Energy storage capacity configuration method for cluster regulation and control of optical storage and charging station

Abstract

The invention discloses an energy storage capacity configuration method for cluster regulation and control of an optical storage and charging station, relates to the technical field of capacity configuration, and effectively solves the problem of balancing economy and robustness of a traditional method through a hierarchical collaborative energy storage capacity configuration mechanism. The established source-load difference uncertainty model accurately quantifies the space-time coupling characteristics of the photovoltaic output and the charging load, and avoids capacity redundancy caused by deterministic planning; independent energy storage is configured on the basis of sub-cluster differentiation requirements, repeated investment of distributed energy storage is reduced through local fluctuation stabilization, and meanwhile, single-point capacity requirements are reduced by using complementary characteristics in clusters; shared energy storage is introduced and a cross-cluster collaborative optimization model is constructed, so that the overall capacity utilization rate is remarkably improved; on the premise that high-proportion photovoltaic consumption and reliable power supply of a charging load are guaranteed, the adaptability to extreme scenes is enhanced through a layered response mechanism, and an extensible solution is provided for a power system.

Description

Technical Field

[0001] The present invention relates to the technical field of capacity configuration, and in particular to a method for configuring energy storage capacity for the regulation and control of a cluster of optical storage and charging stations. Background Art

[0002] The penetration rates of distributed photovoltaic and electric vehicle charging facilities in the distribution network have been continuously increasing, and the integrated optical storage and charging stations have become important nodes in the new power system. However, the strong volatility of photovoltaic power output, the spatio-temporal randomness of charging loads, and the high-cost characteristics of energy storage configuration have led to core challenges such as dynamic imbalance between sources and loads and poor investment economy in the station cluster. In this context, how to achieve collaborative optimization of "source-storage-load" at the cluster level through refined energy storage capacity configuration is not only the key to improving the consumption rate of renewable energy and ensuring the safe operation of the power grid, but also the core technical path to promote the low-carbon transformation of the energy system, with significant engineering application value and strategic significance.

[0003] Current energy storage capacity configuration methods mostly focus on single stations or deterministic scenarios, and there are two significant defects: Firstly, the modeling of photovoltaic power output often uses typical days or average values for simplified processing, failing to effectively characterize its probability distribution characteristics and extreme fluctuation risks, resulting in capacity planning deviating from actual requirements; Secondly, existing strategies often adopt a single mode of "decentralized independent energy storage" or "centralized shared energy storage". The former causes capacity redundancy due to ignoring the spatio-temporal complementarity between stations, while the latter is difficult to adapt to the differentiated needs of the cluster due to the lack of a hierarchical regulation mechanism. In addition, for the capacity allocation model of multi-agent collaboration of optical storage and charging, a global optimization framework that takes into account economy, robustness, and scalability has not yet been formed. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for configuring energy storage capacity for the regulation and control of a cluster of optical storage and charging stations, and to construct an energy storage capacity configuration method that integrates uncertainty modeling, cluster division, and hierarchical optimization. Through the differential clustering of the optical storage and charging station cluster, the collaborative configuration of "local independent energy storage compensation" and "global shared energy storage scheduling" is realized, so as to reduce the single-point investment cost while improving the overall economy by using the spatio-temporal energy complementarity characteristics between clusters. This method needs to synchronously solve key problems such as the randomness modeling of photovoltaic power output, the design of sub-cluster division criteria, and the game allocation of shared energy storage capacity to support the flexible regulation and stable operation of the distribution network in high-proportion renewable energy scenarios; to solve the problems raised in the background art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A method for configuring energy storage capacity for the regulation and control of a cluster of optical storage and charging stations, establishing a source-load capacity difference model of distributed photovoltaic stations and charging stations in the target area based on the randomness of photovoltaic power output;

[0006] Construct a source-load difference scenario set based on the source-load capacity difference model;

[0007] Based on the source-load difference scenario set, the energy storage capacity of the photovoltaic-storage-charging sub-clusters is configured based on the capacity difference of different photovoltaic-storage-charging sub-clusters;

[0008] Configure shared energy storage capacity for multiple clusters.

[0009] According to the above technical solution, the establishment of the difference model includes the following steps:

[0010] S101, collecting and preprocessing multi-source data of distributed photovoltaic stations and charging stations to obtain photovoltaic data and charging load data;

[0011] S102, based on the pre-processed photovoltaic data, analyzing the probability distribution characteristics of distributed photovoltaic output to obtain a photovoltaic output curve; performing multi-dimensional decomposition on the pre-processed charging load data to obtain a charging load curve;

[0012] S103: Based on the photovoltaic output curve and the charging load curve, a source-load capacity difference calculation is performed to obtain a source-load capacity difference model.

[0013] According to the above technical solution, in step S101, first, the historical output data of distributed photovoltaic stations in the target area are collected, covering the output fluctuation characteristics under different seasons, weather types and typical extreme events; the historical operation data of the charging station is simultaneously obtained, including the charging power timing curve, user charging behavior statistics and the association records with the grid time-of-use electricity price; then the original data is cleaned, the outliers caused by sensor failures are eliminated, the missing data is processed by linear interpolation or adjacent time period mean filling, and all data are unified to the same time resolution, and the dimension differences are eliminated by normalization to obtain the pre-processed photovoltaic data and charging load data;

[0014] In the analysis of the probability distribution characteristics of distributed photovoltaic output in step S102, for the photovoltaic output under fine weather, a probability density function fitting method is used to describe the distribution law of the output value; for the photovoltaic output under extreme weather, the upper and lower limits of the photovoltaic output fluctuation are constructed to form a robust photovoltaic output boundary, and finally a photovoltaic output curve is obtained;

[0015] In step S102, in the multi-dimensional decomposition of the charging load data, the basic load component and the elastic load component are identified, and the intra-day and seasonal variation law models are established respectively to obtain the charging load curve; the coupling relationship between user behavior and power grid regulation strategy is further analyzed, such as the impact of the peak-valley electricity price difference on the migration of charging time periods, to build a dynamic prediction framework for charging load;

[0016] In step S103, align the photovoltaic output curve with the charging load curve along the time axis, calculate the power difference between the two for each time period. At the same time, define the positive difference and the negative difference. The positive difference means that the photovoltaic output is greater than the charging load, and storage or power export is required. The negative difference means that the photovoltaic output is less than the charging load, and energy storage or grid supplementation is required. Substitute the probability distribution of the distributed photovoltaic output or the upper and lower limit boundaries of the photovoltaic output fluctuation into the power difference calculation to generate a power difference fluctuation range with confidence, quantify the maximum power deficit or redundancy in extreme scenarios, form a time series of source-load difference containing uncertainty information, and finally obtain the source-load capacity difference model.

[0017] According to the above technical solution, using the historical difference curve formed by the source-load capacity difference model, extract representative scenarios by using the clustering algorithm to construct a typical scenario library. At the same time, use the Monte Carlo simulation method, combined with the photovoltaic output probability model and the random characteristics of the charging load, to batch generate a set of random scenarios covering normal and extreme conditions to obtain the source-load difference scenario set. Finally, verify the coverage and rationality of the scenario set through historical data backtesting to ensure that it can fully reflect the differentiated operation requirements of the photovoltaic-storage-charging station cluster.

[0018] According to the above technical solution, for the configuration of the energy storage capacity of the photovoltaic-storage-charging sub-cluster, it includes

[0019] The following steps:

[0020] S201. Extract the partitioning characteristics of the photovoltaic-storage-charging sub-cluster and determine the clustering basis;

[0021] S202. Use the dynamic clustering algorithm to achieve the partitioning of the photovoltaic-storage-charging sub-cluster;

[0022] S203. Statistically analyze the capacity difference of the photovoltaic-storage-charging sub-cluster and conduct demand analysis;

[0023] S204. Formulate the independent energy storage capacity rules for the photovoltaic-storage-charging sub-cluster;

[0024] S205. Verify and dynamically adjust the results after the energy storage capacity configuration of the photovoltaic-storage-charging sub-cluster.

[0025] According to the above technical solution, in step S201, extract the key features from the generated source-load difference scenario set, including the capacity difference time series curve of each station, geographical location, and operation constraints. Determine "similarity of difference curve shape" and "geographical proximity" as the core clustering basis to ensure that the stations within the same photovoltaic-storage-charging sub-cluster have complementary regulation potential and are convenient for sharing the energy storage layout;

[0026] In step S202, an adaptive weighted clustering algorithm is adopted, assigning a higher weight to the similarity of time series differences and a lower weight to geographical distance. By iteratively optimizing the cluster centers and station affiliations, all photovoltaic-energy storage-charging stations are divided into several photovoltaic-energy storage-charging sub-clusters, such that the difference curves of stations within the same photovoltaic-energy storage-charging sub-cluster exhibit complementary characteristics in terms of time series, while the difference patterns between different clusters are significantly different. The clustering results need to satisfy the constraints of the scale balance of photovoltaic-energy storage-charging sub-clusters and the grid topology connectivity.

[0027] In step S203, for each sub-cluster, aggregate the source-load difference curves of all stations within it, and statistically analyze the time series difference characteristics at the cluster level, including the maximum negative difference (the peak value that needs to be compensated by energy storage discharge), the positive difference duration (the energy storage charging demand window), and the standard deviation of difference fluctuations (characterizing the intensity of uncertainty). Combining the extreme scenario data in the generated source-load difference scenario set, identify the capacity gap risk of the photovoltaic-energy storage-charging sub-cluster under extreme weather or load surges.

[0028] In step S204, based on the difference analysis results of the photovoltaic-energy storage-charging sub-clusters, formulate a hierarchical capacity configuration rule:

[0029] C cluster,i =max(β·ΔP max,i ,α·σ i );

[0030] where C cluster,i is the independent energy storage capacity configuration value of the i-th photovoltaic-energy storage-charging sub-cluster, representing the minimum energy storage capacity to meet the needs of this photovoltaic-energy storage-charging sub-cluster; ΔP max,i is the maximum negative power difference of the i-th photovoltaic-energy storage-charging sub-cluster within the target period, that is, the maximum shortage when the photovoltaic output is insufficient to cover the charging load, which needs to be compensated by energy storage discharge; σ i is the time series standard deviation of the source-load power difference of the i-th photovoltaic-energy storage-charging sub-cluster, characterizing the severity of power fluctuations within the photovoltaic-energy storage-charging sub-cluster; α is the fluctuation safety factor; β is the peak compensation factor;

[0031] In step S205, use historical data to simulate the operating states of each photovoltaic-energy storage-charging sub-cluster with independent energy storage support, and verify whether the configured capacity meets the following indicators:

[0032] 1) Coverage rate: The capacity gap under extreme scenarios does not exceed the threshold;

[0033] 2) Utilization rate: The number of energy storage charge-discharge cycles meets the set threshold to avoid long-term idleness;

[0034] 3) Economy: The investment cost and revenue are balanced;

[0035] For the photovoltaic-energy storage-charging sub-clusters that do not meet the standards, perform dynamic optimization by adjusting the safety factor, re-dividing the clusters, or introducing a shared energy storage interface.

[0036] The specific implementation method for safety factor adjustment is as follows: Based on historical data, dynamically adjust the safety factor in capacity calculation. For example, the fluctuating safety factor is optimized online through a reinforcement learning algorithm to rebalance economy, utilization rate, and coverage rate;

[0037] The specific implementation method for re-clustering the cluster is as follows: If significant spatio-temporal complementary changes occur among multiple sub-clusters, re-execute the clustering algorithm, adjust the number of clustering centers or similarity weights, and generate a better sub-cluster partitioning scheme;

[0038] Introduce a shared energy storage interface: Preset a shared energy storage call interface for high-fluctuation sub-clusters, define the proportion of shared capacity that can be applied for and the call priority, and constrain the capacity allocation rules through a bilateral contract mechanism to ensure seamless connection between local optimization and global scheduling.

[0039] According to the above technical solution, configuring the shared energy storage capacity for multiple clusters includes the following steps:

[0040] S301. Analyze the spatio-temporal complementarity of multiple photovoltaic-storage-charging sub-clusters;

[0041] S302. Construct a shared energy storage optimization model;

[0042] S303. Perform multi-objective collaborative solution and capacity allocation for the optimization model.

[0043] According to the above technical solution, in S301, based on the independent energy storage configuration results of the photovoltaic-storage-charging sub-clusters, collect the time-series source-load difference data of each sub-cluster, and analyze the spatio-temporal complementary characteristics across clusters;

[0044] Identify the following patterns for the analysis of spatio-temporal complementary characteristics:

[0045] 1) There is a "positive-negative" difference complementarity among geographically adjacent photovoltaic-storage-charging sub-clusters in the same time period;

[0046] 2) Cross-time period complementarity;

[0047] Through correlation analysis and energy transmission path evaluation, screen out combinations of photovoltaic-storage-charging sub-clusters with high complementary potential;

[0048] Correlation analysis refers to: Extract the source-load difference curves of each sub-cluster, use the dynamic time warping algorithm to calculate the time alignment similarity between the curves, identify combinations of sub-clusters with significant "positive-negative" difference overlap characteristics, and count the proportion of complementary time periods and the energy gap matching degree;

[0049] Energy transmission path evaluation: Use an improved shortest path algorithm to preferentially screen transmission paths with short electrical distances and sufficient remaining line capacity.

[0050] In S302, a two - layer optimization model is established: the upper - layer model determines the total capacity and location layout of the shared energy storage, considering the grid - node voltage constraints, energy transmission losses, and land construction costs; the lower - layer model defines the capacity allocation rules and scheduling priorities of each photovoltaic - energy - storage - charging sub - cluster for the shared energy storage, embedding the operating constraints of charge - discharge efficiency and life attenuation;

[0051] The first step: Construction of the upper - layer model

[0052] First, based on the grid topology structure and land - resource constraints, determine the candidate placement locations of the shared energy storage and set the initial range of the total capacity; subsequently, aiming at minimizing the investment cost and transmission losses, embed the grid constraints such as node - voltage safety limits and line capacities, and establish a global optimization model for the capacity and location of the shared energy storage.

[0053] The second step: Construction of the lower - layer model

[0054] According to the candidate capacity and location output by the upper layer, for the real - time power deficit data of each sub - cluster, construct a dynamic capacity - allocation model aiming at minimizing the operating cost, embed the operating constraints such as charge - discharge efficiency, life attenuation, and scheduling priority, and generate the shared - energy - storage invocation rules for each sub - cluster.

[0055] The model also integrates the existing capacities of the independent energy storages of the photovoltaic - energy - storage - charging sub - clusters to achieve the collaborative capacity planning of "independent + shared" energy storage; the optimization objective of the shared energy storage can be expressed as:

[0056]

[0057] where E shared is the total energy capacity of the shared energy storage, which is a variable to be optimized and represents the maximum storable electricity that needs to be configured for the shared energy storage; C inv is the investment cost per unit capacity of the shared energy storage, including equipment purchase, land construction, and installation costs; N is the total number of photovoltaic - energy - storage - charging sub - clusters, determined by the clustering result; T is the total number of time periods in the scheduling cycle; c op,i is the unit - power operating cost of the photovoltaic - energy - storage - charging sub - cluster i for invoking the shared energy storage in time period t, including energy loss, life depreciation, and transaction management costs; P shared,i (t) is the charge - discharge power of the shared energy storage allocated to the photovoltaic - energy - storage - charging sub - cluster i in time period t, with discharge being positive and charge being negative, which is a decision variable;

[0058] In step S303, CPLEX is used to solve the optimization model, and the total capacity, placement location of the shared energy storage, and the capacity - allocation ratio of each photovoltaic - energy - storage - charging sub - cluster are output;

[0059] The specific solution process is as follows:

[0060] The first step: Model standardization and interface configuration;

[0061] Convert the shared energy storage optimization model into the standard form of mixed - integer linear programming. Through the Python API of CPLEX, such as docplex, define the decision variables, objective function, and constraint conditions, and configure the solver parameters.

[0062] Step 2: Hierarchical decomposition and iterative solution

[0063] Split the two - layer model into the master problem (upper - layer capacity planning) and the sub - problem (lower - layer scheduling allocation). After the master problem generates the initial capacity plan, call CPLEX to solve the sub - problem to verify the scheduling feasibility. If the sub - problem cannot meet the constraints due to insufficient capacity, generate a cutting plane and feedback it to the master problem, and iteratively correct the capacity configuration until the upper - and lower - layer models converge to the global optimal solution.

[0064] Step 3: Result verification and post - processing

[0065] Output the optimal capacity and scheduling plan that meet all constraints.

[0066] The solution process needs to meet:

[0067] 1) The total capacity of the shared energy storage covers the cross - cluster complementary demand and does not exceed the economic threshold;

[0068] 2) The shared capacity allocated to each photovoltaic - energy storage - charging sub - cluster forms a complement with its independent energy storage, avoiding capacity overlap and waste;

[0069] 3) The power flow security constraints of the power grid are controlled throughout the process.

[0070] According to the above technical solution, the configuration of the shared energy storage capacity for multiple clusters further includes:

[0071] S304. Construct the scheduling strategy of the shared energy storage to meet the following requirements:

[0072] 1) The independent energy storage within the photovoltaic - energy storage - charging sub - cluster responds to the local difference demand first;

[0073] 2) When there is an excess negative difference in the photovoltaic - energy storage - charging sub - cluster, call the shared energy storage capacity according to the priority;

[0074] 3) During the low - electricity - price or photovoltaic - surplus period, coordinate multiple photovoltaic - energy storage - charging sub - clusters to charge the shared energy storage centrally to reduce the power purchase cost;

[0075] Synchronously design the cross - cluster energy trading mechanism to clarify the charging and discharging rights and cost sharing ratio of the shared energy storage.

[0076] Compared with the prior art, the beneficial effects of the present invention are:

[0077] Through a hierarchical collaborative energy storage capacity configuration mechanism, the present invention effectively solves the trade-off problem between economy and robustness in traditional methods. First, an uncertainty model of the source-load difference is established, which accurately quantifies the spatio-temporal coupling characteristics of photovoltaic output and charging load, avoiding capacity redundancy caused by deterministic planning. Second, independent energy storage is configured based on the differentiated needs of sub-clusters. By suppressing local fluctuations, the repeated investment of decentralized energy storage is reduced, and at the same time, the complementary characteristics within the cluster are utilized to reduce the single-point capacity demand. Finally, shared energy storage is introduced and a cross-cluster collaborative optimization model is constructed to significantly improve the overall capacity utilization rate. Under the synergistic effect of the three, while ensuring high-proportion photovoltaic consumption and reliable power supply for charging loads, the system enhances its adaptability to extreme scenarios through the hierarchical response mechanism of "independent + shared" energy storage, providing an expandable solution for the new power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 It is a schematic diagram of the energy storage difference in the embodiment of the present invention;

[0079] Figure 2 It is a schematic diagram of the energy storage configuration in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0080] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0081] The present invention provides a technical solution for an energy storage capacity configuration method for regulating a cluster of photovoltaic-storage-charging stations. A source-load capacity difference model of distributed photovoltaic power stations and charging stations in a target area is established based on the randomness of photovoltaic output;

[0082] The establishment of the difference model includes the following steps:

[0083] S101. Collect multi-source data of distributed photovoltaic power stations and charging stations and perform preprocessing to obtain photovoltaic data and charging load data;

[0084] First, the historical output data of distributed photovoltaic stations in the target area are collected, covering the output fluctuation characteristics under different seasons, weather types and typical extreme events; the historical operation data of the charging station is simultaneously obtained, including the charging power timing curve, user charging behavior statistics and the association records with the grid time-of-use electricity price; then the original data is cleaned, the outliers caused by sensor failures are eliminated, and the missing data are processed by linear interpolation or adjacent time period mean filling, and all data are unified to the same time resolution, and the dimension differences are eliminated by normalization to obtain the pre-processed photovoltaic data and charging load data;

[0085] S102, based on the pre-processed photovoltaic data, analyzing the probability distribution characteristics of distributed photovoltaic output to obtain a photovoltaic output curve; performing multi-dimensional decomposition on the pre-processed charging load data to obtain a charging load curve;

[0086] For the photovoltaic output under sunny weather, the probability density function fitting method is used to describe the distribution law of the output value; for the photovoltaic output under extreme weather, the upper and lower limits of the photovoltaic output fluctuation are constructed. For example, the minimum output value under rainy weather is taken as the lower limit, and the theoretical maximum value under clear sky conditions is taken as the upper limit, forming a robust photovoltaic output boundary, and finally obtaining the photovoltaic output curve;

[0087] Identify the basic load components and elastic load components, and establish models for the intraday and seasonal changes to obtain the charging load curve; further analyze the coupling relationship between user behavior and grid regulation strategy, such as the impact of peak-valley electricity price differences on charging period migration, and build a dynamic prediction framework for charging load;

[0088] S103, based on the photovoltaic output curve and the charging load curve, calculating the source-load capacity difference to obtain a source-load capacity difference model;

[0089] The photovoltaic output curve and the charging load curve are aligned along the time axis, and the power difference between the two is calculated time period by time. At the same time, positive and negative differences are defined. A positive difference means that the photovoltaic output is greater than the charging load and needs to be stored or transmitted. A negative difference means that the photovoltaic output is less than the charging load and needs energy storage or grid supplementation. The probability distribution of distributed photovoltaic output or the upper and lower limits of photovoltaic output fluctuations are substituted into the power difference calculation to generate a power difference fluctuation range with confidence, quantify the maximum power shortage or redundancy in extreme scenarios, form a source-load difference time series containing uncertainty information, and finally obtain a source-load capacity difference model.

[0090] Construct a source-load difference scenario set based on the source-load capacity difference model;

[0091] Using the historical difference curve formed by the source-load capacity difference model, a clustering algorithm is used to extract representative scenarios and construct a typical scenario library. At the same time, using the Monte Carlo simulation method, combined with the photovoltaic output probability model and the random characteristics of the charging load, a set of random scenarios covering normal and extreme conditions is generated in batches to obtain the source-load difference scenario set. Finally, the coverage and rationality of the scenario set are verified through historical data backtesting to ensure that it can fully reflect the differentiated operation requirements of the photovoltaic-storage-charging station cluster.

[0092] Based on the source-load difference scenario set, the energy storage capacity of the photovoltaic-storage-charging sub-cluster is configured according to the capacity difference of different photovoltaic-storage-charging sub-clusters.

[0093] For the configuration of the energy storage capacity of the photovoltaic-storage-charging sub-cluster, the following steps are included:

[0094] S201. Extract the division characteristics of the photovoltaic-storage-charging sub-cluster and determine the clustering basis.

[0095] Extract key features from the generated source-load difference scenario set, including the capacity difference time series curve, geographical location, and operation constraints of each station. Determine "difference curve shape similarity" and "geographical proximity" as the core clustering basis to ensure that stations within the same photovoltaic-storage-charging sub-cluster have complementary regulation potential and are convenient for sharing the energy storage layout.

[0096] S202. Use the dynamic clustering algorithm to achieve the division of the photovoltaic-storage-charging sub-cluster.

[0097] Adopt the adaptive weighted clustering algorithm, giving higher weight to the time series difference similarity and secondary weight to the geographical distance. Through iterative optimization of the cluster center and station attribution, all photovoltaic-storage-charging stations are divided into several photovoltaic-storage-charging sub-clusters, so that the difference curves of stations within the same photovoltaic-storage-charging sub-cluster show complementary characteristics in time series, while the difference patterns between different clusters are significantly different. The clustering results need to meet the constraints of the scale balance of the photovoltaic-storage-charging sub-cluster and the grid topology connectivity.

[0098] S203. Statistically analyze the capacity difference of the photovoltaic-storage-charging sub-cluster and conduct demand analysis.

[0099] For each sub-cluster, aggregate the source-load difference curves of all stations within it, and statistically analyze the cluster-level time series difference characteristics, including the maximum negative difference (the peak value that needs to be compensated by energy storage discharge), the positive difference duration (the energy storage charging demand window), and the standard deviation of the difference fluctuation (characterizing the intensity of uncertainty). Combining the extreme scenario data in the generated source-load difference scenario set, identify the capacity gap risk of the photovoltaic-storage-charging sub-cluster under extreme weather or load surges.

[0100] S204. Formulate the independent energy storage capacity rule for the photovoltaic-storage-charging sub-cluster.

[0101] Based on the difference analysis results of the photovoltaic-storage-charging sub-cluster, formulate a hierarchical capacity configuration rule:

[0102]

[0103] Among them, C cluster,i is the independent energy storage capacity configuration value of the i-th photovoltaic-storage-charging sub-cluster, representing the minimum energy storage capacity to meet the requirements of this photovoltaic-storage-charging sub-cluster; ΔP max,i is the maximum negative power difference of the i-th photovoltaic-storage-charging sub-cluster within the target period, that is, the maximum shortfall when the photovoltaic output is insufficient to cover the charging load, and energy storage discharge compensation is required; σ i is the time series standard deviation of the power difference between the source and load of the i-th photovoltaic-storage-charging sub-cluster, characterizing the severity of power fluctuations within the photovoltaic-storage-charging sub-cluster; α is the fluctuation safety factor; β is the peak compensation factor;

[0104] S205. Verify and dynamically adjust the result after configuring the energy storage capacity of the photovoltaic-storage-charging sub-cluster.

[0105] Use historical data to simulate the operating status of each photovoltaic-storage-charging sub-cluster with independent energy storage support, and verify whether the configured capacity meets the following indicators:

[0106] 1) Coverage rate: The capacity gap does not exceed the threshold under extreme scenarios;

[0107] 2) Utilization rate: The number of energy storage charge and discharge cycles meets the set threshold to avoid long-term idle;

[0108] 3) Economy: The investment cost is balanced with the revenue;

[0109] For the photovoltaic-storage-charging sub-clusters that do not meet the standards, perform dynamic optimization by adjusting the safety factor, re-dividing the clusters or introducing a shared energy storage interface.

[0110] The specific implementation method of adjusting the safety factor is as follows: Based on historical data, dynamically adjust the safety factor (such as the fluctuation safety factor) in the capacity calculation, and online optimize the coefficient value through a reinforcement learning algorithm to achieve rebalancing of economy, utilization rate and coverage rate;

[0111] The specific implementation method of re-dividing the clusters is as follows: If there are significant spatio-temporal complementary changes among multiple sub-clusters, re-execute the clustering algorithm, adjust the number of clustering centers or similarity weights, and generate a better sub-cluster division plan;

[0112] Introduce a shared energy storage interface: Preset a shared energy storage call interface for high-fluctuation sub-clusters, define the proportion of shared capacity that can be applied for and the call priority, and ensure seamless connection between local optimization and global scheduling through a bilateral contract mechanism to constrain the capacity allocation rules.

[0113] Configure the shared energy storage capacity for multiple clusters; including the following steps:

[0114] S301. Analyze the spatio-temporal complementarity of multiple photovoltaic-storage-charging sub-clusters;

[0115] Based on the independent energy storage configuration results of the photovoltaic-storage-charging sub-clusters, collect the time-series source-load difference data of each sub-cluster, and analyze the spatio-temporal complementary characteristics across sub-clusters;

[0116] For the analysis of spatio-temporal complementary characteristics, identify the following patterns:

[0117] 1) There is "positive-negative" difference complementarity among geographically adjacent photovoltaic-storage-charging sub-clusters in the same time period;

[0118] 2) Cross-time period complementarity;

[0119] Through correlation analysis and energy transmission path evaluation, screen out the combinations of photovoltaic-storage-charging sub-clusters with high complementary potential;

[0120] Correlation analysis refers to: Extract the source-load difference curves of each sub-cluster, use the dynamic time warping (DTW) algorithm to calculate the time alignment similarity between the curves, identify the sub-cluster combinations with significant

[0121] "positive-negative" difference overlap characteristics, and count the proportion of complementary time periods and the matching degree of energy gaps;

[0122] Energy transmission path evaluation: Use the improved shortest path algorithm to preferentially screen the transmission paths with short electrical distances and abundant remaining line capacities.

[0123] S302. Construct an optimized model for shared energy storage;

[0124] Establish a two-layer optimization model: The upper-layer model determines the total capacity and location layout of the shared energy storage, considering the grid node voltage constraints, energy transmission losses, and land construction costs; the lower-layer model defines the capacity allocation rules and scheduling priorities of each photovoltaic-storage-charging sub-cluster for the shared energy storage, and embeds the operation constraints of charge-discharge efficiency and life attenuation;

[0125] The first step: Construction of the upper-layer model

[0126] First, based on the grid topology structure and land resource constraints, determine the candidate layout positions of the shared energy storage and set the initial range of the total capacity; subsequently, with the goal of minimizing investment costs and transmission losses, embed grid constraints such as node voltage safety limits and line capacities, and establish a global optimization model for the capacity and location of the shared energy storage.

[0127] The second step: Construction of the lower-layer model

[0128] According to the candidate capacity and location output by the upper layer, for the real-time power deficit data of each sub-cluster, a dynamic capacity allocation model with the goal of minimizing operating costs is constructed, embedding operating constraints such as charge-discharge efficiency, life attenuation, and scheduling priority, and generating the sharing energy storage call rules for each sub-cluster.

[0129] The model simultaneously integrates the existing capacities of the independent energy storages in the optical storage charging sub-clusters to achieve the collaborative capacity planning of "independent + shared" energy storage; the optimization goal of the shared energy storage can be expressed as:

[0130]

[0131] where E shared is the total energy capacity of the shared energy storage, which is a variable to be optimized and represents the maximum chargeable amount of the shared energy storage to be configured; C inv is the unit capacity investment cost of the shared energy storage, including equipment purchase, land construction, and installation costs; N is the total number of optical storage charging sub-clusters, determined by the clustering results; T is the total number of time periods in the scheduling cycle; c op,i is the unit power operating cost of the optical storage charging sub-cluster i calling the shared energy storage in time period t, including energy loss, life depreciation, and transaction management costs; P shared,i (t) is the charge-discharge power of the shared energy storage allocated to the optical storage charging sub-cluster i in time period t, with discharge being positive and charge being negative, which is a decision variable;

[0132] S303. Perform multi-objective collaborative solution and capacity allocation for the optimization model;

[0133] Use CPLEX to solve the optimization model and output the total capacity of the shared energy storage, the placement location, and the capacity allocation ratio of each optical storage charging sub-cluster;

[0134] The specific solution process is as follows:

[0135] The first step: Model standardization and interface configuration

[0136] Convert the shared energy storage optimization model into the standard form of mixed-integer linear programming (MILP). Define decision variables (capacity, power allocation), objective function (cost minimization), and constraint conditions (grid security, equipment life, etc.) through the Python API of CPLEX (such as docplex), and configure the solver parameters (such as optimal gap tolerance, maximum calculation time).

[0137] The second step: Hierarchical decomposition and iterative solution

[0138] Next, the two-layer model is decomposed into the main problem (upper-layer capacity planning) and the sub-problem (lower-layer scheduling and allocation). After the main problem generates the initial capacity plan, CPLEX is called to solve the sub-problem to verify the scheduling feasibility. If the sub-problem cannot meet the constraints due to insufficient capacity, a cutting plane is generated and fed back to the main problem to iteratively correct the capacity configuration until the upper and lower layer models converge to the global optimal solution.

[0139] Step 3: Result verification and post-processing

[0140] Output the optimal capacity and scheduling plan that meet all constraints.

[0141] The solution process must satisfy:

[0142] 1) The total capacity of the shared energy storage covers the cross-cluster complementary demand and does not exceed the economic threshold;

[0143] 2) The shared capacity allocated to each optical storage charging sub-cluster complements its independent energy storage to avoid capacity overlap and waste;

[0144] 3) The power grid power flow security constraints are controlled throughout the process.

[0145] S304. Construct the scheduling strategy of the shared energy storage to meet the following requirements:

[0146] 1) The independent energy storage within the optical storage charging sub-cluster responds to the local differential demand first;

[0147] 2) When there is an excess negative difference in the optical storage charging sub-cluster, call the shared energy storage capacity according to the priority;

[0148] 3) During the low electricity price or photovoltaic surplus period, coordinate multiple optical storage charging sub-clusters to charge the shared energy storage centrally to reduce the electricity purchase cost;

[0149] Synchronously design the cross-cluster energy trading mechanism to clarify the charging and discharging rights and cost sharing ratio of the shared energy storage.

[0150] Steps for designing the cross-cluster energy trading mechanism:

[0151] Design the charging and discharging rights allocation rules. Including:

[0152] Dynamic priority setting: Based on the real-time power deficit degree of the sub-cluster and the grid emergency state, establish a dynamic call priority ranking. For example, a sub-cluster with a deficit exceeding the threshold (such as 30% of the load) is automatically raised to the highest priority.

[0153] Contribution weight accumulation: Statistically calculate the cumulative charging amount of each sub-cluster to the shared energy storage and assign the discharging rights according to the proportion. For example, if sub-cluster A contributes 1 MWh of charging amount, it can exchange for 0.8 MWh of discharging rights to encourage active participation in energy sharing.

[0154] Then, formulate the cost sharing ratio:

[0155] Basic sharing model: Share the investment cost of the shared energy storage according to the ratio of the actual call capacity.

[0156] Reward and punishment correction mechanism: Introduce the "usage efficiency coefficient", charge a maintenance fee for the sub-cluster whose call capacity exceeds the contribution degree, and vice versa, reduce the fee for high-efficiency contributors.

[0157] Embodiment:

[0158] To verify the effectiveness of the present invention, we select a distribution network connected to a photovoltaic-energy storage-charging station as the research object. Photovoltaic capacity: 10 stations, with a total installed capacity of 5 MW, and the output fluctuations of sunny / cloudy days are simulated using the Beta distribution. Charging load: The total peak load is 4.8 MW, including two types of loads, namely commuting vehicles (morning / evening rush hours) and logistics vehicles (midday flat peak). Energy storage parameters: Lithium battery energy storage, with a unit investment cost of 1500 yuan / kWh, a charge-discharge efficiency of 95%, and a cycle life of 6000 times. Grid constraints: Reverse power feeding is allowed, but the power does not exceed 2 MW, and the peak-valley electricity price difference is 0.8 yuan / kWh.

[0159] As Figure 1 shown, Sub-cluster 1 (orange) has a significant positive difference (requiring energy storage charging) due to excessive photovoltaic output near 12:00, while Sub-cluster 2 (blue) has a negative difference (requiring energy storage discharging) due to the charging load peak at 8:00 and 18:00. The spatio-temporal complementarity between the two (filled area) indicates that the shared energy storage can store the excess photovoltaic power of Sub-cluster 1 at noon and supply power to Sub-cluster 2 during the morning and evening rush hours, realizing cross-cluster energy reuse.

[0160] As Figure 2 shown, Independent energy storage (gray): Each sub-cluster is independently configured with energy storage, resulting in a high total cost (8.2 million yuan) and a low photovoltaic accommodation rate (71%). Shared energy storage (blue): Although the cost is reduced to 6.5 million yuan, due to the lack of independent energy storage buffer for sub-clusters, the accommodation rate only increases slightly (83%). The solution of the present invention (orange): Through the "independent + shared" hierarchical configuration, the total cost is further reduced to 5.3 million yuan (a 35.4% reduction compared to the independent solution), and the photovoltaic accommodation rate is increased to 95%. This verifies the dual advantages of the hierarchical cooperation mechanism in terms of cost and performance.

[0161] The combination of the two figures proves that the present invention, through the exploration of sub-cluster complementarity and hierarchical energy storage configuration, significantly improves the photovoltaic accommodation rate while reducing the investment cost, solves the contradiction between "economy - reliability" in the traditional solution, and provides an innovative solution for the photovoltaic-energy storage-charging station cluster.

[0162] In one embodiment, for the synchronous design of the cross-cluster energy trading mechanism, the charging and discharging permissions and cost sharing ratios of the shared energy storage are specified as follows:

[0163] Charge and discharge authority: Since the shortfall of a certain cluster reached 2 MW during the evening peak (with the highest priority), 1.5 MW of shared energy storage was called. Among them, 1.0 MW came from its own historical contribution, and 0.5 MW came from the redundancy of other clusters.

[0164] Cost sharing: 60% of the total cost is shared according to the call volume, and 40% is allocated according to the contribution degree.

[0165] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. A method for configuring energy storage capacity for the cluster control of optical storage and charging stations, characterized in that, Based on the randomness of photovoltaic output, a source-load capacity difference model between distributed photovoltaic stations and charging stations in the target area is established; Construct a source-load difference scenario set based on the source-load capacity difference model; Based on the source-load difference scenario set, the energy storage capacity of the photovoltaic-storage-charging sub-clusters is configured based on the capacity difference of different photovoltaic-storage-charging sub-clusters; Configure shared energy storage capacity for multiple clusters.

2. The energy storage capacity configuration method for the cluster regulation of a photovoltaic energy storage charging station according to claim 1, wherein The establishment of the difference model includes the following steps: S101, collecting and preprocessing multi-source data of distributed photovoltaic stations and charging stations to obtain photovoltaic data and charging load data; S102, based on the pre-processed photovoltaic data, analyzing the probability distribution characteristics of distributed photovoltaic output to obtain a photovoltaic output curve; performing multi-dimensional decomposition on the pre-processed charging load data to obtain a charging load curve; S103: Based on the photovoltaic output curve and the charging load curve, a source-load capacity difference calculation is performed to obtain a source-load capacity difference model.

3. A method for configuring energy storage capacity for the cluster control of a photovoltaic-storage-charging station yard, as claimed in claim 2, wherein In step S101, first, the historical output data of the distributed photovoltaic stations in the target area are collected; the historical operation data of the charging stations are simultaneously obtained; then the original data is cleaned, the outliers caused by sensor failures are eliminated, the missing data are processed by linear interpolation or adjacent time period mean filling, and all data are unified to the same time resolution, and the dimension differences are eliminated by normalization to obtain the pre-processed photovoltaic data and charging load data; In the analysis of the probability distribution characteristics of distributed photovoltaic output in step S102, for the photovoltaic output under fine weather, a probability density function fitting method is used to describe the distribution law of the output value; for the photovoltaic output under extreme weather, the upper and lower limits of the photovoltaic output fluctuation are constructed to form a robust photovoltaic output boundary, and finally a photovoltaic output curve is obtained; In step S102, in multi-dimensional decomposition of charging load data, basic load components and elastic load components are identified, and intra-day and seasonal variation law models are established respectively to obtain a charging load curve; In step S103, the photovoltaic output curve and the charging load curve are aligned along the time axis, and the power difference between the two is calculated time period by time period. At the same time, the positive difference and the negative difference are defined; the probability distribution of distributed photovoltaic output or the upper and lower limits of photovoltaic output fluctuation are substituted into the power difference calculation to generate a power difference fluctuation range with confidence, quantify the maximum power shortage or redundancy in extreme scenarios, form a source-load difference time series containing uncertainty information, and finally obtain a source-load capacity difference model.

4. A method for configuring energy storage capacity for cluster control of a photovoltaic energy storage charging station, according to claim 1, characterized in that: Using the historical difference curve formed by the source-load capacity difference model, a clustering algorithm is used to extract representative scenarios and build a typical scenario library. At the same time, the Monte Carlo simulation method is used to combine the photovoltaic output probability model and the random characteristics of the charging load to batch generate a set of random scenarios covering normal and extreme conditions. The source-load difference scenario set is obtained. Finally, the coverage and rationality of the scenario set are verified through historical data backtesting.

5. A method for configuring energy storage capacity for the cluster control of a photovoltaic energy storage charging station, as claimed in claim 1, wherein The configuration of the energy storage capacity of the photovoltaic storage sub-cluster includes the following steps: S201, extracting the characteristics of the optical storage and charging sub-cluster division and determining the clustering basis; S202, using a dynamic clustering algorithm to realize the division of optical storage and charging sub-clusters; S203. Statistically analyze the capacity difference of the photovoltaic-storage-charging sub-clusters and conduct demand analysis; S204. Formulate the independent energy storage capacity rules for the photovoltaic-storage-charging sub-clusters; S205. Verify and dynamically adjust the results after the energy storage capacity configuration of the photovoltaic-storage-charging sub-clusters.

6. The energy storage capacity configuration method for the cluster control of a photovoltaic energy storage charging station according to claim 5, wherein, In step S201, extract key features from the generated source-load difference scenario set; determine "difference curve form similarity" and "geographical proximity" as the core clustering basis to ensure that the stations within the same photovoltaic-storage-charging sub-cluster have complementary regulation potential and are convenient for sharing the energy storage layout; In step S202, adopt an adaptive weighted clustering algorithm, assign a higher weight to the temporal difference similarity and a lower weight to the geographical distance; through iterative optimization of the cluster center and station attribution, divide all photovoltaic-storage-charging stations into several photovoltaic-storage-charging sub-clusters, so that the difference curves of the stations within the same photovoltaic-storage-charging sub-cluster show complementary characteristics in time series, while the difference patterns between different clusters are significantly different; the clustering results need to meet the constraints of the scale balance of the photovoltaic-storage-charging sub-clusters and the grid topology connectivity; In step S203, for each sub-cluster, aggregate the source-load difference curves of all stations within it, and statistically analyze the cluster-level temporal difference characteristics; combine the extreme scenario data in the generated source-load difference scenario set to identify the capacity gap risk of the photovoltaic-storage-charging sub-clusters under extreme weather or load surges; In step S204, based on the difference analysis results of the photovoltaic-storage-charging sub-clusters, formulate a hierarchical capacity configuration rule: C cluster,i = max(β·ΔP max,i , α·σ i ); Among them, C cluster,i is the independent energy storage capacity configuration value of the i-th photovoltaic-energy storage-charging sub-cluster, representing the minimum energy storage capacity that meets the requirements of this photovoltaic-energy storage-charging sub-cluster; ΔP max,i is the maximum negative power difference of the i-th photovoltaic-energy storage-charging sub-cluster within the target period, that is, the maximum shortfall when the photovoltaic output is insufficient to cover the charging load, and energy storage discharge compensation is required; σ i is the temporal standard deviation of the source-load power difference of the i-th photovoltaic-energy storage-charging sub-cluster, characterizing the severity of power fluctuations within the photovoltaic-energy storage-charging sub-cluster; α is the fluctuation safety factor; β is the peak compensation factor; In step S205, use historical data to simulate the operating status of each photovoltaic-storage-charging sub-cluster with the support of independent energy storage, and verify whether the configured capacity meets the following indicators: 1) Coverage rate: The capacity gap does not exceed the threshold under extreme scenarios; 2) Utilization rate: The charge-discharge cycle times of the energy storage meet the set threshold; 3) Economy: The investment cost and revenue are balanced; For the photovoltaic-storage-charging sub-clusters that do not meet the standards, dynamically optimize them by adjusting the safety factor, re-dividing the clusters or introducing a shared energy storage interface.

7. A method for configuring energy storage capacity for the cluster control of a photovoltaic-storage-charging station yard, characterized in that, The configuration of the shared energy storage capacity for multiple clusters includes the following steps: S301. Analyze the spatio-temporal complementarity of multiple photovoltaic-storage-charging sub-clusters; S302. Construct an optimization model for shared energy storage; S303. Conduct multi-objective collaborative solution and capacity allocation for the optimization model.

8. A method for configuring energy storage capacity for the cluster regulation of a photovoltaic-storage-charging station, according to claim 7, wherein In S301, based on the configuration results of the photovoltaic-storage-charging sub-clusters and their independent energy storage, collect the temporal source-load difference data of each sub-cluster and analyze the spatio-temporal complementary characteristics across clusters; Identify the following patterns for the analysis of spatio-temporal complementary characteristics: 1) There is a "positive-negative" difference complementarity between geographically adjacent photovoltaic-storage-charging sub-clusters at the same time period; 2) Cross-time period complementarity; Through correlation analysis and energy transmission path evaluation, screen out the combinations of photovoltaic-storage-charging sub-clusters with high complementary potential; In S302, establish a two-layer optimization model: The upper layer model determines the total capacity and location layout of the shared energy storage, considering the grid node voltage constraints, energy transmission losses and land construction costs; the lower layer model defines the capacity allocation rules and scheduling priorities of each photovoltaic-storage-charging sub-cluster for the shared energy storage, and embeds the operating constraints of charge-discharge efficiency and life attenuation; The model simultaneously integrates the existing capacity of the independent energy storage of the photovoltaic-storage-charging sub-clusters to achieve the collaborative capacity planning of "independent + shared" energy storage; In step S303, CPLEX is used to solve the optimization model, and the total capacity of the shared energy storage, the placement locations, and the capacity allocation ratios of each optical storage and charging sub-cluster are output.

9. A method for configuring energy storage capacity for the cluster control of a photovoltaic energy storage charging station, according to claim 7, characterized in that The configuration of the shared energy storage capacity for multiple clusters further includes: S304. Construct a scheduling strategy for the shared energy storage to meet the following requirements: 1) The independent energy storage within the optical storage and charging sub-cluster responds to the local difference demand preferentially. 2) When there is an excess negative difference in the optical storage and charging sub-cluster, the shared energy storage capacity is called according to the priority. 3) During the low electricity price or photovoltaic surplus period, multiple optical storage and charging sub-clusters are coordinated to charge the shared energy storage centrally to reduce the electricity purchase cost. Synchronously design a cross-cluster energy trading mechanism to clarify the charging and discharging permissions and cost sharing ratios of the shared energy storage.

Images