Q ij,s,t m Let represent the m-th iteration solution of the active power and reactive power transmitted by line ij at time t in scenario s, respectively.

[0186]

[0187] In the formula, w LS and u LS This represents a quadratic convex function defined according to the loss constraints of a smart soft-switching AC-DC converter.

[0188]

[0189] In the formula, P r,i,s,t CT,m and Q r,i,s,t CT,m Let represent the m-th iteration solution of the active power and reactive power injected into port i of the intelligent soft switch with scheme r at time t in scenario s, respectively.

[0190] A penalty term is added to the optimization model that aims to minimize the sum of investment cost and operating cost. The penalty term is associated with the slack variable of the difference between the first convex function and the second convex function. The modified optimization model is then constructed and solved iteratively.

[0191] This is used to reconstruct the original non-convex constraints when the cone deviation does not meet accuracy requirements. Specifically, the original nonlinear cone constraints (such as the rotated second-order cone constraints or converter loss cone constraints in the power flow equations of a distribution network) are rewritten as the difference between two convex functions. The first convex function maintains the original nonlinear structure, while the second convex function is a linear approximation of the first convex function at the first-order Taylor expansion of the solution in the previous iteration. Through this transformation of subtracting its linear approximation from the convex function, the original non-convex constraints are relaxed to the form of a convex constraint and a penalized non-negative relaxation variable. This allows convex optimization techniques to be used to solve the problem in each iteration. At the same time, by adding a penalty term related to the relaxation variable to the objective function, the relaxation variable is driven to gradually approach zero. Thus, by continuously updating the linearization point and the penalty coefficient, the true boundary of the original non-convex constraints is approximated.

[0192] Right now:

[0193] In the formula, ρ m It is the coefficient of the penalty term in the m-th iteration. ij,s,t PF It is the penalty term variable for line ij at time t in scenario s. r,i,s,t LS Let x be the penalty term variable of the intelligent soft switch port i with scheme r at time t in scenario s. Solving for x yields x. m +1 Verify σ in the previous formula PF and σ LS If the convergence condition is met, stop iterating; otherwise, update ρ. m+1 =min{μρ m ,ρ max}, where μ is the growth rate of the penalty term coefficient, and ρ max It is the maximum value of the penalty term coefficient.

[0194] Update the penalty term coefficients and repeat the iteration until both the first cone deviation value and the second cone deviation value satisfy the convergence condition, then stop the iteration.

[0195] This mechanism controls the advancement and termination of the iterative process. After each iteration, the first and second cone deviation values ​​are calculated based on the current iteration's solution, and it is determined whether both are less than a preset convergence threshold. If the threshold is met, the iteration stops, and the partition boundary data and the deployment and capacity data of the integrated smart soft switch of the energy storage system obtained in the current iteration are output as the final planning scheme. If the threshold is not met, the penalty term coefficient is updated according to a preset growth rate (while ensuring that it does not exceed a preset maximum value). The solution from the previous iteration is used as the new linearization point, the modified optimization model is reconstructed, and the next iteration is performed until the cone deviation meets the convergence condition. This mechanism ensures a balance between accuracy and computational efficiency in the solution process.

[0196] Of course, based on the second-order cone constraints in the flexible distribution network constraint set, the convex relaxation technique can be used for initial solution to obtain the initial solution of the distribution network state variables. Based on the initial solution, the first cone deviation value representing the relaxation deviation of the distribution network power flow constraint and the second cone deviation value representing the relaxation deviation of the intelligent soft-switching converter loss constraint are calculated. When the first cone deviation value or the second cone deviation value is greater than the preset convergence threshold, the original nonlinear cone constraint is equivalently rewritten as the difference between the first convex function and the second convex function, where the second convex function is a linear approximation of the second-order Taylor expansion of the first convex function at the solution of the previous iteration, rather than a linear approximation of the first-order Taylor expansion. A penalty term is added to the optimization model with the goal of minimizing the sum of investment cost and operating cost. The penalty term is associated with the relaxation variable of the difference between the first convex function and the second convex function, and an additional regularization term related to the second-order Hessian matrix of the first convex function is added. The modified optimization model is constructed and iteratively solved. The coefficient of the penalty term is updated and the iteration is repeated until the first cone deviation value and the second cone deviation value both meet the convergence condition, and the iteration stops.

[0197] By replacing the first-order Taylor expansion linear approximation with a second-order Taylor expansion linear approximation and introducing a regularization term related to the Hessian matrix, the curvature characteristics of the original non-convex constraint can be more accurately approximated in each iteration, thereby achieving higher approximation accuracy with the same number of iterations, reducing the number of iterations and accelerating the convergence speed. This is especially suitable for scenarios where the curvature of the original non-convex constraint changes drastically.

[0198] This concludes the description of the collaborative planning method for flexible distribution networks provided in the embodiments of this application.

[0199] Based on the same inventive concept, embodiments of this application also provide a collaborative planning device for flexible power distribution networks.

[0200] Figure 3 This is a schematic diagram of the structure of the collaborative planning device for flexible power distribution networks in the embodiments of this application. Figure 1 See Figure 3As shown, the device may include: The acquisition module 31 is used to acquire the network structure parameters, planning parameters, and typical scenario data of source load output of the target distribution network.

[0201] The first construction module 32 is used to construct an analytical expression of the distribution network partition constraints based on the network parameters and with preset partition conditions as constraints.

[0202] The second construction module 33 is used to construct a flexible distribution network constraint set based on planning parameters and typical source-load output scenario data, with the goal of minimizing the sum of investment cost and operating cost of the target distribution network. The flexible distribution network constraint set includes at least distribution network zoning constraints, intelligent soft-switching planning scheme generation constraints for integrated energy storage systems, intelligent soft-switching planning constraints for integrated energy storage systems, intelligent soft-switching operation constraints for integrated energy storage systems, and flexible distribution network operation constraints.

[0203] Planning module 34 is used to solve the constraint set of the flexible distribution network to obtain the topology and capacity configuration scheme of the integrated intelligent soft switch of the energy storage system, so as to guide the physical network structure upgrade of the target distribution network.

[0204] Furthermore, as a response to Figure 3 In addition to the refinement and expansion of the illustrated device, this application embodiment also provides a collaborative planning device for flexible power distribution networks.

[0205] Figure 4 This is a schematic diagram of the structure of the collaborative planning device for flexible power distribution networks in the embodiments of this application. Figure 2 See Figure 4 As shown, the device may include: The acquisition module 41 is used to acquire the network structure parameters, planning parameters, and typical scenario data of source load output of the target distribution network.

[0206] The first construction module 42 is used to define binary variables that characterize the node partition affiliation, line partition affiliation, and whether a line is a partition boundary line, based at least on the node set, line set, and topology connection relationship in the network parameters. Using binary variables as carriers, and with constraints such as ensuring the unique affiliation of nodes and lines, maintaining the balance of partition size, and ensuring the radial topology structure within the partition, the distribution network partition constraints are constructed in an analytical expression.

[0207] The first construction module 42 is specifically used to construct partition boundary constraints based on the set of lines and the connectivity parameters of the lines, using binary variables to represent whether each line is a partition boundary line, setting that the partition boundary line does not belong to any partition, and maintaining the partition boundary line as a physical line used to transmit partition connection power; based on the set of nodes, the set of lines, and the topology connection relationship, binary variables are used to represent whether each node belongs to a certain partition and whether each line belongs to a certain partition, setting that any node belongs to only one partition, setting that any non-boundary line belongs to only one partition, and setting that a non-boundary line and the nodes at both ends belong to the same partition. First, a partition is defined, and a unique ownership constraint is established for nodes and lines. Based on the node set and the preset number of partitions, the number of nodes in each partition is set to fall within a preset equilibrium range to avoid the generation of single clusters, super-large clusters, or isolated node clusters, thus establishing a partition size equilibrium constraint. Based on the node set, line set, and the partition ownership relationship of each node and line, the difference between the number of nodes and the number of lines in each partition is set to be greater than or equal to 0 and less than or equal to 1, thus establishing a radial topology constraint within the partition. The partition boundary constraint, the unique ownership constraint for nodes and lines, the partition size equilibrium constraint, and the radial topology constraint within the partition are defined as the distribution network partition constraints.

[0208] The second building module 43 includes a first building unit 431 and a second building unit 432.

[0209] The first building unit 431 is used to build an objective function with the goal of minimizing the sum of the investment cost over the entire life cycle of the target distribution network and the annualized operating cost calculated based on typical source and load output scenario data.

[0210] The second building unit 432 is used to construct a set of constraints for a flexible distribution network based on planning parameters and typical source load output scenario data to limit the feasible domain of the objective function.

[0211] The second building unit 432 is specifically used to construct the investment cost by summing the site development cost, equipment purchase cost, and line construction cost based on the intelligent soft switch; to construct the annualized operating cost by weighted summing of the total operating cost under each scenario based on the probability of occurrence of each scenario in the typical source-load output scenario; to uniformly convert the investment cost and operating cost of each future year to the current base year based on the present value coefficient, and to equally allocate the one-time construction investment to each year within the investment recovery period based on the capital recovery coefficient, and to perform equal-annual value processing on the investment cost and annualized operating cost to obtain the objective function expressed in the form of equal-annual value.

[0212] The second building unit 432 is also used to build equipment purchase costs based on the sum of the products of the installed capacity of the AC-DC converter, the DC-DC converter and the energy storage battery and the corresponding unit capacity investment cost; and to build operating costs for the corresponding scenario based on the sum of the line loss cost, voltage over-limit penalty cost, line current over-limit penalty cost and self-balancing deviation penalty cost at each time in any scenario, multiplied by the time interval and accumulated in time sequence.

[0213] The second building unit 432 is specifically used to generate constraints for the integrated smart soft switch planning scheme of the energy storage system, based on the candidate installation locations and candidate scheme set in the planning parameters, limiting each candidate location to a maximum of one planning scheme, and associating the binary variable representing whether a certain planning scheme is selected with the line length required by the scheme. Based on the candidate installation locations, equipment capacity upper and lower limits, and investment cost parameters in the planning parameters, it limits the installation capacity range of the AC-DC converter, DC-DC converter, and energy storage battery, and associates the installation capacity of each component with the binary variable representing whether to deploy a smart soft switch, thus constructing planning constraints for the integrated smart soft switch of the energy storage system. Based on the time-series characteristics in the source-load output typical scenario data, it describes the power balance relationship of each port of the smart soft switch and the time-series evolution relationship of the energy storage battery's state of charge at each time point in each scenario, and limits the power of each port. The system constructs integrated intelligent soft-switching operation constraints for the energy storage system, including the range, lower limit of converter losses, and state-of-charge boundary of energy storage batteries. Based on the node affiliation and line topology within each distributed balance unit defined by the distribution network zoning constraints, it describes the power balance relationship of each node at each time point under each scenario and the analytical relationship between voltage and power at both ends of the line. It introduces the square variables of line current and node voltage, and uses second-order cone-form inequality constraints to replace the nonlinear AC power flow equations. It also limits the square of each node voltage and the square of each line current to be maintained within the preset safety upper and lower limits, respectively, thus constructing flexible distribution network operation constraints. The distribution network zoning constraints, the intelligent soft-switching planning scheme generation constraints of the integrated energy storage system, the intelligent soft-switching planning constraints of the integrated energy storage system, the intelligent soft-switching operation constraints of the integrated energy storage system, and the flexible distribution network operation constraints are determined as the flexible distribution network constraint set.

[0214] Planning module 44 is used to decompose the non-convex constraints in the flexible distribution network constraint set into the difference between the first convex function and the second convex function; to perform successive linearization iterative approximation on the concave terms in the difference at the solution of the previous iteration, and to solve the flexible distribution network constraint set to obtain the topology and capacity configuration scheme of the integrated smart soft switch of the energy storage system, so as to guide the physical network structure upgrade of the target distribution network.

[0215] The planning module 44 includes an initial solution unit 441, a deviation calculation unit 442, a convexity transformation unit 443, a penalty correction unit 444, and an iterative convergence unit 445.

[0216] The initial solution unit 441 is used to perform initial solution based on the second-order cone constraints in the flexible distribution network constraint set, using convex relaxation techniques to obtain the initial solution of the distribution network state variables.

[0217] The deviation calculation unit 442 is used to calculate, based on the initial solution, the first cone deviation value characterizing the power flow constraint relaxation deviation of the distribution network and the second cone deviation value characterizing the loss constraint relaxation deviation of the smart soft-switching converter; The convexity conversion unit 443 is used to convert the original nonlinear cone constraint into the difference between the first convex function and the second convex function when the first cone deviation value or the second cone deviation value is greater than the preset convergence threshold. The second convex function is a linear approximation of the first convex function at the first Taylor expansion of the solution in the previous iteration. The penalty correction unit 444 is used to add a penalty term to the optimization model with the goal of minimizing the sum of investment cost and operating cost. The penalty term is associated with the slack variable of the difference between the first convex function and the second convex function, and the corrected optimization model is constructed for iterative solution. Iterative convergence unit 445 is used to update the penalty term coefficient and repeat the iteration until both the first cone deviation value and the second cone deviation value meet the convergence condition, at which point the iteration stops.

[0218] It should be noted that the description of the above device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0219] Based on the same inventive concept, this application also provides a computer device.

[0220] Figure 5 This is a schematic diagram of the structure of the computer device in an embodiment of this application. See also... Figure 5 As shown, the computer device may include: a memory 51, a processor 52, and a computer program stored on the memory 51, wherein the processor 52 executes the computer program to implement the methods described in the foregoing embodiments.

[0221] It should be noted that the description of the above computer device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the computer device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0222] Based on the same inventive concept, embodiments of this application also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the methods described in the foregoing embodiments.

[0223] It should be noted that the description of the above computer-readable storage medium embodiments is similar to the description of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the computer-readable storage medium embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0224] Based on the same inventive concept, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the methods described in the foregoing embodiments.

[0225] It should be noted that the descriptions of the above computer program product embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the computer program product embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0226] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A collaborative planning method for a flexible distribution network, characterized in that, The method includes: Obtain the network structure parameters, planning parameters, and typical source load output data of the target distribution network; Based on the network parameters, and with preset partitioning conditions as constraints, an analytical expression of the distribution network partitioning constraints is constructed. With the goal of minimizing the sum of investment cost and operating cost of the target distribution network, a flexible distribution network constraint set is constructed based on the planning parameters and the typical source-load output scenario data. The flexible distribution network constraint set includes at least the distribution network zoning constraints, the intelligent soft-switching planning scheme generation constraints of the integrated energy storage system, the intelligent soft-switching planning constraints of the integrated energy storage system, the intelligent soft-switching operation constraints of the integrated energy storage system, and the flexible distribution network operation constraints. Solving the constraint set of the flexible distribution network yields the topology and capacity configuration scheme of the integrated intelligent soft switch of the energy storage system, which guides the physical network structure upgrade of the target distribution network.

2. The method according to claim 1, characterized in that, The step of constructing an analytical expression of distribution network zoning constraints based on the network parameters and using preset zoning conditions as constraints includes: Based at least on the node set, line set, and topology connection relationship in the network parameters, define binary variables to characterize node partition affiliation, line partition affiliation, and whether a line is a partition boundary line; Using the binary variables as carriers, and with constraints such as ensuring the unique ownership of nodes and lines, maintaining the balance of partition size, and ensuring the radial topology within the partition, an analytical expression of the distribution network partition constraints is constructed.

3. The method according to claim 2, characterized in that, The method of constructing analytically expressed distribution network partition constraints, using the binary variables as carriers and constrained by ensuring unique attribution of nodes and lines, maintaining balanced partition size, and guaranteeing a radial topology within a partition, includes: Based on the set of lines and the connectivity parameters of the lines, binary variables are used to represent whether each line is a partition boundary line. The partition boundary line is set to not belong to any partition, and the partition boundary line is kept as a physical line and used to transmit partition communication power, thus constructing partition boundary constraints. Based on the node set, line set, and topology connection relationship, binary variables are used to represent whether each node belongs to a certain partition and whether each line belongs to a certain partition. It is set that any node belongs to only one partition, any non-boundary line belongs to only one partition, and non-boundary lines and their two end nodes belong to the same partition, thus constructing a unique ownership constraint for nodes and lines. Based on the node set and the preset number of partitions, the number of nodes contained in each partition is set to fall within a preset balance range in order to avoid generating a single cluster, an oversized cluster or an isolated node cluster, and to construct a partition size balance constraint. Based on the node set, the line set, and the partition affiliation of each node and line, the difference between the number of nodes and the number of lines in each partition is set to be greater than or equal to 0 and less than or equal to 1, and a radial topological constraint is constructed within the partition. The partition boundary constraints, the unique attribution constraints of nodes and lines, the partition size balancing constraints, and the radial topology constraints within the partitions are determined as the distribution network partition constraints.

4. The method according to claim 1, characterized in that, The objective is to minimize the sum of investment and operating costs of the target distribution network. Based on the planning parameters and typical source-load output scenario data, a set of constraints for the flexible distribution network is constructed, including: An objective function is constructed with the goal of minimizing the sum of the investment cost over the entire life cycle of the target distribution network and the annualized operating cost calculated based on typical source-load output scenario data. Based on the planning parameters and the typical scenario data of the source load output, the feasible domain of the objective function is defined, and the constraint set of the flexible distribution network is constructed.

5. The method according to claim 4, characterized in that, An objective function is constructed with the goal of minimizing the sum of the investment cost over the entire lifecycle of the target distribution network and the annualized operating cost calculated based on typical source-load output scenario data. This function includes: The investment cost is calculated based on the sum of the site development cost, equipment purchase cost, and line construction cost of the aforementioned intelligent soft switch. Based on the probability of occurrence of each scenario in the typical source load output scenarios, the total operating cost under each scenario is weighted and summed to construct the annualized operating cost; Based on the present value factor, the investment costs and operating costs of each future year are uniformly converted to the current base year, and based on the capital recovery factor, the one-time construction investment is equally allocated to each year within the investment recovery period. The investment costs and the annualized operating costs are processed into equal-annual values ​​to obtain the objective function expressed in the form of equal-annual values.

6. The method according to claim 1, characterized in that, The method further includes: The equipment purchase cost is calculated by summing the products of the installed capacity of the AC-DC converter, the DC-DC converter, and the energy storage battery, and the corresponding unit capacity investment cost. The operating cost for any given scenario is constructed by multiplying the sum of line loss cost, voltage over-limit penalty cost, line current over-limit penalty cost, and self-balancing deviation penalty cost at each moment by a time interval and then summing them sequentially.

7. The method according to claim 4, characterized in that, The flexible distribution network constraint set is constructed based on the planning parameters, the typical scenario data of source load output, and the distribution network zoning constraints to limit the feasible domain of the objective function, including: Based on the candidate installation locations and candidate scheme set in the planning parameters, each candidate location is limited to selecting at most one planning scheme, and the binary variable representing whether a certain planning scheme is selected is associated with the line length required by the scheme, thus constructing a smart soft-switching planning scheme generation constraint for the integrated energy storage system. Based on the candidate installation locations, equipment capacity limits and investment cost parameters in the planning parameters, the installation capacity range of AC-DC converters, DC-DC converters and energy storage batteries is limited, and the installation capacity of each component equipment is associated with the binary variable representing whether to deploy the smart soft switch, thus constructing an integrated smart soft switch planning constraint for the energy storage system. Based on the time-series characteristics in the typical source-load output scenario data, the power balance relationship of each port of the intelligent soft switch and the time-series evolution relationship of the state of charge of the energy storage battery at each moment under each scenario are described. The power range of each port, the lower limit of converter loss and the boundary of the state of charge of the energy storage battery are defined to construct the integrated intelligent soft switch operation constraints of the energy storage system. Based on the node affiliation and line topology within each distributed balance unit defined by the distribution network zoning constraints, the power balance relationship of each node at each time under each scenario and the analytical relationship between voltage and power at both ends of the line are described. Line current square variables and node voltage square variables are introduced, and second-order cone form inequality constraints are used to replace nonlinear AC power flow equations. The square of each node voltage and the square of each line current are limited to maintain within the preset safety upper and lower limits, respectively, to construct flexible distribution network operation constraints. The distribution network zoning constraints, the intelligent soft-switching planning scheme generation constraints of the integrated energy storage system, the intelligent soft-switching planning constraints of the integrated energy storage system, the intelligent soft-switching operation constraints of the integrated energy storage system, and the flexible distribution network operation constraints are determined as the flexible distribution network constraint set.

8. The method according to any one of claims 1 to 7, characterized in that, Solving the constraint set of the flexible distribution network includes: The non-convex constraints in the flexible distribution network constraint set are decomposed into the difference between a first convex function and a second convex function; The concave terms in the difference are successively linearized iteratively approximated at the solution of the previous iteration to solve the constraint set of the flexible distribution network.

9. The method according to claim 8, characterized in that, The step of decomposing the non-convex constraints in the flexible distribution network constraint set into the difference between a first convex function and a second convex function includes: Based on the second-order cone constraints in the set of constraints for the flexible distribution network, the initial solution of the distribution network state variables is obtained by using the convex relaxation technique. Based on the initial solution, calculate the first cone deviation value characterizing the power flow constraint relaxation deviation of the distribution network and the second cone deviation value characterizing the loss constraint relaxation deviation of the smart soft-switching converter. When the first cone deviation value or the second cone deviation value is greater than the preset convergence threshold, the original nonlinear cone constraint is transformed into the difference between the first convex function and the second convex function, wherein the second convex function is a linear approximation of the first convex function at the solution of the previous iteration. The step of performing successive linearized iterative approximation on the concave terms in the difference at the solution of the previous iteration, and solving for the constraint set of the flexible distribution network, includes: A penalty term is added to the optimization model that aims to minimize the sum of the investment cost and the operating cost. The penalty term is associated with the slack variable of the difference between the first convex function and the second convex function. The modified optimization model is then constructed and solved iteratively. Update the penalty term coefficients and repeat the iteration until both the first cone deviation value and the second cone deviation value satisfy the convergence condition, then stop the iteration.

10. A collaborative planning device for a flexible power distribution network, characterized in that, The device includes: The acquisition module is used to acquire the network structure parameters, planning parameters, and typical source load output data of the target distribution network. The first construction module is used to construct an analytical expression of the distribution network partition constraints based on the network parameters and with preset partition conditions as constraints. The second construction module is used to construct a flexible distribution network constraint set based on the planning parameters and the typical source-load output scenario data, with the goal of minimizing the sum of the investment cost and operating cost of the target distribution network. The flexible distribution network constraint set includes at least the distribution network zoning constraints, the intelligent soft switch planning scheme generation constraints of the integrated energy storage system, the intelligent soft switch planning constraints of the integrated energy storage system, the intelligent soft switch operation constraints of the integrated energy storage system, and the flexible distribution network operation constraints. The planning module is used to solve the constraint set of the flexible distribution network to obtain the topology and capacity configuration scheme of the integrated intelligent soft switch of the energy storage system, so as to guide the physical network structure upgrade of the target distribution network.

11. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images