Distribution Network Energy Storage Device Topology Configuration and Optimization Method Based on Source-Load Topology Characteristics
By constructing a weighted graph model and calculating source-load topology characteristic indicators, the energy storage configuration scheme is optimized, which solves the problems of low accuracy in energy storage node identification and insufficient dynamic adaptability in the distribution network, and achieves more efficient resource utilization and operational flexibility.
Patent Information
- Application Number
- CN202611315317.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-27
- Publication Date
- 2026-11-17
AI Technical Summary
Existing energy storage configuration schemes for distribution networks are unable to accurately identify key topology nodes, resulting in resource allocation imbalances and insufficient dynamic adaptability, and failing to effectively solve the problem of spatiotemporal mismatch between photovoltaic output and load demand.
By constructing a weighted graph model of the distribution network, calculating source-load topology characteristic indicators, generating a set of candidate nodes and a set of energy storage topology patterns, and combining the objective function and hard constraints for optimization, the energy storage configuration results are obtained, including the optimized configuration of site selection, capacity and topology.
It improves the accuracy of energy storage node identification and utilization efficiency in the distribution network, enhances the resource utilization and operational flexibility of the distribution network, and reduces investment waste and insufficient power support at critical moments.
Smart Images

Figure CN123121634A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of distribution network energy storage planning, and more specifically, to a method for topology configuration and optimization of distribution network energy storage devices based on source-load topology characteristics. Background Technology
[0002] Currently, the penetration rate of renewable energy sources, represented by distributed photovoltaics (DPV), in distribution networks has significantly increased, placing higher demands on the optimized configuration of distributed energy storage systems. However, as a complex network topology, the distribution network exhibits significant mismatches between photovoltaic output and load demand in both spatial (different nodes) and temporal (different time periods). Therefore, comprehensively considering the physical topology of the distribution network and the spatiotemporal distribution characteristics of sources (PV) and loads (loads) to optimize the location, capacity configuration, and topology planning of energy storage devices has become a key technical challenge for improving the resource utilization efficiency of the distribution network.
[0003] Existing energy storage configuration schemes for distribution networks are mainly divided into two categories: one is static planning methods based on empirical rules or simplified power flow constraints, and the other is active dynamic planning methods for distribution networks combined with load forecasting. These methods often struggle to accurately identify key topology nodes (such as electrical center nodes or high-power mismatch nodes), which can easily lead to insufficient support for key nodes or imbalanced resource allocation; moreover, they lack the ability to dynamically adapt to the time-varying characteristics of distribution network sources and loads (such as new photovoltaic grid connections, load center migration, etc.), making it difficult to achieve optimal adjustments. Summary of the Invention
[0004] The purpose of this invention is to provide a method for topology configuration and optimization of energy storage devices in distribution networks based on source-load topology characteristics, so as to alleviate the technical problems of low identification accuracy and poor dynamic adaptability of energy storage nodes to be configured in distribution networks in the prior art.
[0005] In a first aspect, embodiments of the present invention provide a distribution network energy storage configuration method based on source-load topology characteristics, comprising: acquiring distribution network topology data, node time-series power data, and node measurement data, and constructing a weighted graph model of the distribution network; calculating the source-load topology characteristic index of each node based on the weighted graph model and the time-series power data, and generating a candidate node set and a corresponding candidate energy storage topology pattern set based on the source-load topology characteristic index; constructing an energy storage configuration optimization model based on the candidate node set and the candidate energy storage topology pattern set, and solving it in conjunction with an objective function and hard constraints to obtain the energy storage configuration result.
[0006] In some optional implementations, the aforementioned source-load topology characteristic indicators include source-load mismatch indicators. Based on the aforementioned weighted graph model and the aforementioned time-series power data, the source-load topology characteristic indicators of each node are calculated, including: obtaining the photovoltaic active power output sequence and load active power demand sequence of the node under typical operating scenarios from the aforementioned time-series power data; calculating the local energy mismatch intensity of each node under typical operating scenarios based on the aforementioned photovoltaic active power output sequence and the aforementioned load active power demand sequence; and normalizing the aforementioned local energy mismatch intensity to obtain the normalized source-load mismatch indicator.
[0007] In some optional implementations, the aforementioned source-load topology characteristic indicators also include topology reachability indicators; the calculation of source-load topology characteristic indicators for each node based on the aforementioned weighted graph model and the aforementioned time-series power data also includes: obtaining the set of critical loads in the distribution network and the electrical distance matrix between nodes in the aforementioned weighted graph model based on the aforementioned weighted graph model and the aforementioned time-series power data; and calculating the reciprocal of the average normalized electrical distance from each node to the set of critical loads based on the aforementioned electrical distance matrix, as the topology reachability indicator.
[0008] In some optional implementations, the aforementioned source-load topology characteristic indicators also include a voltage support sensitivity weighted centrality indicator. Calculating the source-load topology characteristic indicators for each node based on the aforementioned weighted graph model and time-series power data further includes: obtaining node voltage vectors and voltage sensitivity matrices based on node measurement data, the aforementioned weighted graph model, and a power flow model representing the relationship between node injected power and voltage response; determining the node's active power injection and reactive power injection based on the aforementioned time-series power data; and calculating the comprehensive voltage response intensity corresponding to each node through weighted summation based on the aforementioned node voltage vectors, the aforementioned node active power injection and reactive power injection, the voltage sensitivity matrix, and preset weight adjustment coefficients, as the voltage support sensitivity weighted centrality indicator.
[0009] In some optional implementations, a candidate node set and a corresponding candidate energy storage topology pattern set are generated based on the aforementioned source-load topology feature indicators. This includes: normalizing and weighting multiple of the aforementioned source-load topology feature indicators to generate a corresponding comprehensive priority score; filtering candidates based on the comprehensive priority score using a threshold truncation method or a sorting truncation method to generate a candidate node set; constructing a node similarity matrix based on the electrical distance matrix between nodes in the aforementioned weighted graph model and multiple of the aforementioned source-load topology feature indicators for each node; performing cluster analysis on the aforementioned node similarity matrix to obtain a partition set; and generating at least two topology organization patterns for each partition in the aforementioned partition set to form a candidate energy storage topology pattern set. The aforementioned topology organization patterns include: centralized topology, distributed topology, and / or interconnected aggregation topology.
[0010] In some optional implementations, the aforementioned energy storage configuration optimization model includes a planning layer. Solving the aforementioned energy storage configuration optimization model by combining the objective function and hard constraints includes: inputting the node time-series power data of the aforementioned candidate node set, the aforementioned candidate energy storage topology set, and the typical scenario set into the planning layer of the aforementioned energy storage configuration optimization model; using the minimum life-cycle cost as the objective function and the energy storage equivalent cycle count not less than a preset threshold as the hard constraint, solving the planning layer of the aforementioned energy storage configuration optimization model, and outputting the energy storage configuration results; the aforementioned energy storage configuration results include: energy storage location status, rated energy capacity, rated power, and selected topology mode.
[0011] In some optional implementations, the aforementioned energy storage configuration optimization model further includes an operation layer; solving the aforementioned energy storage configuration optimization model further includes: inputting the aforementioned energy storage configuration results as fixed parameters into the aforementioned operation layer, and inputting the node time-series power data of the aforementioned typical scenario set; based on the physical constraints and equipment constraints of the safe operation of the distribution network, with the objective function of minimizing the comprehensive operating cost under the typical scenario, solving the aforementioned operation layer, and outputting the charging and discharging power sequence of the energy storage unit corresponding to each of the aforementioned typical scenario sets; feeding the aforementioned charging and discharging power sequence back to the aforementioned planning layer, which calculates the equivalent cycle number of energy storage based on the charging and discharging power sequence, and verifies whether the equivalent cycle number of energy storage meets the hard constraint condition of not less than a preset threshold.
[0012] In some optional implementations, the above method further includes: after the above energy storage configuration results are run, recalculating the current source-load topology characteristic indicators of each node within each rolling monitoring window; when the deviation of any of the above source-load topology characteristic indicators from the corresponding indicator benchmark value in the planning stage exceeds a preset threshold, triggering topology dynamic reconstruction and generating a candidate reconstruction action set; the above candidate reconstruction action set includes: service partition affiliation adjustment, logical aggregation relationship update, and physical connection status adjustment; performing short-term rolling optimization evaluation on each action in the above candidate reconstruction action set, selecting the action with the greatest comprehensive benefit as the target action; executing the above target action, and updating the above indicator benchmark value with the new configuration scheme and corresponding indicator value after execution.
[0013] Secondly, embodiments of the present invention provide a distribution network energy storage configuration device based on source-load topology features, comprising: a data acquisition module, used to acquire topology data, node time-series power data, and node measurement data of the distribution network, and construct a weighted graph model of the distribution network; a source-load topology feature extraction module, used to calculate the source-load topology feature index of each node based on the weighted graph model and the time-series power data, and generate a candidate node set and a corresponding candidate energy storage topology pattern set based on the source-load topology feature index; and an energy storage configuration result generation module, used to construct an energy storage configuration optimization model based on the candidate node set and the candidate energy storage topology pattern set, and solve it by combining an objective function and hard constraints to obtain the energy storage configuration result.
[0014] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the steps of the method described in any of the first aspects above.
[0015] This invention provides a method for topology configuration and optimization of energy storage devices in distribution networks based on source-load topology characteristics. The method includes: acquiring topology data, node time-series power data, and node measurement data of the distribution network to construct a weighted graph model of the distribution network; calculating the source-load topology characteristic indicators of each node based on the weighted graph model and time-series power data, and generating a set of candidate nodes and a corresponding set of candidate energy storage topology patterns; constructing an energy storage configuration optimization model based on the candidate node set and the candidate energy storage topology pattern set, and solving it by combining an objective function and hard constraints to obtain the energy storage configuration result. This method solves the technical problem of low accuracy in identifying energy storage nodes to be configured, and improves the accuracy and utilization efficiency of energy storage planning in distribution networks. Attached Figure Description
[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0017] Figure 1 A flowchart illustrating a distribution network energy storage configuration method based on source-load topology characteristics provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of an energy storage configuration optimization model provided in an embodiment of the present invention; Figure 3 A schematic diagram of a distribution network energy storage configuration device based on source-load topology characteristics provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation
[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0019] With the advancement of the "dual carbon" goals, the penetration rate of renewable energy sources, represented by distributed photovoltaic (DPV), in distribution networks is continuously increasing. Photovoltaic power generation exhibits significant volatility, intermittency, and uneven spatial and temporal distribution. Its large-scale integration alters the traditional unidirectional power flow characteristics of distribution networks, leading to increasingly prominent problems such as node voltage exceeding limits, backflow of power flow, and transformer overload. Configuring distributed energy storage systems (ESS) is an effective means to address these issues and improve the absorption capacity of photovoltaic power.
[0020] However, the distribution network is a complex topology, with significant mismatches between the spatial (different nodes) and temporal (different time periods) distribution of photovoltaic (PV) power and loads. Improper configuration of energy storage devices can lead to two typical problems: over-configuration results in resource waste, where excessive capacity is configured at nodes with low PV absorption pressure or low load demand, leading to long-term idle energy storage, low equivalent cycle life, and extremely poor return on investment. Under-configuration leads to functional failure, where insufficient capacity is configured at nodes with high voltage exceedance risks or severe source-load mismatch, resulting in insufficient power support during critical moments (such as peak PV power generation at midday or evening peak load), causing curtailment or insufficient power supply. Therefore, scientifically selecting, determining the capacity of, and optimizing the topology of energy storage devices (i.e., how to group and connect energy storage) based on the physical topology of the distribution network and the spatiotemporal distribution characteristics of sources (PV) and loads (users) has become a key technical challenge for improving the resource utilization rate of the distribution network.
[0021] Existing energy storage configuration schemes for distribution networks are mainly divided into two categories: one is static planning methods based on empirical rules or simplified power flow constraints, and the other is active dynamic planning methods for distribution networks combined with load forecasting. These methods have the following limitations: 1) It is difficult to accurately identify key topology nodes (such as electrical center nodes or high-power mismatch nodes), which can easily lead to insufficient support for key nodes or imbalance in resource allocation; 2) The planning scheme may over-configure energy storage capacity to meet extreme operating conditions, resulting in low equipment utilization, capacity redundancy, and wasted investment; 3) It lacks the ability to dynamically adapt to the time-varying characteristics of distribution network sources and loads (such as new photovoltaic grid connection, load center migration, etc.), making it difficult to achieve optimization and adjustment.
[0022] Based on this, the present invention provides a method for topology configuration and optimization of energy storage devices in distribution networks based on source-load topology features. This method can improve the identification accuracy of energy storage nodes in distribution networks and enhance the utilization efficiency and operational flexibility of distribution networks by extracting source-load topology features.
[0023] To facilitate understanding of this embodiment, a detailed description of a distribution network energy storage configuration method based on source-load topology characteristics disclosed in this invention will be provided first. (See [link to relevant documentation]). Figure 1 The diagram shows a flow chart of a distribution network energy storage configuration method based on source-load topology characteristics. This method can be executed by electronic equipment and mainly includes the following steps S102 to S106: Step S102: Obtain the topology data, node time-series power data, and node measurement data of the distribution network, and construct a weighted graph model of the distribution network. The distribution network topology data can include information on nodes, lines, and switches of the distribution network to be planned. Node time-series power data is derived from typical days or typical operating scenario sets, and can be photovoltaic output curves and load curves obtained through historical operating data. Node measurement data can be time-series sampled data containing node voltage amplitude, voltage phase angle, active power, and reactive power, which can be used for distribution network state estimation, power flow model parameter correction, and online updating of the voltage sensitivity matrix.
[0024] Specifically, the aforementioned node measurement data, together with topology data and node time-series power data, constitute the three-dimensional data foundation for energy storage configuration decisions: topology data defines network connectivity, node time-series power data characterizes the dynamic behavior of source and load, and node measurement data provides real-time state benchmarks and model calibration basis.
[0025] The construction of a weighted graph model for a distribution network specifically includes: obtaining information on the nodes, lines, and switches of the distribution network to be planned, and constructing a weighted graph model. ;in, It is a set of nodes (including at least photovoltaic access points, load points, and tie / section switch nodes). For the set of the line edges, The edge weight (characterized by line impedance, electrical distance, or equivalent resistance, etc.).
[0026] Simultaneously, photovoltaic power output curves under typical days or typical scenarios were collected. With load curve (Include reactive power, power factor, temperature-controlled load, etc., if necessary). Obtain the graphical model. Scene set (Typical daily / seasonal / operating condition labels), node time series dataset .
[0027] As a concrete example, the photovoltaic output curve and load curve mentioned above are preferably determined based on historical operating data. Specifically, node measurement data are collected within a continuous preset period. After outlier removal, missing value completion, time alignment, and extraction of typical scenarios, the photovoltaic active power output of node i at time t is obtained. and load active power demand Where i represents the node number and t represents the time number; when historical data is insufficient, it can be supplemented by combining short-term forecast results or standard typical curves.
[0028] Step S104: Based on the weighted graph model and time series power data, calculate the source-load topology characteristic index of each node, and generate a set of candidate nodes and a set of corresponding candidate energy storage topology patterns according to the source-load topology characteristic index. In one embodiment, the source-load topology characteristic index includes the source-load mismatch index.
[0029] The Source-Load Mismatch Index (SLMI) can be used to represent the degree of energy mismatch between a node's photovoltaic output and load demand.
[0030] As a specific example, the calculation of this source-load mismatch index may include the following steps: First, obtain the photovoltaic active power output sequence and load active power demand sequence of the nodes under typical operating scenarios based on time-series power data; then, calculate the local energy mismatch intensity of each node under typical operating scenarios based on the photovoltaic active power output sequence and load active power demand sequence; and normalize the local energy mismatch intensity to obtain the normalized source-load mismatch index.
[0031] Preferably, for any node With time set (e.g., typical 24-hour day or multi-scene splicing), definition:
[0032] in For time step, To prevent extremely small positive numbers with a denominator of zero, a larger SLMI indicates a stronger mismatch between local power generation and power consumption at that node, making it more likely to experience voltage over-limit (surplus) or insufficient energy supply (shortage), thus making it a priority candidate for energy storage configuration.
[0033] In another embodiment, the source-load topology feature index also includes a topology reachability index.
[0034] The Effective Distance (ED) metric can be used to quantify cross-node voltage support efficiency.
[0035] As a concrete example, the calculation of this topology accessibility index may include the following steps: First, based on the weighted graph model and time-series power data, obtain the set of critical loads in the distribution network and the electrical distance matrix between nodes in the weighted graph model; then, based on the electrical distance matrix, calculate the reciprocal of the average normalized electrical distance from each node to the set of critical loads, as the topology accessibility index.
[0036] Preferred, based on weighted graph model Calculate the electrical distance matrix between nodes (For example, using the shortest path distance based on line impedance, equivalent resistance distance, etc.) first, then... Normalization is performed to obtain the normalized distance. Then, from node i to the critical load set... The reciprocal of the average normalized distance is used as the node-level electrical distance / topology reachability metric, i.e.: ; A larger value indicates a shorter average electrical distance from node i to the critical load and higher cross-node support efficiency. If the critical load is not separately identified, then... The entire set of load nodes can be retrieved.
[0037] In another embodiment, the source-load topology characteristic index also includes a voltage-supported sensitivity-weighted centrality index.
[0038] Voltage Support Sensitivity Weighted Centrality (VSC) can be used to represent the overall influence of a node as a voltage regulation point on the voltage deviation of the entire network.
[0039] As a specific example, the calculation of the voltage support sensitivity weighted centrality index can include the following steps: First, based on nodal measurement data, a weighted graph model, and a power flow model representing the relationship between nodal injected power and voltage response, obtain the nodal voltage vector and voltage sensitivity matrix; then, based on time-series power data, determine the active power injection and reactive power injection amounts at the nodes; and finally, based on the nodal voltage vector, the active and reactive power injection amounts at the nodes, the voltage sensitivity matrix, and preset weight adjustment coefficients, calculate the comprehensive voltage response intensity corresponding to each node through weighted summation, which serves as the voltage support sensitivity weighted centrality index.
[0040] Specifically, the power flow model used to represent the relationship between node injected power and voltage response can be a computational model well-known to those skilled in the art, used to establish a mathematical mapping relationship between node injected power and voltage vectors, including but not limited to DC power flow models, linearized AC power flow models, Jacobian matrix-based sensitivity models, or equivalent circuit models calibrated by parameter identification. Its core function is to provide... V / P i and V / Q i Isovoltage sensitivity information is used to support the quantitative calculation of the voltage-supported sensitivity weighted centrality index.
[0041] The voltage sensitivity matrix reflects the changes in the overall network voltage amplitude and phase angle caused by a unit power disturbance injected by a node, enabling the voltage support sensitivity weighted centrality index to quantify the topology-electrical coordinated support capability of a node for overall network voltage stability. As a concrete example, the voltage sensitivity matrix is generated in real-time by an online power flow model driven by node measurement data, and its elements... V / P i and V / Q i Characterizes the dynamic impact intensity on the voltage amplitude and phase angle of the entire network node j when a unit active / reactive power disturbance is injected at node i.
[0042] Preferably, the impact of node-injected power disturbances on voltage deviation is calculated based on a linearized power flow / Jacobi matrix or voltage sensitivity model, for example:
[0043] in For node voltage vectors, and These represent the active power injection and reactive power injection at node i, respectively. This represents the sensitivity of the entire network voltage vector when a unit disturbance occurs during active power injection at node i. It represents the sensitivity of the entire network voltage vector when a unit disturbance occurs during reactive power injection at node i; α and β are weighting coefficients, and preferably satisfy α+β=1.
[0044] The larger the Voltage Support Sensitivity Weighted Centrality (VSC) index, the more "voltage support topology center" the node is, and configuring energy storage can yield stronger benefits from network-wide voltage regulation.
[0045] In one embodiment, the step S104 above, which generates a candidate node set and a corresponding candidate energy storage topology pattern set based on source-load topology characteristic indicators, may include: (S141) Normalize and weightedly fuse multiple source-load topology feature indicators to generate corresponding comprehensive priority scores; Specifically, the above features are normalized to form a comprehensive priority score:
[0046] in It can be the average electrical distance from the critical load set; or in other embodiments, topological factors such as betweenness centrality and end node marking can be selected.
[0047] Furthermore, critical load sets The following methods can be used to determine the users: Prioritize the identification of important users, users requiring guaranteed power supply, or users with high requirements for continuous power supply that have been identified in the distribution network operation and maintenance ledger.
[0048] Alternatively, when no readily available labels are available, a key load scoring model can be constructed: ; in Indicates the load importance level or power outage loss weight. Sen_j represents the peak load at node j, and Cont_j represents the voltage sensitivity and continuous power supply requirement. Select the load nodes whose KeyScore_j is not lower than the threshold τ_key or are in the top q% as the key load set.
[0049] (S142) Based on the comprehensive priority score, the threshold truncation method or the sorting truncation method is used to filter and generate a set of candidate nodes; Specifically, the threshold truncation method includes: setting a comprehensive priority score threshold and selecting nodes with scores not lower than the threshold; the sorting truncation method includes: sorting all nodes from high to low according to their comprehensive priority scores and selecting the top L nodes, where L is a preset positive integer and satisfies L≤20% of the total number of nodes.
[0050] Output candidate node set (e.g., before retrieving the Score) The system identifies nodes that meet threshold conditions and outputs a feature explanation for each candidate point to facilitate engineering interpretability and review understanding.
[0051] (S143) Based on the electrical distance matrix between nodes in the weighted graph model and multiple source-load topology feature indices of each node, construct a node similarity matrix; To avoid blindly seeking optimization in the entire node space and causing dimensionality explosion, spectral clustering / graph clustering can be used to perform "source-load complementary partitioning" of the distribution network based on the feature index and electrical distance matrix obtained in step S104, and a partition set can be obtained.
[0052] Specifically, the node similarity matrix can be constructed using:
[0053] Where σ is an electrical distance scale parameter, which can be used to measure the sensitivity to differences in physical topological accessibility; specifically, σ can be determined based on the standard deviation, median, or 90th quantile of the set of non-zero elements in the upper triangular region of the normalized electrical distance matrix, and preferably σ = max(σ0, εσ), where σ0 is the statistic and εσ is a preset minimum positive number.
[0054] The electrical distance scale parameter σ can be determined based on the normalized electrical distance matrix, taking the standard deviation or 90th quantile of its upper triangular non-zero element.
[0055] η is a source load feature scale parameter, which can be used to adjust the influence of differences in source load topological features between nodes on similarity. Specifically, this source load feature scale parameter η can be used to first construct the normalized source load topological feature vector F of node i. i =[SLMI i ED i VSC i Then, calculate the set of non-zero Euclidean distances between any two normalized feature vectors of nodes. η can be the standard deviation, median, mean, or 90th quantile of this distance set, and preferably η = max(η0, εη), where η0 is the statistic and εη is a preset minimum positive number. Alternatively, η can be determined by maximizing the clustering silhouette coefficient, modularity, or the comprehensive operating benefit under the validation scenario.
[0056] Each partition in the clustering output is considered a potential energy storage service unit. Subsequently, several topology templates are generated for each partition, forming a candidate topology pattern set.
[0057] (S144) Perform cluster analysis on the node similarity matrix to obtain the partition set; (S145) For each partition in the partition set, generate at least two topology organization modes to form a candidate energy storage topology mode set; wherein, the topology organization modes include: centralized topology, distributed topology, and / or interconnected aggregation topology.
[0058] Specifically, spectral clustering or graph convolutional clustering analysis is performed on the node similarity matrix to obtain the partition set. k is the number of clusters; then for each partition Z in the partition set k Generate at least two topology organization patterns to form a candidate energy storage topology pattern set M. The topology organization patterns may include: Centralized topology, in partition Z k The node with the highest voltage support sensitivity weighted centrality (VSC) value or the highest electrical distance weighted centrality is selected as the sole access point, and the total energy storage capacity required for this partition is concentrated on this node. Distributed topology, in partition Z k Based on the source-load mismatch degree SLMI, sorted from high to low, m is selected. k Each node serves as an access point, m k To satisfy 1 <m k ≤ |Z k | / 3 A positive integer, and the normalized electrical distance between any two access points is not less than the preset minimum isolation distance; the capacity of each access point is allocated according to its SLMI ratio or the local net load fluctuation variance ratio; Interconnected aggregation topology, based on distributed topology, divides partition Z... k The energy storage units at all or part of the access points are associated through a logical aggregation control protocol to form a virtual power plant unit; the logical aggregation control protocol includes at least: an aggregation power command issuance mechanism, a capacity sharing weight allocation rule, a communication delay tolerance constraint, and a handover switch reachability verification.
[0059] Preferably, for each partition Z k Generate at least three candidate patterns according to the template of the preset topology organization pattern: In pattern A, select the partition electrical center node or the node with the largest VSC value as the unique access point, and set the total partition capacity E k Centralized configuration is located on this node; in Mode B, m is selected from the partitions based on SLMI or overall score from high to low. kThere are 10 candidate nodes, and energy storage is distributed in a decentralized manner under the condition of satisfying the minimum electrical distance constraint between nodes. The capacity of each node is allocated according to the local mismatch or local weight ratio. In mode C, an aggregation control relationship is introduced based on mode B, and the partition aggregation power p_k^agg(t)=Σ_{i∈Z_k} p_i(t) is set. The power allocation coefficient γ of each sub-energy storage unit is determined by capacity weight, communication constraint and tie switch reachability constraint. i This is to achieve cross-node collaborative support.
[0060] In the aforementioned candidate energy storage topology set M, each topology explicitly encodes at least one of its partition identifier, access point set, capacity allocation vector, connection relationship matrix, and control coupling parameters, so as to participate in the joint solution as discrete decision variables in the subsequent energy storage configuration optimization model.
[0061] Step S106: Based on the candidate node set and the candidate energy storage topology pattern set, construct an energy storage configuration optimization model, and solve it by combining the objective function and hard constraints to obtain the energy storage configuration result.
[0062] Specifically, the energy storage configuration optimization model constructed in this embodiment is a two-layer optimization model including a planning layer and an operation layer, which can simultaneously constrain the investment economy and actual utilization efficiency of energy storage during the planning stage. As a preferred example, see [link to example]. Figure 2 The diagram shows a structural schematic of an energy storage configuration optimization model, which includes an upper planning layer 210 and a lower operation layer 220. The planning layer 210 can be used to output location, capacity, power, and topology mode; the operation layer 220 can be used to output indicators such as charge and discharge sequence, voltage, curtailment, and grid loss, which are then fed back to the planning layer 210 to calculate the equivalent number of cycles and verify the utilization hard constraints.
[0063] In one embodiment, the energy storage configuration optimization model is solved by combining the objective function and hard constraints, including: inputting the node time-series power data of the candidate node set, the candidate energy storage topology pattern set, and the typical scenario set into the planning layer of the energy storage configuration optimization model; With the minimum life-cycle cost as the objective function and the equivalent cycle number of energy storage not being less than a preset threshold as a hard constraint, the planning layer of the energy storage configuration optimization model is solved, and the energy storage configuration results are output. The energy storage configuration results include: energy storage location status, rated energy capacity, rated power and selected topology mode.
[0064] Specifically, the planning layer (Leader) is located at the upper level of the two-level optimization model, and the decision variables at this layer include: location. ;capacity ;power Topology mode selection (Can be represented by discrete variables); the objective function is to minimize the total life cycle cost.
[0065] in, It is a binary decision variable representing the choice of physical access location for energy storage. Specifically, =1 indicates that an energy storage device is configured in candidate node i; =0 indicates that no energy storage device is configured on this node.
[0066] Furthermore, to avoid over-configuration that results in "installed but not used," hard constraints on utilization can be added, for example, by defining energy storage... Equivalent number of loops (or throughput utilization) within the evaluation period.
[0067] And apply hard constraints: ; in, This is the charge / discharge power sequence obtained from the optimal operation at the lower level. This is the utilization threshold (which can be set by the equivalent number of cycles per year). This hard constraint is used to directly exclude low-utilization scenarios during the planning phase, thus suppressing over-provisioning at its source.
[0068] The following is a further supplement to the above embodiments, providing a calculation method for ensuring that the energy storage equivalent cycle number is not less than a preset threshold: The equivalent number of cycles for the i-th energy storage unit within the evaluation period can be defined as:
[0069] Among them, P i (t) represents the charging and discharging power of the i-th energy storage unit at time t, obtained from the solution of the operating layer. The positive and negative signs indicate discharging and charging, respectively. |P i (t)|△t represents the charging and discharging throughput energy during this period; E i N represents the rated energy capacity of the i-th energy storage unit; Δt represents the time step; N i To evaluate the equivalent number of complete cycles within a period. Since a complete charge-discharge cycle includes two energy processes, charging and discharging, when using the absolute value of bidirectional charge-discharge power to statistically calculate throughput, 2E is used. i As a normalization benchmark, for all nodes with configured energy storage, the following should be satisfied:
[0070] Where, x i =1 indicates that node i is selected for energy storage configuration, N min This is the preset minimum equivalent number of loops threshold.
[0071] As a preferred example, to avoid division by zero when xi=0, the above constraint can be equivalently rewritten as a lower limit constraint on throughput in the actual optimization model:
[0072] Alternatively, in models with two variables, the Big M form can be used:
[0073] and cooperate use.
[0074] If charging power and discharging power are modeled separately in the operating layer, for example: Then the formula can be changed to:
[0075] The above formula avoids the nonlinearity or additional linearization caused by absolute values, therefore it is more efficient than using |P|. i (t) is more suitable for optimizing model representation.
[0076] As another preferred example, for the case of a "typical scenario set", the following formula can be used:
[0077] Where ω is the typical scene number, ρ ω This represents the annualized weight or the frequency weight corresponding to this typical scenario.
[0078] In one embodiment, solving the energy storage configuration optimization model further includes: inputting the energy storage configuration result as a fixed parameter into the operation layer, and inputting the node time-series power data of typical scenario sets; based on the physical constraints and equipment constraints of the safe operation of the distribution network, solving the operation layer with the objective function of minimizing the comprehensive operating cost under typical scenarios, and outputting the charging and discharging power sequence of the energy storage unit in each typical scenario set; feeding the charging and discharging power sequence back to the planning layer, which calculates the equivalent cycle number of energy storage based on the charging and discharging power sequence, and verifies whether the equivalent cycle number of energy storage meets the hard constraint condition of not less than a preset threshold.
[0079] Specifically, the run-up layer (Follower) is located at the lower layer of the two-layer optimization model, given the upper-layer solution. Then, for each typical scenario Solving for the optimal run yields the following results. ,Voltage Indicators such as light waste.
[0080] Preferably, the runtime layer can be constructed as follows:
[0081] in, This represents the network active power loss at time t. Let V(t) represent the power of light discarded at time t, and let V(t) represent the node voltage vector at time t. Represents the reference voltage vector. This represents the change in switch state between adjacent time points. , , and These represent the network loss cost coefficient, the curtailment penalty coefficient, the voltage deviation penalty coefficient, and the switching action penalty coefficient, respectively. If ω is calculated based on a typical scenario, scenario weights may also be included. .
[0082] Operational layer constraints include at least: power flow balance, upper and lower limits of line current / power, upper and lower limits of voltage, state of charge (SOC) constraints of energy storage, charging and discharging power constraints, mutual exclusion constraints of charging and discharging, equipment capacity constraints, and topology reconfiguration switching states. And restrictions on the number of actions, etc.
[0083] The solution strategy can be implemented by using genetic algorithms / particle swarm optimization for upper-level discrete variables; convex optimization such as QP / SOCP can be used for fast solution of lower-level continuous optimization; when introducing discrete switching variables, "discrete search and convex optimization evaluation" or mixed integer convex programming can be used.
[0084] In another embodiment, the above method may further include: (S107) After the energy storage configuration results are run, recalculate the current source-load topology characteristic index of each node in each rolling monitoring window; (S108) When the deviation of any source-load topology characteristic index from the benchmark value of the corresponding index in the planning stage exceeds the preset threshold, such as the voltage over-limit rate, power curtailment rate, and line overload rate, at least one of the operation indicators continuously deteriorates beyond the preset threshold, triggering topology dynamic reconfiguration and generating a candidate reconfiguration action set; the candidate reconfiguration action set includes: service partition affiliation adjustment, logical aggregation relationship update, and physical connection status adjustment; (S109) Perform short-term rolling optimization evaluation on each action in the candidate reconstruction action set, and select the action with the largest comprehensive benefit as the target action; (S110) Execute the target action and update the indicator baseline value with the new configuration scheme and corresponding indicator value after execution.
[0085] In other words, this embodiment employs a closed-loop mechanism during operation: "monitoring, triggering, generating candidate actions, rapid evaluation, execution, and updating the baseline." Specifically, within each rolling monitoring window, updated SLMI, ED, VSC, voltage overrun rate, curtailment rate, and line overload rate are calculated. When the deviation of any indicator from the planned baseline exceeds a preset threshold, or when voltage overrun / curtailment degradation occurs for several consecutive windows, reconfiguration is triggered.
[0086] Preferably, the candidate actions include at least service partition assignment adjustment, aggregation relationship update, energy storage power allocation weight adjustment, and tie / segment switch status adjustment. Subsequently, short-term rolling optimization or reduced-order power flow evaluation is used to assess the comprehensive benefits of each candidate action. If the optimal candidate action improves the benefits relative to the current solution by more than a threshold while satisfying safety constraints and action count constraints, then the action is executed and the baseline state is updated.
[0087] The final optimal solution may include: the energy storage site selection set S={i|x_i=1}, and the capacity of each site. With power Partitioning and topology organization (centralized / distributed / interconnected aggregation), and corresponding typical scenario operation strategies (charge and discharge curves, curtailment suppression and voltage support effects, utilization rate evaluation reports, etc.).
[0088] Based on the same inventive concept, this invention also provides a distribution network energy storage configuration device based on source-load topology characteristics, see [link to relevant documentation]. Figure 3 As shown, the device mainly includes the following parts: The data acquisition module 310 is used to acquire the topology data, node time-series power data and node measurement data of the distribution network, and to construct the weighted graph model of the above distribution network. The source-load topology feature extraction module 320 is used to calculate the source-load topology feature index of each node based on the above weighted graph model and the above time-series power data, and generate a set of candidate nodes and a set of corresponding candidate energy storage topology patterns based on the above source-load topology feature index. The energy storage configuration result generation module 330 is used to construct an energy storage configuration optimization model based on the above-mentioned candidate node set and the above-mentioned candidate energy storage topology pattern set, and solve it by combining the objective function and hard constraints to obtain the energy storage configuration result.
[0089] In one embodiment, the above-mentioned device may further include: a utilization verification module, used to calculate the equivalent number of energy storage cycles based on the charge and discharge power sequence output by the operating layer, and verify whether the equivalent number of energy storage cycles is not lower than a preset threshold; and a dynamic reconstruction module, used to recalculate the source-load topology characteristic index within a rolling monitoring window, and generate, evaluate and execute candidate reconstruction actions when the triggering conditions are met.
[0090] The energy storage configuration device for distribution networks based on source-load topology characteristics provided in this invention embodiment can be specific hardware on the device or software or firmware installed on the device. The implementation principle and technical effects of the device provided in this invention embodiment are the same as those in the foregoing method embodiments. For the sake of brevity, any parts not mentioned in the device embodiment section can be referred to the corresponding content in the foregoing method embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.
[0091] Based on the same inventive concept, embodiments of the present invention also provide an electronic device, specifically, the electronic device includes a processor and a storage device; the storage device stores a computer program, and the computer program, when run by the processor, executes the method described in any of the above embodiments.
[0092] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 400 includes: a processor 410, a memory 420, a communication interface 430, and a bus 440. The memory 420 stores machine-readable instructions that can be executed by the processor 410. When the electronic device is running, the processor 410 communicates with the memory 420 through the bus 440. The processor 410 executes the machine-readable instructions to perform the steps of the method described above.
[0093] Specifically, the memory 420 and processor 410 can be general-purpose memory and processor, without any specific limitations. When the processor 410 runs the computer program stored in the memory 420, it can execute the above method.
[0094] Processor 410 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 410 or by instructions in software form. The processor 410 may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory 420, and processor 410 reads the information from memory 420 and, in conjunction with its hardware, completes the steps of the above method.
[0095] Corresponding to the above method, this embodiment of the invention also provides a computer-readable storage medium storing machine-executable instructions. When the computer-executable instructions are called and run by a processor, the computer-executable instructions cause the processor to perform the steps of the above method.
[0096] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.
[0097] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0098] Furthermore, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.
[0099] It should be noted that if the functionality is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0100] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.
[0101] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A power distribution network energy storage device topology configuration and optimization method based on source load topology characteristics, characterized in that, include: Obtain the topology data, node time-series power data, and node measurement data of the distribution network, and construct a weighted graph model of the distribution network. Based on the weighted graph model and the time-series power data, the source-load topology characteristic index of each node is calculated, and a set of candidate nodes and a set of corresponding candidate energy storage topology patterns are generated according to the source-load topology characteristic index. Based on the candidate node set and the candidate energy storage topology set, an energy storage configuration optimization model is constructed, and the model is solved by combining the objective function and hard constraints to obtain the energy storage configuration result.
2. The method of claim 1, wherein, The source-load topology characteristic index includes the source-load mismatch index; Based on the weighted graph model and the time-series power data, the source-load topology characteristic indices of each node are calculated, including: Based on the time-series power data acquisition node, the photovoltaic active power output sequence and load active power demand sequence are obtained under typical operating scenarios. Based on the photovoltaic active power output sequence and the load active power demand sequence, calculate the local energy mismatch intensity of each node under typical operating scenarios; The local energy mismatch intensity is normalized to obtain the normalized source-load mismatch index.
3. The method of claim 2, wherein, The source-load topology characteristic index also includes the topology reachability index; Based on the weighted graph model and the time-series power data, the calculation of source-load topology characteristic indices for each node also includes: Based on the weighted graph model and the time-series power data, the key load set of the distribution network and the electrical distance matrix between nodes in the weighted graph model are obtained. Based on the electrical distance matrix, the reciprocal of the average normalized electrical distance from each node to the set of critical loads is calculated as a topology reachability index.
4. The method of claim 2, wherein, The source-load topology characteristic index also includes a voltage-supported sensitivity weighted centrality index; Based on the weighted graph model and the time-series power data, the calculation of source-load topology characteristic indices for each node also includes: Based on node measurement data, the weighted graph model, and the power flow model used to represent the relationship between node injected power and voltage response, the node voltage vector and voltage sensitivity matrix are obtained. Based on the time-series power data, determine the active power injection amount and reactive power injection amount of the node; Based on the node voltage vector, the node active power injection and reactive power injection, the voltage sensitivity matrix, and the preset weight adjustment coefficient, the comprehensive voltage response intensity corresponding to each node is calculated by weighted summation, which serves as the weighted centrality index of voltage support sensitivity.
5. The method according to claim 1, characterized in that, Based on the source-load topology characteristic indicators, a set of candidate nodes and a corresponding set of candidate energy storage topology patterns are generated, including: The multiple source-load topology feature indicators are normalized and weighted and fused to generate a corresponding comprehensive priority score; Based on the comprehensive priority score, a candidate node set is generated by using either a threshold truncation method or a sorting truncation method for filtering. Based on the electrical distance matrix between nodes in the weighted graph model and multiple source-load topology feature indices of each node, a node similarity matrix is constructed. Cluster analysis is performed on the node similarity matrix to obtain a set of partitions; For each partition in the partition set, at least two topology organization patterns are generated to form a candidate energy storage topology pattern set; The topology organization modes include: centralized topology, distributed topology, and / or interconnected aggregation topology.
6. The method according to claim 1, characterized in that, The energy storage configuration optimization model includes a planning layer; the energy storage configuration optimization model is solved by combining the objective function and hard constraints, including: The node time-series power data of the candidate node set, the candidate energy storage topology pattern set, and the typical scenario set are input into the planning layer of the energy storage configuration optimization model; With the minimum life-cycle cost as the objective function and the equivalent cycle number of energy storage not being less than a preset threshold as a hard constraint, the planning layer of the energy storage configuration optimization model is solved, and the energy storage configuration result is output. The energy storage configuration results include: energy storage location status, rated energy capacity, rated power, and selected topology.
7. The method according to claim 6, characterized in that, The energy storage configuration optimization model also includes an operation layer; solving the energy storage configuration optimization model further includes: The energy storage configuration result is input into the operation layer as a fixed parameter, and the node timing power data of the typical scenario set is also input. Based on the physical and equipment constraints of the safe operation of the distribution network, the operation layer is solved with the objective function of minimizing the comprehensive operation cost under typical scenarios, and the charging and discharging power sequence of the energy storage unit in each typical scenario set is output. The charging and discharging power sequence is fed back to the planning layer, which calculates the equivalent number of energy storage cycles based on the charging and discharging power sequence and verifies whether the equivalent number of energy storage cycles meets the hard constraint condition of not being lower than a preset threshold.
8. The method according to claim 1, characterized in that, The method further includes: After the energy storage configuration results are implemented, the current source-load topology characteristic indicators of each node are recalculated within each rolling monitoring window; When the deviation of any of the source-load topology feature indicators from the benchmark value of the corresponding indicator in the planning stage exceeds a preset threshold, dynamic topology reconstruction is triggered, and a set of candidate reconstruction actions is generated. The set of candidate reconstruction actions includes: service partition affiliation adjustment, logical aggregation relationship update, and physical connection status adjustment. A short-term rolling optimization evaluation is performed on each action in the candidate reconstruction action set, and the action with the greatest comprehensive benefit is selected as the target action. Execute the target action, and update the benchmark value of the indicator with the new configuration scheme and corresponding indicator value after execution.
9. A distribution network energy storage configuration device based on source-load topology characteristics, characterized in that, include: The data acquisition module is used to acquire the topology data, node time-series power data and node measurement data of the distribution network, and to construct a weighted graph model of the distribution network. The source-load topology feature extraction module is used to calculate the source-load topology feature index of each node based on the weighted graph model and the time-series power data, and generate a set of candidate nodes and a set of corresponding candidate energy storage topology patterns based on the source-load topology feature index. The energy storage configuration result generation module is used to construct an energy storage configuration optimization model based on the candidate node set and the candidate energy storage topology pattern set, and solve it by combining the objective function and hard constraints to obtain the energy storage configuration result.
10. The apparatus according to claim 9, characterized in that, The device further includes: The utilization rate verification module is used to calculate the equivalent number of energy storage cycles based on the charge and discharge power sequence output by the operation layer, and to verify whether the equivalent number of energy storage cycles is not lower than a preset threshold. The dynamic reconstruction module is used to recalculate the source-load topology characteristic indicators within the scrolling monitoring window, and to generate, evaluate, and execute candidate reconstruction actions when the triggering conditions are met.
