A power distribution network multi-power router planning method based on time sequence region division

Abstract

The present application relates to the technical field of smart grid, in particular to a power distribution network multi-energy router planning method based on time sequence region division. The present application is applied to an energy router, comprising constructing a simplified energy router model, which comprises equivalent internal loss of the energy router to a virtual external line connected between a port of the energy router and a virtual node representing the port of the energy router; adopting lower planning in a double-layer planning model, based on the simplified energy router model, determining capacity configuration of the energy router connection node and port through optimal power flow calculation. Compared with the prior art, the present application simplifies problem solving complexity, and compared with single energy router planning, can better improve the economy and operation performance of the power distribution network, and can better solve the problems of voltage deviation and insufficient distributed power supply accommodation capacity caused by random and volatility of new energy unit output and disorderly access of distributed power supply in the power distribution network.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart grid, and particularly relates to a power distribution network multi-power router planning method based on time sequence region division. BACKGROUND

[0002] With the rapid development of power systems, the traditional power distribution network is limited by the topological constraints of radial operation, and it is difficult to properly handle the node voltage deviation problem and power transmission loss problem in the network. As a new component of the power system, distributed generation such as photovoltaic and wind power is connected to the power distribution network on a large scale, which aggravates the voltage deviation problem and introduces new problems such as power flow return and low wind power / photovoltaic accommodation level. Therefore, finding an efficient power distribution network control method suitable for new power systems has become a research focus.

[0003] The power router is a kind of multi-port flexible converter capable of flexibly adjusting power, which provides a solution to the power transmission loss and radial topology constraint problem of the power distribution network. Similar to the router in the Internet, the power router can flexibly control the power transmission in the power grid. The existing architecture of the power router is initially used for the interconnection of power transmission systems across voltage levels, and its typical structure is composed of AC / DC converter ports, an isolation stage and an internal DC bus.

[0004] The prior art discloses an intelligent control system and method for energy optimization scheduling of a power router (publication number: CN116247722A), which is composed of a device parameter input module, a data acquisition module, a data prediction module, a power router short-time energy scheduling decision module, a power router energy scheduling database module and a data terminal display module. The data prediction module performs time sequence data prediction on user electricity consumption data and photovoltaic power generation data. The power router short-time energy scheduling decision module, based on the input data and the built-in scheduling algorithm, optimizes the scheduling step length, takes the optimal economic benefit of the power router as the target, formulates an energy storage energy management optimization scheduling strategy, and issues specific control commands to each device unit of the power router. The invention can intelligently realize the storage charging and discharging control strategy according to user electricity consumption habits, real-time electricity prices and user photovoltaic power generation prediction power. Although the invention has advantages in improving the performance and economy of the power grid based on the power router, its application in the power distribution network still has limitations. On the one hand, the isolation stage in the existing power router model may cause voltage matching problems of the power router port, and on the other hand, the existence of the isolation stage and the DC bus makes the power flow calculation need to be alternately performed between the nodes of the power distribution network and the internal ports of the power router, increasing the calculation complexity. In addition, the planning of the power router in the power distribution network is complex, and the above factors lead to the problems of voltage deviation and insufficient distributed power supply accommodation capacity caused by the random nature and volatility of the distributed power supply and the output of new energy units. SUMMARY

[0005] The application aims to solve the problems of voltage deviation and insufficient distributed power accommodation capacity caused by random and fluctuation of new energy unit output and disorderly access of distributed power in power distribution network, and the problem of overly complex power flow calculation, and provides a power distribution network multi-energy router planning method based on time sequence region division.

[0006] The application provides a power distribution network multi-energy router planning method based on time sequence region division, applied to an energy router, comprising:

[0007] A simplified energy router model is constructed, which comprises equivalent of internal loss of the energy router to a virtual external line connected between a port of the energy router and a virtual node representing the port of the energy router;

[0008] A lower layer planning in a double-layer planning model is adopted, and capacity configuration of a connected node and port of the energy router is determined through optimal power flow calculation based on the simplified energy router model.

[0009] Preferably, the optimal power flow calculation adopted by the lower layer planning specifically comprises distribution of capacity of the port of the energy router, and removal of the port of the energy router with the capacity of 0 to determine capacity configuration of the node and port connected to the port of the energy router.

[0010] Further preferably, the optimal power flow calculation is based on an optimal power flow model of a power distribution network containing the energy router, comprising:

[0011] Node power balance:

[0012]

[0013] In the formula: Pj,t is active power absorbed from the power grid by the virtual node j at time t, Qj,t is reactive power absorbed from the power grid by the virtual node j at time t; Pj,t is active load demand of the virtual node j at time t, Qj,t is reactive load demand of the virtual node j at time t; Pj,t is predicted output of the distributed photovoltaic of the virtual node j at time t, Pj,t is curtailment power of the distributed photovoltaic of the virtual node j at time t, and the difference between the two represents actual output of the distributed photovoltaic at the current time; Pj,t is active power purchased from the upper-level power grid by the virtual node j at time t, Qj,t is reactive power purchased from the upper-level power grid by the virtual node j at time t; Pj,t is discharge power of the energy storage device deployed at the virtual node j at time t, Pchargej,tis the charging power of the energy storage device deployed at the virtual node j at time t; Pportsetis the set of power router ports, Pgridsetis the set of power grid ports;

[0014] Line transmission balance:

[0015]

[0016] where: Ptransj,k,tis the active power transmitted from node j to node k at time t, Qtransj,k,tis the reactive power transmitted from node j to node k at time t; Plossi,j,tis the transmission loss active power of line (i, j) at time t, Qlossi,j,tis the transmission loss reactive power of line (i, j) at time t, Ii,j,tis the current of line (i, j) at time t, Ri,jis the resistance of line (i, j), Xi,jis the reactance value of line (i, j); Ptransi,j,tis the active power transmitted by line (i, j) at time t, Qtransi,j,tis the reactive power transmitted by line (i, j) at time t;

[0017] Node voltage balance:

[0018]

[0019] where: Vi,tis the voltage of power grid node i at time t, Vj,tis the voltage of virtual node j at time t;

[0020] Power grid operation constraints:

[0021]

[0022]

[0023]

[0024] where: Ci,jis the capacity of line (i, j), , Ii,jminand Ii,jmaxare the lower and upper limits of the current transmitted by line (i, j), respectively; , Vi,minand Vi,maxare the lower and upper limits of the voltage of node i, respectively;

[0025] Power router operation and configuration constraints:

[0026]

[0027] wherein: is a capacity configuration coefficient of the power router port corresponding to the virtual node j, is a unit port reference capacity of the power router, is an active power absorbed from the power grid by the virtual node j at time t, is a reactive power absorbed from the power grid by the virtual node j at time t; is a port set of the i-th power router.

[0028] Preferably, the power router is applied to the bi-level programming, comprising:

[0029] S1, based on the power router candidate deployment node, using the upper level programming in the bi-level programming model, determining the position and quantity of the power router, combining the capacity configuration of the power router connection node and port determined by the lower level programming, obtaining a certain number of power router planning schemes;

[0030] S2, taking the minimum total annual cost as the objective function, screening the power router planning scheme to obtain the final power router planning scheme, wherein the total annual cost includes the annual deployment cost of the power router, the configuration cost of the power router and the annual operation cost of the distribution network.

[0031] Further preferably, the S1 further comprises:

[0032] Step A, dividing the typical daily operation cycle of the distribution network into a certain number of snapshots, and abstracting the distribution network into the time sequence network model, the time sequence network model using coupling matrix to represent the connectivity between different nodes in the time sequence network model and the electrical distance information within the snapshot;

[0033] Step B, obtaining the coupling matrix corresponding to the snapshot through the coupling matrix power flow weight and the coupling matrix voltage weight of the time sequence network model;

[0034] Step C, based on the snapshot and its corresponding coupling matrix, performing time sequence regional division on the distribution network, obtaining sub-zones and complete zones;

[0035] Step D, based on the source and load gravity center identification analysis, obtaining the centralized distribution relationship of the power supply and load in the sub-zone, combining the complete zone and the centralized distribution relationship to determine the power router candidate deployment node;

[0036] Further preferably, the Step C, based on the snapshot and its corresponding coupling matrix, performing time sequence regional division on the distribution network, specifically includes two steps of static local optimization and time sequence incremental update;

[0037] The static local optimization is specifically to divide the time series area based on the local adaptive function to obtain a set number of sub-partitions, and to evaluate the strength of the sub-partitions using the modularity function to obtain the optimal initial partition, that is, the snapshot at the initial time. Partition results within ;

[0038] Specifically, the time series incremental update is to adjust the belonging relationship between the sub-partitions and the boundary nodes based on the optimal initial partition by analyzing the changes of nodes and edges between the snapshots to obtain the complete partition.

[0039] Further preferably, the source-load gravity center identification and analysis in Step D is specifically to identify and analyze the power source gravity center and the load gravity center in the sub-area;

[0040] The centralized distribution relationship of the power supply and the load in the sub-area is determined based on the magnitude relationship between the center of gravity distance of the power supply center of gravity and the center of gravity distance of the load center of gravity. The centralized distribution relationship includes two types, as follows:

[0041] The load gravity center is ahead of the power gravity center, and the power of the power node in the sub-area is first transmitted in reverse to the load gravity center, and then continues to be transmitted in reverse to the first node of the sub-area;

[0042] The power center of gravity is ahead of the load center of gravity, and power reverse transmission exists in the portion from the power center of gravity to the first node of the sub-partition.

[0043] Further preferably, the annual deployment cost of the power router in the total annual cost is calculated by the upper-level planning, specifically as follows:

[0044]

[0045] Where: is the annual deployment cost of the power router, is the number of deployed power routers, is the unit deployment cost of the power router, is the discount rate of the power router, is the life span of the power router.

[0046] Where: is the configuration cost of the power router, is the unit port capacity configuration cost of the power router, is the capacity configuration coefficient of the power router port corresponding to the virtual node j, Unit construction cost of new line for the power router, Binary variable for the power router construction of the new line, 1 when the virtual node j capacity configuration of the power router port is not zero, otherwise 0; Total resistance of the new line for the power router, wherein the power grid node i is a power grid node connected to the virtual node j; Discount rate of the new line, Lifetime of the new line;

[0047]

[0048] In the formula: Annual operation cost of the power grid, Network loss of the power grid line at time t, Network loss of the new line of the power router at time t, Network loss unit price; Power purchase amount of the power grid from the upper-level power grid at time t, Electricity price of the power grid from the upper-level power grid at time t; Distributed photovoltaic power purchase amount at time t, Distributed photovoltaic electricity price at time t; Distributed photovoltaic abandoned power at time t, Abandoned light penalty unit price of distributed photovoltaic at time t.

[0049] Further preferably, the S2 specifically screens the planning scheme generated by the bi-level planning model through the objective function, and takes the annual total cost minimum as the final planning scheme of the power router, and the objective function is specifically as follows:

[0050]

[0051] In the formula: Annual deployment cost of the power router, Configuration cost of the power router, Annual operation cost of the power grid.

[0052] Compared with the prior art, the application has the beneficial effects that: the problem solving complexity is simplified by time sequence region division and source load center identification analysis, and the solving efficiency is greatly improved; secondly, the application can be applied to the optimal power flow calculation of the power distribution network by constructing a simplified electric energy router model, which greatly reduces the calculation complexity compared with the prior art; in addition, the method is suitable for multi-electric energy router planning in the power distribution network, which can better solve the problems of voltage deviation and insufficient distributed power supply accommodation capacity caused by the random and volatility of new energy unit output and the disorderly access of distributed power supply in the power distribution network compared with the single electric energy router planning of the traditional method, and can better improve the economy and operation performance of the power distribution network. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 The flowchart of the power distribution network multi-electric energy router planning method based on time sequence region division in embodiment 1.

[0054] Figure 2 The circuit diagram of the simplified electric energy router model in embodiment 1.

[0055] Figure 3 The power distribution network model containing electric energy routers in embodiment 1.

[0056] Figure 4 The equivalent virtual node model of the electric energy router in embodiment 1.

[0057] Figure 5 The IEEE-33 node system structure diagram in embodiment 3.

[0058] Figure 6 The simulation parameter diagram in embodiment 3.

[0059] Figure 7 The IEEE-33 node partition result diagram in embodiment 3.

[0060] Figure 8 The electric energy router parameter configuration diagram in embodiment 3.

[0061] Figure 9 The cost and power comparison result diagram of the three schemes in embodiment 3.

[0062] Figure 10 The voltage deviation rate result diagram of the three schemes in embodiment 3.

[0063] Figure 11 The ER working efficiency diagram of the two schemes in embodiment 3.

[0064] Figure 12 The solving time and total cost diagram of the two schemes under different allowable errors in embodiment 3. DETAILED DESCRIPTION

[0065] The application will be described in further detail below with reference to the embodiments. However, it should be understood that the scope of the above subject matter of the application is not limited to the following embodiments, and any technology achieved based on the content of the application falls within the scope of the application.

[0066] In the description of the embodiments of the application, the terms indicating the orientation or positional relationship such as "upper", "lower", "left", "right", "center", "inner", "outer", etc. are expressed based on the orientation or positional relationship shown in the drawings or the orientation or positional relationship in which the product / device / apparatus of the application is usually placed. These terms of orientation or positional relationship are only used for the convenience of describing the application or simplifying the description in the embodiments to facilitate the understanding of the scheme by the skilled person, and are not intended to indicate or imply that a specific device / component / element must have a specific orientation or be constructed and operated in a specific positional relationship, and therefore cannot be understood as a limitation on the application.

[0067] In addition, the terms "first", "second", "third", etc. appearing in the terms are only used to distinguish the description of the same or similar components, and should not be understood as emphasizing or implying the relative importance of the specific components.

[0068] In addition, in the description of the embodiments of the application, "several", "a plurality of", "several" represent at least 2. It can be 2, 3, 4, 5, 6, 7, 8, 9, etc. in any case, and even more than 9.

[0069] In addition, in the description of the technical scheme of the application, unless otherwise specified / limited / limited, the terms "set", "install", "connect", "connect", "set", "lay", "arrange" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected, such as welding, riveting, bolting, screwing, etc. The connection means commonly used in the art. The connection can be mechanical connection, electrical connection or communication connection; it can be directly connected or indirectly connected through an intermediate medium; it can be the communication between two elements.

[0070] Embodiment 1

[0071] To solve the problems of voltage deviation and insufficient distributed power supply accommodation capacity caused by the random access of distributed power supply and the randomness and volatility of new energy unit output in the power distribution network, the embodiment provides a power distribution network multi-energy router planning method based on time sequence region division, and the flow chart is as shown in Figure 1 The steps are as follows:

[0072] S1, dividing a typical daily operation cycle of a power distribution network into a set number of snapshots, and abstracting the power distribution network into a time-series network model, the time-series network model representing connectivity between different nodes in the time-series network model and electrical distance information within the snapshot by a coupling matrix;

[0073] S2, obtaining a coupling matrix corresponding to the snapshot through a coupling matrix power flow weight and a coupling matrix voltage weight of the time-series network model;

[0074] S3, performing time-series regional division on the power distribution network based on the snapshot and the coupling matrix corresponding thereto, to obtain a complete partition;

[0075] S4, obtaining a centralized distribution relationship of power sources and loads within the sub-partition based on source and load gravity identification analysis, and determining a candidate deployment node of an energy router in combination with the complete partition and the centralized distribution relationship;

[0076] S5, simplifying a model of the energy router, canceling a voltage isolation level device in the energy router, and introducing a virtual node to represent a connection relationship between nodes in the power distribution network and the energy router, the virtual node and a newly added line representing connection between a port of the energy router and a node of the power distribution network and corresponding transmission loss of the connection;

[0077] S6, adopting a bi-level programming model, the bi-level programming model specifically including an upper level programming and a lower level programming, the upper level programming determining a number and a position of the energy router based on a sub-partition gravity node of the power distribution network obtained through the time-series regional division and the source and load gravity identification analysis, and the lower level programming optimizing a capacity of a connection node and a port of the energy router through optimal power flow calculation;

[0078] S7, taking a total annual cost minimum as an objective function, and screening a result of the bi-level programming model to obtain a final planning scheme of the energy router, wherein the total annual cost includes an annual deployment cost of the energy router, a configuration cost of the energy router, and an annual operation cost of the power distribution network.

[0079] A calculation process of the coupling matrix power flow weight, the coupling matrix voltage weight, and the coupling matrix corresponding to the snapshot is specifically as follows:

[0080] Coupling matrix power flow weight:

[0081] For the tthsnapshot, the coupling matrix power flow weight is defined as:

[0082]

[0083] In the formula: is the coupling matrix power flow weight, is the active power flow transmitted on line (i, j) of the distribution network at time t, is the reactive power flow transmitted on line (i, j) of the distribution network at time t, is the power factor of the distribution network at time t, is the total reactive power injected into the distribution network at the current moment, is the total active power injected into the distribution network at the current moment, and normal represents the absolute value;

[0084] Coupling matrix voltage weights:

[0085] The voltage weight of the coupling matrix is ​​derived based on the voltage sensitivity in the network sensitivity matrix. The network sensitivity matrix reflects the impact of node active power and reactive power changes on node voltage amplitude and phase:

[0086]

[0087] Where: Inject active power changes into the node, Inject reactive power changes into the node, is the phase angle change of the node voltage, is the change in node voltage amplitude; is the active-voltage sensitivity coefficient, is the reactive-voltage sensitivity coefficient;

[0088] Based on this inference, the calculation method of the coupling matrix voltage weight is obtained:

[0089]

[0090] In the formula is the coupling matrix voltage weight, is the active power-voltage sensitivity of distribution network node i to node j in the t-th snapshot, is the reactive power-voltage sensitivity of distribution network node i to node j in the t-th snapshot, is the active power injected into the grid by the distribution network node i in the t-th snapshot, The reactive power injected into the grid by the distribution network node i in the t-th snapshot.

[0091] The coupling matrix corresponding to the snapshot is:

[0092] The t-th snapshot is obtained according to the coupling matrix power flow weight and the coupling matrix voltage weight. The coupling matrix of , the specific calculation process is as follows:

[0093]

[0094] In the formula: is the corresponding element of the ith row and jth column of the coupling matrix of the tth snapshot, is a weighting coefficient, specifically 0.5, is the power flow weight of the coupling matrix, is the voltage weight of the coupling matrix.

[0095] The S3 performs time sequence regional division on the power distribution network based on the snapshots and the corresponding coupling matrices, specifically including two steps of static local optimization and time sequence incremental update;

[0096] (1) Static local optimization

[0097] In the static local optimization stage, the time sequence partition method takes the local self-adaptation function of the sub-partition as the basis for division. For a sub-partition g, the expression of the local self-adaptation function is:

[0098]

[0099] In the formula, is the local self-adaptation of the sub-partition g, the higher the value, the higher the structural strength of the sub-partition g, and the better the partition effect, is the internal weight of the sub-partition g, is the external weight of the sub-partition g, is a resolution parameter that affects the number of sub-partitions, the higher the value, the more the number of sub-partitions;

[0100] For the partition result, the modularity function is used to evaluate the partition strength:

[0101]

[0102] In the formula: Q is the modularity value of the partition result, is the degree of node i, equal to the sum of the coupling weights of all lines connected to the node, and m is the sum of the coupling weights of all lines of the power distribution network, is a binary variable, which is 1 when there is a line connection between nodes i and j, otherwise it is 0; the modularity value reflects the quality of the partition result, and its value range is between (-1, 1), the higher the value, the better the partition effect.

[0103] The steps of using the local self-adaptation function and the modularity function for static initial partitioning are as follows:

[0104] Step 1: Initialization: Set the initial resolution value, and randomly select a seed node from the power supply nodes as the starting point of partitioning.

[0105] Step 2: Partition expansion: For the current partition, evaluate the degrees of adjacent external nodes. If all degrees are negative, stop expanding; otherwise, incorporate the node with the highest degree into the partition to form a new partition.

[0106] Step 3: Partition optimization: Recalculate the degree of each node in the new partition, remove nodes with negative degrees, and obtain the final partition.

[0107] Step 4: Iterative process: Update the current partition and repeatedly select seed nodes to form other partitions until all nodes have partitions.

[0108] Step 5: Optimal partition selection: Increase the resolution parameter and repeat the above steps to obtain the partition results and modularity values ​​at different resolutions. Select the partition with the highest modularity as the optimal initial partition, where the optimal initial partition is the snapshot at the initial time. Partition results within .

[0109] (2) Time series incremental update

[0110] Based on the optimal initial partition, the time incremental update method dynamically adjusts the affiliation between nodes and subpartitions by analyzing the changes of nodes and edges between different snapshots.

[0111] In a temporal network, boundary nodes are located at the outermost edges of subpartitions and directly connected to other subpartitions. Therefore, they are directly affected by network changes. When the network changes, the ownership of boundary nodes must be reevaluated. Time incremental updates focus on changes in the ownership of boundary nodes within each subpartition, as well as state transitions between internal nodes and boundary nodes. The conditions for updating boundary nodes are as follows:

[0112]

[0113] Where: is the degree of belonging of the boundary node a to the sub-partition in the t-1 snapshot, is the update threshold of the boundary node;

[0114] After the snapshot changes, if the degree of a boundary node's belonging to the adjacent partition exceeds its belonging to the current partition, and the relative size exceeds the update threshold, the boundary node is included in the adjacent partition. Otherwise, the partition belonging of the boundary node remains unchanged. The value affects the update frequency of the boundary node. The larger the value, the less likely the boundary node is to change its ownership. The smaller the value, the more frequently the partition ownership of the boundary node is updated as the snapshot changes.

[0115] Snapshot Partition results The steps for performing a time series incremental update are:

[0116] Step 1: Obtain initial time snapshot Partition result in and boundary node index ;

[0117] Step 2: Calculate coupling weight matrix of current time snapshot ;

[0118] Step 3: Calculate the change of ownership of each boundary node in based on and update the boundary condition, and determine whether to update the ownership of the boundary node: If the relative change of ownership is less than the update threshold, keep the current partition ownership of the boundary node unchanged, and keep the current boundary node in the boundary node index;

[0119] If the relative change of ownership is greater than the update threshold, include the node in the adjacent sub-partition, update the partition result

[0120] , and update the boundary node index .

[0121] Step 4: Repeat step 3 until the boundary node index no longer changes;

[0122] Step 5: Output the partition result and the boundary node index of the current snapshot .

[0123] Steps 1 to 5 update the boundary nodes and sub-partition results. Apply this process to each time snapshot to dynamically update the partition results of the time network model, and finally generate the complete partition of the power distribution network model.

[0124] Each sub-partition can be equivalently represented as a mass point according to the results of the time sequence area division. In this representation, the load demand of the node corresponds to the mass of the load, and the power supply capacity of the power supply node represents the mass of the power supply; by applying the calculation principle of the center of gravity of the particle system in the uniform gravitational field, multiplying the mass of each part by its position vector, adding the results, and then dividing by the total mass, a specific method for determining the center of gravity of the power supply and the center of gravity of the load in the sub-partition can be derived:

[0125]

[0126] In the formula: is the distance of the center of gravity of the power supply of the sub-partition g, ​​The load center distance of the sub-partition g represents the electrical distance between the power center and the load center of the sub-partition, and the sub-partition first node is defined as the node in the sub-partition that has the shortest electrical distance from the root node of the power distribution network;

[0127] The source load center identification analysis in the S4 is specifically identification analysis on the power center and the load center in the sub-partition;

[0128] According to the size relationship between the center distance of the power center and the center distance of the load center, the concentrated distribution relationship of the power and the load in the sub-partition is judged, and the concentrated distribution relationship includes two kinds, which are specifically as follows:

[0129] The load center is ahead of the power center, and the power of the power node in the sub-partition is first transmitted reversely to the load center and then reversely transmitted to the sub-partition first node;

[0130] The power center is ahead of the load center, and the power reverse transmission exists in the part from the power center to the sub-partition first node.

[0131] The S5 simplifies the model of the electric energy router to cooperate with the optimal power flow calculation in the S6;

[0132] The model simplification is specifically that the voltage isolation device in the electric energy router is cancelled, and a virtual node is introduced to represent the connection relationship between the nodes in the power distribution network and the electric energy router. The virtual node and the newly added line represent the connection between the port of the electric energy router and the node of the power distribution network and the corresponding transmission loss of the connection. The simplified electric energy router model circuit diagram is as shown in Figure 2 , in which i is a power distribution network node connected with the port of the electric energy router, is the active power of the power distribution network node i, is the reactive power of the power distribution network node i, j is a virtual node representing the port of the electric energy router, is a newly added line, and the equivalent impedance of the newly added line includes two parts: is the impedance of the line connecting the power distribution network node i and the port j of the electric energy router, is the equivalent impedance of the internal loss of the electric energy router, which is used to represent the loss generated when the power is transmitted in the electric energy router; is the active power of the virtual node j of the port of the electric energy router, is the reactive power of the virtual node j of the port of the electric energy router. Figure 3 is a power distribution network model diagram containing an electric energy router, and in Figure 3 , ER is an electric energy router, each number represents a node number in the power distribution network, the connection lines of the nodes represent the physical connection relationship between the nodes, and the solid circles of 1-18 are actual nodes.Figure 4 An equivalent virtual node model diagram of the power router is shown in Figure 4 In the diagram, each number represents a node number in the power distribution network, and the connection lines between the nodes represent the physical connection relationship between the nodes. The solid circles 1-18 represent actual nodes, and the hollow circles 19-21 represent virtual nodes.

[0133] The power absorbed by the virtual node from the power grid represents the power injected by the power distribution network node into the corresponding port of the power router. The virtual node must satisfy the following constraint conditions:

[0134]

[0135] In the formula: Pj is the active power absorbed by the node of the virtual node j, Qj is the reactive power absorbed by the node of the virtual node j; Pi is the active power transmitted by the power distribution network node i to the port of the power router, Qi is the reactive power transmitted by the power distribution network node i to the port of the power router; I is the current on the new line, R is the equivalent resistance of the new line, X is the equivalent reactance of the new line; Cj is the port capacity of the virtual node j.

[0136] The bi-level programming model in S6 is constructed based on the virtual nodes. Through the bi-level programming model, a certain number of power router planning schemes are obtained, specifically:

[0137] The number and position of the power routers obtained by the upper-level planning, and the connection nodes and ports of the power routers obtained by the lower-level planning are taken as the power router planning scheme;

[0138] The upper-level planning also needs to calculate the annual deployment cost of the power routers, and the lower-level planning also needs to calculate the configuration cost of the power routers and the annual operation cost of the power distribution network.

[0139] The upper-level planning specifically includes:

[0140] The deployment number of the power routers is set, and the power routers are arranged in the candidate deployment nodes of the power routers obtained in S4 to determine the number and position of the power routers;

[0141] The annual deployment cost of the power routers is calculated, specifically as follows:

[0142]

[0143] Where: is the annual deployment cost of the power router, is the number of deployed power routers, is the unit deployment cost of the power router, is the discount rate of the power router, is the life span of the power router.

[0144] The lower-level planning in S5 specifically includes:

[0145] Allocating the capacity of the power router ports by the optimal power flow calculation, removing the power router ports with a capacity of 0, determining the nodes and capacities connected to the power router ports, and obtaining a set number of power router planning schemes;

[0146] The optimal power flow calculation model for the distribution network including the power router is as follows:

[0147] Node power balancing:

[0148]

[0149] Where: is the active power absorbed by virtual node j from the grid at time t, is the reactive power absorbed by virtual node j from the grid at time t; is the active load demand of virtual node j at time t, is the reactive load demand of virtual node j at time t; is the predicted output of distributed photovoltaic power generation of virtual node j at time t, is the abandoned power of distributed photovoltaic of virtual node j at time t, and the difference between the two represents the actual output of distributed photovoltaic at the current moment; is the active power purchased by virtual node j from the upper power grid at time t, is the reactive power purchased by virtual node j from the upper power grid at time t; is the discharge power of the energy storage device deployed at virtual node j at time t, is the charging power of the energy storage device deployed at virtual node j at time t; is the set of ports of the power router, is a collection of distribution network ports.

[0150] Line transmission balance:

[0151]

[0152] Where: active power transmitted from node j to node k at time t, reactive power transmitted from node j to node k at time t; active power loss of line (i, j) at time t, reactive power loss of line (i, j) at time t, current of line (i, j) at time t, resistance of line (i, j), reactance of line (i, j); active power transmitted by line (i, j) at time t, reactive power transmitted by line (i, j) at time t.

[0153] Node voltage balance:

[0154]

[0155] wherein: voltage of distribution network node i at time t, voltage of virtual node j at time t.

[0156] Distribution network operation constraints:

[0157]

[0158]

[0159]

[0160] wherein: capacity of line (i, j), , upper and lower limits of current transmitted by line (i, j), respectively; , upper and lower limits of voltage of node i, respectively.

[0161] Energy router operation and configuration constraints:

[0162]

[0163] wherein: capacity configuration coefficient of the energy router port corresponding to the virtual node j, unit port reference capacity of the energy router, active power absorbed by virtual node j from the power grid at time t, reactive power absorbed by virtual node j from the power grid at time t; set of ports of the i-th energy router.

[0164] The lower layer planning calculates the configuration cost of the power router and the annual operation cost of the power distribution network, specifically as follows:

[0165]

[0166] In the formula: is the configuration cost of the power router, is the unit port capacity configuration cost of the power router, is the capacity configuration coefficient of the power router port corresponding to the virtual node j, is the unit construction cost of the new line of the power router, is a binary variable of the new line of the power router, when the virtual node j capacity configuration of the power router port is not zero, it is 1, otherwise it is 0; is the total resistance of the new line of the power router, wherein the power distribution network node i is the power distribution network node connected with the virtual node j; is the discount rate of the new line, is the life span of the new line;

[0167]

[0168] In the formula: is the annual operation cost of the power distribution network, is the network loss of the power distribution network line at t time, is the network loss of the new line of the power router at t time, is the network loss unit price; is the power purchase amount of the power distribution network to the upper-level power grid at t time, is the electricity price of the power distribution network to the upper-level power grid at t time; is the distributed photovoltaic power purchase amount at t time, is the distributed photovoltaic electricity price at t time; is the abandoned power of the distributed photovoltaic at t time, is the abandoned light penalty unit price of the distributed photovoltaic at t time.

[0169] Preferably, the S2 specifically screens the planning scheme generated by the bi-level planning model through the objective function, and takes the annual total cost minimum as the final planning scheme of the power router, and the objective function is specifically as follows:

[0170]

[0171] In the formula: is the annual deployment cost of the power router, the configuration cost of the electric energy router, the annual operation cost of the power distribution network.

[0172] Embodiment 2:

[0173] A comprehensive evaluation system is established for the bi-level programming result for multi-dimensional analysis. In this embodiment, by quantifying the indicators, the economic benefit and performance of the electric energy router planning scheme are evaluated from the aspects of grid economy, grid operation performance, electric energy router use efficiency, and grid center of gravity offset.

[0174] (1) Grid economy

[0175] In the evaluation of grid economy, two indicators, cost reduction rate (CRR) and return on investment (ROI), are designed to reflect the effect of the electric energy router in reducing the operation cost of the power distribution network and its investment benefit.

[0176]

[0177] In the formula: CRR is the cost reduction rate indicator, C is the cost after planning the electric energy router, C0 is the cost without the electric energy router; ROI is the return on investment indicator, Cp is the annual deployment cost of the electric energy router, Cc is the configuration cost of the electric energy router.

[0178] (2) Grid operation performance

[0179] The grid operation performance evaluation mainly focuses on two key indicators: distributed photovoltaic consumption rate and node voltage offset rate. They respectively reflect the utilization capacity of the grid for photovoltaic power generation and the stability of the voltage at each node in the grid.

[0180]

[0181] In the formula: PVCR is the photovoltaic consumption rate, Pj,t is the predicted output of the distributed photovoltaic at virtual node j at time t, Dj,t is the discarded power of the distributed photovoltaic at virtual node j at time t, D is the set of distributed photovoltaic nodes, P is the set of ports of the power distribution network; NVR is the node voltage offset rate, Vj,t is the voltage of virtual node j at time t, V0 is the standard voltage.

[0182] (3) Electric energy router working efficiency

[0183] Load processing efficiency is used to characterize the utilization efficiency of the power router, which reflects the adaptability and operation effect of the power router under different load levels.

[0184]

[0185] Where: is the load handling efficiency of the power router, is the capacity configuration coefficient of the power router port corresponding to the virtual node j, is the active power absorbed by virtual node j from the grid at time t, is the reactive power absorbed by virtual node j from the grid at time t, is the benchmark capacity of the unit port of the power router, and T is the total operating time.

[0186] Example 3:

[0187] To verify the effectiveness of the proposed power router optimization configuration model, this embodiment is based on the IEEE-33 node system design simulation example. The IEEE-33 node system has a base power of 10 MVA and a base voltage of 12.66 kV. The IEEE-33 node system structure diagram is shown in the figure below. Figure 5 As shown, in Figure 5 In the figure, the symbol "PV" in the circle indicates the node connected to distributed photovoltaic (PV), and the battery symbol indicates the node connected to the energy storage device; the numbers 1-33 are the node numbers in the IEEE-33 node system, and the lines connecting the nodes indicate the physical connection relationship between the nodes. The simulation parameter diagram is shown in Figure 6 As shown, Figure 6 middle, is the ER configuration cost per ohm of line, is the ER efficiency coefficient, is the ER power regulation coefficient, is the ER unit port benchmark capacity, is the energy storage charging efficiency coefficient, is the energy storage discharge efficiency coefficient, is the energy storage self-loss coefficient, is the maximum energy capacity of energy storage, is the minimum energy capacity of energy storage, is the maximum charging power of energy storage, is the maximum discharge power of energy storage, is the basic cost of a single ER device, is the additional cost of a single ER port, is the unit network loss penalty cost coefficient, is the construction cost per unit photovoltaic output, is the penalty cost per unit of photovoltaic curtailment.

[0188] To verify the effectiveness of the proposed method for planning multiple energy routers in a distribution network based on time-series region partitioning, three different scenarios were set up: 1. The original scenario with no energy routers planned; 2. Planning energy routers using a two-stage energy router planning method that considers optimal distribution network operation; and 3. Optimizing the configuration of energy routers using the method of this embodiment. The effectiveness and superiority of the proposed method were evaluated by comparing distribution network performance indicators under these three scenarios.

[0189] In the IEEE-33 node system, the area division and sub-partition center of gravity identification are first carried out based on the typical daily operation data. On this basis, the power router is optimized and configured. The IEEE-33 node partition result is shown in the figure below. Figure 7 As shown, Figure 7 In the figure, the symbol "PV" in a circle indicates the node connected to distributed photovoltaic (PV), and the battery symbol indicates the node connected to the energy storage device; the solid rectangular symbol indicates the power center node in the IEEE-33 node system, and the solid triangle symbol indicates the load center node in the IEEE-33 node system. The numbers 1-33 are node numbers in the IEEE-33 node system, and the lines connecting each node indicate the physical connection relationship between the nodes. Based on the typical daily data of distributed photovoltaic output and load demand, the distribution network is divided into five areas. The load center nodes and power center nodes of each area are [7, 24, 17, 29, 20] and [9, 23, 19, 13, 32] respectively. Based on the results of the above area division and center identification, the power router is optimized. It was finally determined that two power router devices need to be deployed: the first power router is deployed at node 17 in the 4th partition, and the second power router is located at node 29 in the 3rd partition. The power router parameter configuration diagram is shown in the figure below. Figure 8 shown.

[0190] The two-stage planning method for power routers, which considers the optimal operation of the distribution network, configures the power router. A single 33-port power router is deployed at each node in the distribution network, and the optimal plan is selected based on the highest CRR. The annual cost of the power router at node 29 is the lowest, at 7,658,602.68, achieving a CRR of 10.5%. Therefore, according to the two-stage planning method, the power router will be deployed at node 29. At this point, the power router will connect to nodes 1, 11, and 29, with port capacities of 650, 350, and 300 kW, respectively.

[0191] Figure 9The cost and power consumption comparison results of the three schemes are shown in the figures, which specifically show the cost composition of the distribution network under the planning scheme of the power router proposed in the embodiment, the two-stage planning method of the power router considering the optimal operation of the distribution network, and the distribution network without the power router. In the original case without the power router, the total annual cost of the distribution network is 8553846.93 yuan, the annual power supply is 48659.55 kWh, the proportion of photovoltaic power generation and main grid power purchase is 46% and 52% respectively, and the photovoltaic consumption rate is 91.88%. After using the two-stage planning method of the power router considering the optimal operation of the distribution network, the total annual cost of the distribution network is reduced to 7658602.68 yuan, with a cost reduction rate of 10.47%, the annual power purchase is 50446.00 kWh, the proportion of photovoltaic power generation and main grid power purchase is 48.2% and 51.8% respectively, the photovoltaic consumption rate is increased to 100%, and the investment return rate of the power router is 6.44. After using the method proposed in the embodiment, the total annual cost of the distribution network is further reduced to 7644685.59 yuan, with a cost reduction rate of 10.63%, which is 1.5 percentage points higher than the two-stage planning method of the power router considering the optimal operation of the distribution network, the annual power purchase is 50333 kWh, the proportion of photovoltaic power generation and main grid power purchase is 48.3% and 51.7% respectively, the photovoltaic consumption rate remains at 100%, and the investment return rate of the power router is further increased to 6.61, which is 2.6% higher than the two-stage planning method of the power router considering the optimal operation of the distribution network.

[0192] Figure 10 The voltage deviation rate results of the three schemes are shown in the figures, which specifically show the node voltage deviation level of the distribution network at each time under the three conditions. The results show that after planning the power router under the three conditions, the voltage deviation levels of the two schemes are significantly lower than the original case without the power router. Among them, the results of the method proposed in the embodiment are better than the two-stage planning method of the power router considering the optimal operation of the distribution network in terms of node voltage deviation level at most times. The maximum voltage deviation level of the three planning schemes all occurs at 22 o'clock. Specifically, compared with the original case without the power router, the two-stage planning method of the power router considering the optimal operation of the distribution network reduces the voltage deviation level by 49%, while the method proposed in the embodiment reduces the deviation level by 63% (compared with the original case). In addition, the method proposed in the embodiment further reduces the voltage deviation level by 8% compared with the two-stage planning method of the power router considering the optimal operation of the distribution network.

[0193] In terms of the efficiency of the electric energy router, the working efficiency of the electric energy router considering the optimal operation of the power distribution network is 88.86%, and the working efficiency of the method proposed in the embodiment is 98.29%, which is 10.61% higher than that of the electric energy router considering the optimal operation of the power distribution network. The ER working efficiency diagrams of the two schemes are shown in Figure 11 The working efficiency of the electric energy router proposed in the embodiment is better than or equal to that of the electric energy router considering the optimal operation of the power distribution network at all time points. In addition, the working efficiency of the electric energy router reaches 100% in most time periods, which means that all ports of the electric energy router are transmitting power at the maximum capacity, achieving full utilization. However, the working efficiency of the electric energy router is relatively low at 8 am and 6 pm, which is 85.57% and 91.62%, respectively. Compared with the electric energy router considering the optimal operation of the power distribution network, the method proposed in the embodiment not only has higher working efficiency of the electric energy router, but also achieves longer full-load operation time in the time period from 8 am to 19 pm. In addition, the electric energy router considering the optimal operation of the power distribution network reaches the lowest working efficiency at 9 am, which is only 55.31%, and the lowest working efficiency of the method proposed in the embodiment is 54.7% higher than that of the electric energy router considering the optimal operation of the power distribution network.

[0194] The number of fixed electric energy routers deployed is 1, and the solving time of the method proposed in the embodiment and the electric energy router considering the optimal operation of the power distribution network is compared under different solving precisions. The solving time and total cost of the two schemes under different allowable errors are shown in Figure 12 The results show that the annual total cost of the two methods changes little under different solving precisions. The cost of the electric energy router considering the optimal operation of the power distribution network is slightly lower than that of the method proposed in the embodiment under the same precision, and the economy is slightly better, but the difference between the two is within 0.5%. However, there is a significant difference in solving time between the two methods. The solving time of the electric energy router considering the optimal operation of the power distribution network is significantly higher than that of the method proposed in the embodiment under each precision, and the difference between the two gradually increases as the precision improves. Under 1% error, the solving time of the electric energy router considering the optimal operation of the power distribution network is 48 times that of the method proposed in the embodiment; when the error is reduced to 0.01%, the solving time is 100 times that of the method proposed in the embodiment.

[0195] The huge difference in solving time is due to the difference in variable scale between the two methods. For a power system containing n nodes, the two-stage planning method of the power router considering the optimal operation of the distribution network uses an n-port power router, which is deployed at each node in turn, and determines the final scheme by economic optimization. This method introduces multiple n-dimensional discrete variables or binary variables in the planning model to represent the state of the power router port, and needs to be traversed n times to solve the optimal result. Due to the involvement of a large number of discontinuous variables, the solving time is significantly increased. As the size of the power grid increases, the value of n rises, and the variable scale of a single solution and the total number of cycles are significantly increased, leading to a sharp increase in overall time consumption.

[0196] Unlike the two-stage planning method of the power router considering the optimal operation of the distribution network, the parameter setting of the power router in the present embodiment is not completely dependent on the size of the power system. Specifically, the present embodiment first divides the power grid into several regions and identifies the gravity nodes, and then deploys and configures the parameters of the power router with the gravity nodes as the target. Taking the IEEE-33 node system as an example, after regional division, 5 sub-zones and 10 gravity nodes are obtained. At this time, the planning of a single power router only needs to traverse and solve a 5-port power router 5 times, significantly reducing the scale of the single solving problem and the total number of cycles, and having less dependence on the size of the power system nodes. Although the present embodiment method has a slight gap in economy compared with the complete traversal scheme of the two-stage planning method of the power router considering the optimal operation of the distribution network, the improvement in solving speed is extremely significant.

Claims

1. A power distribution network multi-energy router planning method based on timing area division, characterized in that, The application is applied to an electric energy router, comprising: A simplified electric energy router model is constructed, including equivalent internal loss of the electric energy router to a virtual external line connected between a port of the electric energy router and a virtual node representing the port of the electric energy router; A lower planning in a bi-level planning model is adopted to determine capacity configuration of the electric energy router connection node and port through optimal power flow calculation based on the simplified electric energy router model; The electric energy router is applied to the bi-level planning model, comprising: S1, based on the candidate deployment node of the electric energy router, using the upper planning in the bi-level planning model, determining the position and quantity of the electric energy router, combining the capacity configuration of the electric energy router connection node and port determined by the lower planning, obtaining a certain number of electric energy router planning schemes; S2, taking the minimum total annual cost as the objective function, screening the electric energy router planning scheme to obtain the final electric energy router planning scheme, wherein the total annual cost includes annual deployment cost of the electric energy router, configuration cost of the electric energy router and annual operation cost of the distribution network; The annual deployment cost of the electric energy router in the total annual cost is calculated by the upper planning, specifically as follows: wherein: annual deployment cost of the electrical energy router, number of deployments of the electrical energy router, unit deployment cost of the electrical energy router, discount rate of the electrical energy router, lifetime of the electrical energy router; The configuration cost of the electric energy router and the annual operation cost of the distribution network in the total annual cost are calculated by the lower planning, specifically as follows: In the formula: is the configuration cost of the power router, is the unit port capacity configuration cost of the power router, is the capacity configuration coefficient of the power router port corresponding to the virtual node j, is the unit construction cost of the new line of the power router, is a binary variable of the power router on the new line, which is 1 when the virtual node j capacity configuration of the power router port is not zero, otherwise it is 0; is the total resistance of the new line of the power router, wherein the power grid node i is a power grid node connected with the virtual node j; is the discount rate of the new line, is the life span of the new line; In the formula: is the annual operation cost of the power distribution network, is the network loss of the power distribution network line at time t, is the network loss of the new line of the power energy router at time t, is the network loss unit price; is the electricity purchase quantity of the power distribution network from the upper-level power grid at time t, is the electricity price of the power distribution network from the upper-level power grid at time t; is the distributed photovoltaic electricity purchase quantity at time t, is the distributed photovoltaic electricity price at time t; is the abandoned power of the distributed photovoltaic at time t, is the abandoned light penalty unit price of the distributed photovoltaic at time t.

2. The power distribution network multi-energy router planning method based on timing area division of claim 1, wherein, The optimal power flow calculation adopted by the lower planning is specifically to allocate the capacity of the electric energy router port, remove the electric energy router port with the capacity of 0, and determine the capacity configuration of the node and port connected to the electric energy router port.

3. The power distribution network multi-energy router planning method based on timing area partitioning of claim 2, wherein, The optimal power flow calculation is based on a distribution network optimal power flow model containing the electric energy router, comprising: Node power balance: wherein: Pj,t is the active power absorbed from the grid by virtual node j at time t, Qj,t is the reactive power absorbed from the grid by virtual node j at time t; Pj,t is the active power absorbed from the grid by virtual node j at time t, Qj,t is the reactive power absorbed from the grid by virtual node j at time t; Pj,t is the active power absorbed from the grid by virtual node j at time t, Pj,t is the active power absorbed from the grid by virtual node j at time t, Pj,t is the active power absorbed from the grid by virtual node j at time t, Qj,t is the reactive power absorbed from the grid by virtual node j at time t; Pj,t is the active power absorbed from the grid by virtual node j at time t, Pj,t is the active power absorbed from the grid by virtual node j at time t; Pj,t is the active power absorbed from the grid by virtual node j at time t, Pj,t is the active power absorbed from the grid by virtual node j at time t, Line transmission balance: wherein: Pij(t) is the active power transmitted from node j to node k at time t, Qij(t) is the reactive power transmitted from node j to node k at time t; Pij(t) is the active power transmitted from node j to node k at time t, Qij(t) is the reactive power transmitted from node j to node k at time t; Iij(t) is the current of line (i,j) at time t, Rij is the resistance of line (i,j), Xij is the reactance value of line (i,j); Pij(t) is the active power transmitted from node j to node k at time t, Qij(t) is the reactive power transmitted from node j to node k at time t; Node voltage balance: In the formula: is the voltage of the power distribution network node i at time t, is the voltage of the virtual node j at time t; Distribution network operation constraint: wherein: Ci,jis the capacity of line (i,j), , are respectively the upper and lower limits of the current transmitted by line (i,j); , are respectively the upper and lower limits of the voltage at node i; Electric energy router operation and configuration constraint: wherein: is a capacity configuration coefficient of the power router port corresponding to the virtual node j, is a unit port reference capacity of the power router, is the active power absorbed from the grid by the virtual node j at time t, is the reactive power absorbed from the grid by the virtual node j at time t; is the set of ports of the i-th power router.

4. The power distribution network multi-energy router planning method based on timing area partitioning of claim 1, wherein, The S1 further comprises: Step A, dividing a typical daily operation cycle of the distribution network into a certain number of snapshots, and abstracting the distribution network into a time sequence network model, which uses a coupling matrix to represent the connectivity between different nodes in the time sequence network model and the electrical distance information within the snapshot; Step B, obtaining the coupling matrix corresponding to the snapshot through the coupling matrix power flow weight and the coupling matrix voltage weight of the time sequence network model; Step C, performing time sequence regional division on the distribution network based on the snapshot and the corresponding coupling matrix to obtain sub-zones and complete zones; Step D, obtaining the centralized distribution relationship of power supply and load in the sub-zone based on source and load gravity center identification analysis, and determining the candidate deployment node of the electric energy router in combination with the complete zone and the centralized distribution relationship.

5. The power distribution network multi-energy router planning method based on timing area partitioning of claim 4, wherein, The time sequence regional division on the distribution network based on the snapshot and the corresponding coupling matrix in the Step C specifically comprises two steps of static local optimization and time sequence incremental update; The static local optimization is specifically to divide the time series area based on the local adaptive function to obtain a set number of sub-partitions, and to evaluate the strength of the sub-partitions using the modularity function to obtain the optimal initial partition, that is, the snapshot at the initial time. Partition results within ; The time sequence incremental update is specifically adjusting the belonging relationship of the sub-partitions and boundary nodes by analyzing the changes of nodes and edges between the snapshots based on the optimal initial partition to obtain the complete partition.

6. The power distribution network multi-energy router planning method based on timing area partitioning of claim 4, wherein, The source load center identification analysis in the StepD is specifically identifying and analyzing the power supply center and the load center in the sub-partition. According to the size relationship between the center distance of the power supply center and the center distance of the load center, the concentrated distribution relationship of the power supply and the load in the sub-partition is judged, and the concentrated distribution relationship includes two kinds, which are specifically as follows: The load center is ahead of the power supply center, and the power supply node power in the sub-partition is first reversely transmitted to the load center and then reversely transmitted to the first node of the sub-partition. The power supply center is ahead of the load center, and the power reverse transmission exists in the power supply center to the first node of the sub-partition.

7. The power distribution network multi-energy router planning method based on timing area partitioning of claim 1, wherein, The S2 specifically screens the planning scheme generated by the double-layer planning model through the target function, takes the planning scheme with the minimum total annual cost as the final planning scheme of the electric energy router, and the target function is specifically as follows: wherein: Cannualdeploymentis the annual deployment cost of the electrical energy router, Cconfigis the configuration cost of the electrical energy router, Cannualoperationis the annual operation cost of the electrical energy router.

Images