New energy power distribution network collaborative optimization system based on distributed regulation and control architecture

Abstract

The invention relates to a new energy power distribution network collaborative optimization system based on a distributed regulation and control framework. The system comprises a cluster module, a virtual power plant module, a power distribution network module and an optimization module. The cluster module gathers a plurality of small-capacity dispersedly accessed photovoltaic power supplies to form a plurality of photovoltaic power supply clusters; the virtual power plant modules form a plurality of virtual power plants, each virtual power plant comprises a high-capacity photovoltaic power station, an energy storage device and a load, and the virtual power plants serve as an optimization main body to supply power to the internal load and reversely transmit electric quantity to the power distribution network module so as to generate benefits; and the power distribution network module is used for connecting a plurality of photovoltaic power supply clusters and a plurality of virtual power plants. The invention relates to the technical field of new energy power distribution networks. Dispersed small-capacity photovoltaic power supplies are divided into photovoltaic clusters, large-capacity photovoltaic power stations are combined with energy storage and loads to form a virtual power plant, it is guaranteed that the photovoltaic power supplies connected in different forms fully participate in collaborative optimization, and the adverse effect of photovoltaic power generation intermittency and randomness on a power distribution network is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy distribution networks, and in particular to a new energy distribution network collaborative optimization system based on a distributed control architecture. Background Art

[0002] In order to cope with a series of increasingly severe problems such as global energy shortage, environmental pollution and climate deterioration, it has become a consensus of the international community to accelerate the development and utilization of renewable energy. Solar energy is a new type of renewable clean energy, and photovoltaic power generation is one of its main forms of utilization. Compared with traditional power generation technology, photovoltaic power generation has the advantages of abundant resources, wide distribution, clean and pollution-free, no geographical restrictions, and can generate electricity locally, etc., and has received more and more attention and development.

[0003] Compared with large-scale photovoltaic power stations, the characteristics of distributed photovoltaic power generation include low threshold, low construction and installation costs, small installed capacity, short construction period and low transmission and distribution losses. It adheres to the principles of nearby power generation, nearby grid connection, nearby conversion and nearby utilization. While providing the same scale of photovoltaic power station power generation, it reduces or even solves the energy loss caused by cross-regional transmission of electricity, and provides stable electricity within a certain range to make up for the lack of stability of large power grids.

[0004] However, photovoltaic power generation output is intermittent and random. When it is connected to the distribution network in a large-scale, small-capacity, and decentralized manner, the distribution network faces new problems in terms of power flow distribution, network loss, power quality, and safe and stable operation of the system, which brings challenges to the traditional distribution network system. To this end, the present invention proposes a new energy distribution network collaborative optimization system based on a distributed control architecture. Summary of the invention

[0005] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a new energy distribution network collaborative optimization system based on a distributed control architecture to solve the above technical problems.

[0006] The above technical objectives of the present invention are achieved through the following technical solutions:

[0007] A new energy distribution network collaborative optimization system based on a distributed control architecture, including cluster modules, virtual power plant modules, distribution network modules and optimization modules;

[0008] The cluster module gathers a number of small-capacity decentralized photovoltaic power sources to form a number of photovoltaic power source clusters;

[0009] The virtual power plant module is composed of several virtual power plants, which include large-capacity photovoltaic power stations, energy storage equipment and loads. As an optimization subject, the virtual power plant (VPP) not only supplies power to internal loads, but also sends power back to the distribution network module to generate revenue;

[0010] The distribution network module is used to connect several photovoltaic power clusters and several virtual power plants, and collect the operation information of the cluster modules and the virtual power plant modules, including the active and reactive output of the photovoltaic power clusters and the internal power consumption of the VPP;

[0011] The optimization module controls the photovoltaic power inverter according to the active and reactive output of the photovoltaic power cluster to provide a certain reactive power support for the system, and the optimization module supplies power to the distribution network module according to the internal power consumption of the VPP to generate revenue.

[0012] In a preferred example, the present invention can be further configured as follows: the method for the cluster module to divide the photovoltaic power cluster is:

[0013] S1. Obtain relevant network parameters of the distribution network, such as active and reactive power regulation capacity and load data of photovoltaic power sources;

[0014] S2, selecting and calculating the photovoltaic power cluster division index value;

[0015] S3, normalizing the index values ​​and calculating the similarity values;

[0016] S4, during initialization, each photovoltaic power source node in the network is treated as a cluster separately, and the modularity function value at this time is calculated;

[0017] S5, randomly select the cluster where node i in the network is located and merge it with other clusters to form a new cluster, and calculate the incremental value ΔQ of the network's modularity function respectively max , if ΔQ max > 0, the cluster where node i is located is connected to ΔQ max The corresponding clusters are merged into new clusters, otherwise no new cluster structure is formed;

[0018] S6. The newly formed cluster structure is used as a node. The similarity between nodes is the sum of the similarities between the two clusters. Step S5 is repeated until the modularity function value Q of the entire network no longer increases. This is the optimal cluster division result.

[0019] In a preferred example, the present invention can be further configured as follows: the calculation method for normalizing the index value is:

[0020]

[0021] Among them, x i,m is the mth index value of the ith node, max(x i,m ) is the maximum value among the m indices of node i, min(x i,m ) is the minimum value among the m indices of node i, X′ i,mis the mth index value of the ith node after normalization;

[0022] The method for calculating the similarity value is:

[0023]

[0024] Among them, d′ ij is the normalized electrical distance, λ1,λ2 are weights, s′ ij is the similarity value.

[0025] In a preferred example, the present invention can be further configured as follows: the modularity function value is calculated at this time, and the specific method is as follows:

[0026]

[0027] In the above formula, Q is the modularity function value, s ij is the similarity between nodes i and j, s i is the sum of the similarities between node i and other nodes, and s is the sum of the similarities between any two nodes in the network.

[0028] In a preferred example, the present invention can be further configured as follows: the optimization module establishes a VPP optimization operation model, and the objective function is:

[0029] F=f1+f 3+ f4―f2

[0030] Among them, f1 is the electricity purchase cost of the virtual power plant, f2 is the income from the access to the grid of photovoltaic power, f3 is the operating cost of the energy storage system, f4 is the cost of flexible load optimization, and F is the total cost of VPP;

[0031] By imposing constraints on the objective function, the VPP optimization operation model is obtained to minimize the total operation cost of the VPP.

[0032] In a preferred example, the present invention can be further configured as follows: the calculation method of the power purchase cost of the virtual power plant is:

[0033] When the photovoltaic power source cannot meet the internal power supply of the VPP, it is necessary to purchase electricity from the external power grid:

[0034]

[0035] Among them, c t is the electricity purchase price of the virtual power plant, S is the amount of electricity purchased by the virtual power plant from the power grid, T is the number of time periods in the optimization cycle, and Δt is the length of a time period;

[0036] When the photovoltaic power source can meet the internal power supply of VPP, f1 is 0;

[0037] The calculation method of the photovoltaic power grid-connected income is:

[0038] When the photovoltaic power source can guarantee the internal power supply of the VPP, it can reversely supply power to the distribution network to generate income:

[0039]

[0040] in, is the on-grid electricity price of photovoltaic power source, S is the on-grid electricity of photovoltaic power source, T is the number of time periods in the optimization cycle, and Δt is the length of a time period;

[0041] When the photovoltaic power source cannot guarantee the internal power supply of VPP, f2 is 0;

[0042] The calculation method of the energy storage operation cost is:

[0043]

[0044] In the above formula, is the energy storage system loss, C loss is the equivalent operating cost of a single cycle of the energy storage system, Δt is the duration of a period;

[0045] The calculation method of the flexible load regulation cost is:

[0046]

[0047] In the above formula, is the electricity purchase price of the load in period t, is the power of the i-th load in period t, N is the number of loads, ξ p is the translation compensation coefficient of the user of the translatable load, μ p is the state of the translatable load, P t,p is the load capacity that can be translated during period t, ξ q is the reduction compensation coefficient of the load reduction users, μ q For a state where the load can be reduced, P t,q is the load capacity reduced in period t, P is the number of users who can shift load, Q is the number of users who can reduce load, and Δy is the length of a period.

[0048] In a preferred example, the present invention can be further configured as follows: the constraints imposed on the objective function include power balance constraints, energy storage constraints within the VPP, flexible load capacity limitation constraints, photovoltaic power generation constraints, node voltage constraints and branch flow constraints.

[0049] In a preferred example, the present invention can be further configured as follows: the constraint method of the power balance constraint is:

[0050]

[0051] In the above formula, and are the active and reactive power exchanged between VPP and the distribution network during period t, and are all active and reactive loads in VPP during period t, and are respectively the active and reactive losses in VPP during period t, and They are respectively the active and reactive outputs of the photovoltaic power source in VPP during period t.

[0052] In a preferred example, the present invention can be further configured as follows: the constraint method of the photovoltaic power generation power constraint is:

[0053]

[0054] In the above formula, P max and P min are the upper and lower limits of the active output of the photovoltaic power source, P is the actual output active power, Q max and Q min They are the upper and lower limits of the reactive output of the photovoltaic power source, and Q is the actual output reactive power;

[0055] The constraint method of the node voltage constraint is:

[0056] V min ≤V i ≤V max

[0057] In the above formula, V max and V min are the upper and lower limits of the voltage amplitude of node i, V i is the voltage amplitude of node i, i=1,2,3,…,N K , N K is the number of summary points.

[0058] In a preferred example, the present invention can be further configured as follows: the constraint method of the branch power flow constraint is:

[0059] S i ≤S imax

[0060] In the above formula, S i is the power flow of branch i, S imax is the upper limit of the power flow, i∈N L , N L is the total number of branches.

[0061] In summary, the present invention includes at least one of the following beneficial technical effects:

[0062] 1. The present invention divides dispersed small-capacity photovoltaic power sources into photovoltaic clusters, and large-capacity photovoltaic power stations combine energy storage and loads to form virtual power plants, ensuring that photovoltaic power sources connected to the grid in different forms fully participate in collaborative optimization, reducing the adverse effects of intermittent and random photovoltaic power generation on the distribution network;

[0063] 2. When performing cluster division, the present invention uses node net load and active and reactive regulation capacity of photovoltaic power sources as clustering indicators, and applies the modularity evaluation function in complex networks, which can not only describe the characteristics of the photovoltaic power source itself, but also reflect the relationship between the photovoltaic power source and the load when it is connected to the distribution network node, thereby improving the optimization effect and being able to quickly and accurately divide photovoltaic power source clusters when large-scale photovoltaic power sources are connected;

[0064] 3. The present invention aims to maximize the VPP revenue and establishes a VPP optimization operation model, which reduces the electricity cost and increases the income from photovoltaic power access while ensuring the quality of electricity consumption for users. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0066] Figure 1 It is a system diagram of the present invention. DETAILED DESCRIPTION

[0067] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0068] Example:

[0069] Reference Figure 1 , the new energy distribution network collaborative optimization system based on distributed control architecture disclosed in the present invention includes a cluster module, a virtual power plant module, a distribution network module and an optimization module;

[0070] The cluster module gathers a number of small-capacity decentralized photovoltaic power sources to form a number of photovoltaic power source clusters;

[0071] The virtual power plant module is composed of several virtual power plants, which include large-capacity photovoltaic power stations, energy storage equipment and loads. As an optimization subject, the virtual power plant (VPP) not only supplies power to internal loads, but also sends power back to the distribution network module to generate revenue;

[0072] The distribution network module is used to connect several photovoltaic power clusters and several virtual power plants, and collect the operation information of the cluster modules and the virtual power plant modules, including the active and reactive output of the photovoltaic power clusters and the internal power consumption of the VPP;

[0073] The optimization module controls the photovoltaic power inverter according to the active and reactive output of the photovoltaic power cluster to provide a certain reactive power support for the system, and the optimization module supplies power to the distribution network module according to the internal power consumption of the VPP to generate revenue.

[0074] The present invention divides dispersed small-capacity photovoltaic power sources into photovoltaic clusters, and large-capacity photovoltaic power stations combine energy storage and loads to form a virtual power plant, ensuring that photovoltaic power sources connected to the grid in different forms fully participate in collaborative optimization and reducing the adverse effects of the intermittent and random nature of photovoltaic power generation on the distribution network.

[0075] In order to effectively control and utilize large-scale wind farms, Germany first proposed the concept of distributed power clusters. The cluster system has the following two characteristics:

[0076] (1) When participating in the optimization of the distribution network, the cluster is a dynamically changing independent entity. It adjusts the output of the photovoltaic power generation units in the cluster according to the optimization requirements of the distribution network to reduce the calculation pressure of the distribution network;

[0077] (2) Each cluster is composed of photovoltaic power generation units with similar operating characteristics and relatively close distances. The clusters are reasonably arranged to participate in regulation according to the optimization objectives.

[0078] The process of cluster participation optimization can be briefly described as follows: the distribution network sends instructions to the cluster control center according to the formulated optimization target. The center first decomposes the target and distributes it to each cluster. After each cluster completes its own target in a distributed manner, it returns to the control center, and the control center coordinates all cluster results to complete the established target. Therefore, the cluster can effectively handle the communication and computing pressure brought about by the large-scale access of photovoltaic power to the distribution network, and can improve the distribution network's control ability over it.

[0079] When dividing photovoltaic power clusters, it is necessary to fully consider the control method, regulation capability, geographical location, regulation speed, communication method and other characteristics of photovoltaic power. The existing cluster division algorithms mainly include clustering algorithms and intelligent algorithms. The specific analysis is as follows:

[0080] Commonly used clustering algorithms include: partitioning method, hierarchical method, density-based method, grid-based method and model-based method. And clustering algorithms are not limited to hard clustering, that is, each data belongs to only one category. Fuzzy clustering is also a widely used clustering algorithm.

[0081] This embodiment uses the Fast Unfolding clustering algorithm to divide clusters. The algorithm combines the complex network modularity function on the basis of hierarchical clustering to avoid the problem of setting the threshold k. The algorithm is divided into two stages. In the first stage, each node in the network is treated as a cluster separately, and two nodes are randomly selected to merge into a new cluster so that the modularity function value Q continues to increase; in the second stage, the newly formed cluster is equivalent to a node, and new clusters are continuously divided until the modularity function value Q no longer increases. At this time, the optimal clustering result can be obtained. The algorithm can realize the rapid processing of large-scale node data.

[0082] The specific method for the cluster module to divide the photovoltaic power cluster is:

[0083] S1. Obtain relevant network parameters of the distribution network, such as active and reactive power regulation capacity and load data of photovoltaic power sources;

[0084] S2, selecting and calculating the photovoltaic power cluster division index value;

[0085] S3, normalizing the index values ​​and calculating the similarity values;

[0086] The calculation method for normalizing the index value is:

[0087]

[0088] Among them, x i,m is the mth index value of the ith node, max(x i,m ) is the maximum value among the m indices of node i, min(x i,m ) is the minimum value among the m indices of node i, X′ i,m is the mth index value of the ith node after normalization;

[0089] The method for calculating the similarity value is:

[0090]

[0091] Among them, d′ ij is the normalized electrical distance, λ1,λ2 are weights, s′ ij is the similarity value.

[0092] S4, during initialization, each photovoltaic power source node in the network is treated as a cluster separately, and the modularity function value at this time is calculated;

[0093] The conventional method for calculating the modularity function value at this time is as follows:

[0094]

[0095] Among them, Q is the modularity function value, A ij is the weight of the edge connecting nodes i and j, k i is the sum of the weights of all edges connected to node i, and m is the sum of the weights of all edges in the network;

[0096] If nodes i and j are in the same partition, function is 0, otherwise it is not 0.

[0097] The above formula is an evaluation function for community division in complex networks. The traditional modularity function considers the edges between nodes and their weights. However, this embodiment has described the relationship between nodes through node similarity, so it is necessary to replace the weight of the edge between nodes with the calculated value of the similarity between nodes. The improved modularity function is as follows:

[0098]

[0099] In the above formula, Q is the modularity function value, s ij is the similarity between nodes i and j, s i is the sum of the similarities between node i and other nodes, and s is the sum of the similarities between any two nodes in the network.

[0100] S5, randomly select the cluster where node i in the network is located and merge it with other clusters to form a new cluster, and calculate the incremental value ΔQ of the network's modularity function respectively max , if ΔQ max > 0, the cluster where node i is located is connected to ΔQ max The corresponding clusters are merged into new clusters, otherwise no new cluster structure is formed;

[0101] S6. The newly formed cluster structure is used as a node. The similarity between nodes is the sum of the similarities between the two clusters. Step S5 is repeated until the modularity function value Q of the entire network no longer increases. This is the optimal cluster division result.

[0102] This embodiment uses the node net load, active and reactive regulation capacity of photovoltaic power sources as clustering indicators, applies the modularity evaluation function in complex networks, and divides photovoltaic power source clusters through the Fast Unfolding clustering algorithm. The proposed clustering indicators can not only describe the characteristics of the photovoltaic power source itself, but also reflect its relationship with the load when it is connected to the distribution network node.

[0103] In the optimization operation strategy of virtual power plants containing photovoltaic power sources and flexible loads, the photovoltaic power source access income, energy storage operation costs and flexible load regulation costs are used as sub-goals to construct an optimization operation model of virtual power plants with the highest overall economic benefits. For the convenience of description, the highest VPP benefit is converted into the lowest VPP total cost. The virtual power plant is regarded as a stakeholder and actively participates in the power market and distribution network to trade electricity. On the premise of meeting the internal electricity demand, it supplies electricity to the distribution network to obtain benefits.

[0104] The objective function of the VPP optimization operation model is:

[0105] F=f1+f 3+ f4―f2

[0106] Among them, f1 is the electricity purchase cost of the virtual power plant, f2 is the grid-connected income of photovoltaic power source, f3 is the operating cost of the energy storage system, f4 is the flexible load optimization cost, and F is the total cost of VPP.

[0107] By imposing constraints on the objective function, the VPP optimization operation model is obtained to minimize the total operation cost of the VPP.

[0108] Virtual power plants make full use of the fast and flexible regulation characteristics of energy storage systems, add flexible loads to the optimization process, reduce the impact of intermittent photovoltaic power output on virtual power plants, improve load curves, achieve peak load shifting, and promote the effective use of photovoltaic power. Establish a flexible and economical virtual power plant optimization model, coordinate various optimizable resources, maximize the consumption of photovoltaic power, and reduce the operating costs of virtual power plants.

[0109] The calculation method of the power purchase cost of the virtual power plant is:

[0110] When the photovoltaic power source cannot meet the internal power supply of the VPP, it is necessary to purchase electricity from the external power grid:

[0111]

[0112] Among them, c t is the electricity purchase price of the virtual power plant, S is the amount of electricity purchased by the virtual power plant from the power grid, T is the number of time periods in the optimization cycle, and Δt is the length of a time period;

[0113] When the photovoltaic power source can meet the internal power supply of VPP, f1 is 0.

[0114] The calculation method of the photovoltaic power grid-connected income is:

[0115] When the photovoltaic power source can guarantee the internal power supply of the VPP, it can reversely supply power to the distribution network to generate income:

[0116]

[0117] in, is the on-grid electricity price of photovoltaic power source, S is the on-grid electricity of photovoltaic power source, T is the number of time periods in the optimization cycle, and Δt is the length of a time period;

[0118] When the photovoltaic power source cannot guarantee the internal power supply of VPP, f2 is 0.

[0119] The cycle life of the energy storage system during operation is closely related to the depth of discharge. The greater the depth of discharge, the shorter the cycle life. This embodiment takes into account the number of charge and discharge times, state conversion loss and charge and discharge energy loss of the energy storage system in the objective function, and minimizes the energy storage operation cost while ensuring normal and reasonable electricity consumption of users. The energy storage operation cost expression is as follows:

[0120]

[0121] In the above formula, is the energy storage system loss, C loss is the equivalent operating cost of a single cycle of the energy storage system, and Δt is the duration of a period.

[0122] Through the price incentive of time-of-use electricity price and flexible flexible load compensation incentive mechanism, users are encouraged to actively participate in the coordination and optimization process of virtual power plants, promote users to appropriately adjust the power consumption structure and power consumption mode, reduce users' power purchase costs, improve load curves, thereby achieving the purpose of peak shaving and valley filling, reducing the power fluctuation of photovoltaic power sources, and increasing their online benefits. The calculation method of the flexible load regulation cost is:

[0123]

[0124] In the above formula, is the electricity purchase price of the load in period t, is the power of the i-th load in period t, N is the number of loads, ξ p is the translation compensation coefficient of the user of the translatable load, μ p is the state of the translatable load, P t,p is the load capacity that can be translated during period t, ξ q is the reduction compensation coefficient of the load reduction users, μ q For a state where the load can be reduced, P t,q is the load capacity reduced in period t, P is the number of users who can shift load, Q is the number of users who can reduce load, and Δt is the length of a period.

[0125] The constraints imposed on the objective function include power balance constraints, energy storage constraints within the VPP, flexible load capacity limit constraints, photovoltaic power generation power constraints, node voltage constraints and branch flow constraints.

[0126] The constraint method of the power balance constraint is:

[0127]

[0128] In the above formula, and are the active and reactive power exchanged between VPP and the distribution network during period t, and are all active and reactive loads in VPP during period t, and are respectively the active and reactive losses in VPP during period t, and They are respectively the active and reactive outputs of the photovoltaic power source in VPP during period t.

[0129] The constraint method of the photovoltaic power generation power constraint is:

[0130]

[0131] In the above formula, P max and P min are the upper and lower limits of the active output of the photovoltaic power source, P is the actual output active power, Q max and Q min They are the upper and lower limits of the reactive output of the photovoltaic power source, and Q is the actual output reactive power;

[0132] The constraint method of the node voltage constraint is:

[0133] V min ≤V i ≤V max

[0134] In the above formula, V max and V min are the upper and lower limits of the voltage amplitude of node i, V i is the voltage amplitude of node i, i=1,2,3,…,N K , N K is the number of summary points.

[0135] The constraint method of the branch power flow constraint is:

[0136] S i ≤S imax

[0137] In the above formula, S i is the power flow of branch i, S imax is the upper limit of the power flow, i∈N L , N L is the total number of branches.

[0138] This embodiment takes into account the user's electricity cost, network loss cost, and load compensation fee, and establishes a VPP optimization operation model with the goal of maximizing VPP benefits. While ensuring the quality of user electricity, it reduces electricity costs and increases the benefits of photovoltaic power grid access.

[0139] The implementation principle of the above embodiment is:

[0140] The cluster module gathers several small-capacity decentralized photovoltaic power sources to form several photovoltaic power clusters; the virtual power plant module consists of several virtual power plants, which include large-capacity photovoltaic power stations, energy storage equipment and loads. As an optimization subject, the virtual power plant (VPP) not only supplies power to internal loads, but also sends electricity back to the distribution network module to generate revenue; the distribution network module is used to connect several photovoltaic power clusters and several virtual power plants, and collect the operation information of the cluster module and the virtual power plant module, including the active and reactive output of the photovoltaic power cluster and the internal power consumption of the VPP; the optimization module controls the photovoltaic power inverter according to the active and reactive output of the photovoltaic power cluster to provide a certain reactive power support for the system, and the optimization module sends power to the distribution network module according to the internal power consumption of the VPP to generate revenue.

[0141] The present invention divides dispersed small-capacity photovoltaic power sources into photovoltaic clusters, and large-capacity photovoltaic power stations combine energy storage and loads to form a virtual power plant, ensuring that photovoltaic power sources connected to the grid in different forms fully participate in collaborative optimization and reducing the adverse effects of the intermittent and random nature of photovoltaic power generation on the distribution network.

[0142] The embodiments of this specific implementation method are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A new energy distribution network collaborative optimization system based on a distributed control architecture, characterized by: Includes cluster module, virtual power plant module, distribution network module and optimization module; The cluster module gathers a number of small-capacity decentralized photovoltaic power sources to form a number of photovoltaic power source clusters; The virtual power plant module is composed of several virtual power plants, which include large-capacity photovoltaic power stations, energy storage equipment and loads. As an optimization subject, the virtual power plant (VPP) not only supplies power to internal loads, but also sends power back to the distribution network module to generate revenue; The distribution network module is used to connect several photovoltaic power clusters and several virtual power plants, and collect the operation information of the cluster modules and the virtual power plant modules, including the active and reactive output of the photovoltaic power cluster and the internal power consumption of the VPP; The optimization module controls the photovoltaic power inverter according to the active and reactive output of the photovoltaic power cluster to provide a certain reactive power support for the system, and the optimization module supplies power to the distribution network module according to the internal power consumption of the VPP to generate revenue.

2. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 1 is characterized in that: The method for the cluster module to divide the photovoltaic power cluster is: S1. Obtain relevant network parameters of the distribution network, such as active and reactive power regulation capacity and load data of photovoltaic power sources; S2, selecting and calculating the photovoltaic power cluster division index value; S3, normalizing the index values ​​and calculating the similarity values; S4, during initialization, each photovoltaic power source node in the network is treated as a cluster separately, and the modularity function value at this time is calculated; S5, randomly select the cluster where node i in the network is located and merge it with other clusters to form a new cluster, and calculate the incremental value ΔQ of the network's modularity function respectively max , if ΔQ max > 0, the cluster where node i is located is connected to ΔQ max The corresponding clusters are merged into new clusters, otherwise no new cluster structure is formed; S6. The newly formed cluster structure is used as a node. The similarity between nodes is the sum of the similarities between the two clusters. Step S5 is repeated until the modularity function value Q of the entire network no longer increases. This is the optimal cluster division result.

3. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 2 is characterized in that: The calculation method for normalizing the index value is: Among them, x i,m is the mth index value of the ith node, max(x i,m ) is the maximum value among the m indices of node i, min(x i,m ) is the minimum value among the m indices of node i, X′ i,m is the mth index value of the ith node after normalization; The method for calculating the similarity value is: Among them, d′ ij is the normalized electrical distance, λ1,λ2 are weights, s′ ij is the similarity value.

4. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 3 is characterized in that: The specific method for calculating the modularity function value at this time is as follows: In the above formula, Q is the modularity function value, s ij is the similarity between nodes i and j, s i is the sum of the similarities between node i and other nodes, and s is the sum of the similarities between any two nodes in the network.

5. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 4 is characterized in that: The optimization module establishes a VPP optimization operation model, and the objective function is: F=f1+f 3+ f4―f2 Among them, f1 is the electricity purchase cost of the virtual power plant, f2 is the income from the access to the grid of photovoltaic power, f3 is the operating cost of the energy storage system, f4 is the cost of flexible load optimization, and F is the total cost of VPP; By imposing constraints on the objective function, the VPP optimization operation model is obtained to minimize the total operating cost of the VPP.

6. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 5 is characterized in that: The calculation method of the power purchase cost of the virtual power plant is: When the photovoltaic power source cannot meet the internal power supply of the VPP, it is necessary to purchase electricity from the external power grid: Among them, c t is the electricity purchase price of the virtual power plant, S is the amount of electricity purchased by the virtual power plant from the power grid, T is the number of time periods in the optimization cycle, and Δt is the length of a time period; When the photovoltaic power source can meet the internal power supply of VPP, f1 is 0; The calculation method of the photovoltaic power grid-connected income is: When the photovoltaic power source can guarantee the internal power supply of the VPP, it can reversely supply power to the distribution network to generate income: in, is the on-grid electricity price of photovoltaic power source, S is the on-grid electricity of photovoltaic power source, T is the number of time periods in the optimization cycle, and Δt is the length of a time period; When the photovoltaic power source cannot guarantee the internal power supply of VPP, f2 is 0; The calculation method of the energy storage operation cost is: In the above formula, is the energy storage system loss, C loss is the equivalent operating cost of a single cycle of the energy storage system, Δt is the duration of a period; The calculation method of the flexible load regulation cost is: In the above formula, is the electricity purchase price of the load in period t, is the power of the i-th load in period t, N is the number of loads, ξ p is the translation compensation coefficient of the user of the translatable load, μ p is the state of the translatable load, P t,p is the load capacity that can be translated during period t, ξ q is the reduction compensation coefficient of the load reduction users, μ q For a state where the load can be reduced, P t,q is the load capacity reduced in period t, P is the number of users who can shift load, Q is the number of users who can reduce load, and Δt is the length of a period.

7. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 6 is characterized in that: The constraints imposed on the objective function include power balance constraints, energy storage constraints within the VPP, flexible load capacity limit constraints, photovoltaic power generation power constraints, node voltage constraints and branch flow constraints.

8. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 7 is characterized in that: The constraint method of the power balance constraint is: In the above formula, and are the active and reactive power exchanged between VPP and the distribution network during period t, and are all active and reactive loads in VPP during period t, and are respectively the active and reactive losses in VPP during period t, and They are respectively the active and reactive outputs of the photovoltaic power source in VPP during period t.

9. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 8 is characterized in that: The constraint method of the photovoltaic power generation power constraint is: In the above formula, P max and P min are the upper and lower limits of the active output of the photovoltaic power source, P is the actual output active power, Q max and Q min They are the upper and lower limits of the reactive output of the photovoltaic power source, and Q is the actual output reactive power; The constraint method of the node voltage constraint is: In min ≤V i ≤V max In the above formula, V max and V min are the upper and lower limits of the voltage amplitude of node i, V i is the voltage amplitude of node i, i=1,2,3,…,N K , N K is the number of summary points.

10. The new energy distribution network collaborative optimization system based on distributed control architecture according to claim 9 is characterized in that: The constraint method of the branch power flow constraint is: S i ≤S imax In the above formula, S i is the power flow of branch i, S imax is the upper limit of the power flow, i∈N L , N L is the total number of branches.

Images