An online dispatching method and system based on residential photovoltaic and energy storage integrated machine clusters

Abstract

This invention discloses an online dispatching method and system based on a residential photovoltaic-storage integrated unit cluster. The method includes: acquiring basic data of a low-voltage distribution network; preprocessing uncertainties; establishing a distribution network model; calculating the dispatchable power or power deficit within nodes and dispatching energy among the photovoltaic-storage integrated units within the nodes; establishing upper and lower layer models for the online dispatching strategy and optimizing the input information of the upper and lower layers of the distribution network model; integrating the optimized input information to obtain the input power of the distribution network reference node, the energy dispatching situation between photovoltaic-storage integrated units, and the reactive power output of the photovoltaic-storage nodes. This invention employs a two-layer optimization model for residential photovoltaic-storage integrated unit clusters, achieving local power quality optimization and internal mutual support. It adapts to the characteristics of distributed photovoltaic resources being numerous and widely distributed, enabling on-site photovoltaic power consumption and helping photovoltaic-storage nodes participate in the operation of low-voltage distribution networks, thus improving the stability of the distribution network.

Description

Technical Field

[0001] This invention belongs to the field of energy consumption in the power industry, specifically relating to an online dispatching method and system based on a residential photovoltaic-storage integrated cluster. Background Technology

[0002] In low-voltage distribution networks in rural towns and villages, user-side photovoltaic (PV) energy storage is a suitable approach. Residential PV-energy storage integrated units are self-built in homes, capable of simultaneously connecting to PV, energy storage, and charging ports. They provide flexible energy storage and respond to grid active and reactive power regulation. Through effective operation models and large-scale integration of PV-energy storage units, market-based benefits can be realized for multiple stakeholders, while grid companies can exercise refined management of large-scale rural distributed PV systems.

[0003] However, residential integrated photovoltaic and energy storage systems still suffer from significant drawbacks, including a lack of mature operation mechanisms, high costs, large size, and low operating efficiency. Large-scale distributed photovoltaic systems are often haphazardly connected to the low-voltage distribution network in the distribution area, resulting in poor power quality, low photovoltaic utilization, and difficulty in forming effective aggregation zones to support upper-level control.

[0004] Therefore, the online flexible allocation technology for residential photovoltaic-storage integrated machine clusters, which takes into account the differences between power generation and energy storage, has high application value. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to provide an online dispatching method and system based on a residential photovoltaic-storage integrated cluster, which can achieve real-time balance between energy storage and controllable capacity, respond to upper-level management and control strategies, make full use of the characteristics of energy storage, participate in grid frequency regulation, peak-valley filling and other services, and provide flexible support for the distribution network system.

[0006] Technical solution: The present invention provides an online allocation method based on a residential photovoltaic-storage integrated machine cluster, comprising:

[0007] Obtain basic data of the low-voltage distribution network;

[0008] Preprocessing is performed on the uncertainties in the basic data regarding the operation of the power distribution network;

[0009] A basic distribution network model is established based on the power flow constraints and safe operation constraints of the distribution network, combined with the preprocessed basic data.

[0010] Based on the calculated dispatchable power or deficit power within the node to which the photovoltaic and energy storage integrated machine is connected, it is determined whether the node to which the photovoltaic and energy storage integrated machine is connected needs to absorb power from the distribution network. Based on the determination result, the energy allocation of the photovoltaic and energy storage integrated machine within the node is carried out, and the dispatchable power constraint of the photovoltaic and energy storage node is added to the basic distribution network model to obtain the distribution network model.

[0011] A higher-level model for the online dispatching strategy is established, and the Benders decomposition algorithm is used to optimize the input information of the upper-level distribution network model.

[0012] A lower-level model for the online dispatching strategy is established, and the particle swarm optimization algorithm is used to optimize the input information of the lower level of the distribution network model.

[0013] By integrating the input information from the upper layer of the optimized distribution network model and the input information from the lower layer of the distribution network model, we can obtain the input power of the distribution network reference node, the energy allocation between the photovoltaic and energy storage integrated units, and the reactive power output of the photovoltaic and energy storage nodes.

[0014] Furthermore, the basic distribution network model includes:

[0015] Power flow constraints in distribution networks:

[0016] The node voltage relationship between nodes i and j can be expressed as:

[0017]

[0018] Among them, V j V represents the actual voltage at node j. i I represents the actual voltage at node i; ij R represents the current in branch ij; ij With X ij These are the impedance and inductive reactance on branch ij, respectively; P ij With Q ij P represents the active and reactive power on branch ij; jl With Q jl P represents the active and reactive power on branch jl; j and Q j Let represent the active and reactive power injected at node j; the constraints are as follows:

[0019]

[0020] π(j) and σ(j) are the sets of parent and child nodes of node j, and the following active and reactive power calculation constraints must be satisfied:

[0021]

[0022] Safety operation constraints of power distribution networks:

[0023] The current in branch ij and the voltage at node j must be within safe operating ranges:

[0024] V min ≤V j ≤V max

[0025]

[0026] Furthermore, based on the calculated dispatchable power or deficit power within the node where the integrated photovoltaic and energy storage unit is connected, it is determined whether the node needs to absorb power from the distribution network. Based on the determination result, energy allocation for the integrated photovoltaic and energy storage unit within the node is performed, including:

[0027] The process iterates through each user. When a user's photovoltaic power generation exceeds their load, the excess power is calculated and stored in the user's energy storage or allocated to other users on the same node. When a user's photovoltaic power generation is less than their load, the power deficit is calculated. If the user's energy storage has sufficient reserves, the power deficit is first compensated from the user's energy storage; otherwise, it is compensated from other users on the same node. This process continues until the power calculation for all photovoltaic and energy storage users in the node is completed, yielding the maximum dispatchable power output from the node or the deficit power that needs to be absorbed from other nodes. This completes the energy allocation of the integrated photovoltaic and energy storage unit within the node.

[0028] Furthermore, the objective function F of the upper-level model of the online allocation strategy u for:

[0029]

[0030] Where β and γ are the weights of the active power input and the difference in user energy storage capacity, respectively; P ref This represents the active power input to node 1; This represents the difference between the energy storage capacity of the u-th user at node j and the average energy storage capacity of all users in the distribution network.

[0031] Furthermore, the upper-level model of the online dispatching strategy utilizes the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model, including:

[0032] The Benders decomposition algorithm is used to optimize the active power of the photovoltaic-storage nodes, including the input power of the reference node and the power input between the photovoltaic and storage nodes. The input power of the reference node and the power input between the photovoltaic and storage nodes are decomposed into a main problem and sub-problems. The input power of the reference node is obtained by solving the main problem, and the power input and output between the photovoltaic and storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users.

[0033] Furthermore, by solving the main problem, the input power of the reference node is obtained, and by solving the sub-problems, the input and output power between the optical and energy storage nodes is obtained, thereby balancing the energy storage capacity among users, including:

[0034] Initialize algorithm parameters, and initialize upper and lower bounds;

[0035] Initialize the main problem: Randomly generate a solution to the main problem within the runtime range, i.e., the total input power F of the reference nodes. m =P ref ;

[0036] Solving the subproblem: Substituting the solution to the main problem into the node information of the distribution network model, and under the constraint conditions, calculating the power flow distribution of the distribution network, thereby obtaining the solution to the subproblem and the corresponding objective value. The solution to the subproblem represents the input and output of user power between the photovoltaic and energy storage nodes, and the corresponding objective function is the energy storage differential. Based on the direction of minimizing the objective value, the input and output of user power between optical storage nodes are optimized through iteration, thereby obtaining the optimal solution of the subproblem and the objective function value minimized thereto;

[0037] Update the main problem by adding the resulting Benders cut to the constraints of the main problem based on the cutting planes fed back from the subproblems, and adjusting the solution of the main problem, including:

[0038] The shadow price of the objective solution is calculated using the Lagrange multiplier method in linear programming. The shadow price is typically represented by the Lagrange multiplier λ corresponding to the constraints at each node. j , representing the sensitivity to the solution of the main problem, that is, the rate at which the objective function of the subproblem changes when the solution of the main problem undergoes a small change. The calculation formula is:

[0039]

[0040] P j Let be the power of node j;

[0041] The Benders cut is calculated using shadow prices to approximate the solution space of the main problem, assuming the solution to the current subproblem is... Then Benders cut is:

[0042]

[0043] This inequality will be added to the main problem to constrain the solution to the main problem, so that future solutions get closer and closer to the global optimum.

[0044] The solution to the new principal problem is obtained by taking values ​​from the newly generated solution space of the principal problem. Among the updated main problems is the new Benders cut;

[0045] Substitute the solution to the new main problem into the subproblem to solve it, and repeat the above steps until the solutions to the main problem and the subproblem converge.

[0046] Furthermore, the objective function F of the lower-level model of the online allocation strategy lfor:

[0047] F l =min∑ j (V j -V j_ref ) 2

[0048] Among them, V j and V j_ref These are the actual voltage and reference voltage at node j, respectively.

[0049] Furthermore, the lower-level model of the online dispatching strategy utilizes particle swarm optimization to optimize the input information of the lower-level distribution network model, including:

[0050] Initialize the number of particles N in the particle swarm; each particle represents a possible solution; the dimension of the particle is equal to the number of variables to be optimized, i.e., the number of nodes in the integrated photovoltaic and energy storage system; the position of each particle represents its output reactive power Q. j Initialize the particle positions within the feasible range of reactive power; assign an initial velocity v to each particle. j The initial position of the particle is taken as its initial individual optimal solution p. j In the initial particle swarm, the particle with the smallest objective function value is selected as the global optimal solution g. j In the t-th iteration, the particle velocity is updated according to the formula:

[0051]

[0052] Where ω is the inertia weight; c1 and c2 are acceleration constants, chosen in the range of 0.1 to 0.3; r1 and r2 are random numbers; the particle position is adjusted using the updated velocity. If the location exceeds the boundary, it will be restricted to the boundary.

[0053] In the t-th iteration, the objective function value (i.e., the impact of reactive power distribution on system voltage) is calculated for the new position of each particle. The objective function value corresponding to the particle's new position is compared with that of its historical best position. Based on the objective function value of the particle's new position and the objective function value of its historical best position, the particle's new position and the better position among the historical best positions are determined. If the particle's new position is better, the particle's individual optimal solution is updated using the new position.

[0054] Compare the individual optimal solutions of all particles, determine the global optimal solution for the current iteration from the individual optimal solutions of all particles, and update the global optimal solution based on the global optimal solution of the current iteration and the global optimal solution of the previous iteration. Check convergence conditions: If the number of iterations reaches the maximum value, or the global optimal solution no longer changes in multiple iterations, the algorithm is considered to have converged; if it has not converged, return to continue iterating; after the algorithm converges, it outputs the global optimal solution, which is the optimal distribution of reactive power.

[0055] Based on the same inventive concept, the present invention provides an online dispatching system for a residential photovoltaic and energy storage integrated cluster, comprising:

[0056] The data acquisition module is used to acquire basic data of the low-voltage distribution network;

[0057] The data preprocessing module is used to preprocess the uncertainties in the basic data during the operation of the power distribution network.

[0058] The distribution network model building module is used to build a basic distribution network model based on the power flow constraints and safe operation constraints of the distribution network, combined with the preprocessed basic data.

[0059] The energy allocation module is used to determine whether the nodes connected to the photovoltaic-storage integrated machine need to absorb power from the distribution network based on the calculated dispatchable power or deficit power within the nodes. Based on the determination result, the module allocates energy for the photovoltaic-storage integrated machine within the node and forms the dispatchable power constraint of the photovoltaic-storage node, which is then added to the basic distribution network model to obtain the distribution network model.

[0060] The model optimization module is used to build the upper-level model of the online dispatching strategy. The upper-level model of the online dispatching strategy uses the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model.

[0061] The model optimization module is used to establish a lower-level model of the online dispatching strategy. The lower-level model of the online dispatching strategy uses the particle swarm optimization algorithm to optimize the input information of the lower level of the distribution network model.

[0062] The information integration module is used to integrate the input information of the upper layer of the optimized distribution network model and the input information of the lower layer of the distribution network model to obtain the input power of the distribution network reference node, the energy allocation between the photovoltaic and energy storage integrated unit, and the reactive power output of the photovoltaic and energy storage node.

[0063] Furthermore, the basic distribution network model includes:

[0064] Power flow constraints in distribution networks:

[0065] The node voltage relationship between nodes i and j can be expressed as:

[0066]

[0067] Among them, V j V represents the actual voltage at node j. iI represents the actual voltage at node i; ij R represents the current in branch ij; ij With X ij These are the impedance and inductive reactance on branch ij, respectively; P ij With Q ij P represents the active and reactive power on branch ij; jl With Q jl P represents the active and reactive power on branch jl; j and Q j Let represent the active and reactive power injected at node j; the constraints are as follows:

[0068]

[0069] π(j) and σ(j) are the sets of parent and child nodes of node j, and the following active and reactive power calculation constraints must be satisfied:

[0070]

[0071] Safety operation constraints of power distribution networks:

[0072] The current in branch ij and the voltage at node j must be within safe operating ranges:

[0073] V min ≤V j ≤V max

[0074]

[0075] Furthermore, based on the calculated dispatchable power or deficit power within the node where the integrated photovoltaic and energy storage unit is connected, it is determined whether the node needs to absorb power from the distribution network. Based on the determination result, energy allocation for the integrated photovoltaic and energy storage unit within the node is performed, including:

[0076] The process iterates through each user. When a user's photovoltaic power generation exceeds their load, the excess power is calculated and stored in the user's energy storage or distributed to other users on the same node. When a user's photovoltaic power generation is less than their load, the power deficit is calculated. If the user's energy storage has sufficient reserves, the power deficit is compensated from the user's energy storage first; otherwise, it is compensated from other users on the same node. This process continues until the power calculation for all photovoltaic and energy storage users in the node is completed, yielding the maximum dispatchable power output of the node or the deficit power that needs to be absorbed from other nodes. This completes the energy allocation of the integrated photovoltaic and energy storage unit within the node.

[0077] Furthermore, the objective function F of the upper-level model of the online allocation strategy u for:

[0078]

[0079] Where β and γ are the weights of the active power input and the difference in user energy storage capacity, respectively; P ref This represents the active power input to node 1; This represents the difference between the energy storage capacity of the u-th user at node j and the average energy storage capacity of all users in the distribution network.

[0080] Furthermore, the upper-level model of the online dispatching strategy utilizes the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model, including:

[0081] The Benders decomposition algorithm is used to optimize the active power of the photovoltaic-storage nodes, including the input power of the reference node and the power input between the photovoltaic and storage nodes. The input power of the reference node and the power input between the photovoltaic and storage nodes are decomposed into a main problem and sub-problems. The input power of the reference node is obtained by solving the main problem, and the power input and output between the photovoltaic and storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users.

[0082] Furthermore, by solving the main problem, the input power of the reference node is obtained, and by solving the sub-problems, the input and output power between the optical and energy storage nodes is obtained, thereby balancing the energy storage capacity among users, including:

[0083] Initialize algorithm parameters, and initialize upper and lower bounds;

[0084] Initialize the main problem: Randomly generate a solution to the main problem within the runtime range, i.e., the total input power F of the reference nodes. m =P ref ;

[0085] Solving the subproblem: Substituting the solution to the main problem into the node information of the distribution network model, and under the constraint conditions, calculating the power flow distribution of the distribution network, thereby obtaining the solution to the subproblem and the corresponding objective value. The solution to the subproblem represents the input and output of user power between the photovoltaic and energy storage nodes, and the corresponding objective function is the energy storage differential. Based on the direction of minimizing the objective value, the input and output of user power between optical storage nodes are optimized through iteration, thereby obtaining the optimal solution of the subproblem and the objective function value minimized thereto;

[0086] Update the main problem by adding the resulting Benders cut to the constraints of the main problem based on the cutting planes fed back from the subproblems, and adjusting the solution of the main problem, including:

[0087] The shadow price of the objective solution is calculated using the Lagrange multiplier method in linear programming. The shadow price is typically represented by the Lagrange multiplier λ corresponding to the constraints at each node.j , representing the sensitivity to the solution of the main problem, that is, the rate at which the objective function of the subproblem changes when the solution of the main problem undergoes a small change. The calculation formula is:

[0088]

[0089] P j Let be the power of node j;

[0090] The Benders cut is calculated using shadow prices to approximate the solution space of the main problem, assuming the solution to the current subproblem is... Then Benders cut is:

[0091]

[0092] This inequality will be added to the main problem to constrain the solution to the main problem, so that future solutions get closer and closer to the global optimum.

[0093] The solution to the new principal problem is obtained by taking values ​​from the newly generated solution space of the principal problem. Among the updated main problems is the new Benders cut;

[0094] Substitute the solution to the new main problem into the subproblem to solve it, and repeat the above steps until the solutions to the main problem and the subproblem converge.

[0095] Furthermore, the objective function F of the lower-level model of the online allocation strategy l for:

[0096] F l =min∑ j (V j -V j_ref ) 2

[0097] Among them, V j and V j_ref These are the actual voltage and reference voltage at node j, respectively.

[0098] Furthermore, the lower-level model of the online dispatching strategy utilizes particle swarm optimization to optimize the input information of the lower-level distribution network model, including:

[0099] Initialize the number of particles N in the particle swarm; each particle represents a possible solution; the dimension of the particle is equal to the number of variables to be optimized, i.e., the number of nodes in the integrated photovoltaic and energy storage system; the position of each particle represents its output reactive power Q. j Initialize the particle positions within the feasible range of reactive power; assign an initial velocity v to each particle. j The initial position of the particle is taken as its initial individual optimal solution p. jIn the initial particle swarm, the particle with the smallest objective function value is selected as the global optimal solution g. j In the t-th iteration, the particle velocity is updated according to the formula:

[0100]

[0101] Where ω is the inertia weight; c1 and c2 are acceleration constants; the range is 0.1 to 0.3; r1 and r2 are random numbers; the particle position is adjusted using the updated velocity. If the location exceeds the boundary, it will be restricted to the boundary.

[0102] In the t-th iteration, the objective function value (i.e., the impact of reactive power distribution on system voltage) is calculated for the new position of each particle. The objective function value corresponding to the particle's new position is compared with that of the individual best position in history. Based on the objective function value of the particle's new position and the objective function value of the particle's historical best position, the new position and the best position among the historical best positions are determined. If the particle's new position is optimal, the individual optimal solution of the particle is updated using the new position.

[0103] Compare the individual optimal solutions of all particles, determine the global optimal solution for the current iteration from the individual optimal solutions of all particles, and update the global optimal solution based on the global optimal solution of the current iteration and the global optimal solution of the previous iteration. Check convergence conditions: If the number of iterations reaches the maximum value, or the global optimal solution no longer changes in multiple iterations, the algorithm is considered to have converged; if it has not converged, return to continue iterating; after the algorithm converges, it outputs the global optimal solution, which is the optimal distribution of reactive power.

[0104] Based on the same inventive concept, the present invention provides an online dispatching device based on a residential photovoltaic and energy storage integrated cluster, comprising a processor and a memory. The memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the above-described online dispatching method based on a residential photovoltaic and energy storage integrated cluster.

[0105] Based on the same inventive concept, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-described online allocation method based on a residential optical storage integrated machine cluster.

[0106] Beneficial effects: Compared with the prior art, the significant technical effects of the present invention are as follows:

[0107] This invention enables the formation of a cluster of photovoltaic and energy storage integrated units for residential use in a distribution area. Initially, power distribution is achieved within the cluster. When the photovoltaic and energy storage capacity of some users is insufficient to support their load, other photovoltaic and energy storage integrated units are prioritized to meet the load demand. This not only improves the utilization rate of photovoltaic power in rural areas but also achieves reactive power and harmonic management of rural power distribution networks, improves the power quality of rural areas, and enhances user satisfaction.

[0108] This invention enables an optimized configuration model for photovoltaic and energy storage resource scheduling with the goal of economic efficiency. It employs a two-layer model based on algorithmic characteristics. In the upper-layer model, because two objectives need to be optimized simultaneously, the Benders algorithm, which decomposes the primary and secondary problems, is selected to improve the optimization efficiency. The invention also incorporates the energy allocation process within the integrated photovoltaic and energy storage units in residential applications, thereby improving the absorption efficiency of distributed photovoltaic energy, balancing the differences in energy storage capacity among users, enhancing the stability of the weak power grid in the face of extreme weather, and suppressing voltage fluctuations in low-voltage distribution networks.

[0109] This invention balances the differences in energy storage capacity among users at each node, thereby reducing the waste of photovoltaic power generation energy and increasing the overall energy storage capacity in the distribution network, thus improving the stability of the low-voltage distribution network.

[0110] This invention integrates photovoltaic and energy storage clusters in low-voltage power distribution networks into an effective whole, facilitating upward response to dispatch commands from higher levels. Attached Figure Description

[0111] Figure 1 This is a flowchart illustrating an online allocation method based on a residential photovoltaic-storage integrated machine cluster disclosed in an embodiment of the present invention;

[0112] Figure 2 This is a flowchart illustrating a Benders algorithm disclosed in an embodiment of the present invention;

[0113] Figure 3 This is a flowchart illustrating a PSO algorithm disclosed in an embodiment of the present invention;

[0114] Figure 4 This is a schematic diagram of the structure of an online dispatching system based on a residential photovoltaic and energy storage integrated cluster disclosed in an embodiment of the present invention;

[0115] Figure 5 This is a schematic diagram of the structure of an online dispatching device based on a residential photovoltaic and energy storage integrated cluster disclosed in an embodiment of the present invention. Detailed Implementation

[0116] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. Those skilled in the art will understand that the objectives and advantages achievable with the present invention are not limited to those specifically described above, and that the above and other objectives achievable by the present invention will become clearer from the following detailed description.

[0117] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application, design, and conditions of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0118] In this invention, the term "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0119] Example 1

[0120] Please see Figure 1 , Figure 1 This is a flowchart illustrating an online allocation method based on a residential photovoltaic-storage integrated machine cluster, as disclosed in an embodiment of the present invention. Wherein, Figure 1 The described online allocation method is applied to power systems, such as for the flexible online allocation of residential photovoltaic-storage integrated machine cluster resources. This invention is not limited to specific applications. Figure 1 As shown, the online deployment method based on residential photovoltaic-storage integrated machine clusters may include the following operations:

[0121] S1. Obtain basic data of the low-voltage distribution network.

[0122] Input the basic data of the power distribution network, such as network structure, branch impedance, and line capacity.

[0123] S2. Preprocess the uncertainties in the basic data of the distribution network operation process, eliminate outliers in the distribution network data, and fill in missing values;

[0124] S3. Based on the power flow constraints and safe operation constraints of the distribution network, and combined with the preprocessed basic data, establish a basic distribution network model.

[0125] The basic distribution network model includes:

[0126] Power flow constraints in distribution networks:

[0127] The node voltage relationship between nodes i and j can be expressed as:

[0128]

[0129] Among them, V j V represents the actual voltage at node j. i I represents the actual voltage at node i; ij R represents the current in branch ij; ij With X ij These are the impedance and inductive reactance on branch ij, respectively; P ij With Q ij P represents the active and reactive power on branch ij; jl With Q jl P represents the active and reactive power on branch jl; j and Q j Let represent the active and reactive power injected at node j; the constraints are as follows:

[0130]

[0131] π(j) and σ(j) are the sets of parent and child nodes of node j, and the following active and reactive power calculation constraints must be satisfied:

[0132]

[0133] The safety operation constraints of the power distribution network are:

[0134] The current in branch ij and the voltage at node j must be within safe operating ranges:

[0135] V min ≤V j ≤V max

[0136]

[0137] S4. Based on the calculated dispatchable power or deficit power within the node to which the photovoltaic-storage integrated unit is connected, determine whether the node to which the photovoltaic-storage integrated unit is connected needs to absorb power from the distribution network. Based on the determination result, perform energy allocation for the photovoltaic-storage integrated unit within the node and form the dispatchable power constraint of the photovoltaic-storage node, which is then added to the basic distribution network model to obtain the distribution network model.

[0138] The process iterates through each user. When a user's photovoltaic power generation exceeds their load, the excess power is calculated and stored in the user's energy storage or allocated to other users on the same node. When a user's photovoltaic power generation is less than their load, the power deficit is calculated. If the user's energy storage has sufficient reserves, the power deficit is compensated from the user's energy storage first; otherwise, it is compensated from other users on the same node. This process continues until the power calculation for all photovoltaic and energy storage users in the node is completed, yielding the maximum dispatchable power that the node can output or the deficit power that needs to be absorbed from other nodes. This completes the energy allocation of the integrated photovoltaic and energy storage unit within the node.

[0139] S5, such as Figure 2 As shown, an upper-level model for the online dispatching strategy is established. The upper-level model of the online dispatching strategy uses the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model.

[0140] In this embodiment, the objective function F of the upper-level model of the online allocation strategy is... u for:

[0141]

[0142] Where β and γ are the weights of the active power input and the difference in user energy storage capacity, respectively; P ref This represents the active power input to node 1; This represents the difference between the energy storage capacity of the u-th user at node j and the average energy storage capacity of all users in the distribution network.

[0143] It needs to be explained that when the algorithm optimizes the problem, it will cause the objective function F of the upper-level model of the online allocation strategy to change. u The calculated result gradually decreases, and this value is minimized. This value is a weighted sum of the difference between a reference power and energy storage capacity. Node 1 represents the first node of the distribution network, and is generally called the reference node.

[0144] Among them, the upper-level model of the online dispatching strategy uses the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model, as follows:

[0145] The Benders decomposition algorithm is used to optimize the active power of the photovoltaic-storage nodes, including the input power of the reference node and the power input between the photovoltaic and storage nodes. The input power of the reference node and the power input between the photovoltaic and storage nodes are decomposed into a main problem and sub-problems. The input power of the reference node is obtained by solving the main problem, and the power input and output between the photovoltaic and storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users.

[0146] In this embodiment, the input power of the reference node is obtained by solving the main problem, and the input and output power between the optical and energy storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users. This includes the following:

[0147] Initialize algorithm parameters, and initialize upper and lower bounds;

[0148] Initialize the main problem: Randomly generate a solution to the main problem (i.e., the total input power F of the reference node) within the operating range. m =P ref );

[0149] Solving the subproblem: Substituting the solution to the main problem into the node information of the distribution network model, and under the constraint conditions, the power flow distribution of the distribution network is calculated, thereby obtaining the solution to the subproblem and the corresponding objective value. The solution to the subproblem represents the input and output of user power between the photovoltaic and energy storage nodes, and the corresponding objective function is the energy storage differential. Based on the direction of minimizing the objective value, the input and output of user power between optical storage nodes are optimized through iteration, thereby obtaining the optimal solution of the subproblem and the objective function value minimized thereto;

[0150] Update the main problem by adding the resulting Benders cut to the constraints of the main problem based on the cutting planes fed back from the subproblems, and adjusting the solution of the main problem, including:

[0151] The shadow price of the objective solution is calculated using the Lagrange multiplier method in linear programming. The shadow price is typically represented by the Lagrange multiplier λ corresponding to the constraints at each node. j , representing the sensitivity to the solution of the main problem, that is, the rate at which the objective function of the subproblem changes when the solution of the main problem undergoes a small change. The calculation formula is:

[0152]

[0153] P j Let be the power of node j;

[0154] The Benders cut is calculated using shadow prices to approximate the solution space of the main problem, assuming the solution to the current subproblem is... Then Benders cut is:

[0155]

[0156] This inequality will be added to the main problem to constrain the solution to the main problem, so that future solutions get closer and closer to the global optimum.

[0157] The solution to the new principal problem is obtained by taking values ​​from the newly generated solution space of the principal problem. Among the updated main problems is the new Benders cut;

[0158] Substitute the solution to the new main problem into the subproblem to solve it, and repeat the above steps until the solutions to the main problem and the subproblem converge.

[0159] In this scheme, the objective function of the Benders algorithm consists of two parts: the former is defined as the main problem, and the latter as a subproblem. Under other operating conditions of the distribution network, the power of the reference node of the distribution network is obtained by solving the main problem. Then, based on the solution of the main problem, the subproblems are solved to obtain the active power output of the photovoltaic-storage integrated machine and the value of the objective function. It is then determined whether the solution of the subproblem satisfies the constraints. If not, a cut function is generated and added to the main problem. The cut function iteratively approximates the feasible region of the main problem variables. If the constraints are satisfied, the upper and lower bounds are updated. The expressions for the upper and lower bounds are:

[0160] U B =min(U B ,F u )

[0161] L B =max(L B ,F u )

[0162] S6, such as Figure 3 As shown, a lower-level model of the online dispatching strategy is established. The lower-level model of the online dispatching strategy uses the particle swarm optimization algorithm to optimize the input information of the lower level of the distribution network model.

[0163] In this embodiment, the objective function F of the lower-level model of the online allocation strategy is... l for:

[0164] F l =minΣ j (V j -V j_ref ) 2

[0165] Among them, V j and V j_ref These are the actual voltage and reference voltage at node j, respectively.

[0166] It needs to be explained that the objective function F l This means minimizing the node voltage deviation.

[0167] Among them, the lower-level model of the online dispatching strategy uses the particle swarm optimization algorithm to optimize the input information of the lower-level distribution network model, as follows:

[0168] The reactive power of the photovoltaic-storage integrated machine node is optimized using the particle swarm optimization algorithm to reduce voltage fluctuations at each node in the distribution network.

[0169] Initialize the number of particles N in the particle swarm; each particle represents a possible solution; the dimension of the particle is equal to the number of variables to be optimized, i.e., the number of nodes in the integrated photovoltaic and energy storage system; the position of each particle represents its output reactive power Q. j Initialize the particle positions within the feasible range of reactive power; assign an initial velocity v to each particle. j The initial position of the particle is taken as its initial individual optimal solution p. j In the initial particle swarm, the particle with the smallest objective function value is selected as the global optimal solution g. j In the t-th iteration, the particle velocity is updated according to the formula:

[0170]

[0171] Where ω is the inertial weight; c1 and c2 are acceleration constants, selected from 0.1 to 0.3; and r1 and r2 are random numbers (between 0 and 1). The particle position is adjusted using the updated velocity. If the location exceeds the boundary, it will be restricted to the boundary.

[0172] In the t-th iteration, the objective function value, i.e., the impact of reactive power distribution on system voltage (the sum of squares of voltage fluctuations), is calculated for the new position of each particle. The objective function value corresponding to the particle's new position is compared with the objective function value corresponding to the individual optimal position in history (individual optimal). Based on the objective function value of the particle's new position and the objective function value of the particle's historical optimal position, the new position of the particle and the optimal position among the historical optimal positions are determined. If the particle's new position is optimal, the individual optimal solution of the particle is updated using the particle's new position.

[0173] Compare the individual optimal solutions of all particles, determine the global optimal solution for the current iteration from the individual optimal solutions of all particles, and update the global optimal solution based on the global optimal solution of the current iteration and the global optimal solution of the previous iteration. Convergence check: The algorithm is considered converged if the number of iterations reaches its maximum value, or if the global optimal solution no longer changes in multiple iterations (the change is less than a set threshold). If it does not converge, it returns to continue iterating. After convergence, the algorithm outputs the global optimal solution, which is the optimal distribution of reactive power.

[0174] This invention employs a two-layer optimization model. The upper layer uses the Benders algorithm to optimize the balance between the active power input of the distribution network and the user's energy storage capacity, thereby obtaining the output active power of the residential photovoltaic-storage integrated cluster. The lower layer uses the PSO algorithm to optimize the voltage deviation of the distribution network nodes, thereby obtaining the output reactive power of the residential photovoltaic-storage integrated cluster.

[0175] In this scheme, the initial position of the particle swarm is the reactive power output of the PESIE node, and the initial velocity of the particle swarm is set. For each particle, its fitness, i.e., the voltage deviation of the distribution network, is calculated, thereby updating the particle's velocity and position.

[0176] S7. Integrate the input information of the upper layer of the optimized distribution network model and the input information of the lower layer of the distribution network model to obtain the input power of the distribution network reference node, the energy allocation between the photovoltaic and energy storage integrated unit, and the reactive power output of the photovoltaic and energy storage node.

[0177] It should be noted that a residential PV-storage integrated system (PV-SHSI) allows users within a node to simultaneously possess the characteristics of PV power generation, energy storage, and user load. The difference between these three aspects determines the user's external performance. Residential PV-SHSI clusters must meet the following constraints:

[0178]

[0179] in and These are the minimum and maximum allowable energy storage capacity for the u-th user at node j during operation, respectively. P is the power that user u is allowed to output. j,u,PV and P j,u,load This represents the user's photovoltaic power generation and load. s is a state function; its value is 1 when the user's photovoltaic power generation exceeds its required load, and 0 otherwise. The total output power of node j is... Depend on And the shortfall in user solar power capacity compared to their own load. Decide.

[0180] Photovoltaic-storage integrated power storage clusters can also achieve reactive power output. Constrained by active power output and inverter capacity:

[0181]

[0182] Where α is the maximum power factor angle output by the integrated photovoltaic and energy storage unit.

[0183] The integrated photovoltaic and energy storage system manifests itself in the following ways: when a user's photovoltaic power generation exceeds the user's load, the user will not absorb power from the distribution network; the excess power will be absorbed by the energy storage or transmitted to other integrated photovoltaic and energy storage systems. When a user's photovoltaic power generation is less than the user's load, the shortfall will be provided by other integrated photovoltaic and energy storage systems or energy storage.

[0184] Within the same node, photovoltaic and energy storage units prioritize mutual support. When a user's photovoltaic power generation is lower than their electricity consumption, the excess power is preferentially absorbed from users within the same node whose photovoltaic power generation exceeds their electricity consumption and whose energy storage capacity is high. After calculating the power situation within the same node, the allocation between different nodes is carried out, prioritizing the supply of power to insufficient nodes from nodes with high energy storage capacity and whose overall photovoltaic power generation exceeds their electricity consumption.

[0185] This invention adopts a two-layer residential photovoltaic-storage integrated cluster optimization model, which can not only achieve local power quality optimization and internal mutual support, adapting to the characteristics of distributed photovoltaic resources being numerous and widespread, and realizing on-site photovoltaic consumption, but also help photovoltaic and storage nodes participate in the operation of low-voltage distribution networks, improving the stability of distribution networks.

[0186] Example 2

[0187] Please see Figure 4 , Figure 4 This is a schematic diagram of an online allocation system based on a residential photovoltaic-storage integrated machine cluster, as disclosed in an embodiment of the present invention. This system enables flexible online allocation of residential photovoltaic-storage integrated machine cluster resources used by residents, specifically including:

[0188] The data acquisition module is used to acquire basic data of the low-voltage distribution network;

[0189] Data preprocessing is used to preprocess uncertainties in the basic data of the distribution network operation process, eliminate outliers in the distribution network data, and fill in missing values;

[0190] The distribution network model building module builds a distribution network model based on the power flow constraints and safe operation constraints of the distribution network, combined with the preprocessed data.

[0191] The energy allocation module is used to calculate the dispatchable power or deficit power within the node, allocate the energy of the photovoltaic-storage integrated unit within the node, and form the dispatchable power constraint of the photovoltaic-storage node, which is then added to the distribution network model.

[0192] The model optimization module is used to build the upper-level model of the online dispatching strategy. The upper-level model of the online dispatching strategy uses the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model.

[0193] The model optimization module is used to establish a lower-level model of the online dispatching strategy. The lower-level model of the online dispatching strategy uses the particle swarm optimization algorithm to optimize the input information of the lower level of the distribution network model.

[0194] The information integration module is used to integrate the upper and lower layer input information of the optimized distribution network model to obtain the input power of the distribution network reference node, the energy allocation between the photovoltaic and energy storage integrated unit, and the reactive power output of the photovoltaic and energy storage node.

[0195] In one optional implementation, the online dispatching method based on a residential photovoltaic-storage integrated unit cluster includes: a) acquiring basic data of the low-voltage distribution network; b) preprocessing uncertainties, eliminating outliers in the distribution network data, and filling in missing values; c) establishing a distribution network model based on the power flow constraints and safe operation constraints of the distribution network, combined with the preprocessed data; d) calculating the dispatchable power or deficit power within the nodes and performing energy dispatching for the photovoltaic-storage integrated units within the nodes; e) establishing an upper-level model of the online dispatching strategy, which uses the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model; f) establishing a lower-level model of the online dispatching strategy, which uses the particle swarm optimization algorithm to optimize the input information of the lower-level distribution network model; g) integrating the optimized upper-level and lower-level input information of the distribution network model to obtain the input power of the distribution network reference node, the energy dispatching status between the photovoltaic-storage integrated units, and the reactive power output of the photovoltaic-storage node.

[0196] In one embodiment, the basic distribution network model includes:

[0197] Power flow constraints in distribution networks:

[0198] The node voltage relationship between nodes i and j can be expressed as:

[0199]

[0200] Among them, V j V represents the actual voltage at node j. i I represents the actual voltage at node i; ij R represents the current in branch ij; ij With X ij These are the impedance and inductive reactance on branch ij, respectively; P ij With Q ij P represents the active and reactive power on branch ij; jl With Q jl P represents the active and reactive power on branch jl; j and Q j Let represent the active and reactive power injected at node j; the constraints are as follows:

[0201]

[0202] π(j) and σ(j) are the sets of parent and child nodes of node j, and the following active and reactive power calculation constraints must be satisfied:

[0203]

[0204] Safety operation constraints of power distribution networks:

[0205] The current in branch ij and the voltage at node j must be within safe operating ranges:

[0206] V min ≤V j ≤V max

[0207]

[0208] In one embodiment, based on the calculated dispatchable power or deficit power within the node to which the integrated photovoltaic and energy storage unit is connected, it is determined whether the node needs to absorb power from the distribution network, and energy allocation for the integrated photovoltaic and energy storage unit within the node is performed according to the determination result, including:

[0209] The process iterates through each user. When a user's photovoltaic power generation exceeds their load, the excess power is calculated and stored in the user's energy storage or distributed to other users on the same node. When a user's photovoltaic power generation is less than their load, the power deficit is calculated. If the user's energy storage has sufficient reserves, the power deficit is compensated from the user's energy storage first; otherwise, it is compensated from other users on the same node. This process continues until the power calculation for all photovoltaic and energy storage users in the node is completed, yielding the maximum dispatchable power output of the node or the deficit power that needs to be absorbed from other nodes. This completes the energy allocation of the integrated photovoltaic and energy storage unit within the node.

[0210] In one embodiment, the objective function F of the upper-level model of the online allocation strategy is... u for:

[0211]

[0212] Where β and γ are the weights of the active power input and the difference in user energy storage capacity, respectively; P ref This represents the active power input to node 1; This represents the difference between the energy storage capacity of the u-th user at node j and the average energy storage capacity of all users in the distribution network.

[0213] In one embodiment, the upper-level model of the online dispatching strategy utilizes the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model, including:

[0214] The Benders decomposition algorithm is used to optimize the active power of the photovoltaic-storage nodes, including the input power of the reference node and the power input between the photovoltaic and storage nodes. The input power of the reference node and the power input between the photovoltaic and storage nodes are decomposed into a main problem and sub-problems. The input power of the reference node is obtained by solving the main problem, and the power input and output between the photovoltaic and storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users.

[0215] In one embodiment, the input power of the reference node is obtained by solving the main problem, and the input and output power between the optical and energy storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users, including:

[0216] Initialize algorithm parameters, and set initial upper and lower bounds;

[0217] Initialize the main problem: Randomly generate a solution to the main problem within the runtime, i.e., the total input power F of the reference nodes. m =P ref ;

[0218] Solving the subproblem: Substituting the solution to the main problem into the node information of the distribution network model, and under the constraint conditions, calculating the power flow distribution of the distribution network, thereby obtaining the solution to the subproblem and the corresponding objective value. The solution to the subproblem represents the input and output of user power between the photovoltaic and energy storage nodes, and the corresponding objective function is the energy storage differential. Based on the direction of minimizing the objective value, the input and output of user power between optical storage nodes are optimized through iteration, thereby obtaining the optimal solution of the subproblem and the objective function value minimized thereto;

[0219] Update the main problem by adding the resulting Benders cut to the constraints of the main problem based on the cutting planes fed back from the subproblems, and adjusting the solution of the main problem, including:

[0220] The shadow price of the objective solution is calculated using the Lagrange multiplier method in linear programming. The shadow price is typically represented by the Lagrange multiplier λ corresponding to the constraints at each node. j , representing the sensitivity to the solution of the main problem, that is, the rate at which the objective function of the subproblem changes when the solution of the main problem undergoes a small change. The calculation formula is:

[0221]

[0222] P j Let be the power of node j;

[0223] The Benders cut is calculated using shadow prices to approximate the solution space of the main problem, assuming the solution to the current subproblem is... Then Benders cut is:

[0224]

[0225] This inequality will be added to the main problem to constrain the solution to the main problem, so that future solutions get closer and closer to the global optimum.

[0226] The solution to the new principal problem is obtained by taking values ​​from the newly generated solution space of the principal problem. Among the updated main problems is the new Benders cut;

[0227] Substitute the solution to the new main problem into the subproblem to solve it, and repeat the above steps until the solutions to the main problem and the subproblem converge.

[0228] In one embodiment, the objective function F of the lower-level model of the online allocation strategy is... l for:

[0229] F l =min∑ j (V j -V j_ref ) 2

[0230] Among them, V j and V j_ref These are the actual voltage and reference voltage at node j, respectively.

[0231] In one embodiment, the lower-level model of the online dispatching strategy utilizes a particle swarm optimization algorithm to optimize the input information of the lower-level distribution network model, including:

[0232] Initialize the number of particles N in the particle swarm; each particle represents a possible solution; the dimension of the particle is equal to the number of variables to be optimized, i.e., the number of nodes in the integrated photovoltaic and energy storage system; the position of each particle represents its output reactive power Q. j Initialize the particle positions within the feasible range of reactive power; assign an initial velocity v to each particle. j The initial position of the particle is taken as its initial individual optimal solution p. j In the initial particle swarm, the particle with the smallest objective function value is selected as the global optimal solution g. j In the t-th iteration, the particle velocity is updated according to the formula:

[0233]

[0234] Where ω is the inertia weight; c1 and c2 are acceleration constants; the range is 0.1 to 0.3; r1 and r2 are random numbers; the particle position is adjusted using the updated velocity. If the location exceeds the boundary, it will be restricted to the boundary.

[0235] In the t-th iteration, the objective function value (i.e., the impact of reactive power distribution on system voltage) is calculated for the new position of each particle. The objective function value corresponding to the particle's new position is compared with that of the individual best position in history. Based on the objective function value of the particle's new position and the objective function value of the particle's historical best position, the new position and the best position among the historical best positions are determined. If the particle's new position is optimal, the individual optimal solution of the particle is updated using the new position.

[0236] Compare the individual optimal solutions of all particles, determine the global optimal solution for the current iteration from the individual optimal solutions of all particles, and update the global optimal solution based on the global optimal solution of the current iteration and the global optimal solution of the previous iteration. Check convergence conditions: If the number of iterations reaches the maximum value, or the global optimal solution no longer changes in multiple iterations, the algorithm is considered to have converged; if it has not converged, return to continue iterating; after the algorithm converges, it outputs the global optimal solution, which is the optimal distribution of reactive power.

[0237] Example 3

[0238] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of an online dispatching device based on a residential photovoltaic-storage integrated machine cluster, as disclosed in an embodiment of the present invention. Figure 5 The described equipment can be applied to power systems, such as for online flexible allocation of residential photovoltaic-storage integrated machine cluster resources, etc., and the embodiments of the present invention are not limited thereto.

[0239] like Figure 5 As shown, the device may include a processor and a memory, the memory storing computer instructions, and the processor executing the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the method described in the above embodiments and achieves the same technical effect as the above method.

[0240] The memory may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the memory may be used to read and write non-removable, non-volatile magnetic media (commonly referred to as a "hard disk drive"). A program / utility having a set (at least one) of program modules may be stored in, for example, memory. Such program modules include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. The program modules typically perform the functions and / or methods described in the embodiments of the present invention.

[0241] The processor executes various functional applications and data processing by running programs stored in memory, such as the method provided in Embodiment 1 of the present invention.

[0242] Example 4

[0243] Embodiment 4 of the present invention also provides a computer-readable storage medium storing a computer program thereon. When the program is executed by a processor, it implements the steps of the method described in the above embodiments and achieves the same technical effect as the above method.

[0244] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0245] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0246] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0247] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0248] Of course, the computer-executable instructions provided in the embodiments of the present invention are not limited to the above-described method operations, but can also perform related operations in the methods provided in any embodiment of the present invention.

[0249] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An online dispatching method based on a residential photovoltaic-storage integrated machine cluster, characterized in that, include: Obtain basic data of the low-voltage distribution network; Preprocessing is performed on the uncertainties in the basic data regarding the operation of the power distribution network; Based on the power flow constraints and safe operation constraints of the distribution network, a basic distribution network model is established using preprocessed basic data. Based on the calculated dispatchable power or deficit power within the node to which the photovoltaic and energy storage integrated machine is connected, it is determined whether the node to which the photovoltaic and energy storage integrated machine is connected needs to absorb power from the distribution network. Based on the determination result, the energy allocation of the photovoltaic and energy storage integrated machine within the node is carried out, and the dispatchable power constraint of the photovoltaic and energy storage node is added to the basic distribution network model to obtain the distribution network model. A higher-level model for the online dispatching strategy is established. This higher-level model utilizes the Benders decomposition algorithm to optimize the input information of the distribution network model. The objective function of the higher-level model for the online dispatching strategy is... for: ； in, and These are the weights for the difference between active power input and user energy storage capacity, respectively. This represents the active power input to node 1; Indicates at node The The difference between the energy storage capacity of an individual user and the average energy storage capacity of all users in the distribution network; A lower-level model for the online dispatching strategy is established, which utilizes particle swarm optimization to optimize the input information of the lower-level distribution network model. The objective function of the lower-level model for the online dispatching strategy is... for: ； in, and They are nodes The actual voltage and the reference voltage; By integrating the input information from the upper layer of the optimized distribution network model and the input information from the lower layer of the distribution network model, we can obtain the input power of the distribution network reference node, the energy allocation between the photovoltaic and energy storage integrated units, and the reactive power output of the photovoltaic and energy storage nodes.

2. The online dispatching method based on a residential photovoltaic-storage integrated machine cluster according to claim 1, characterized in that, The basic power distribution network model includes: Power flow constraints in distribution networks: Node and The node voltage relationship between them is expressed as follows: ； in, Represents a node The actual voltage, Represents a node The actual voltage; Indicates a branch Current in; and They are branch roads The impedance and inductive reactance on the surface; and Indicates a branch Active power and reactive power and Indicates a branch Active power and reactive power; and Represents a node Injected active and reactive power; constraints are as follows: ； ； and It is a node The set of parent and child nodes must satisfy the following active and reactive power calculation constraints: ； Safety operation constraints of power distribution networks: branch road Current and nodes in The voltage must be within the safe operating range: ； 。 3. The online dispatching method based on a residential photovoltaic-storage integrated machine cluster according to claim 1, characterized in that, Based on the calculated dispatchable or deficit power within the node where the integrated photovoltaic and energy storage unit is connected, it is determined whether the node needs to absorb power from the distribution network. Based on the determination result, energy allocation within the node for the integrated photovoltaic and energy storage unit is performed, including: The process iterates through each user. When a user's photovoltaic power generation exceeds their load, the excess power is calculated and stored in the user's energy storage or allocated to other users on the same node. When a user's photovoltaic power generation is less than their load, the power deficit is calculated. If the user's energy storage has sufficient reserves, the power deficit is first compensated from the user's energy storage; otherwise, it is compensated from other users on the same node. This process continues until the power calculation for all photovoltaic and energy storage users in the node is completed, yielding the maximum dispatchable power output from the node or the deficit power that needs to be absorbed from other nodes. This completes the energy allocation of the integrated photovoltaic and energy storage unit within the node.

4. The online dispatching method based on a residential photovoltaic-storage integrated machine cluster according to claim 1, characterized in that, The upper-level model of the online dispatch strategy utilizes the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model, including: The Benders decomposition algorithm is used to optimize the active power of the photovoltaic-storage nodes, including the input power of the reference node and the power input between the photovoltaic and storage nodes. The input power of the reference node and the power input between the photovoltaic and storage nodes are decomposed into a main problem and sub-problems. The input power of the reference node is obtained by solving the main problem, and the power input and output between the photovoltaic and storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users.

5. The online dispatching method based on a residential photovoltaic-storage integrated machine cluster according to claim 4, characterized in that, The input power of the reference node is obtained by solving the main problem, and the input and output power between the optical and energy storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users, including: Initialize algorithm parameters, and initialize upper and lower bounds; Initialize the main problem: Randomly generate a solution to the main problem within the operating range, i.e., the total input power of the reference nodes. ; Solving the subproblem: Substituting the solution to the main problem into the node information of the distribution network model, and under the constraint conditions, calculating the power flow distribution of the distribution network, thereby obtaining the solution to the subproblem and the corresponding objective value. The solution to the subproblem represents the input and output of user power between the photovoltaic and energy storage nodes, and the corresponding objective function is the energy storage differential. Based on the direction of minimizing the objective value, the input and output of user power between optical storage nodes are optimized through iteration, thereby obtaining the optimal solution of the subproblem and the objective function value minimized thereto. Update the main problem by adding the resulting Benders cut to the constraints of the main problem based on the cutting planes fed back from the subproblems, and adjusting the solution of the main problem, including: The shadow price of the objective solution is calculated using the Lagrange multiplier method in linear programming. The shadow price is typically represented by the Lagrange multiplier corresponding to the constraint at each node. , representing the sensitivity to the solution of the main problem, that is, the rate at which the objective function of the subproblem changes when the solution of the main problem undergoes a small change. The calculation formula is: ； For nodes The power; The Benders cut is calculated using shadow prices to approximate the solution space of the main problem, assuming the solution to the current subproblem is... Then Benders cut is: ； This inequality will be added to the main problem to constrain the solution to the main problem, so that future solutions get closer and closer to the global optimum. The solution to the new principal problem is obtained by taking values ​​from the newly generated solution space of the principal problem. The updated main problem includes the new Benders cut; Substitute the solution to the new main problem into the subproblem to solve it, and repeat the above steps until the solutions to the main problem and the subproblem converge.

6. The online dispatching method based on a residential photovoltaic-storage integrated machine cluster according to claim 1, characterized in that, The lower-level model of the online dispatch strategy utilizes the particle swarm optimization algorithm to optimize the input information of the lower-level distribution network model, including: Initialize the number of particles N in the particle swarm; each particle represents a possible solution; the dimension of a particle is equal to the number of variables to be optimized, i.e., the number of nodes in the integrated photovoltaic and energy storage system; the position of each particle represents its output reactive power. Initialize the particle positions within the feasible range of reactive power; assign an initial velocity to each particle. The initial position of the particle is taken as its initial individual optimal solution. In the initial particle swarm, the particle with the smallest objective function value is selected as the global optimal solution. ; in the In the next iteration, the particle velocity is updated according to the formula: ； in, It is inertial weight; and It is the acceleration constant; the range is 0.1 to 0.

3. and It is a random number; the particle position is adjusted using the updated velocity. If it exceeds the boundary, its position will be restricted to the boundary. In the In each iteration, the objective function value (i.e., the impact of reactive power distribution on system voltage) is calculated for the new position of each particle. The objective function value corresponding to the particle's new position is compared with that of the individual's best position in history. Based on the objective function value of the particle's new position and the objective function value of the particle's historical best position, the better position among the new and historical best positions is determined. If the particle's new position is better, the individual optimal solution of the particle is updated using the new position. ; Compare the individual optimal solutions of all particles, determine the global optimal solution for the current iteration from the individual optimal solutions of all particles, and update the global optimal solution based on the global optimal solution of the current iteration and the global optimal solution of the previous iteration. Check the convergence condition: if the number of iterations reaches the maximum value, or the global optimal solution no longer changes in multiple iterations, the algorithm is considered to have converged; if it has not converged, return to continue iterating; after the algorithm converges, it outputs the global optimal solution, which is the optimal distribution of reactive power.

7. An online dispatching system based on a residential photovoltaic-storage integrated machine cluster, characterized in that, include: The data acquisition module is used to acquire basic data of the low-voltage distribution network; The data preprocessing module is used to preprocess the uncertainties in the basic data during the operation of the power distribution network. The distribution network model building module is used to build a basic distribution network model based on the power flow constraints and safe operation constraints of the distribution network, combined with the preprocessed basic data. The energy allocation module is used to determine whether the nodes connected to the photovoltaic-storage integrated machine need to absorb power from the distribution network based on the calculated dispatchable power or deficit power within the nodes. Based on the determination result, the module allocates energy for the photovoltaic-storage integrated machine within the node and forms the dispatchable power constraint of the photovoltaic-storage node, which is then added to the basic distribution network model to obtain the distribution network model. The model optimization module is used to establish the upper-level model of the online dispatching strategy. This upper-level model utilizes the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model. The objective function of the upper-level model of the online dispatching strategy is... for: ； in, and These are the weights for the difference between active power input and user energy storage capacity, respectively. This represents the active power input to node 1; Indicates at node The The difference between the energy storage capacity of an individual user and the average energy storage capacity of all users in the distribution network; The model optimization module is used to establish a lower-level model of the online dispatching strategy. This lower-level model utilizes a particle swarm optimization algorithm to optimize the input information of the lower-level distribution network model. The objective function of the lower-level model of the online dispatching strategy is... for: ； in, and They are nodes The actual voltage and the reference voltage; The information integration module is used to integrate the input information of the upper layer of the optimized distribution network model and the input information of the lower layer of the distribution network model to obtain the input power of the distribution network reference node, the energy allocation between the photovoltaic and energy storage integrated unit, and the reactive power output of the photovoltaic and energy storage node.

8. The online dispatching system based on a residential photovoltaic-storage integrated machine cluster according to claim 7, characterized in that: The basic power distribution network model includes: Power flow constraints in distribution networks: Node and The node voltage relationship between them is expressed as follows: ； in, Represents a node The actual voltage, Represents a node The actual voltage; Indicates a branch Current in; and They are branch roads The impedance and inductive reactance on the surface; and Indicates a branch Active power and reactive power and Indicates a branch Active power and reactive power; and Represents a node Injected active and reactive power; constraints are as follows: ； ； and It is a node The set of parent and child nodes must satisfy the following active and reactive power calculation constraints: ； Safety operation constraints of power distribution networks: branch road Current and nodes in The voltage must be within the safe operating range: ； 。 9. The online dispatching system based on a residential photovoltaic-storage integrated machine cluster as described in claim 7, characterized in that: Based on the calculated dispatchable or deficit power within the node where the integrated photovoltaic and energy storage unit is connected, it is determined whether the node needs to absorb power from the distribution network. Based on the determination result, energy allocation within the node for the integrated photovoltaic and energy storage unit is performed, including: The process iterates through each user. When a user's photovoltaic power generation exceeds their load, the excess power is calculated and stored in the user's energy storage or distributed to other users on the same node. When a user's photovoltaic power generation is less than their load, the power deficit is calculated. If the user's energy storage has sufficient reserves, the power deficit is compensated from the user's energy storage first; otherwise, it is compensated from other users on the same node. This process continues until the power calculation for all photovoltaic and energy storage users in the node is completed, yielding the maximum dispatchable power output of the node or the deficit power that needs to be absorbed from other nodes. This completes the energy allocation of the integrated photovoltaic and energy storage unit within the node.

10. The online dispatching system based on a residential photovoltaic-storage integrated machine cluster according to claim 7, characterized in that: The upper-level model of the online dispatch strategy utilizes the Benders decomposition algorithm to optimize the input information of the upper-level distribution network model, including: The Benders decomposition algorithm is used to optimize the active power of the photovoltaic-storage nodes, including the input power of the reference node and the power input between the photovoltaic and storage nodes. The input power of the reference node and the power input between the photovoltaic and storage nodes are decomposed into a main problem and sub-problems. The input power of the reference node is obtained by solving the main problem, and the power input and output between the photovoltaic and storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users.

11. The online dispatching system based on a residential photovoltaic-storage integrated machine cluster as described in claim 10, characterized in that: The input power of the reference node is obtained by solving the main problem, and the input and output power between the optical and energy storage nodes is obtained by solving the sub-problems, thereby balancing the energy storage capacity among users, including: Initialize algorithm parameters, and initialize upper and lower bounds; Initialize the main problem: Randomly generate a solution to the main problem within the operating range, i.e., the total input power of the reference nodes. ; Solving the subproblem: Substituting the solution to the main problem into the node information of the distribution network model, and under the constraint conditions, calculating the power flow distribution of the distribution network, thereby obtaining the solution to the subproblem and the corresponding objective value. The solution to the subproblem represents the input and output of user power between the photovoltaic and energy storage nodes, and the corresponding objective function is the energy storage differential. Based on the direction of minimizing the objective value, the input and output of user power between optical storage nodes are optimized through iteration, thereby obtaining the optimal solution of the subproblem and the objective function value minimized thereto. Update the main problem by adding the resulting Benders cut to the constraints of the main problem based on the cutting planes fed back from the subproblems, and adjusting the solution of the main problem, including: The shadow price of the objective solution is calculated using the Lagrange multiplier method in linear programming. The shadow price is typically represented by the Lagrange multiplier corresponding to the constraint at each node. , representing the sensitivity to the solution of the main problem, that is, the rate at which the objective function of the subproblem changes when the solution of the main problem undergoes a small change. The calculation formula is: ； For nodes The power; The Benders cut is calculated using shadow prices to approximate the solution space of the main problem, assuming the solution to the current subproblem is... Then Benders cut is: ； This inequality will be added to the main problem to constrain the solution to the main problem, so that future solutions get closer and closer to the global optimum. The solution to the new principal problem is obtained by taking values ​​from the newly generated solution space of the principal problem. The updated main problem includes the new Benders cut; Substitute the solution to the new main problem into the subproblem to solve it, and repeat the above steps until the solutions to the main problem and the subproblem converge.

12. The online dispatching system based on a residential photovoltaic-storage integrated machine cluster as described in claim 7, characterized in that: The lower-level model of the online dispatch strategy utilizes the particle swarm optimization algorithm to optimize the input information of the lower-level distribution network model, including: Initialize the number of particles N in the particle swarm; each particle represents a possible solution; the dimension of a particle is equal to the number of variables to be optimized, i.e., the number of nodes in the integrated photovoltaic and energy storage system; the position of each particle represents its output reactive power. Initialize the particle positions within the feasible range of reactive power; assign an initial velocity to each particle. The initial position of the particle is taken as its initial individual optimal solution. In the initial particle swarm, the particle with the smallest objective function value is selected as the global optimal solution. ; in the In the next iteration, the particle velocity is updated according to the formula: ； in, It is inertial weight; and It is the acceleration constant; the range is 0.1 to 0.

3. and It is a random number; the particle position is adjusted using the updated velocity. If it exceeds the boundary, its position will be restricted to the boundary. In the In each iteration, the objective function value, i.e., the impact of reactive power distribution on system voltage, is calculated for the new position of each particle. The objective function value corresponding to the particle's new position is compared with that of the individual best position in history. Based on the objective function value of the particle's new position and the objective function value of the particle's historical best position, the new position and the best position among the historical best positions are determined. If the particle's new position is optimal, the individual optimal solution of the particle is updated using the new position. ; Compare the individual optimal solutions of all particles, determine the global optimal solution for the current iteration from the individual optimal solutions of all particles, and update the global optimal solution based on the global optimal solution of the current iteration and the global optimal solution of the previous iteration. Check the convergence condition: if the number of iterations reaches the maximum value, or the global optimal solution no longer changes in multiple iterations, the algorithm is considered to have converged; if it has not converged, return to continue iterating; after the algorithm converges, it outputs the global optimal solution, which is the optimal distribution of reactive power.

13. An online dispatching device based on a residential photovoltaic-storage integrated machine cluster, characterized in that, The device includes a processor and a memory, wherein the memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the device implements the steps of the online dispatching method based on a residential optical storage integrated machine cluster as described in any one of claims 1 to 6.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the online allocation method based on a residential optical storage integrated machine cluster as described in any one of claims 1 to 6.

Images