A spongy power grid system

Abstract

The application belongs to the technical field of power system control, and particularly relates to a sponge power grid system. A centralized method has potential defects in privacy protection and calculation efficiency. The application provides a sponge power grid system, which comprises a network layer, an agent layer and a virtual energy storage layer arranged in sequence. The network layer comprises a plurality of nodes. The agent layer comprises a plurality of agent modules. The virtual energy storage layer comprises a plurality of energy storage components. The nodes, the agent modules and the energy storage components correspond to each other in sequence. The nodes and the agent modules interact with each other. The agent modules and the energy storage components interact with each other. The agent modules can interact with each other. The nodes call energy storage resources of the energy storage components through the agent modules. Local control of voltage is realized, and the potential of massive energy storage is fully utilized. The completely distributed mode effectively avoids the problems of large calculation consumption and privacy leakage caused by centralized massive data.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power system control, and particularly relates to a sponge power grid system. BACKGROUND

[0002] With the continuous development of social economy, the construction of modern power systems puts forward higher requirements for the safe and stable operation of power grids. Emerging challenges include, but are not limited to, voltage fluctuation, cascading trip-out failure and other voltage stability problems that need to be solved. With the increasing proportion of flexible controllable resources in smart grids, voltage regulation problems should be reconsidered. Two most important tasks are involved: one is what resources to use to regulate voltage, and the other is what way to regulate voltage.

[0003] Large photovoltaic and wind power plants are considered to be effective voltage regulation resources. However, the intermittency of wind and photovoltaic power generation leads to significant variability and uncertainty, bringing various challenges to voltage regulation. By integrating distributed energy storage, load-side scheduling capabilities can be further utilized. In the future, more energy bi-directional flow resources can be aggregated to form a more powerful virtual energy storage regulation system.

[0004] In addition, voltage regulation often adopts centralized control, and the control center uniformly manages and schedules all adjustable resources in the region. However, as the number of participants increases, it needs to integrate information from all nodes and then issue control commands. The centralized method has potential defects in privacy protection and computational efficiency. SUMMARY

[0005] 1. Technical problems to be solved

[0006] Based on the problem that voltage regulation often adopts centralized control, and the control center uniformly manages and schedules all adjustable resources in the region. However, as the number of participants increases, it needs to integrate information from all nodes and then issue control commands. The centralized method has potential defects in privacy protection and computational efficiency, the application provides a sponge power grid system.

[0007] 2. Technical solutions

[0008] To achieve the above object, the application provides a sponge power grid system, which comprises a network layer, an agent layer and a virtual energy storage layer arranged in sequence, the network layer comprises a plurality of nodes, the agent layer comprises a plurality of agent modules, the virtual energy storage layer comprises a plurality of energy storage components, the nodes, the agent modules and the energy storage components correspond in sequence, the nodes and the agent modules interact with each other, the agent modules and the energy storage components interact with each other, the agent modules can interact with each other, and the nodes call the energy storage resources of the energy storage components through the agent modules; the sponge power grid system is a decentralized control mode, the energy storage resources of the energy storage components are called through the information interaction between the agent modules corresponding to the nodes, so that the voltage corresponding to the nodes can be controlled within a safe range.

[0009] Another embodiment provided by the application is that the network layer comprises a plurality of photovoltaic mechanisms and fans, the photovoltaic mechanisms and the fans are connected through physical lines, and a plurality of nodes are arranged on the physical lines.

[0010] Another embodiment provided by the application is that the physical lines are radial lines, the end nodes on the physical lines are balance nodes, and the voltage of the balance nodes is taken as a standard value of 1.0 p.u.

[0011] Another embodiment provided by the application is that the energy storage components comprise charging stations, multi-energy conversion stations, electric vehicles, roof photovoltaic houses, intelligent houses or energy storage devices.

[0012] Another embodiment provided by the application is that the agent modules can control the corresponding energy storage components, and the target function of the control is to minimize the voltage difference between the nodes and the balance nodes.

[0013] Another embodiment provided by the application is that the control constraint conditions comprise power balance constraints, virtual energy storage component constraints and network power and voltage safety constraints, and the network power and voltage constraints are coupled with different agent modules.

[0014] Another embodiment provided by the application is that the power balance constraint is wherein, and p i (t) is the discharging and charging power of agent module i at time t, p i (t) is the injection power of the agent module to the power grid, p i (t) is the load of the agent module at time t, and p i (t) is the inverter efficiency of agent module i for converting direct current into alternating current. i,t i,t

[0015] Another embodiment provided by the application is that the nodes and the agent modules correspond one by one.​​​

[0016] Another embodiment provided by the present application is that the agent module is provided with a data processing unit, which outputs after processing information.

[0017] Another embodiment provided by the present application is that the agent module is connected with a display, which is used for displaying voltage fluctuation heat map, total power demand, actual input power and charge-discharge power of the agent module.

[0018] 3. Beneficial effects

[0019] Compared with the prior art, the sponge power grid system provided by the present application has the beneficial effects that:

[0020] The sponge power grid system provided by the present application is an implementation method for the voltage distribution control problem of the urban distribution network under the condition of containing a large number of energy storage devices.

[0021] The sponge power grid system provided by the present application realizes local voltage regulation through multiple virtual energy storages. The proposed model fully considers the multi-time scale of power system scheduling and power flow distribution. The dispersed voltage regulation is realized through limited information transmission between agents, realizing local control of voltage, and fully exerting the potential of a large number of energy storages.

[0022] The sponge power grid system provided by the present application is analytically expressed in all calculation processes, which ensures that it can be efficiently completed under the condition of a large number of energy storages. In the case of further expansion of the scale of future urban distribution networks, the system can realize local regulation, which is beneficial to maintaining good operation of the power system.

[0023] The sponge power grid system provided by the present application is designed to be decentralized, which reduces the operation and maintenance cost of centralized systems, and at the same time, the completely distributed mode effectively avoids the problems of large calculation consumption and privacy leakage caused by the concentration of a large amount of data. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 FIG. 1 is a structural schematic diagram of the sponge power grid system of the present application;

[0025] Figure 2 FIG. 2 is a structural schematic diagram of the virtual energy storage of the present application;

[0026] Figure 3 FIG. 3 is a structural schematic diagram of the radial distribution network of the present application;

[0027] Figure 4 FIG. 4 is a schematic diagram of the average and maximum voltage fluctuation with iteration of the present application;

[0028] Figure 5This is a schematic diagram of voltage fluctuations during a day for Agent 6 and Agent 12 of this application;

[0029] Figure 6 It is a schematic diagram of the power and charge and discharge conditions at different times of this application;

[0030] Figure 7 This is the voltage fluctuation situation of all agents in this application at all times. DETAILED DESCRIPTION

[0031] Hereinafter, specific embodiments of the present application will be described in detail with reference to the accompanying drawings. Based on these detailed descriptions, those skilled in the art will be able to clearly understand the present application and implement the present application. Without violating the principles of the present application, the features of different embodiments may be combined to obtain new implementations, or certain features of certain embodiments may be substituted to obtain other preferred implementations.

[0032] See also Figures 1-7 The present application provides a sponge power grid system, comprising a network layer, an agent layer and a virtual energy storage layer arranged in sequence, wherein the network layer comprises a plurality of nodes, the agent layer comprises a plurality of agent modules, and the virtual energy storage layer comprises a plurality of energy storage components. The nodes, the agent modules and the energy storage components correspond to each other in sequence, the nodes exchange information with the agent modules, the agent modules exchange information with the energy storage components, and the agent modules can exchange information with each other. The nodes call the energy storage resources of the energy storage components through the agent modules. The sponge power grid system adopts a decentralized control mode. Through information interaction between the agent modules, the energy storage resources of the energy storage components are called by the corresponding nodes, so that the voltage of the corresponding nodes can be controlled within a safe range.

[0033] A sponge is a tool made of a soft, porous material that can flexibly absorb and release water. Inspired by the sponge, the sponge grid is built on distributed energy storage components. Its primary goal is to achieve effective distributed control of grid parameters by building flexible communication between numerous energy storage systems. It mobilizes various forms of energy to meet local energy needs while maintaining power system stability and achieving efficient and equitable operation.

[0034] The physical topology of the sponge power grid is consistent with the existing power grid structure, that is, it meets the whole process of existing power system generation, transmission and distribution. The traditional power distribution station calculates the total power generation demand according to the actual power consumption of the downstream, and arranges the power generation plan of each power plant according to certain rules. However, due to the large-scale introduction of distributed energy storage components, the operation mode of the power distribution link has changed. Distributed energy storage components are scattered in various locations, and their size and application scenarios are different, but when they are connected to the same power network structure, the working characteristics of each energy storage component will have different effects on the power grid. At the same time, based on the characteristics of the two-way flow of energy storage, a safe distributed control is realized through the information exchange between energy storage components, so that the energy storage components can maintain the power transmission of the power grid and the node voltage within a stable range. In this way, the entire power grid system is like a sponge, which can flexibly cope with fluctuations caused by load or power generation, thereby stabilizing operation.

[0035] The sponge power grid is a method for constructing a city power distribution network. The sponge power grid designs a decentralized control method, which controls the voltage of the corresponding node within a safe range through limited information exchange between agent modules and calls virtual energy storage resources. The decentralized control method can ensure the balance of power supply and demand within the node, and optimize the scheduling of virtual energy storage components to minimize voltage fluctuations. Agent modules exchange power values and voltage values of corresponding nodes, obtain information related to network power and voltage constraints and other agent modules, and thus realize localized solution of the problem.

[0036] Virtual energy storage represents the sum of different types of energy storage resources under the corresponding node of the agent module, rather than a single energy storage. If a user under the agent has energy storage resources and can provide a certain energy calling space to the agent module, and this calling is bidirectional, that is, both charging and discharging are possible. Then this user transmits the controllable time and capacity to the agent module, and the agent module integrates the energy storage resources of multiple users to form a virtual energy storage resource.

[0037] The sponge power grid system in the present application can fully mobilize the regulation potential of dispersed energy storage components for power networks, and the designed distributed control method can avoid the problems of large calculation amount and privacy invasion caused by centralized calculation, thereby providing a new way for city power distribution network construction.

[0038] Further, the network layer includes a plurality of photovoltaic mechanisms and wind machines, the photovoltaic mechanisms and the wind machines are connected through physical lines, and a plurality of nodes are arranged on the physical lines.

[0039] Further, the physical line end node is a balanced node, and the voltage of the balanced node is taken as a standard value of 1.0 p.u. The physical line of the power distribution network is a radial line, and the first end node is a balanced node, which is used to meet the power demand presented by downstream nodes.

[0040] Further, the energy storage assembly includes a charging station, a multi-energy conversion station, an electric vehicle, a roof photovoltaic house, a smart house, or an energy storage device.

[0041] Further, the agent module can control the corresponding energy storage assembly, and the target function of the control is to minimize the voltage difference between the corresponding node and the balanced node. All agent modules are distributed nodes without difference, and the virtual energy storage owned by the agent modules is controlled.

[0042] Further, the control constraint includes a power balance constraint, a virtual energy storage assembly constraint, and a network power and voltage safety constraint, and the network power and voltage constraint is coupled to different agent modules.

[0043] Further, the power balance constraint is wherein, and p is the discharging and charging power of agent module i at time t, p i,t is the injection power of the agent module to the power grid, is the inverter efficiency of agent module i for converting direct current to alternating current, d i,t is the load of the agent module at time t.

[0044] Further, the node corresponds to the agent module one by one.

[0045] The number of agent modules in the sponge power grid is the same as the number of nodes in the power distribution network, and each agent module has a set of virtual energy storage.

[0046] Further, the agent module is provided with a data processing unit, and the data processing unit outputs after processing information.

[0047] Further, the agent module is connected with a display, and the display is used to display the voltage fluctuation heat map of the agent module, the total power demand, the actual input power, and the charging and discharging power.

[0048] Embodiment

[0049] The specific structure of the sponge power grid is as follows Figure 1As shown in the figure. The upper layer represents the physical circuitry of the power grid, showing a 33-node radial system. Node 1 is the distribution network balancing node, responsible for meeting the load demands of downstream nodes. Each downstream node contains virtual energy storage of varying sizes. This not only meets its own flexible needs, but also responds to the grid voltage. Much like a sponge absorbs and releases water to maintain a stable moisture content, virtual energy storage maintains a flexible and stable grid voltage.

[0050] The introduction of distributed energy storage not only improves the elasticity of the sponge grid distribution layer, but also makes the lower power consumption system with nodes as the boundary highly elastic. The energy storage represented by the node in the sponge grid is the sum of different types of energy storage in its subordinate power consumption system. Various types of energy storage are the core of the sponge grid. If each lower-level energy storage can provide a certain amount of energy call space, and this call is bidirectional, that is, it can be charged and discharged. Then, for Figure 1 For each node in the network, the energy storage under it is a virtual energy storage component, which can integrate various types of energy storage devices in the lower layer to form a virtual energy storage resource with high storage capacity and high power. Figure 2 The structure of virtual energy storage was demonstrated, and six different energy storage usage scenarios were further detailed within three virtual energy storage systems. In multi-energy conversion stations, energy storage plays a transitional role in the energy conversion process, reducing energy waste. Public charging stations and battery swap stations store a large amount of energy storage equipment to efficiently charge electric vehicles. In smart homes, users have energy storage components installed to reduce peak electricity costs or for emergency use. Users with rooftop photovoltaic systems often also have energy storage. When photovoltaic power generation is sufficient, energy storage can be used to store electricity for later use, effectively reducing curtailment. Electric vehicles themselves contain one or more batteries, making them part of the energy storage system. The increasing number of electric vehicles means that mobile energy storage plays a significant role.

[0051] Precisely because nodes contain abundant energy storage resources, while power supply and demand still need to be balanced in real time, the "sponge" properties of energy storage components mean that load fluctuations do not require an immediate change in the power transmitted by the generation node. Instead, the distributed energy storage components within the nodes provide a buffer and balance, thereby reducing network line security risks and power losses. In other words, sponge grids trade space for time, reducing the strict time constraints of real-time power supply and demand balancing.

[0052] Establishment of sponge grid regulation model

[0053] (1) Power and energy storage constraints

[0054] Consider a power network with N agents, denoted as the set N:={1,2,…,N}. T:={1,2,…,T} denotes a time set. The power balance constraint of the agent i is denoted as:

[0055]

[0056] where, and are the discharging and charging power of agent i at time t. i,t is the injection power of agent i to the grid. is the inverter efficiency of agent i, which is used to convert DC to AC. i,t is the load of agent i at time t. The energy storage model is described as follows

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063] where, (2)-(3) are the power constraints of charging and discharging, (4) is the energy storage capacity constraint, (5) is the dynamic constraint of charging and discharging, and (6) is the reservation constraint. SOC i,final is the lower limit of the final capacity value of the energy storage after one day, which is generally taken as the average value of the maximum capacity, i.e. (7) indicates that the energy storage cannot be charged and discharged at the same time.

[0064] (2) Network constraints

[0065] In a given distribution network, N agents and a distributed system dispatcher are given. Assuming that the system is a balanced radial system, a tree graph Γ:={B,Λ} can be used to represent the system, represented by the line set Λ and the node set B. B:={0,1,…,N}, node 0 is the distributed system dispatcher. Therefore, the total number of lines is |L|=N, and the total number of nodes is |B|=N+1. In this application, node i and agent i represent the same subject. Let (i,j) represent a line from i to j, i.e., the active and reactive power from i to j can be represented as P ij,t and Q ij,t . p i,t and q i,t are the injected active and reactive power of node i. Let vi The square of the voltage of the representative node i, while v0 is the voltage of the distributed system dispatcher, which is a fixed value.

[0066] Based on Ohm's law and Kirchhoff's law, ignoring the network loss, the power flow of the power grid can be expressed as

[0067]

[0068]

[0069]

[0070] where Ψ(j) represents the child node of node i, that is, the line (i, j) can also be represented as (i, Ψ(i)). Each line and each node has operating constraints, which are expressed as

[0071]

[0072]

[0073]

[0074]

[0075] where (11)~(12) are the injected power constraints, (13) is the voltage amplitude constraint, and (14) is the complex power constraint of the line. is the maximum complex power on the line (i, j).

[0076] (3) System objective function

[0077] The main role of the sponge network is to smooth voltage fluctuations. Since the regulation of voltage is limited by the entire network reactive power injection, it becomes challenging to establish a local solution. Therefore, the voltage regulation problem is considered to be the superposition of local voltage fluctuations

[0078]

[0079] The goal of this optimization problem is to make the square of the voltage close to the standard value. How to make the constraints also become localized, and then solve P1 in a completely distributed manner.

[0080] The introduction of a large number of virtual energy storage brings challenges to voltage regulation. If a centralized scheduling is adopted, not only will it bring a large amount of calculation, but also the centralized data collection will bring privacy problems. Therefore, it is particularly important to design a self-regulating scheduling scheme, that is, a decentralized method to realize the benign operation of the system.

[0081] (1) Compact form of variables

[0082] Regarding notation, italic (roman) bold letters represent column vectors (matrices). Calligraphic letters represent sets, and R denotes the real space. For convenience of representation, the variables of each agent are compressed on the top, i.e., p i = [p i,1 ,…,p i,T ] T ,d i = [d i,1 ,…,d i,T ] T , v = [v i ,…,v i,1 ] i,T , y = [p T i ,…,p i,1 Sd ,…,p i,T Sd ,…,p i,1 Sc ,…,p i,T Sc ] T . In this way, equations (1)-(6) can be represented as

[0083]

[0084]

[0085]

[0086] where B i ∈ T×2T and C i ∈ T×2T .I is the identity matrix, and tril(·) is the lower triangular matrix. is represented as

[0087]

[0088] where SOC i,0 denotes the initial energy storage capacity of user i.

[0089] (2) Reorganize the network

[0090] The network is reorganized so that all parameters can be solved locally. The main consideration is the injection power and the line power. The injection power includes the part of the agent and the part of the dispatcher. The injection power of the agent is solved by the user direct iteration algorithm, and then the dispatcher balances the injection power of all agents.

[0091]

[0092] ​For line power, each non-feeder node i in the radial network has a unique upstream node denoted by (i). According to graph theory, the distribution line connecting node i and its upstream node (i) is also unique. Except for the dispatcher represented by bus 0, any node connected to an upstream line and this upstream line in the distributed network are a unit, as shown in Figure 3 .

[0093] For a line l between (i) and i, the graph incidence matrix is defined as G e N×(N+1) . G(l,(i)) = 1, G(l,i) = ~1. Then, (8) can be expressed as

[0094]

[0095] By removing node 0 from the above equation, we can obtain G A T P = p A , G A T ∈ N×N , P = [P1…P N T , p A = [p1…p N ] T Let The active power flowing through line i can be expressed as

[0096]

[0097] where, Similarly, (9) can be replaced by

[0098]

[0099] where, Further, (10) and (14) can be replaced by

[0100]

[0101]

[0102] Note 1. The upper end of all buses is unique, while the lower end can be multiple. Therefore, a node and its upstream line are selected to form a unit.

[0103] Note 2. Bus 0 is a balanced bus, which only needs to meet the power demand of other agents. It does not need to consider whether the power of the line is overloaded, because the line connected to it belongs to the unit of the lower end node of the bus, which is irrelevant to it.

[0104] (3) Decentralized algorithm based on information transmission​

[0105] Decision table aggregation into x i = [p i T , q i T , v i T , y i T , P i , Q i ] T The inequality constraints (11)-(13), (17), (18) can be written as

[0106]

[0107] where H i is a block matrix representing the variable coefficients in the inequality constraints. H i = diag(I pi , I qi , I vi , H yi , I Pi , I Qi ). Here, diag(·) is a diagonal matrix. Except for y, all variables are box constraints with no coefficients. The inequality constraints can be written concisely as

[0108]

[0109] where (25) and (7) are expressed as

[0110]

[0111]

[0112] Let where 1∈ T×1 is a column vector with all elements equal to 1. Then P1 becomes

[0113]

[0114] where, is the indicator function of x i , i.e., if x i ∈X i , then otherwise it is infinite. The augmented Lagrangian of (29) is expressed as

[0115]

[0116] where ρ > 0 is the penalty factor, λ = [λ 1iT ,λ 2i T ,λ 3i T ,λ 4i T ,λ 5i T ] T is a multiplier. It can be found that all variables are local except e i and s i . The calculation of e i and s i requires information from other agents. When the topology of the network is given, is known. The distributed algorithm based on information passing is shown in Algorithm 1, k is the iteration counter. in lines 6-8 is the projection of x i onto i , which is given in Algorithm 2.

[0117]

[0118]

[0119]

[0120] Numerical results

[0121] (1) Algorithm performance

[0122] The proposed model is applied to the modified IEEE 33-bus distribution system in Figure 1 . Node 0 represents the slack bus with a voltage reference of 12.7 kV. The upper and lower voltage limits are 1.1 p.u. and 0.94 p.u., T = 24, η i inv = 0.95, η i c and η i d are all taken as 0.9. The replicated algorithm is used for comparison, which is explained as follows. The injected power is replicated to the agent modules and the dispatcher, so that

[0123]

[0124] where is the replicated variable to the dispatcher, which is denoted as The constraints in the model are partly local to the agents and partly to the dispatcher, so that

[0125]

[0126] L DSO= {l | (8) - (14)} (34)

[0127] where m i = {p i , y i} is the decision variable set of agent i. is the decision variable set of dispatchers.

[0128] This problem is then solved by the ADMM algorithm

[0129]

[0130] First, the results are compared with four solutions:

[0131] Solution 1: P1 is solved by the centralized algorithm of Gurobi 9.

[0132] Solution 2: Solved by the algorithm in the present application, but the users do not contain energy storage.

[0133] Solution 3: Solved by the copy algorithm.

[0134] Solution 4: Solved by the algorithm in the present application.

[0135] The value of p is le-4 (k = 1-70), 5e-2 (k = 71-150), and 4e-1 (k > 150). is the original iteration error, is the dual iteration error. The error tolerance is 0.01. Figure 4 The average value and maximum fluctuation of the voltage with iterations are shown in Table 1. The algorithm in the present application and the copy algorithm have similar results, and after 700 iterations, the average voltage fluctuation is stabilized at 0.0064 (p.u.), while the result of the centralized algorithm is 0.0059 (p.u.). The algorithm without energy storage reaches 0.0071 (p.u.) of the average voltage fluctuation. In terms of the average value, these algorithms can all ensure that the voltage does not violate the safety constraint. Figure 4 (b) The change of the maximum voltage fluctuation of all agents in 24 hours is introduced. The centralized algorithm produces a maximum deviation of 0.0490 (p.u.). The algorithm without energy storage can reach stability after 300 iterations, but the maximum voltage fluctuation reaches 0.071 (p.u.), which means that some voltages can be lower than the lower limit of 0.94 (p.u.) and violate the safety constraint. The algorithm in the present application and the copy algorithm both converge after 500 iterations, but the algorithm in the present application controls the maximum deviation to 0.0570 (p.u.), which is slightly better than the 0.0600 (p.u.) of the copy algorithm.

[0136] Figure 5The voltage profiles of agent 6 and 12 are analyzed. Agent 6 is located upstream of a line. The centralized algorithm takes a global perspective and the chosen strategy is to keep the voltage of agent 6 at a high level to ensure that the total voltage fluctuation is minimized. Compared with the centralized algorithm, the voltage of agent 6 calculated by the algorithm in the present application is lower. This is because, in the calculation process, the agent performs local calculation and is more inclined to the goal of minimizing its voltage deviation. This result is also reasonable under the distributed control mode. Agent 12 is located downstream of a line and the voltage of agent 12 is less than 1.0 p.u. throughout the day. The results of the algorithm in the present application and the replicator algorithm are similar, both of which ensure that the voltage fluctuation is within the safe range. Without energy storage, the network regulation is insufficient and the voltage of agent 12 drops to 0.9398 (p.u.) at 8 a.m.

[0137] In addition, the algorithm in the present application can achieve the equivalent results of the replicator algorithm. However, in terms of operation mode, the algorithm in the present application is superior to the replicator algorithm. In the replicator algorithm, the dispatcher needs to wait for all agents to upload their local parameters each time before updating the network parameters. At the same time, the information of the agents needs to be received under the execution of the network boundary conditions before the local operation is performed. In the algorithm in the present application, the network constraints are decomposed and this decentralized method enables the agents to regulate the voltage and local scheduling. This shows that the agents can ensure that the local scheduling meets the safety limit. If the latest message is not available, the algorithm in the present application can obtain the allowed suboptimal solution according to the existing information. In addition, in the replicator algorithm, the dispatcher collects the information of all users in each iteration to update the replica variables, and the dispatcher will inevitably seek excessive benefits through extreme information asymmetry conditions. In contrast, the algorithm in the present application completes the fully distributed calculation by sharing limited information among users and is more fair.

[0138] (2) Sponge power grid characteristics

[0139] After introducing the performance of the algorithm, the characteristics of the sponge power grid will be analyzed in detail. Figure 6 The total electricity demand of all users, the actual input power, and the charge and discharge power are shown. It can be seen that the sponge power grid, which aims to smooth the voltage fluctuation, can effectively peak shaving. In the peak period, the energy storage relieves the load demand. From 7 a.m. to 9 a.m., the energy storage supplements 21.16 MW of power, reducing the original load demand of 124.75 MW to 103.59 MW. From 18:00 to 21:00, the energy storage supplements 33.90 MW of power, reducing the original load demand of 203.83 MW to 169.93 MW. In other periods, the agents are continuously charged to meet the capacity constraint, and this charging also aims to smooth the load. In addition, from Figure 6 It can be seen that the energy storage does not charge and discharge at the same time, which also verifies the effectiveness of the proposed projection method.

[0140] Figure 7 Comparison of the voltage of each agent in different time periods with and without energy storage. Figure 7 The top half of the figure shows a heat map of voltage fluctuations across various agents over the course of a day. As can be seen, the heat map is flatter when energy storage is present than when it is absent. This demonstrates that energy storage regulates voltage in a bidirectional manner. This regulation is effective for different users at different times, demonstrating that the sponge grid optimizes and improves the grid by leveraging its temporal and spatial flexibility. Figure 7 At the bottom is the voltage distribution diagram for all users in each time period. When the load is low, the voltage distribution is mostly around 1 (pu). When the load is high, the voltage fluctuates widely. However, energy storage can reduce voltage deviations. For example, at 8:00 PM, without energy storage, the maximum voltage is 1.0556 (pu) and the minimum is 0.9237 (pu). With the help of energy storage, the maximum voltage is 1.0436 (pu) and the minimum is 0.9446 (pu). During high-load periods, the average voltage level without energy storage is below 1 (pu), indicating that the voltage corresponding to most agents is undervoltage. In contrast, the average level with energy storage is generally around 1, making the voltage distribution in the network more stable. This shows that voltage fluctuations can be mitigated by energy storage.

[0141] Although the present application has been described above with reference to specific embodiments, it should be understood by those skilled in the art that many modifications may be made to the configurations and details disclosed herein within the principles and scope of the present application. The scope of protection of the present application is determined by the appended claims, and the claims are intended to cover all modifications encompassed by the literal meaning or scope of equivalents of the technical features in the claims.

Claims

1. A sponge grid system, characterized by: The invention comprises a network layer, an agent layer and a virtual energy storage layer arranged in sequence, wherein the network layer comprises a plurality of nodes, the agent layer comprises a plurality of agent modules, and the virtual energy storage layer comprises a plurality of energy storage components. The nodes, the agent modules and the energy storage components correspond to each other in sequence, the nodes exchange information with the agent modules, the agent modules exchange information with the energy storage components, and the agent modules can exchange information with each other. The nodes call the energy storage resources of the energy storage components through the agent modules. The sponge grid system adopts a decentralized control mode. Through information interaction between the agent modules, the energy storage resources of the energy storage components are called by the corresponding nodes, so that the voltage of the corresponding nodes can be controlled within a safe range. The network layer comprises a plurality of photovoltaic mechanisms and wind turbines. The photovoltaic mechanisms are connected to the wind turbines through physical lines, and a plurality of nodes are arranged on the physical lines. The physical lines are radial lines, and the end nodes on the physical lines are balance nodes. The voltage of the balance nodes is taken as a per-unit value of 1.

0. pu; the proxy module is capable of controlling the corresponding energy storage component, and the objective function of the control is to minimize the difference between the voltage of the corresponding node and the voltage of the balance node; the control constraints include power balance constraints, virtual energy storage component constraints and network power and voltage safety constraints, and the network power and voltage constraints couple different proxy modules.

2. The sponge grid system according to claim 1, wherein: The energy storage components include charging stations, multi-energy conversion stations, electric vehicles, rooftop photovoltaic houses, smart houses or energy storage devices.

3. The sponge grid system according to claim 1, wherein: The power balance constraint is ;in, and It is a proxy module exist Discharge and charge power at each moment, is the power injected into the grid by the proxy module, It is a proxy module The inverter efficiency, used to convert DC to AC, Is the agent module at time load.

4. The sponge grid system according to claim 1, wherein: The nodes correspond to the agent modules one by one.

5. The sponge grid system according to claim 1, wherein: The proxy module is provided with a data processing unit, which processes the information and then outputs it.

6. The sponge grid system according to claim 5, wherein: The proxy module is connected to a display, and the display is used to display a voltage fluctuation heat map, a total power demand, an actual input power, and a charge and discharge power of the proxy module.

Images