Power distribution network partition coordinated optimization method considering on-site consumption of new energy and flexible regulation and control of voltage

Abstract

The invention provides a power distribution network partition coordinated optimization method considering on-site consumption of new energy and flexible regulation and control of voltage, which comprises the following steps of: firstly, introducing comprehensive indexes of a source load mismatching degree considering 5G base station access, i.e., a maximum net load rate and a station area minimum net load, screening transformer heavy load and a DPV output surplus station area, and then, carrying out power distribution network partition coordinated optimization on the transformer heavy load and the DPV output surplus station area; the method comprises the steps of calculating voltage-power sensitivity of a power distribution network according to a load flow calculation result, defining a dominant node selection basis, screening voltage dominant nodes of the power distribution network, then optimizing a partition scheme by adopting a genetic algorithm, dividing the power distribution network into a plurality of associated physical sub-regions, and finally establishing a power distribution network partition coordinated optimization model containing a 5G base station and a DPV. And carrying out distributed solution by adopting an SADMM algorithm to realize partition coordinated optimization operation of the power distribution network. The method can effectively solve the problem that the voltage treatment capability in the region is limited, and improves the new energy on-site absorption capability of the power distribution network and the voltage regulation and control flexibility of the power distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid control, and particularly to a method for coordinated optimization of distribution network zoning considering local consumption of new energy and flexible voltage regulation. Background Technique

[0002] With the rapid advancement of the "dual carbon" goal, the distribution network will connect a large number of multi-type sources and loads such as distributed photovoltaics (DPVs) and 5G base stations. High-penetration DPVs and high-energy-consuming devices will be a major feature of the new power system. However, the high-dimensional access of DPVs and 5G base stations to the distribution network will seriously increase the computing burden on the central processor of the distribution network, bringing new challenges to aspects such as economic dispatch and power quality of the distribution network. Conducting research on the method for coordinated optimization of distribution network zoning is of great significance for aspects such as local consumption of new energy and voltage quality governance.

[0003] There are numerous DPVs and 5G base stations, but it is difficult to make full use of them. The difficulty lies in that the increase in the amount of decision-making is likely to cause problems such as large centralized computing burden and poor reliability. Therefore, it is necessary to flexibly and effectively control the controllable resources in the distribution network through a suitable distributed coordinated optimization method. However, the existing distributed active and reactive power coordinated optimization model for the distribution network does not refer to the distribution network indicators regarding active and reactive power, making it difficult to fully regulate resources, unable to ensure that each zone has sufficient reserves to respond to reactive power compensation, and the voltage governance ability of each region is limited, resulting in large voltage deviations. At the same time, the existing research does not consider the source-load mismatch problem brought by high-energy-consuming base station equipment in the controllable resources, which is likely to increase the power loss of the transformers in each distribution substation area. Therefore, the distributed coordinated optimization method for the distribution network needs to introduce the source-load mismatch degree as a zoning indicator to ensure the safe and stable operation of the distribution network. Summary of the Invention

[0004] To solve the technical problems that the traditional distributed coordinated optimization method for the distribution network is difficult to fully regulate resources and has poor voltage deviation governance ability, the present invention provides a method for coordinated optimization of distribution network zoning considering local consumption of new energy and flexible voltage regulation.

[0005] To solve the above technical problems, the present invention adopts the following technical method: A method for coordinated optimization of distribution network zoning considering local consumption of new energy and flexible voltage regulation, including:

[0006] Step S1, taking the comprehensive index of source-load mismatch degree and the comprehensive index of voltage-power sensitivity as the comprehensive evaluation index for regional division, constructing an optimal zoning model for the distribution network that comprehensively considers local consumption of new energy and flexible voltage regulation, using the genetic algorithm to perform optimization solution on the optimal zoning model for the distribution network, obtaining the distribution network regional division result, and dividing the distribution network into several sub-regions;

[0007] Step S2: According to the distribution network area division result obtained in Step S1, a distributed power optimization control coupling model is constructed with the state variables of the coupling branches in adjacent areas of the distribution network being equal as the constraint.

[0008] Step S3: An optimal distribution network loss and average voltage deviation are used as the objective function to construct a distribution network partition coordination optimization model including 5G base stations and distributed photovoltaics. The constraint conditions of this distribution network partition coordination optimization model include: power system power flow constraint, power security constraint, grouped switched capacitor bank constraint, static var compensator constraint, 5G base station constraint, distributed photovoltaic constraint, regional power balance constraint, and partition regulation constraint; among them, the partition regulation constraint is determined according to the distributed power optimization control coupling model.

[0009] Step S4: The SADMM algorithm is used to perform distributed solution on the coordination optimization model corresponding to each sub-region of the distribution network, and the real-time operating power and voltage distribution of the distribution network are obtained to achieve the partition coordination optimization operation of the distribution network.

[0010] Furthermore, in Step S1, the process of constructing the optimal distribution network partition model includes:

[0011] S101: The maximum load rate of the substation area transformer and the minimum net load are introduced to determine the comprehensive index χ of the source-load mismatch 1 , as follows:

[0012]

[0013] χ 1 = max{ω T T load,i + ω P P pure,i} (2)

[0014] In the formula: P load,i,t , Q load,i,t are the active and reactive power loads of substation area i at time t respectively; P DPV,i,t is the actual output of the photovoltaic power station installed in substation area i at time t; P 5G.BS,i,t is the total load of 5G base stations in substation area i at time t; S T is the capacity of each substation area transformer; T load,i , P pure,i represent the maximum load rate and minimum net load of the transformer in substation area i respectively. When T load,i > 0.8, it means that there is a heavy load situation of the transformer in substation area i, which is classified into the set Φ 11 , otherwise it is classified into the set Φ 12 . When P pure,i < 0, it means that there is an excessive output situation of distributed photovoltaics in substation area i, which is classified into the set Φ21 , otherwise it is classified into the set Φ 22 ; ω T , ω P are the weight coefficients of the maximum load rate of the substation area transformer and the minimum net load of the substation area respectively;

[0015] S102, determine the basis for selecting the dominant nodes of the distribution network;

[0016] First, determine the voltage - active power sensitivity and voltage - reactive power sensitivity of each branch of the distribution network;

[0017]

[0018] In the formula: S U-P,ij is the voltage - active power sensitivity of branch ij; l represents branch ij; C 0,i represents the set of line nodes from node 0 to node i; r ij is the resistance of the distribution network line; V j is the voltage of node j; V i is the voltage of node i; P j is the active power flowing out of node j; P k is the active power flowing through branch k; P T,loss is the active power loss of the transformer; S U-Q,ij is the voltage - reactive power sensitivity of branch ij; x ij is the reactance of the distribution network line; Q j , Q k are the reactive powers of node j and node k respectively; Q T,loss is the reactive power loss of the transformer;

[0019] Then define the observability index and controllability index of the dominant nodes;

[0020]

[0021] In the formula: i is the number of the dominant node; is the set of all nodes in the distribution network; j is the number of other nodes in the area; O i is the observability index of the dominant node i; S U-P,ij / S U-P,jj represents the influence degree of the dominant node voltage on the active voltage of other nodes in the area; S U-Q,ij / S U-Q,jj represents the influence degree of the dominant node voltage on the reactive voltage of other nodes in the area; M i is the controllability index of the dominant node i;

[0022] Finally, determine the basis for selecting the dominant nodes of the distribution network;

[0023] χ 2= max{ω o O i + ω M M i} (7)

[0024] Where: χ 2 is the comprehensive index of voltage - power sensitivity; ω o , ω M are the observability and controllability weight coefficients;

[0025] S103. Taking the comprehensive index of source - load mismatch and the comprehensive index of voltage - power sensitivity as the comprehensive evaluation index for regional division, a distribution network optimization zoning model that comprehensively considers new - energy local consumption and voltage flexible regulation is constructed as follows:

[0026] F = ω 1 χ 1 + ω 2 χ 2 (8)

[0027] Where: F represents the objective function of the distribution network optimization zoning model; ω 1 , ω 2 are the weight coefficients corresponding to each index, and ω 1 + ω 2 = 1.

[0028] Furthermore, the distributed power optimization control coupling model is as follows:

[0029]

[0030] X a,ij :={P a,ij , Q a,ij , U a,i , U a,j}(10)

[0031] X a,ij = X b,ij , e ij ∈ O(11)

[0032] Where: O represents the set of coupling branches of the distribution network; e ij represents the coupling branch between adjacent regions of the distribution network; E a , E b represent the sets of branches in regions a and b respectively, and regions a and b are adjacent; R represents the set of physically connected sub - regions into which the distribution network is divided; X a,ij , X b,ij represent the state variables of the coupling branches in regions a and b respectively; P a,ij represents the active power transmitted by the coupling branch in region a; Q a,ijReactive power transmitted by the coupling branch in area a; U a,i , U a,j respectively represent the nodal voltages at nodes i and j at both ends of the coupling branch in area a.

[0033] Furthermore, the objective function of the distribution network zoning coordination optimization model is as follows:

[0034] min f = λ LIN f LIN + λ SYN f SYN (12)

[0035]

[0036] γ LIN + γ SYN = 1 (15)

[0037]

[0038] In the formula: min f represents the objective function of the distribution network zoning coordination optimization model; f LIN , f SYN are respectively the distribution network power loss and the average voltage deviation; λ LIN , λ SYN are respectively the weight factors of the distribution network power loss and the average voltage deviation; γ LIN , γ SYN are respectively the weight coefficients of the distribution network power loss and the average voltage deviation; are respectively the initial values of the distribution network power loss and the average voltage deviation before optimization; c LIN,t is the unit power loss price at time t; I ij is the branch current between nodes i and j; Δt is the duration of a single time period; N bus is the number of nodes in the distribution network; V N is the reference voltage; V i,t is the voltage of node j at time t; T is an optimization period;

[0039] Preferably, the power system power flow constraint is as follows:

[0040]

[0041] In the formula, p j , q j are respectively the active and reactive injection powers of node j; P jk , Q jk are respectively the active and reactive powers flowing from node j to the next node k; P ij , Q ij are respectively the active and reactive powers flowing into node j from the previous node i; Iij is the branch current between node i and node j; g j , b j are the ground conductance and susceptance of node j, which are constants; V i , V j are the voltages of nodes i and j respectively;

[0042] The power security constraints are as follows:

[0043]

[0044] In the formula, I ij.max , I ij.min are the upper and lower limits of the current of the branch between nodes i and j respectively; are the interactive active and reactive powers with the superior power grid respectively; are the maximum and minimum interactive active powers allowed to pass through the connection branch between the distribution network and the superior power grid respectively; are the maximum and minimum reactive interactive powers allowed to pass through the connection branch between the distribution network and the superior power grid respectively; V j.max , V j.min are the upper and lower limits of the voltage of node j respectively; is the square of the branch current between nodes i and j; is the square of the voltage of node j;

[0045] The constraints for the switched capacitor banks are as follows:

[0046]

[0047] In the formula: y CB,j,t is the number of switched capacitor banks in operation, which is a discrete variable value; Y CB,max,j is the upper limit of the number of capacitor banks connected to node j; Q CB,step,j is the compensation power of each capacitor bank, which is a constant; N CB,max,j is the upper limit of the number of operations; Q CB,j,t is the reactive power of the switched capacitor banks; t and T are an optimization time period and an optimization cycle respectively;

[0048] The constraints for the static var compensator are as follows:

[0049] Q SVG,min,j ≤Q SVG,j ≤Q SVG,max,j (20)

[0050] In the formula: Q SVG,j is the reactive power of the static var compensator; Q SVG,min,j , Q SVG,max,j are the lower and upper limits of the reactive power of the static var compensator respectively;

[0051] The 5G base station constraints include:

[0052] 1) 5G base station load model constraint:

[0053]

[0054] P D,i,t = βP D,max (22)

[0055] In the formula, P 5G.BS,t is the total load of the 5G base station at time t; ε represents the working state identifier of the 5G base station. When ε is 1, the 5G base station is in the active state. When it is 0, the 5G base station is in the sleep state; P ACT,i,t and P SLE,i,t are the active state load and sleep state load of the i-th 5G base station at time t respectively; P S,i,t is the static load of the i-th 5G base station at time t; P D,i,t represents the communication load of the i-th 5G base station at time t, which is a dynamic load; α is the load scale factor of the i-th 5G base station; β is the coefficient reflecting the communication data of mobile users; P D,max represents the predicted maximum value of the dynamic load of the 5G base station, which is the radio frequency output power corresponding to the active antenna unit in the 5G base station communication device when the communication load of the mobile user is predicted;

[0056] 2) 5G base station backup energy storage safety power backup capacity model constraint:

[0057]

[0058] In the formula: E REM,t is the safety power backup capacity required by the 5G base station in the t time period; T res,min is the shortest power backup time of the 5G base station;

[0059] 3) 5G base station backup energy storage power control model constraint:

[0060]

[0061]

[0062] In the formula, represents or is the active power charging identifier of the 5G base station backup energy storage, which is a variable of 0 or 1. When the identifier is 0, no charging is performed. When the identifier is 1, charging is performed; is the active power discharging identifier of the 5G base station backup energy storage, which is a variable of 0 or 1. When the identifier is 0, no discharging is performed. When the identifier is 1, discharging is performed; P t5G,ch / dis The active power absorbed or released by the energy storage battery at time t is; denoted as P t 5G,ch / dis the minimum value of; is P t 5G,ch / dis the maximum value of; is the actual operating capacity of the 5G base station's backup energy storage during period t; δ is the self-discharge rate of the 5G base station's backup energy storage; are respectively the charging and discharging efficiencies of the 5G base station's backup energy storage; are respectively the maximum and minimum actual operating capacities of the 5G base station's backup energy storage; are respectively the actual operating capacities of the 5G base station's backup energy storage at the initial and final periods of the scheduling cycle; P inn,t is the actual charging and discharging power of the 5G base station's energy storage during period t;

[0063] 4) Constraints of the 5G base station's backup energy storage SOC model:

[0064]

[0065] In the formula: are respectively the SOC of the 5G base station's backup energy storage during period t and its maximum and minimum values;

[0066] The constraints of the distributed photovoltaic are as follows:

[0067] P DPV,j,t = P DPV,PRE,j,t (29)

[0068] In the formula, P DPV,j,t , P DPV,PRE,j,t are respectively the actual and predicted power outputs of the photovoltaic power station installed at node j during the t-th period;

[0069] The power balance constraints in the area are as follows:

[0070]

[0071] In the formula, P IN , Q IN are respectively the sum of the active power and the sum of the reactive power injected into each node of the distribution network; P load , Q load are respectively the total active load and the total reactive load; P 5G,BS represents the total load of the 5G base stations; P GRI represents the total amount of active power purchased by the distribution network, which is obtained by adding the P t grid in each period; Q GRI is the total amount of reactive power purchased by the distribution network, which is obtained by adding the in each period; P5G is the total charge and discharge amount of the backup energy storage battery for the 5G base station, which is obtained by adding the active power P absorbed in each period t 5G,ch and the active power P released t 5G,dis ; P DPV and Q CB and Q SVG are respectively the total output of DPV, the total reactive power of the switched capacitor banks in groups, and the total reactive power of the static var compensator;

[0072] The partition regulation constraints are as shown in formulas (9) to (11).

[0073] The distribution network partition coordination optimization method considering new energy local consumption and voltage flexible regulation proposed by the present invention. First, an integrated index of source-load mismatch considering the access of 5G base stations, namely the maximum net load rate and the minimum net load of the substation area, is introduced to screen the overloaded transformers and the areas with excessive DPV output. Then, according to the power flow calculation results, the voltage-power sensitivity of the distribution network is calculated, and the basis for selecting the dominant nodes is defined to screen the voltage dominant nodes of the distribution network. Next, the genetic algorithm is used to optimize the partition scheme, dividing the distribution network into multiple connected physical sub-regions. Finally, a distribution network partition coordination optimization model including 5G base stations and DPV is established, and the SADMM algorithm is used for distributed solution to realize the partition coordination optimization operation of the distribution network. Based on the integrated index of voltage-power sensitivity, the present invention introduces an integrated index of source-load mismatch, avoiding the phenomenon of no adjustable resources in the region caused by single-index partitioning, solving the problem of limited voltage governance ability in the region, and improving the new energy local consumption ability of the distribution network and the flexibility of distribution network voltage regulation. Brief Description of the Drawings

[0074] Figure 1 is the flow chart of the distribution network partition coordination optimization method considering new energy local consumption and voltage flexible regulation proposed by the present invention;

[0075] Figure 2 is the flow chart for solving the distribution network partition coordination optimization model proposed by the present invention;

[0076] Figure 3 is the topology diagram of the IEEE 33-node distribution network in the embodiment of the present invention;

[0077] Figure 4 is the schematic diagram of the load and DPV output prediction data in the embodiment of the present invention;

[0078] Figure 5 is the distribution diagram of the minimum net load of the substation area in the embodiment of the present invention;

[0079] Figure 6It is the distribution diagram of the maximum net load rate of the substation area transformer in the embodiment of the present invention;

[0080] Figure 7 It is the distribution diagram of the voltage-power sensitivity matrix in the embodiment of the present invention;

[0081] Figure 8 It is the schematic diagram of the distribution area of the distribution network obtained by using the method of the present invention in the embodiment of the present invention;

[0082] Figure 9 It is the schematic diagram of the distribution area of the distribution network obtained by using the traditional method in the embodiment of the present invention;

[0083] Figure 10 It is the schematic diagram of the action plan of the reactive power compensation device obtained by using the method of the present invention in the embodiment of the present invention;

[0084] Figure 11 It is the schematic diagram of the node voltage distribution in Scenario 1 in the embodiment of the present invention;

[0085] Figure 12 It is the schematic diagram of the node voltage distribution in Scenario 2 in the embodiment of the present invention. Specific embodiments

[0086] For the convenience of understanding by those skilled in the art, the present invention will be further described below in conjunction with the embodiments and the accompanying drawings. The content mentioned in the embodiments does not limit the present invention.

[0087] As Figure 1 shown, a distribution network zoning coordination optimization method considering local consumption of new energy and flexible voltage regulation mainly includes the following steps.

[0088] Step S1, construct and solve an optimal distribution network zoning model that comprehensively considers local consumption of new energy and flexible voltage regulation.

[0089] S101, introduce a source-load mismatch index considering the access of 5G base stations, that is, the maximum load rate of the substation area transformer and the minimum net load of the substation area, and screen the overloaded transformers and the areas with excessive DPV output. Specifically, as shown in formula (1):

[0090]

[0091] In the formula: P load,i,t , Q load,i,t are the active and reactive power loads of substation area i at time t respectively; P DPV,i,t is the actual output of the photovoltaic power station installed in substation area i at time t; P 5G.BS,i,t is the total load of 5G base stations in substation area i at time t; S T is the capacity of each substation area transformer; T load,i , P pure,irespectively represent the maximum load rate and minimum net load of the transformer in substation area i. When T load,i > 0.8, it indicates that there is a heavy load situation of the transformer in substation area i, which is classified into the set Φ 11 , otherwise it is classified into the set Φ 12 , when P pure,i < 0, it indicates that there is an excess distributed photovoltaic power output situation in substation area i, which is classified into the set Φ 21 , otherwise it is classified into the set Φ 22 . The substations with large differences between source and load can be screened according to the above 4 sets.

[0092] Define the comprehensive index χ of the source-load mismatch degree 1 as:

[0093] χ 1 = max{ω T T load,i + ω P P pure,i}(2)

[0094] In the formula: ω T , ω P are the weight coefficients of the maximum load rate of the transformer in the substation area and the minimum net load of the substation area respectively. In this embodiment, ω T = ω P = 0.5.

[0095] S102. Determine the basis for selecting the dominant nodes of the distribution network.

[0096] First, determine the voltage-active power sensitivity and voltage-reactive power sensitivity of each branch of the distribution network;

[0097]

[0098] In the formula: S U-P,ij is the voltage-active power sensitivity of branch ij; l represents branch ij; C 0,i represents the set of line nodes from node 0 to node i; r ij is the resistance of the distribution network line; V j is the voltage of node j; V i is the voltage of node i; P j is the active power flowing out of node j; P k is the active power flowing through branch k; P T,loss is the active power loss of the transformer; S U-Q,ij is the voltage-reactive power sensitivity of branch ij; x ij is the reactance of the distribution network line; Q j , Q k are the reactive powers of node j and node k respectively; Q T,loss is the reactive power loss of the transformer.

[0099] Then, based on the influence of the dominant node voltage on the voltages of other nodes in the partition, the observability index of the dominant node is defined;

[0100]

[0101] In the formula: i is the number of the dominant node; is the set of all nodes in the distribution network; j is the number of other nodes in the region; O i is the observability index of the dominant node i; S U-P,ij / S U-P,jj represents the degree of influence of the dominant node voltage on the active voltage of other nodes in the region; S U-Q,ij / S U-Q,jj represents the degree of influence of the dominant node voltage on the reactive voltage of other nodes in the region.

[0102] Next, the controllability represents the influence of the change in the dominant node power on the voltages of other nodes in the partition, and the voltage level of the entire partition is controlled by controlling the dominant node power, and the controllability index is defined;

[0103]

[0104] In the formula: M i is the controllability index of the dominant node i.

[0105] Finally, the basis for selecting the dominant nodes of the distribution network is determined;

[0106] χ 2 = max{ω o O i + ω M M i} (7)

[0107] In the formula: χ 2 is the comprehensive voltage-power sensitivity index; ω o , ω M are the observability and controllability weight coefficients. In this embodiment, ω o = ω M = 0.5.

[0108] S103. Construct an optimal partitioning model for the distribution network.

[0109] Taking into account the local consumption of new energy and the flexible voltage regulation, the comprehensive index of source-load mismatch and the comprehensive voltage-power sensitivity index are used as the comprehensive evaluation indexes for regional division, and an optimal partitioning model for the distribution network that comprehensively considers the local consumption of new energy and the flexible voltage regulation is constructed. The objective function of this model is as follows:

[0110] F = ω 1 χ 1 + ω2 χ 2 (8)

[0111] In the formula: F represents the objective function of the optimal partition model of the distribution network; ω 1 , ω 2 are the weight coefficients corresponding to each index, ω 1 +ω 2 = 1. By adjusting the weight coefficients, partition results for different objectives can be obtained. Since the present invention is faced with a distribution network with a high proportion of 5G base stations and DPVs connected, the phenomenon of source-load mismatch is prominent. Therefore, ω 1 can be set to be relatively large. In this embodiment, ω 1 is taken as 0.65.

[0112] S104. Solve the optimal partition model of the distribution network using a genetic algorithm, and divide the distribution network into several connected physical sub-regions.

[0113] Taking the comprehensive evaluation index for the distribution network area division (i.e., formula (8)) as the fitness function, restricting the upper limit of the number of distribution network partitions, considering the electrical distance characteristics between nodes, quantifying the similarity of the operating characteristics between nodes through the node similarity calculation formula, and performing genetic algorithm-based partition optimization. Among them, the concept of the complex network modularity function is introduced during the partition process, and the number of partitions and the specific nodes included in each partition are obtained by optimizing the modularity function. Since using a genetic algorithm for distribution network partition optimization is a conventional technique in the art, it will not be elaborated here.

[0114] Step S2. According to the distribution network area division result obtained in step S1, construct a distributed power optimization control coupling model with the constraint that the state variables of the coupling branches of adjacent areas of the distribution network are equal.

[0115] Each area is equipped with a terminal controller. Each terminal controller only measures data for the area it controls and only collects the boundary coordination information of adjacent terminal controllers. The coupling part between adjacent areas is the branch on the area boundary, which is called the coupling branch e ij . The set of coupling branches is represented by formula (9) as follows:

[0116]

[0117] In the formula: O represents the set of distribution network coupling branches; e ij represents the coupling branch of adjacent areas of the distribution network; E a , E b represent the branch sets of areas a and b respectively, and areas a and b are adjacent; R represents the set of connected physical sub-regions into which the distribution network is divided.

[0118] The coupling branch e ijThe state variables include the power transmitted by the branch and the square of the node voltage. The state variables of the coupled branch in area a are expressed by Equation (10) as follows:

[0119] X a,ij :={P a,ij ,Q a,ij ,U a,i ,U a,j}(10)

[0120] Where: X a,ij represents the state variables of the coupled branch in area a, P a,ij represents the active power transmitted by the coupled branch in area a; Q a,ij represents the reactive power transmitted by the coupled branch in area a; U a,i ,U a,j represent the node voltages of nodes i and j at both ends of the coupled branch in area a, respectively.

[0121] In order to make the problem of sub-region solution equivalent to the original problem, the state X a,ij of the coupled branch obtained from the sub-problem in area a and the state X b,ij of the coupled branch obtained from the sub-problem in the adjacent area b must be equal. The coupling constraint of the distributed power optimization control based on partition coordination is shown in Equation (11) as follows:

[0122] X a,ij =X b,ij ,e ij ∈O(11)

[0123] Where: X b,ij represents the state variables of the coupled branch in area b.

[0124] Step S3, construct a partition coordination optimization model for the distribution network containing 5G base stations and distributed photovoltaics.

[0125] The purpose of partition coordination optimization is to optimize the comprehensive voltage level of the distribution network. The comprehensive voltage level of the distribution network is evaluated by the voltage reference deviation. Based on this, a partition coordination optimization model for the distribution network containing 5G base stations and distributed photovoltaics is constructed with the goal of optimizing the distribution network power loss and average voltage deviation.

[0126] The objective function of this distribution network partition coordination optimization model is as follows:

[0127] min f = λ LIN f LIN +λ SYN f SYN (12)

[0128]

[0129] γ LIN+γ SYN = 1 (15)

[0130] Where: minf represents the objective function of the distribution network partition coordination optimization model; f LIN and f SYN are the power loss and average voltage deviation of the distribution network respectively; λ LIN and λ SYN are the weight factors of the power loss and average voltage deviation of the distribution network respectively; γ LIN and γ SYN are the weight coefficients of the power loss and average voltage deviation of the distribution network respectively; are the initial values of the power loss and average voltage deviation of the distribution network before optimization, as follows:

[0131]

[0132] Where: c LIN,t is the unit power loss price at time t; I ij is the branch current between node i and node j; Δt is the duration of a single time period; N bus is the number of nodes in the distribution network; V N is the reference voltage; V i,t is the voltage of node j at time t; T is an optimization period.

[0133] The constraint conditions of the distribution network partition coordination optimization model include: power system power flow constraint, power security constraint, grouped switched capacitor bank constraint, static var compensator constraint, 5G base station constraint, distributed photovoltaic constraint, regional power balance constraint, partition regulation constraint.

[0134] ① The power system power flow constraint is as follows:

[0135]

[0136] Where, p j and q j are the active and reactive injection powers of node j respectively; P jk and Q jk are the active and reactive powers flowing from node j to the next node k respectively; P ij and Q ij are the active and reactive powers flowing into node j from the previous node i respectively; I ij is the branch current between node i and node j; g j and b j are the ground conductance and susceptance of node j, which are constants; V i and V j are the voltages of nodes i and j respectively.

[0137] ② The power security constraint is as follows:

[0138]

[0139] Wherein, I ij.max and I ij.min are respectively the upper and lower limits of the current of the branch between nodes i and j; are respectively the active and reactive powers of the interaction with the superior power grid; are respectively the maximum and minimum active powers of the interaction allowed to pass through the connection branch between the distribution network and the superior power grid; are respectively the maximum and minimum reactive powers of the interaction allowed to pass through the connection branch between the distribution network and the superior power grid; V j.max and V j.min are respectively the upper and lower limits of the voltage of node j; is the square of the current of the branch between nodes i and j; is the square of the voltage of node j.

[0140] ③ The constraints for switching capacitor banks (CB) in groups are as follows:

[0141]

[0142] Wherein: y CB,j,t is the number of capacitor banks in operation, which is a discrete variable value; Y CB,max,j is the upper limit of the number of capacitor banks connected to node j; Q CB,step,j is the compensation power of each capacitor bank, which is a constant; N CB,max,j is the upper limit of the number of operations; Q CB,j,t is the reactive power of the switched capacitor banks in groups; t and T are an optimization time period and an optimization cycle respectively.

[0143] ④ The constraints for static var generators (SVG) are as follows:

[0144] Q SVG,min,j ≤Q SVG,j ≤Q SVG,max,j (20)

[0145] Wherein: Q SVG,j is the reactive power of the static var generator; Q SVG,min,j and Q SVG,max,j are respectively the lower and upper limits of the reactive power of the static var generator.

[0146] ⑤ Constraints for 5G base stations

[0147] 1) The constraint of the 5G base station load model refers to the linear equation of the 5G base station load, expressed as:

[0148]

[0149] P D,i,t = βP D,max (22)

[0150] In the formula, P 5G.BS,t is the total load of the 5G base station at time t; ε represents the working status identifier of the 5G base station. When ε is 1, the 5G base station is in the active state. When it is 0, the 5G base station is in the sleep state; P ACT,i,t , P SLE,i,t are respectively the active state load and the sleep state load of the i-th 5G base station at time t; P S,i,t is the static load of the i-th 5G base station at time t; P D,i,t represents the communication load of the i-th 5G base station at time t, which is a dynamic load; α is the load scale factor of the i-th 5G base station; β is the coefficient reflecting the communication data of mobile users; P D,max represents the predicted maximum value of the dynamic load of the 5G base station, which is the radio frequency output power corresponding to the predicted value of the mobile user communication load in the active antenna unit of the 5G base station communication device.

[0151] 2) The safety backup power capacity of the 5G base station's backup energy storage in each period can be calculated from the above-mentioned total base station load. Therefore, the constraint of the 5G base station's backup energy storage safety backup power capacity model is:

[0152]

[0153] In the formula: E REM,t is the safety backup power capacity required by the 5G base station in the t period; T res,min is the shortest backup power time of the 5G base station.

[0154] 3) In order to fully reflect the flexibility of the backup energy storage device, a power factor control method is adopted to constrain the active / reactive power by the maximum charge / discharge capacity of the converter. Therefore, the constraints of the 5G base station's backup energy storage power control model include:

[0155]

[0156] In the formula: represents or is the active charging identifier of the 5G base station's backup energy storage, which is a variable of 0 or 1. When the identifier is 0, no charging is performed. When the identifier is 1, charging is performed; is the active discharging identifier of the 5G base station's backup energy storage, which is a variable of 0 or 1. When the identifier is 0, no discharging is performed. When the identifier is 1, discharging is performed; P t 5G,ch / dis is the active power absorbed or released by the energy storage battery at time t; is Pt 5G,ch / dis The minimum value; is P t 5G,ch / dis The maximum value.

[0157] The backup energy storage capacity of 5G base stations is directly related to the charge-discharge power. At the same time, it is necessary to ensure that the backup energy storage capacity does not exceed the limit during the entire cycle, and the backup energy storage capacity at the beginning and end of the cycle is equal. Therefore, the constraints of the 5G base station backup energy storage power control model also include:

[0158]

[0159] In the formula: is the actual operating capacity of the 5G base station backup energy storage at time t; δ is the self-discharge rate of the 5G base station backup energy storage; are the charging and discharging efficiencies of the 5G base station backup energy storage respectively; are the maximum and minimum actual operating capacities of the 5G base station backup energy storage respectively; are the actual operating capacities of the 5G base station backup energy storage at the initial and final time periods of the scheduling cycle respectively.

[0160] There is a certain power loss in the actual charge-discharge process of the energy storage battery, which causes a slight difference between the actual charge-discharge power of the base station backup energy storage and its internal power during the entire cycle. Therefore, the constraints of the 5G base station backup energy storage power control model also include:

[0161]

[0162] In the formula: P inn,t is the actual charge-discharge power of the 5G base station energy storage at time t.

[0163] 4) Constraints of the 5G base station backup energy storage SOC model:

[0164]

[0165] In the formula: are the SOC of the 5G base station backup energy storage and its maximum and minimum values respectively within time t.

[0166] ⑥ Distributed photovoltaic constraints

[0167] In the actual power grid, most DPVs are active power uncontrollable - reactive power controllable resources, and the active power output is the predicted output of DPV, as follows:

[0168] P DPV,j,t = P DPV,PRE,j,t (29)

[0169] In the formula, P DPV,j,t 、P DPV,PRE,j,tThe actual and predicted power outputs of the PV power stations installed for node j respectively in the t-th period.

[0170] ⑦ The power balance constraints in the region are as follows:

[0171]

[0172] In the formula, P IN , Q IN are respectively the sum of the active power and the sum of the reactive power injected into each node of the distribution network; P load , Q load are respectively the total active load and the total reactive load; P 5G,BS represents the total load of the 5G base stations; P GRI represents the total active power purchase amount of the distribution network, which is obtained by adding the P t grid in each period; Q GRI is the total reactive power purchase amount of the distribution network, which is obtained by adding the Q t grid in each period; P 5G is the total charge and discharge amount of the backup energy storage battery of the 5G base station, which is obtained by adding the active power P t 5G,ch absorbed in each period and the active power P t 5G,dis released; P DPV , Q CB , Q SVG are respectively the total DPV output amount, the total reactive power amount of the switched capacitor banks, and the total reactive power amount of the static var compensator.

[0173] ⑧ Sub-region regulation constraints

[0174] The distribution network sub-region coordinated optimization model needs to satisfy the sub-region regulation constraints of the state variables including the coupling branches, such as formulas (9) to (11).

[0175] Step S4, use the synchronous alternating direction method of multipliers (SADMM algorithm) to perform distributed solution on the coordinated optimization models corresponding to each sub-region of the distribution network, obtain the real-time operating power and voltage distribution of the distribution network, and realize the sub-region coordinated optimization operation of the distribution network.

[0176] The distribution network partition coordination optimization model reasonably partitions the distribution network by combining the voltage-power sensitivity of the distribution network with the source-load mismatch index to achieve voltage deviation control in each partition. Next, the distributed solution process of the SADMM algorithm is briefly described using the coordinated sub-problems of regions a and b as an example (since using the SADMM algorithm for distribution network partition coordination optimization is a conventional technique in this field and will not be elaborated here), and the solution process is referred to Figure 2 。

[0177] 1) Each region establishes its own sub-optimization problem. The augmented Lagrangian functions corresponding to the objective functions of the sub-problems of regions a and b are constructed and And through appropriate transformations, they are respectively transformed into equations (31) and (32).

[0178]

[0179] In the formula: x a , x b are the decision variables of the sub-problems of regions a and b respectively; f a (x a ), f b (x b ) are the objective functions of the sub-problems of regions a and b respectively; t is the number of iterations; ρ is the penalty parameter; are the vectors composed of the dual variables of regions a and b respectively. For example, λ a = [λ a1 , λ a2 , λ a3 , λ a4 corresponds to the coupling branch state X a,ij : = {P a,ij , Q a,ij , U a,i , U a,j}. Are the fixed reference values for the (t + 1)-th iteration of regions a and b respectively, and take the average value of the coupling branch states of regions a and b in the t-th iteration, as shown in equation (33).

[0180]

[0181] After the objective function is determined, the constraint conditions of each region are determined as equations (17) to (18), equations (21) to (30).

[0182] 2) At the (t + 1)-th iteration, calculate and obtain the decision variable values that minimize the augmented Lagrangian functions and , as shown in equations (34) to (25), and at the same time obtain the coupling branch states of each region.

[0183]

[0184] Where: are the decision variables of the sub-problems in regions a and b updated in the t+1th iteration respectively.

[0185] 3) The average value of the coupling branch state is calculated according to the coupling branch state of each region as the fixed reference value for the next iteration, as shown in the following formula:

[0186]

[0187] 4) Each region updates its dual variable separately, as follows:

[0188]

[0189] 5) Algorithm iteration convergence judgment. The algorithm iteration convergence criterion is whether the square of the second norm of the difference between the coupling branch states obtained in adjacent regions meets the conditions, and δ is the convergence accuracy. When the condition of formula (39) is met, the iteration ends.

[0190]

[0191] The present invention calculates the voltage-power sensitivity of the distribution network and reasonably partitions the distribution network in combination with the source-load mismatch index to construct a distribution network partition coordination optimization model. The distribution network partition coordination optimization model is a convex programming model with a separable objective function and linear boundary coupling constraints. The above-mentioned SADMM algorithm can be used to achieve distributed solution and realize voltage deviation management in each partition.

[0192] In order to verify the effectiveness and superiority of the method involved in the present invention, this embodiment is based on Figure 3 Taking the distribution network system shown as an example, the method of the present invention is used to formulate a 5G base station backup energy storage power control strategy and control the distribution network flow. To meet the research needs of the present invention, nodes 4, 11, 16, 22, and 32 are connected to DPV, nodes 9 and 13 are connected to CB, nodes 21 and 30 are connected to SVG, and nodes 6, 15, 19, and 28 are connected to the 5G base station system. Each device is numbered with Arabic numerals according to the above access order. The 24-hour load and DPV predicted total output curve is shown in the figure. Figure 4 The relevant parameters of the example are shown in Tables 1 and 2.

[0193] Table 1 DPV grid-connected inverter capacity at each node

[0194]

[0195]

[0196] In order to verify the effectiveness and accuracy of the method proposed in this invention, two planning methods and scenarios are set.

[0197] Scenario 1: The traditional distribution network zoning coordination optimization method is adopted, in which only voltage-power sensitivity is considered for distribution network zoning.

[0198] Scenario 2: The distribution network zoning coordination optimization method involved in the present invention is adopted.

[0199] Table 3 shows the comparison of various costs and the comprehensive cost of the two scenarios under the IEEE 33-node case study. In terms of network loss optimization, the network loss value under the action of the method of the present invention is smaller, reducing by 22.41% compared with Scenario 1; in terms of the comprehensive operation cost, the comprehensive cost of the method of the present invention is reduced by 7.27% compared with Scenario 1; in terms of voltage deviation, the average voltage deviation of the method of the present invention is 0.786, which is the lowest among the two scenarios. Therefore, the method proposed in the present invention has a more obvious voltage regulation effect, ensuring the safe and economic operation of the distribution network.

[0200] Table 3 Results of Each Scenario

[0201]

[0202] Based on the optimization results of the method of the present invention, the distribution network source-load mismatch index is obtained as Figures 5 - 6 shown. According to this figure, the sets Φ 11 , Φ 12 , Φ 21 and Φ 22 are obtained, and thus the areas with overloaded transformers and excessive DPV output are screened out. The areas with overloaded transformers are matched with the areas with excessive DPV output to form a set Ω of area-nearby mutual assistance combinations = {(3,4), (6,7), (8,9), (13,14), (16,17), (20,21), (23,24), (24,25), (24,26), (28,29), (31,32)}.

[0203] The voltage-power sensitivity matrix distribution of the IEEE33-node distribution system is as Figure 7 shown. Nodes 17, 21, and 32 are located at the end of the line, and the voltage drop caused by power change is greater. Therefore, they are respectively divided into different regions. Considering the system source-load mismatch degree and the voltage governance ability of the distribution network comprehensively, the distribution network is zoned by using the minimum net load of the area, the maximum net load rate of the area transformer, and voltage-power sensitivity as Figure 8 shown, which are Region 1 {33, 1-7, 19-24}, Region 2 {5, 25-32}, Region 3 {7-17}, and the coupling regions are Nodes 5 and 7. From Figure 9It can be seen that the traditional zoning method only considers voltage-power sensitivity but does not consider the regional source-load distribution, resulting in no adjustable resources that can be matched in Region 4 and relatively weak regional autonomy. At the same time, the traditional zoning method makes the adjustable resources in this region as available for itself as possible, divides the distribution network into more regions, resulting in too many coupling regions and increasing the burden of algorithm solving.

[0204] Figure 10 This is the action plan of the reactive power compensation device obtained by the method of the present invention. The calculated action costs of CB and SVG are 1,131.4 yuan, and the proportion in the comprehensive operation cost is 31.72%. Table 4 shows the calculation results of the voltage distribution under each scenario. Figures 11 - 12 They are the voltage distribution diagrams of each scenario. In the traditional method, the maximum voltage of the distribution network is 1.036, the minimum voltage is 0.992, the overall voltage deviation is 10.472, and there are not enough controllable resources in Region 3 and Region 5 to participate in voltage regulation, and the voltage governance ability in the region is limited. Therefore, the node voltages in Region 3 and Region 5 are significantly higher than those in other regions. In the method of the present invention, the backup energy storage of the 5G base station absorbs part of the DPV output during the high DPV stage and buffers the increase in the distribution network voltage. At the same time, the remaining capacities of the DPV and the backup energy storage of the 5G base station are fully exploited. The overall voltage deviation of the distribution network in the method of the present invention is 6.227, which is 68.84% lower than that in Scenario 1 and meets the national standard requirement within the range of ±7%.

[0205] Table 4 Voltage distribution under each scenario

[0206]

[0207] The above embodiments are the preferred implementation solutions of the present invention. In addition, the present invention can also be implemented in other ways. Any obvious replacement without departing from the concept of the technical solution of the present invention is within the protection scope of the present invention.

[0208] In order to make it more convenient for those of ordinary skill in the art to understand the improvements of the present invention over the prior art, some drawings and descriptions of the present invention have been simplified. For the sake of clarity, some other elements have also been omitted in this application document. Those of ordinary skill in the art should be aware that these omitted elements can also constitute the content of the present invention.

Claims

1. A distribution network zoning coordination optimization method considering local consumption of new energy and flexible voltage regulation, characterized in that: include: Step S1, taking the comprehensive index of source-load mismatch and the comprehensive index of voltage-power sensitivity as comprehensive evaluation indexes for regional division, constructing an optimal distribution network partitioning model that comprehensively considers local consumption of new energy and flexible voltage regulation, and using a genetic algorithm to optimize and solve the optimal distribution network partitioning model to obtain a distribution network regional division result, and dividing the distribution network into several sub-regions; Step S2, based on the distribution network area division result obtained in step S1, a distributed power optimization control coupling model is constructed with the equality of state variables of coupling branches in adjacent areas of the distribution network as a constraint; Step S3, taking the optimization of distribution network loss and average voltage deviation as the objective function, constructing a distribution network partition coordination optimization model containing 5G base stations and distributed photovoltaics, the constraints of which include: power system flow constraints, power safety constraints, group switching capacitor group constraints, static VAR compensator constraints, 5G base station constraints, distributed photovoltaic constraints, regional power balance constraints, and partition regulation constraints; wherein the partition regulation constraints are determined according to the distributed power optimization control coupling model; Step S4, using the synchronous alternating direction multiplier method to perform distributed solution on the coordinated optimization model corresponding to each sub-area of ​​the distribution network, obtain the real-time operating power and voltage distribution of the distribution network, and realize the zoned coordinated optimization operation of the distribution network.

2. The method for coordinated optimization of distribution network zoning considering local consumption of new energy and flexible voltage regulation according to claim 1 is characterized in that: The process of constructing the distribution network optimization partition model in step S1 includes: S101, introduce the maximum load rate and minimum net load of the transformer in the substation area, and determine the comprehensive index of source-load mismatch χ1, as follows: x1=max{ω T T load,i +oh P P pure,i } (2) Where: P load,i,t , Q load,i,t are respectively the active and reactive loads of the station area i in period t; P DPV,i,t is the actual output of the photovoltaic power station installed in the area i at time t; P 5G.BS,i,t is the total load of 5G base stations in area i at time t; S T is the transformer capacity of each area; T load,i , P pure,i They represent the maximum load rate and minimum net load of transformer in area i respectively. load,i When >0.8, it means that there is a transformer overload in the area i, and it is classified into the set Φ 11 , otherwise they are classified into the set Φ 12 , when P pure,i When <0, it means that there is a surplus of distributed photovoltaic power output in the station area i, and it is included in the set Φ 21 , otherwise they are classified into the set Φ 22 ;ω T ,ω P They are the weight coefficients of the maximum load rate of transformer in the substation area and the minimum net load in the substation area respectively; S102, determining the basis for selecting the leading node of the distribution network; First determine the voltage-active power sensitivity and voltage-reactive power sensitivity of each branch of the distribution network; Where: S U-P,ij is the voltage-active power sensitivity of branch ij; l represents branch ij; C 0,i Represented as the set of line nodes from node 0 to node i; r ij is the line resistance of the distribution network; V j is the voltage at node j; V i is the voltage at node i; P j is the active power flowing out of node j; P k is the active power flowing through branch k; P T,loss is the transformer active power loss; S U-Q,ij is the voltage-reactive power sensitivity of branch ij; x ij is the line reactance of the distribution network; Q j , Q k are the reactive powers of nodes j and k respectively; Q T,loss is the reactive power loss of the transformer; Then define the observability index and controllability index of the dominant node; Where: i is the dominant node number; is the set of all nodes in the distribution network; j is the number of other nodes in the area; O i is the observability index of the dominant node i; S U-P,ij / S U-P,jj Indicates the influence of the dominant node voltage on the active voltage of other nodes in the region; S U-Q,ij / S U-Q,jj Indicates the influence of the dominant node voltage on the reactive voltage of other nodes in the region; M i is the controllability index of the dominant node i; Finally, the basis for selecting the leading nodes of the distribution network is determined; x2=max{ω o The i +oh M M i } (7) Where: χ2 is the comprehensive index of voltage-power sensitivity; ω o ,ω M is the weight coefficient of observability and controllability; S103, taking the comprehensive index of source-load mismatch and the comprehensive index of voltage-power sensitivity as comprehensive evaluation indicators for regional division, construct an optimal distribution network partition model that comprehensively considers local consumption of new energy and flexible voltage regulation, as follows: F=ω1χ1+ω2χ2 (8) Where: F represents the objective function of the distribution network optimization partition model; ω1 and ω2 are the weight coefficients corresponding to each indicator, ω1+ω2=1.

3. The method for coordinated optimization of distribution network zoning considering local consumption of new energy and flexible voltage regulation according to claim 2 is characterized in that: The distributed power optimization control coupling model is as follows: X a,ij :={P a,ij ,Q a,ij ,U a,i ,U a,j } (10) X a,ij =X b,ij ,e ij ∈O (11) Where: O represents the set of coupled branches of the distribution network; e ij Represents the coupling branch of the adjacent area of ​​the distribution network; E a 、E b They represent the branch sets of area a and area b respectively, and area a and area b are adjacent to each other; R represents the set of related physical sub-areas into which the distribution network is divided; X a,ij , X b,ij represent the state variables of the coupling branches in regions a and b respectively; P a,ij represents the active power transmitted by the coupling branch in region a; Q a,ij represents the reactive power transmitted by the coupling branch in area a; U a,i ,U a,j They respectively represent the node voltages of nodes i and j at both ends of the coupling branch in region a.

4. The method for coordinated optimization of distribution network zoning considering local consumption of new energy and flexible voltage regulation according to claim 3 is characterized in that: The objective function of the distribution network partition coordination optimization model is as follows: minf=λ LIN f LIN +λ SYN f SYN (12) c LIN +g SYN =1 (15) Where: minf represents the objective function of the distribution network partition coordination optimization model; f LIN 、f SYN are distribution network loss and average voltage deviation respectively; LIN , SYN are the weight factors of distribution network loss and average voltage deviation respectively; γ LIN , γ SYN are the weight coefficients of distribution network loss and average voltage deviation respectively; are the initial values ​​of distribution network loss and average voltage deviation before optimization; c LIN,t is the unit network loss price at time t; I ij is the branch current between node i and node j; Δt is the duration of a single period; N bus is the number of distribution network nodes; V N is the reference voltage; V i,t is the voltage of node j at time t; T is an optimization period.

5. The method for coordinated optimization of distribution network zoning considering local consumption of new energy and flexible voltage regulation according to claim 4 is characterized in that: The power system flow constraints are as follows: In the formula, p j ,q j are the active and reactive injected powers of node j respectively; P jk , Q jk are the active and reactive power flowing from node j to the next node k; P ij , Q ij are the active and reactive power flowing from the previous node i to node j; I ij is the branch current between node i and node j; g j , b j is the conductance and susceptance of node j to ground, which are constants; V i 、V j are the voltages of nodes i and j respectively; The power security constraints are as follows: In the formula, I ij.max ,I ij.min are the upper and lower limits of the current in the branch between node i and node j respectively; P t grid , They are the interactive active and reactive power of the upper power grid respectively; They are the maximum and minimum interactive active powers allowed to pass through the connecting branch between the distribution network and the upper-level power grid; are the maximum and minimum reactive interaction powers allowed to pass through the distribution network and the upper grid connection branch; V j.max 、V j.min are the upper and lower limits of the voltage at node j respectively; is the square of the branch current between node i and node j; is the square of the voltage at node j; The group switching capacitor bank constraints are as follows: Where: y CB,j,t is the number of capacitor groups in operation, which is a discrete variable value; Y CB,max,j The upper limit of the number of capacitors connected to node j; Q CB,step,j is the compensation power of each group of capacitors, which is a constant; N CB,max,j is the upper limit of the number of operations; Q CB,j,t is the reactive power of the grouped switched capacitor bank; t and T are an optimization time period and an optimization cycle respectively; The static VAR compensator constraints are as follows: Q SVG,min,j ≤Q SVG,j ≤Q SVG,max,j (20) Where: Q SVG,j is the reactive power of static VAR compensator; Q SVG,min,j , Q SVG,max,j They are the lower and upper limits of the reactive power of the static VAR compensator respectively; The 5G base station constraints include: 1) 5G base station load model constraints: P D,i,t =βP D,max (22) Where P 5G.BS,t is the total load of the 5G base station at time t; ε represents the working status of the 5G base station. When ε is 1, the 5G base station is in an active state, and when it is 0, the 5G base station is in a sleep state; P ACT,i,t , P SLE,i,t are the activation state load and sleep state load of the i-th 5G base station at time t respectively; P S,i,t is the static load of the i-th 5G base station at time t; P D,i,t represents the communication load of the i-th 5G base station at time t, which is a dynamic load; α is the load scale factor of the i-th 5G base station; β is the coefficient reflecting the communication data of mobile users; P D,max Indicates the predicted maximum value of the dynamic load of the 5G base station, which is the RF output power corresponding to the active antenna unit in the 5G base station communication device when the mobile user communication load is predicted; 2) 5G base station backup energy storage safety backup capacity model constraints: Where: E REM,t The safe backup power capacity required by the 5G base station in period t; T res,min The shortest power backup time for 5G base stations; 3) 5G base station backup energy storage power control model constraints: In the formula, express or It is the active charging flag of the backup energy storage of the 5G base station, which is a variable of 0 or 1. When the flag is 0, no charging is performed, and when the flag is 1, charging is performed; is the active discharge flag of the 5G base station backup energy storage, which is a variable of 0 or 1. When the flag is 0, no discharge is performed, and when the flag is 1, discharge is performed; P t 5G,ch / dis is the active work absorbed or released by the energy storage battery at time t rate; P t 5G,ch / dis The minimum value of P t 5G,ch / dis The maximum value of is the actual operating capacity of the 5G base station backup energy storage in period t; δ is the self-discharge rate of the 5G base station backup energy storage; They are the charging and discharging efficiency of 5G base station backup energy storage; They are the maximum and minimum actual operating capacities of 5G base station backup energy storage; P is the actual operating capacity of the 5G base station backup energy storage in the initial period and the final period of the scheduling cycle respectively; inn,t is the actual charging and discharging power of the 5G base station energy storage during period t; 4) 5G base station backup energy storage SOC model constraints: Where: They are the 5G base station backup energy storage SOC and its maximum and minimum values ​​in period t respectively; The distributed photovoltaic constraints are as follows: P DPV,j,t =P DPV,PRE,j,t (29) Where P DPV,j,t , P DPV,PRE,j,t are the actual and predicted outputs of the PV power station installed at node j in period t, respectively; The power balance constraints within the region are as follows: Where P IN , Q IN are the sum of active power and reactive power injected into each node of the distribution network; P load , Q load are respectively the total active load and the total reactive load; P 5G.BS is the total load of 5G base stations; P GRI Represents the total amount of active power purchased by the distribution network, which is determined by P in each period. t grid Add together to get; Q GRI is the total amount of reactive power purchased by the distribution network, which is determined by the Add together to get; P 5G is the total charge and discharge amount of the 5G base station backup energy storage battery, which is the active power P absorbed in each period t 5G,ch And the released active power P t 5G,dis Add together to get; P DPV is the total DPV output; Q CB Q is the total reactive power of the grouped switched capacitor bank; SVG is the total reactive power of the static VAR compensator; The partition control constraints are as shown in equations (9) to (11).

Images