A method for day-ahead and day-ahead operation optimization of energy storage system based on energy supply and demand relationship scene

Abstract

The application discloses a kind of energy storage system day-to-day operation optimization methods based on energy supply and demand relationship scene, wherein the method includes: complex network of power distribution system based on source and load node energy supply and demand similar characteristics is constructed;Similar source and load nodes in different time and space scenes are clustered based on Louvain algorithm maximum modularity, to form optimal community;Further form the day-ahead operation optimization model of composite energy storage;Balance the composite energy storage of battery energy storage and pumped storage, to form the day-ahead operation optimization model of composite energy storage.This application fully considers the day-ahead operation of composite energy storage in dynamic energy supply and demand topology mapping relationship scene, can solve the peak shaving problem through the full adjustment of pumped storage in the day-ahead stage, improve economy;Combined with battery energy storage, pumped storage can realize fast response in the day, more effectively balance energy supply on both sides of supply and demand, tap the potential of composite energy storage.

Description

Technical Field

[0001] This invention relates to the field of scheduling and operation optimization technology for composite energy storage systems, and in particular to a day-ahead-intraday operation optimization method for energy storage systems based on energy supply and demand relationship scenarios. Background Technology

[0002] With the continuous development of new power systems and the large-scale integration of distributed energy resources, the distribution network landscape has undergone significant changes, increasing the demand for flexibility and stability. Energy storage technology, as a flexible energy conversion and regulation method, can effectively address the characteristics of high volatility and intermittency of renewable energy, smoothing power fluctuations on the user side, and is gradually becoming an indispensable and important component of the power system. As energy storage is applied in distribution networks in more and more scenarios, the requirements for energy storage are also becoming higher. It must respond to rapid dynamic changes in the system and meet medium- and long-term load adjustment needs. Single types of energy storage are often insufficient to cope with such mixed scenarios. Therefore, by combining the energy supply and consumption characteristics of various energy storage systems, combining energy storage systems with fast response characteristics and systems with large-capacity energy storage characteristics to form composite energy storage for coordinated control, a more efficient approach is adopted.

[0003] Currently, energy storage systems mainly include: capacity-type energy storage, such as pumped hydro storage, primarily used in peak shaving and off-grid energy storage scenarios; power-type energy storage, such as supercapacitors or superconducting magnetic energy storage, mainly used in situations requiring instantaneous high power input and output, such as assisting AGC frequency regulation or smoothing intermittent power fluctuations; and energy-type energy storage, which falls between capacity-type and power-type energy storage, such as battery energy storage, generally used in composite energy storage scenarios, providing multiple functions such as peak shaving, frequency regulation, and emergency backup. Backup energy storage is mainly used in off-grid black-start scenarios, providing emergency power as an uninterruptible power supply.

[0004] Currently, the technical composition of composite energy storage systems and the selection and configuration of energy storage equipment are addressed in two ways: one is through mathematical models and optimization algorithms to plan and configure composite energy storage systems; the other is by focusing on the operation and control strategies of composite energy storage systems, using intelligent algorithms and control methods to optimize their operation. However, there are relatively few methods for planning, configuring, and optimizing the operation of composite energy storage systems in scenarios with dynamic energy supply and demand topology mapping.

[0005] Patent document CN103248064B discloses a composite energy charging and storage system and its method. The system collects necessary parameters, establishes an economic dispatch mathematical model for the composite energy charging and storage system, and uses an improved particle swarm optimization algorithm to obtain the optimal dispatch command. This reduces the impact on the power grid, ensures stable operation, and achieves high operating efficiency, thus realizing economical operation. However, it also suffers from the problem of not planning the composite energy storage system under dynamic energy supply and demand topology mapping scenarios, making it difficult to achieve a balance between energy supply and demand.

[0006] Patent document CN110009262A discloses a two-stage optimization scheduling method for active distribution networks (ADN) from day-ahead to intraday. This method establishes a mathematical optimization model for day-ahead economic scheduling, intraday regional local power correction, and system-wide voltage adjustment, and uses a harmony search algorithm to solve the scheduling model. Finally, through numerical examples, it is verified that the proposed scheduling method is applicable to both day-ahead and intraday stages, ensuring the economical and safe operation of the ADN. However, it also has the following limitations: 1. It does not include planning for composite energy storage systems under dynamic energy supply and consumption topology mapping scenarios; 2. The day-ahead planning does not employ a two-layer optimization model, resulting in insufficient exploitation of the potential of composite energy storage. Summary of the Invention

[0007] The purpose of this invention is to provide a day-ahead and intraday operation optimization method for energy storage systems based on energy supply and demand scenarios. This method can guide the day-ahead and intraday operation of composite energy storage systems, solve peak-shaving problems through sufficient adjustment of pumped hydro storage during the day-ahead phase, and meet the needs of rapid response by combining battery energy storage and pumped hydro storage during the intraday phase, thereby tapping the potential of composite energy storage.

[0008] This invention provides a day-ahead-intraday operation optimization method for energy storage systems based on energy supply and demand scenarios. The method employs a composite energy storage system combining battery energy storage and pumped hydro storage, and includes the following steps:

[0009] S1. Analyze the spatiotemporal energy supply and consumption similarity characteristics of source and load nodes in the region, and construct a complex power distribution system network based on the energy supply and consumption similarity characteristics.

[0010] S2. Based on the complex network of the power distribution system, the Louvain algorithm is used to maximize the modularity of clustering similar source-load nodes in different spatiotemporal scenarios to form the optimal community.

[0011] S3. Based on the optimal community, a day-ahead operation optimization model for composite energy storage is formed;

[0012] S4. Based on the SOC capacity and charge / discharge constraints of the composite energy storage, the intraday time window is dynamically optimized to balance the composite energy storage consisting of battery energy storage and pumped hydro storage, forming an intraday optimization scheduling model for composite energy storage.

[0013] Furthermore, the topology construction in S1, based on the similar characteristics of power supply and consumption, of a complex power distribution system includes the following steps:

[0014] S11. Obtain the energy supply and consumption similarity characteristics of any two source load nodes, expressed by the formula:

[0015]

[0016] Where S(a,b) and S C(a,b), S L (a,b), S P (a,b), S σ (a,b), S D (a, b) represent the similarity characteristics of energy supply and consumption, category similarity, spatial proximity, average energy supply and consumption level, power fluctuation, and response time requirement between source and load nodes a and b, respectively; x a x b y a y b These are the coordinates of the source load positions represented by nodes a and b. This represents the average coordinate of the source load nodes; σ represents the average output or load of source load nodes a and b, respectively; a σ b τ represents the fluctuation degree of the output and load at source load nodes a and b, respectively; a τ b These represent the response time requirements of source and load nodes a and b for distributed renewable energy consumption and load demand, respectively.

[0017] S12. Calculate the similarity characteristics of power supply and consumption among all source load nodes, determine the topological connection relationship, and obtain the complex network of the power distribution system based on the similarity characteristics of power supply and consumption.

[0018] The topology connection is based on the following: if the energy supply and consumption similarity characteristic S(a,b) between source load nodes a and b is greater than a given threshold, it indicates that there is similarity between the two nodes, and an edge connection is made; otherwise, no edge connection is made.

[0019] Furthermore, based on the importance of source and load nodes in the complex network of the power distribution system, a node centrality evaluation index is established, expressed by the formula:

[0020]

[0021] Among them, D C (a) represents the centrality evaluation index of node a; n is the total number of source load nodes in the network; A ab Let A be a network adjacency element, representing the connection between nodes a and b. If nodes a and b are connected, then A... ab It is 1 if it is true, otherwise it is 0.

[0022] Furthermore, in S2, the Louvain algorithm is used to maximize the modularity of clustering similar source-load nodes in different spatiotemporal scenarios to form communities, including the following steps:

[0023] S21. Treat each node as a separate community and calculate the initial modularity value Q of the entire network.

[0024] S22. For each node, place it in the community of adjacent nodes, calculate the modularity gain after moving it, and move the node to the community that maximizes the modularity; repeat the above process until it is no longer possible to obtain a greater modularity gain by moving nodes.

[0025] S23. Merge all the communities where the nodes are located to form a new super node; based on the merged super node, recalculate the modularity Q of the entire network;

[0026] S24. For the newly formed supernode, repeat the iterative optimization process in step 22 to continue searching for the community that maximizes modularity; repeat the iterative process until no greater modularity gain can be found or the preset number of iterations is reached.

[0027] S25. Output the partitioning results and obtain the optimal community.

[0028] Furthermore, the formula for calculating the modularity value Q is:

[0029]

[0030] Where Q is the modularity, M is the number of edges in the network; k a k b δ(c) represents the degree of node a and node b, respectively, and the number of edges directly connected to nodes a and b, respectively; a ,c b The value indicates whether nodes a and b are in the same community. If they are, the value is 1; otherwise, the value is 0.

[0031] Furthermore, the day-ahead optimization model in S3 aims to minimize regional energy costs, improve user day-ahead energy supply satisfaction, and maximize the fulfillment of user energy supply and demand. It employs an upper-level objective function and a lower-level objective function to construct a day-ahead two-layer optimization model, and performs constrained optimization. The steps include:

[0032] S31. With the goal of minimizing regional energy costs and maximizing user satisfaction with daily energy supply, an upper-level objective function is established, the formula of which is:

[0033] min C TOTAL_DH =C COST_DH +SAT COST_DH (10)

[0034] C COST_DH =C G_DH +C RE_DH +M ES_DH -I G_DH (11)

[0035]

[0036]

[0037] Among them, C TOTAL_DH Total regional energy supply and consumption cost; SAT COST_DH User satisfaction with energy supply as of today; C COST_DH Indicates day-ahead energy cost; C SAT_DH Indicates the penalty cost for regional energy consumption satisfaction; C G_DH Indicates the cost of purchasing electricity from the grid; C RE_DH This represents the penalty cost of abandoning renewable energy; M ES_DH The charging, discharging, operation, and maintenance costs of composite energy storage; I G_DH This indicates the revenue generated from feeding electricity into the power grid; This represents the cost of purchasing electricity from the grid at time t; This represents the day-ahead electricity price at time t. Let t be the amount of photovoltaic power generation discarded. c represents the amount of wind power waste at time t; pun Penalty cost per unit of renewable energy discard; The charging / discharging cost and start / stop cost of the composite energy storage at time t are respectively. c represents the charging and discharging power of battery energy storage and pumped hydro storage at time t, respectively; BES c PES These are the unit operating costs for battery energy storage and pumped hydro storage, respectively. These represent the start-stop transition states of battery energy storage and pumped hydro storage at time t, respectively; that is, 1 if their charging / discharging state is the same as the previous time, and 0 if they are different. BES_SS c PES_SS The unit start-up and shutdown cost for battery energy storage and pumped hydro storage; η is the feed power at time t; G This refers to the feed-in electricity price factor to the power grid.

[0038] S32. To maximize the satisfaction of users' daily energy supply and consumption needs, a lower-level objective function is established, with the following formula:

[0039] min SAT COST_DH =Cap ES c cap +Tim ES c tim (18)

[0040]

[0041] Among them, Cap ES This refers to the capacity mismatch in composite energy storage; c cap The cost mismatch between the capacity and the composite energy storage; TimES c represents the response time mismatch in composite energy storage. tim Costs due to mismatched response time in composite energy storage; The difference between the discharge of the composite energy storage at time t and the load or charging of the nth community and the output of DG; Let N_ec be the load or DG output of the nth community at time t; N_ec is the total number of communities in the cluster. Let t be the difference between the response time of the composite energy storage and the average demand time of the nodes in the nth community; Let T be the average demand time for the nth community at time t; BES With T PES These are the response times for battery energy storage and pumped hydro storage, respectively.

[0042] S33. Apply power balance constraints and composite energy storage constraints to the day-ahead two-layer optimization scheduling model;

[0043] The power balance constraint is expressed by the following formula:

[0044]

[0045] in, η is the charging and discharging power of the battery at time t; BES The charging and discharging efficiency of battery energy storage; η is the charging and discharging power of pumped hydro storage at time t; PES The charging and discharging efficiency of pumped-storage hydroelectric power. The output of wind power and photovoltaic power at time t; Let t be the regional load at time t.

[0046] The formula for the constraint of composite energy storage is:

[0047]

[0048] in, The state of the battery energy storage charging and discharging at time t is a 0-1 variable; The state of pumped water storage charging and discharging at time t is a 0-1 variable; These are the upper and lower limits of maximum charge and discharge for battery energy storage and pumped hydro storage, respectively. Let t represent the state of charge of the battery energy storage and pumped hydro storage at time t. This indicates the upper and lower limits of the state of charge for battery energy storage and pumped hydro storage.

[0049] Furthermore, in S4, based on the SOC capacity of the composite energy storage and the charge / discharge constraints, the intraday time window is dynamically optimized, as expressed by the formula:

[0050]

[0051] in, The time window for rolling optimization of pumped-storage energy storage at time t within the day; The charge / discharge efficiency of battery energy storage; η in_PES Cap is the window coefficient; PES This refers to the capacity of pumped-storage hydroelectric power.

[0052] Furthermore, the composite energy storage system in S4, which balances battery energy storage and pumped hydro storage, includes:

[0053] Based on the day-ahead operation optimization model, with the objective of minimizing the total cost during the intraday time window scheduling period, a composite energy storage system combining battery energy storage and pumped hydro storage is balanced, as shown in the following formula:

[0054]

[0055] in, The cost of purchasing electricity from the grid at time T during the day; This represents the penalty cost of forgoing renewable energy at time T; and This indicates the operating and maintenance costs of battery energy storage and pumped hydro storage; This indicates the cost of the power deviation between the current day and the previous day; Deviation error coefficient; and These represent the amount of electricity purchased from the power grid on the current day and at time T before the current day, respectively.

[0056] The beneficial effects of this invention are:

[0057] 1. This invention fully considers the day-ahead and intraday operation of composite energy storage in scenarios with dynamic energy supply and demand topology mapping. It focuses on the coordination of composite energy storage technology, which combines energy-type (battery storage) and capacity-type (pumped hydro storage), in energy supply and demand scenarios with a wide distribution of various distributed resources. This guides the day-ahead and intraday operation of composite energy storage systems. During the day-ahead phase, the peak-shaving problem can be solved by fully adjusting pumped hydro storage, improving economic efficiency. Combined with battery storage, it can work with pumped hydro storage to achieve rapid response to demand during the day, more effectively balancing energy supply on both the supply and demand sides and tapping the potential of composite energy storage.

[0058] 2. This invention establishes a day-ahead operation optimization model with the goal of minimizing regional energy costs, improving user satisfaction with day-ahead energy supply and maximizing the fulfillment of user energy supply and demand. It adopts a day-ahead two-layer optimization model built with upper-level objective functions and lower-level objective functions, and performs constraint optimization so that the composite energy storage can fully cooperate with the source load and demand in the day-ahead stage, and can fully solve the peak-shaving problem through the adjustment of pumped storage in the day-ahead stage, thereby improving economic efficiency.

[0059] 3. Based on the SOC capacity and charge / discharge constraints of composite energy storage, this invention dynamically optimizes the intraday time window, balances the composite energy storage composed of battery energy storage and pumped hydro storage, and forms an intraday optimized scheduling model for composite energy storage. This model can effectively balance the long-term and short-term scheduling ranges of the two types of energy storage, focusing on both local adjustment ranges and global objectives, resulting in better economic efficiency and effectively improving the grid's ability to absorb wind power. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating a day-ahead-intraday operation optimization method for an energy storage system based on a power supply and demand relationship scenario, according to the present invention.

[0061] Figure 2 This is a schematic diagram of the application architecture of the composite energy storage system of the present invention;

[0062] Figure 3 This is a schematic diagram of the source load and energy storage distribution in a certain area of ​​Anhui Province, as described in an embodiment of the present invention.

[0063] Figure 4 This is a schematic diagram of the distributed source-load community layout at different times according to the present invention;

[0064] Figure 5 This is a schematic diagram of the optimized scheduling results of the present invention;

[0065] Figure 6 This is a schematic diagram of the intraday optimized scheduling results of the present invention;

[0066] Figure 7 This is a schematic diagram of the electronic device structure of the present invention. Detailed Implementation

[0067] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar symbols denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0068] Currently, to more effectively balance the uncertainties on both the supply and demand sides, a method of day-ahead unified optimization and intraday rolling optimization is often adopted. However, energy-type energy storage and capacity-type energy storage have significant differences in response speed and capacity potential. When combined into composite energy storage, day-ahead optimization needs to address the fluctuation levels of source load to tap the potential of composite energy storage. In the intraday phase, the traditional fixed rolling optimization time window often fails to balance the long-term and short-term dispatch ranges of the two types of energy storage. A longer rolling optimization time window is beneficial for leveraging the large-capacity storage potential of capacity-type energy storage, but it also limits the short-term regulation effect of energy-type energy storage, causing the role of composite energy storage to mainly focus on peak shaving and valley filling, neglecting its role in smoothing rapid fluctuations. Short-term rolling optimization, on the other hand, forces capacity-type energy storage to focus only on local regulation ranges, failing to utilize its peak-shaving capabilities and making it difficult to consider global objectives, resulting in frequent start-ups and shutdowns and impacting economic efficiency.

[0069] To address the above problems, this invention provides a day-ahead-intraday operation optimization method for energy storage systems based on energy supply and demand relationship scenarios. Figure 1 This is a flowchart illustrating the day-ahead-intraday scheduling and operation optimization method for a composite energy storage system provided in an embodiment of the present invention.

[0070] Based on the similarity of source and load, communities are divided and corresponding to different day-ahead dispatch ranges for capacity-type and energy-type energy storage, establishing an optimization model for day-ahead composite energy storage. Simultaneously, based on the capacity, charge / discharge rate, and other characteristics of the two types of energy storage, a dynamic intraday dispatch time window is defined for intraday rolling optimization. The combination of pumped hydro storage with its large-capacity peak-shaving function and battery energy storage with its short-term intraday charge / discharge capabilities forms composite energy storage. This allows for sufficient peak-shaving adjustments through pumped hydro storage during the day-ahead phase, improving economic efficiency. Battery energy storage can cooperate with pumped hydro storage intraday to meet rapid response needs. The optimization method includes the following steps:

[0071] S1. Analyze the spatiotemporal energy supply and consumption similarity characteristics of source and load nodes in the region, and construct a complex power distribution system network based on the energy supply and consumption similarity characteristics.

[0072] like Figure 2 , 3 As shown, firstly, the energy supply and consumption similarity characteristics of any two source load nodes are obtained, expressed by the formula:

[0073]

[0074] Where S(a,b) and S C (a,b), S L (a,b), S P (a,b), S σ (a,b), S D(a, b) represent the similarity characteristics of energy supply and consumption, category similarity, spatial proximity, average energy supply and consumption level, power fluctuation, and response time requirement between source and load nodes a and b, respectively; S C (a, b) are 0-1 variables; they are 1 when a and b belong to the same category set U, and 0 otherwise. a x b y a y b These are the coordinates of the source load positions represented by nodes a and b. This represents the average coordinate of the source load nodes; σ represents the average output or load of source load nodes a and b, respectively; a σ b τ represents the fluctuation degree of the output and load at source load nodes a and b, respectively; a τ b These represent the response time requirements of source and load nodes a and b for distributed renewable energy consumption and load demand, respectively.

[0075] Then, the similarity characteristics of power supply and consumption among all source load nodes are calculated, the topological connection relationship is determined, and the complex network of the power distribution system based on the similarity characteristics of power supply and consumption is obtained.

[0076] like Figure 4 As shown, the topology connection is based on the following: if the energy supply and consumption similarity characteristic S(a,b) between source load nodes a and b is greater than a given threshold, it indicates that there is similarity between the two nodes, and an edge connection is made; otherwise, no edge connection is made.

[0077] To evaluate the importance of a node representing a source load in a complex power distribution system network with similar energy supply and demand characteristics, a node centrality evaluation index is established. This index measures the importance of a node in the network by calculating the number of edges connected to it, expressed by the following formula:

[0078]

[0079] Among them, D C (a) represents the centrality evaluation index of node a; n is the total number of source load nodes in the network; A ab Let A be a network adjacency element, representing the connection between nodes a and b. If nodes a and b are connected, then A... ab It is 1 if it is true, otherwise it is 0.

[0080] The right side of the equation represents the total number of edges connecting node a to the other n-1 nodes (excluding node b). The more edges a node connects to other nodes, the higher its centrality, indicating that it can represent the local characteristics of the community and is a representative energy supply node.

[0081] S2. Based on the complex network of the power distribution system, the Louvain algorithm is used to maximize the modularity of clustering similar source-load nodes in different spatiotemporal scenarios to form the optimal community.

[0082] Distributed energy sources or loads with similar energy supply and consumption characteristics will cluster into communities in the network. Each community represents a typical energy supply and consumption pattern. However, there are connections between communities, so the community partitioning result cannot be obtained directly. Community partitioning can be implemented based on Newman's maximum modularity algorithm, and the Louvain algorithm is used to find the optimal community partitioning, including the following steps:

[0083] Initialization: Treat each node as a separate community and calculate the initial modularity value Q of the entire network;

[0084] Iterative optimization involves placing each node in a community of neighboring nodes, calculating the modularity gain after moving it, and moving the node to the community that maximizes modularity. This process is repeated until no further modularity gain can be obtained by moving nodes.

[0085] Merge communities by combining the communities where all nodes reside to form a new supernode; based on the merged supernode, recalculate the modularity Q of the entire network;

[0086] Repeat the iteration process for newly formed supernodes, repeating the single-node iterative optimization process to continue searching for the community that maximizes modularity. Continue the iterative process until no greater modularity gain can be found or the preset number of iterations is reached, then output the partitioning result and obtain the optimal community.

[0087] Modularity is an indicator for evaluating and analyzing community structure, as shown in the following formula. The larger the indicator value, the better the community division result.

[0088] The formula for calculating the modularity value Q is as follows:

[0089]

[0090] Where Q is the modularity, M is the number of edges in the network; ka and kb are the degrees of nodes a and b, respectively, and represent the number of edges directly connected to nodes a and b, respectively; δ(c a ,c b The value indicates whether nodes a and b are in the same community. If they are, the value is 1; otherwise, the value is 0.

[0091] By continuously optimizing the community affiliation of source energy nodes through the Louvain algorithm, the algorithm finds the community partition that maximizes modularity through an iterative process. By merging the communities of nodes to form super nodes, and then performing iterative optimization, the optimal community partition of the network can be effectively found. The optimal community partition will provide a basis for the charging and discharging combination of the collaborative optimization strategy of composite energy storage.

[0092] S3. Based on the optimal community, a day-ahead operation optimization model for composite energy storage is formed.

[0093] The current dispatch needs to consider the overall energy economy of the regional system and reduce regional energy costs. At the same time, due to the differences in capacity and response time of composite energy storage, the satisfaction of users at both the upper and lower levels with the energy supply and use of composite energy storage should also be considered.

[0094] The day-ahead optimization model aims to minimize regional energy costs, improve user satisfaction with day-ahead energy supply, and maximize the fulfillment of user energy needs. It employs an upper-level objective function and a lower-level objective function to construct a two-layer day-ahead optimization model, followed by constrained optimization. The steps include:

[0095] A higher-level objective function is established. The objective of day-ahead energy optimization scheduling is to minimize regional energy costs and user satisfaction with day-ahead energy supply and consumption. Regional energy costs include regional electricity purchase costs, penalty costs for renewable energy abandonment, operation and maintenance costs of composite energy storage, and revenue from power feedback to the grid. The formula is as follows:

[0096] min C TOTAL_DH =C COST_DH +SAT COST_DH (10)

[0097] C COST_DH =C G_DH +C RE_DH +M ES_DH -I G_DH (11)

[0098]

[0099] Among them, C TOTAL_DH Total regional energy supply and consumption cost; SAT COST_DH User satisfaction with energy supply as of today; C COST_DH Indicates day-ahead energy cost; C SAT_DH Indicates the penalty cost for regional energy consumption satisfaction; C G_DH Indicates the cost of purchasing electricity from the grid; C RE_DH This represents the penalty cost of abandoning renewable energy; M ES_DH The charging, discharging, operation, and maintenance costs of composite energy storage; I G_DH This indicates the revenue generated from feeding electricity into the power grid; This represents the cost of purchasing electricity from the grid at time t; This represents the day-ahead electricity price at time t. Let t be the amount of photovoltaic power generation discarded. c represents the amount of wind power waste at time t; pun Penalty cost per unit of renewable energy discard; The charging / discharging cost and start / stop cost of the composite energy storage at time t are respectively. c represents the charging and discharging power of battery energy storage and pumped hydro storage at time t, respectively; BES c PES These are the unit operating costs for battery energy storage and pumped hydro storage, respectively. These represent the start-stop transition states of battery energy storage and pumped hydro storage at time t, respectively; that is, 1 if their charging / discharging state is the same as the previous time, and 0 if they are different. BES_SS c PES_SS The unit start-up and shutdown cost for battery energy storage and pumped hydro storage; η is the feed power at time t; G This refers to the feed-in electricity price factor to the power grid.

[0100] A lower-level objective function is established, primarily aimed at maximizing the satisfaction of users' daily energy supply and consumption needs, i.e., minimizing the energy supply and consumption deviation. Based on the aforementioned community partitioning results with similar energy supply and consumption characteristics, the lower-level optimization variables are the charging and discharging power of each user community, the day-ahead period, the scheduling timescale, user energy requirements, and the energy consumption requirements of distributed renewable energy sources. These variables are mainly related to the energy storage charging and discharging capacity and response time. The formula is as follows:

[0101] min SAT COST_DH =Cap ES c cap +Tim ES c tim (18)

[0102]

[0103] Among them, Cap ES This refers to the capacity mismatch in composite energy storage; c cap The cost mismatch between the capacity and the composite energy storage; Tim ES c represents the response time mismatch in composite energy storage. tim Costs due to mismatched response time in composite energy storage; The difference between the discharge of the composite energy storage at time t and the load or charging of the nth community and the output of DG; Let N_ec be the load or DG output of the nth community at time t; N_ec is the total number of communities in the cluster. Let t be the difference between the response time of the composite energy storage and the average demand time of the nodes in the nth community; Let T be the average demand time for the nth community at time t; BES With T PES These are the response times for battery energy storage and pumped hydro storage, respectively.

[0104] Then, power balance constraints and composite energy storage constraints are applied to the day-ahead two-level optimization scheduling model;

[0105] The power balance constraint is expressed by the following formula:

[0106]

[0107] in, η is the charging and discharging power of the battery at time t; BES The charging and discharging efficiency of battery energy storage; η is the charging and discharging power of pumped hydro storage at time t; PES The charging and discharging efficiency of pumped-storage hydroelectric power. The output of wind power and photovoltaic power at time t; Let t be the regional load at time t.

[0108] The formula for the constraint of composite energy storage is:

[0109]

[0110] in, The state of the battery energy storage charging and discharging at time t is a 0-1 variable; The state of pumped water storage charging and discharging at time t is a 0-1 variable; These are the upper and lower limits of maximum charge and discharge for battery energy storage and pumped hydro storage, respectively. Let t represent the state of charge of the battery energy storage and pumped hydro storage at time t. This indicates the upper and lower limits of the state of charge for battery energy storage and pumped hydro storage.

[0111] S4. Based on the SOC capacity and charge / discharge constraints of the composite energy storage, the intraday time window is dynamically optimized to balance the composite energy storage consisting of battery energy storage and pumped hydro storage, forming an intraday optimization scheduling model for composite energy storage.

[0112] For intraday optimization, the main problem lies in the long response time of pumped hydro storage, and frequent start-ups and shutdowns will increase the cost of pumped hydro storage. Therefore, it is necessary to dynamically adjust the intraday scheduling time window of the hybrid energy storage to balance the advantages of energy-type and capacity-type energy storage.

[0113] First, based on the SOC capacity and charge / discharge constraints of the composite energy storage, the intraday time window is dynamically optimized, as expressed by the formula:

[0114]

[0115] in, The time window for rolling optimization of pumped-storage energy storage at time t within the day; The charge / discharge efficiency of battery energy storage; η in_PES Cap is the window coefficient; PES This refers to the capacity of pumped-storage hydroelectric power.

[0116] Based on the day-ahead operation optimization model, with the objective of minimizing the total cost during the intraday time window scheduling period, a combined energy storage system consisting of battery energy storage and pumped hydro storage is formed, resulting in an intraday optimization scheduling model for combined energy storage. The formula is as follows:

[0117]

[0118] in, The cost of purchasing electricity from the grid at time T during the day; This represents the penalty cost of forgoing renewable energy at time T; and This indicates the operating and maintenance costs of battery energy storage and pumped hydro storage; This indicates the cost of the power deviation between the current day and the previous day; Deviation error coefficient; and These represent the amount of electricity purchased from the power grid on the current day and at time T before the current day, respectively.

[0119] The following analysis uses examples to illustrate the scenario generation method of this invention:

[0120] A simplified verification example was conducted using a region in Anhui Province. Figure 2 This is a capacity-energy hybrid energy storage application architecture for the power distribution network in this region. Figure 3 The diagram shows the power generation and energy storage distribution in the region. The region includes a 1MW / 5MWh pumped hydro storage system, a 500kW / 1.5MWh battery storage system, three 1MW distributed photovoltaic systems, and one 1MW and one 500kW distributed wind turbine. There are 12 residential users, 4 mixed-use residential / commercial users, and 3 industrial users. Some energy storage parameters are shown in Table 1 below.

[0121] Table 1 Parameters of the Composite Energy Storage System

[0122]

[0123] Based on the method for generating complex networks in power distribution systems, aggregated distributed source-load communities, such as... Figure 4 As shown:

[0124] At t=3h, the user load and distributed renewable energy are mainly clustered into two communities. The community represented by node 2 includes industrial user loads and wind power nodes where the energy supply and consumption status is relatively stable in the early morning and night.

[0125] When t=8h, the residential load node is at the peak of daytime energy consumption, the industrial load remains stable, the wind power output decreases, and the photovoltaic and mixed commercial and residential load nodes are all on the rise in energy supply and consumption. Therefore, they are clustered into 4 communities. It can be seen that during the peak period of energy supply and consumption, more communities have emerged due to the increase in energy consumption methods.

[0126] At t=15h, the demand of both residential and commercial loads gradually decreases, and the residential and commercial load nodes begin to aggregate into communities. The residential load nodes and commercial load nodes contain two energy consumption modes.

[0127] At t=23h, all loads showed a downward trend. Due to its anti-peak-shaving characteristics, wind power output began to rise, while photovoltaic power output was not increasing. The three different energy supply and consumption modes converged into three communities.

[0128] Comparing the energy supply and consumption characteristics network and community segmentation at four different time points reveals that energy consumption behavior diversifies with the emergence of peak energy consumption periods, the number of communities gradually increases, and the differences in energy consumption among users widen. Furthermore, as load increases and time progresses, the degree centrality of each community is relatively high during peak energy supply and consumption periods in the daytime, while the degree centrality of communities shows a diversified trend in the evening and afternoon when energy consumption decreases. This indicates that the growth of source load exhibits a similar energy supply and consumption pattern, while the energy consumption pattern begins to diversify when overall energy consumption starts to decline.

[0129] Figure 5 (a) and (b) represent the day-ahead forecasts for residential, commercial, and industrial loads, as well as the day-ahead forecasts for photovoltaic and wind power outputs. To study the day-ahead dispatch of the integrated energy storage system, this invention aims to minimize the total regional operating cost by solving the day-ahead dispatch model, obtaining the optimized day-ahead dispatch results for the integrated energy storage system over a 24-hour period and the corresponding costs. At this time, the dispatch output of grid-purchased electricity, distributed renewable energy, and various energy storage devices is as follows: Figure 5 As shown in (c), the day-ahead SOC changes for battery energy storage and pumped hydro storage are as follows: Figure 5 As shown in (d).

[0130] As shown in the graph, wind power output is at its peak between midnight and 6 AM, while electricity prices are at their lowest, resulting in a relatively stable power load. Pumped hydro storage continuously stores electricity, while battery storage adjusts its electricity sales and generates profit through charging and discharging. The SOC curve also shows that battery storage flexibly meets load demand through deep charging and discharging, while pumped hydro storage only discharges to its limit during peak load periods at night. Furthermore, the charging and discharging of pumped hydro storage is often delayed compared to battery storage, achieving a synergistic effect of combined energy storage.

[0131] from Figure 6 (a) The intraday fluctuations in photovoltaic and wind power output are solved using an intraday dynamic time window optimization model for the composite energy storage system, yielding the intraday dispatch output, SOC, and power purchase difference of the composite energy storage system, as shown below. Figure 6 As shown in (b) to (d), the SOC changes of the composite energy storage and the difference between intraday and day-ahead power purchase and sale rates demonstrate that the frequent charging and discharging of battery energy storage smooths the SOC curve of pumped hydro storage, reducing the impact of frequent pump start-ups and shutdowns on economic efficiency and pump lifespan. Furthermore, the introduction of composite energy storage has significantly improved the peak-valley load in the region, playing a role in mitigating and absorbing the volatility of distributed renewable energy and load fluctuations.

[0132] Table 2. Cost Comparison of Various Energy Storage Options

[0133] type Current operating costs (RMB) Daily operating costs (RMB) HESS 26785.84 27816.35 BESS 27813.45 28651.23 PESS 27522.43 28763.12

[0134] Table 2 compares the costs of adding composite energy storage, using only a single type of energy storage, and not adding any energy storage. The table shows that composite energy storage is effective in reducing daily operating costs, minimizing day-to-day deviations, and improving user satisfaction. In contrast, single-type energy storage has both significant advantages and disadvantages. Single-type battery energy storage has insufficient capacity to effectively smooth peak and valley loads, and single pumped hydro storage has limited response speed, making it difficult to achieve rapid distributed renewable energy and smooth load fluctuations.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A day-ahead-intraday operation optimization method for an energy storage system based on energy supply and demand scenarios, characterized in that: Including the following steps: S1. Analyze the spatiotemporal energy supply and consumption similarity characteristics of source and load nodes in the region, and construct a complex power distribution system network based on the energy supply and consumption similarity characteristics. S2. Based on the complex network of the power distribution system, the Louvain algorithm is used to maximize the modularity of clustering similar source-load nodes in different spatiotemporal scenarios to form the optimal community. S3. Based on the optimal community, a day-ahead operation optimization model for composite energy storage is formed; S4. Based on the SOC capacity and charge / discharge constraints of composite energy storage, dynamically optimize the intraday time window, balance the composite energy storage composed of battery energy storage and pumped hydro storage, and form an intraday optimization scheduling model for composite energy storage. The topology construction in S1, based on the similar characteristics of power supply and consumption, is a complex network of a power distribution system, including the following steps: S11. Obtain the energy supply and consumption similarity characteristics of any two source load nodes, expressed by the formula: （1） （2） （3） （4） （5） （6） in, , , , , , These represent the similarity characteristics of energy supply and consumption, category similarity, spatial proximity, average energy supply and consumption level, power fluctuation, and response time requirement between source load nodes a and b, respectively. , , , These are the coordinates of the source load positions represented by nodes a and b. , This represents the average coordinate of the source load nodes; , These represent the average output or load of source load nodes a and b, respectively. , These represent the fluctuations in output and load at source nodes a and b, respectively. , These represent the response time requirements of source-load nodes a and b for distributed renewable energy consumption and load demand, respectively. S12. Calculate the similarity characteristics of power supply and consumption among all source load nodes, determine the topological connection relationship, and obtain the complex network of the power distribution system based on the similarity characteristics of power supply and consumption. The topology connection is based on the similarity of energy supply and consumption characteristics between source load nodes a and b. If the similarity between the two nodes is greater than a given threshold, an edge connection is established; otherwise, no edge connection is established.

2. The day-ahead-intraday operation optimization method for an energy storage system based on a power supply and demand relationship scenario, as described in claim 1, is characterized in that... Based on the importance of source and load nodes in the complex network of the power distribution system, a node centrality evaluation index is established, expressed by the formula: （7） in, Let be the centrality evaluation index for node a; n is the total number of source load nodes in the network. This is a network adjacency element, representing the connection between nodes a and b. If nodes a and b are connected, then... It is 1 if it is true, otherwise it is 0.

3. The day-ahead-intraday operation optimization method for an energy storage system based on a power supply and demand relationship scenario, as described in claim 1, is characterized in that... The steps in S2, which maximize the modularity of clustering similar source-load nodes in different spatiotemporal scenarios based on the Louvain algorithm to form communities, include: S21. Treat each node as a separate community and calculate the initial modularity value Q of the entire network; S22. For each node, place it in the community of adjacent nodes, calculate the modularity gain after moving it, and move the node to the community that maximizes the modularity; repeat the above process until it is no longer possible to obtain a greater modularity gain by moving nodes. S23. Merge all the communities where the nodes are located to form a new super node; based on the merged super node, recalculate the modularity Q of the entire network; S24. For the newly formed supernode, repeat the iterative optimization process in step 22 to continue searching for the community that maximizes modularity; repeat the iterative process until no greater modularity gain can be found or the preset number of iterations is reached. S25. Output the partitioning results and obtain the optimal community.

4. The day-ahead-intraday operation optimization method for an energy storage system based on a power supply and demand relationship scenario, as described in claim 3, is characterized in that... The formula for calculating the modularity value Q is: （8） （9） Where Q is the modularity and M is the number of edges in the network; , , where represents the degree of node a and node b respectively, and represents the number of edges directly connected to nodes a and b respectively; This indicates whether nodes a and b are in the same community. If they are, the value is 1; otherwise, the value is 0.

5. The day-ahead-intraday operation optimization method for an energy storage system based on a power supply and demand relationship scenario, as described in claim 3, is characterized in that... The day-ahead optimization model in S3 aims to minimize regional energy costs, improve user satisfaction with day-ahead energy supply, and maximize the fulfillment of user energy needs. It employs an upper-level objective function and a lower-level objective function to construct a two-layer day-ahead optimization model, and performs constrained optimization. The steps include: S31. With the goal of minimizing regional energy costs and maximizing user satisfaction with daily energy supply, an upper-level objective function is established, the formula of which is: （10） （11） （12） （13） （14） （15） （16） （17） in, Total energy supply and consumption costs for the region; To assess user satisfaction with current energy supply; Indicates day-ahead energy costs; This represents the penalty cost for regional energy consumption satisfaction. This indicates the cost of purchasing electricity from the grid; This indicates the penalty cost of abandoning renewable energy; The charging, discharging, operation, and maintenance costs of composite energy storage; This indicates the revenue generated from feeding electricity into the power grid; This represents the cost of purchasing electricity from the grid at time t; This represents the day-ahead electricity price at time t. Let t be the amount of photovoltaic power generation discarded. Let t be the amount of wind power generation waste. Penalty cost per unit of renewable energy discard; , The charging / discharging cost and start / stop cost of the composite energy storage at time t are respectively. , These represent the charging and discharging power of battery energy storage and pumped hydro storage at time t, respectively. , These are the unit operating costs for battery energy storage and pumped hydro storage, respectively. , These represent the start-stop switching states of battery energy storage and pumped hydro storage at time t, respectively. That is, 1 is the same as the previous time when the charging and discharging states are the same, and 0 is the different times. , The unit start-up and shutdown cost for battery energy storage and pumped hydro storage; Let be the power supplied at time t; The feed-in electricity price coefficient to the power grid; S32. To maximize the satisfaction of users' daily energy supply and consumption needs, a lower-level objective function is established, with the following formula: （18） （19） （20） （21） （22） in, This refers to the capacity mismatch in composite energy storage. The cost is mismatched with the capacity of the composite energy storage system. This refers to the response time mismatch in composite energy storage. Costs due to mismatched response time in composite energy storage; The difference between the discharge of the composite energy storage at time t and the load or charging of the nth community and the output of DG; For the nth community load or DG output at time t; The total number of communities in the cluster; Let t be the difference between the response time of the composite energy storage and the average demand time of the nodes in the nth community; Let t be the average demand time for the nth community; and These are the response times for battery energy storage and pumped hydro storage, respectively. S33. Apply power balance constraints and composite energy storage constraints to the day-ahead two-layer optimization scheduling model; The power balance constraint is expressed by the following formula: （23） in, , The charging and discharging power of the battery at time t; The charging and discharging efficiency of battery energy storage; , The charging and discharging power of the pumped hydro storage at time t; The charging and discharging efficiency of pumped-storage hydroelectric power. , The output of wind power and photovoltaic power at time t; Let t be the regional load at time t; The formula for the constraint of composite energy storage is: （24） （25） （26） （27） （28） （29） （30） （31） （32） in, , The state of the battery energy storage charging and discharging at time t is a 0-1 variable; , The state of pumped water storage charging and discharging at time t is a 0-1 variable; , , , These are the upper and lower limits of maximum charge and discharge for battery energy storage and pumped hydro storage, respectively. , The state of charge of the battery energy storage and pumped hydro storage at time t; , , , This indicates the upper and lower limits of the state of charge for battery energy storage and pumped hydro storage.

6. The day-ahead-intraday operation optimization method for an energy storage system based on a power supply and demand relationship scenario, as described in claim 1, is characterized in that... In S4, based on the SOC capacity of the composite energy storage and the charge / discharge constraints, the intraday time window is dynamically optimized, as expressed by the formula: （33） （34） in, The time window for rolling optimization of pumped-storage energy storage at time t within the day; The charging and discharging efficiency of battery energy storage; The window coefficient; This refers to the capacity of pumped-storage hydroelectric power.

7. The day-ahead-intraday operation optimization method for an energy storage system based on a power supply and demand relationship scenario, as described in claim 1, is characterized in that... The composite energy storage system in S4, which combines battery energy storage and pumped hydro storage, includes: Based on the day-ahead operation optimization model, with the objective of minimizing the total cost during the intraday time window scheduling period, a composite energy storage system combining battery energy storage and pumped hydro storage is balanced, as shown in the following formula: （35） （36） in, The cost of purchasing electricity from the grid at time T during the day; This represents the penalty cost of forgoing renewable energy at time T; and This indicates the operating and maintenance costs of battery energy storage and pumped hydro storage; This indicates the cost of the power deviation between the current day and the previous day; Deviation error coefficient; and These represent the amount of electricity purchased from the power grid on the current day and at time T before the current day, respectively.

Images