Medium-voltage power distribution network source-grid-storage double-layer planning method considering flexibility

Abstract

The present invention relates to a medium-voltage power distribution network source-grid-storage double-layer planning method considering flexibility. The method comprises the following steps: step 1, a power distribution network source-grid-storage double-layer planning framework based on flexibility assessment indicators; step 2, a power distribution network source-grid-storage double-layer planning model; step 3, solving of the power distribution network source-grid-storage double-layer planning model; and step 4, use analysis of demonstration cases. The present invention has the advantages of establishing assessment indicators for flexibility level from the aspects of flexibility supply-demand matching and network transmission capacity, and improving the economy and flexibility of expansion planning and system operation.

Description

A two-layer planning method for source, grid and storage of medium voltage distribution network considering flexibility Technical Field

[0001] The present invention belongs to the technical field of medium-voltage distribution networks, distribution network source-grid-storage coordinated planning, and distribution network flexibility improvement under high-penetration distributed power sources and large-scale power load access, and specifically relates to a medium-voltage distribution network source-grid-storage two-layer planning method taking flexibility into consideration. Background Art

[0002] In recent years, rural electrification has been vigorously promoted, resulting in a large increase in electricity load, making the original distribution network grid structure unable to adapt to its large-scale development; at the same time, the large-scale access of distributed photovoltaic (DPV) and loads has aggravated the problems of uneven source and load distribution and mismatch of source and load characteristics in the distribution network. On the one hand, it increases the period of voltage upper and lower limits in the distribution network, that is, when the DPV output is high during the day and the load power demand is low, the node voltage is prone to exceed the upper limit, forming a reverse flow, affecting the network loss and safe operation of the distribution system; at night, when the DPV output is zero and the load demand is high, the node voltage is prone to exceed the lower limit, which also affects the system safety; on the other hand, due to the large differences in the geographical access of sources and loads, some nodes have less load access and more photovoltaic access, making photovoltaic absorption difficult, while other nodes have more load access and less photovoltaic access, resulting in insufficient capacity, heavy overload and low voltage problems; current research has focused on the problems caused by a large number of load access to the distribution network and the DPV absorption problem The solution to this problem still relies mainly on the configuration and optimized operation of equipment, including the addition of energy storage systems (ESS), reactive power compensation devices, automatic control systems, demand-side management, etc., while research on grid flexibility is relatively small. In fact, the grid structure provides a path for the flow calculation of equipment in the system and is also one of the basic elements of the distribution network. It is urgent to make reasonable adjustments to the grid to improve the grid power supply capacity. Currently, distribution network planning can be divided into substation planning, grid planning, DG planning, ESS planning, etc. according to the planning object. The existing technology is a distribution network planning carried out for a single object. The actual distribution network planning is affected by multiple factors such as the grid, DG, ESS, etc. It is difficult for a single object planning to ensure the overall economic efficiency of the planning. Therefore, it is very necessary to provide a medium-voltage distribution network source-grid-storage two-layer planning method that takes flexibility into consideration, establishes flexibility level evaluation indicators from the flexibility supply and demand matching and network transmission capacity, and improves the economy and flexibility of expansion planning and system operation.

[0003] Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a medium-voltage distribution network source-network-storage two-layer planning method that takes flexibility into consideration, which establishes flexibility level evaluation indicators from flexibility supply and demand matching and network transmission capacity, and improves the economy and flexibility of expansion planning and system operation.

[0005] The object of the present invention is achieved by: a medium voltage distribution network source-network-storage two-layer planning method considering flexibility, the method comprising the following steps:

[0006] Step 1: A two-layer planning framework of source, grid and storage for distribution network based on flexibility evaluation index;

[0007] Step 2: Distribution network source-grid-storage two-layer planning model;

[0008] Step 3: Solve the source-grid-storage two-level programming model of the distribution network;

[0009] Step 4: Demonstration application analysis.

[0010] The distribution network source-grid-storage two-layer planning framework based on flexibility evaluation index in step 1 includes the following steps:

[0011] Step 1.1: Construction of flexibility indicators;

[0012] Step 1.2: Distribution network source-grid-storage two-layer planning model framework.

[0013] The construction of the flexibility index in step 1.1 includes the following steps:

[0014] Step 1.11: Net load fluctuation index: The net load fluctuation curve is affected by both load demand and DPV penetration. The net load fluctuation index refers to the system's ability to withstand uncertainty and volatility under large-scale loads and high-penetration DPV access. It can clearly characterize the system's flexibility requirements from a time scale. The smaller the index, the flatter the overall net load fluctuation curve, indicating that the system is more adaptable to large loads and DPV access, the higher the overall source-load matching of the system, the better the overall flexibility of the distribution network, and the less flexibility resources need to be added or modified in the planning scheme. Conversely, the worse the system flexibility, the more flexibility resources need to be supplied in the planning scheme. The net load fluctuation index F in the sth scenario need (s) can be expressed as: Where, P NL,t (s) is the net load size of period t under scenario s; P NL,ave (s) is the average net load of the system under scenario s; T is the total number of time periods in the scheduling cycle;

[0015] Step 1.12: Line load margin balance index: To balance the supply and demand of distribution network flexibility, in addition to the net load fluctuation rate index, reasonable requirements for grid flexibility must also be made. Although grid flexibility resources cannot directly provide flexibility to the distribution network, they can provide access for flexibility resources and avoid the occurrence of "flexibility resource unavailability". Therefore, it is necessary to propose line load margin balance to characterize the balance of distribution system grid power flow. The load margin of a single line is described by the real-time load margin of the line transmission channel, reflecting the degree of line transmission congestion. The calculation formula is: Where, F ij,t I is the load margin of a single line between nodes i and j in period t; ij,t is the line transmission current between nodes i and j during period t; I ij,max is the upper limit of the line transmission current between nodes i and j; F ij,t The larger the index value, the lower the possibility of line blockage, so the line load margin can be used to evaluate the grid transmission capacity; when F ij,t ≥0, indicating that the line has sufficient load margin and can adapt to load power fluctuations; when F ij,t When ≤0, it indicates that the line load margin is insufficient and line congestion will occur; therefore, the load margin balance index F of the power supply line in the s scenario is net (s) can be expressed as: Where, F ave (s) is the mean value of the load margin of the power supply line; L is the set of system lines.

[0016] The distribution network source-grid-storage two-layer planning model in step 2 includes the following steps:

[0017] Step 2.1: Upper level planning model;

[0018] Step 2.2: Lower-level planning model.

[0019] The upper-level planning model in step 2.1 is specifically as follows: the upper-level planning model objective function is composed of the annual comprehensive cost of the distribution network, the net load fluctuation index and the line load margin balance index, and the decision variables are whether the line is newly built, the DPV location and capacity, and the ESS location and capacity; the objective function is: f1 = {minC total ,minF net ,minF need}(6), C total =C inv +C mai +C ope (7), Where f1 is the objective function of the upper-level planning model; C total is the annual comprehensive cost of the distribution network; C inv is the annual investment cost of the equipment; C mai is the annual operation and maintenance cost of the equipment; C ope is the annual operation and maintenance cost of the distribution system, including the annual network loss cost of the distribution system and the annual power purchase cost of the upper power grid; λ is the equipment maintenance cost coefficient; Λ is the lower-level operation scenario set; is the probability of scene s appearing; C ope (s) is the annual operation and maintenance cost of the distribution system under scenario s; α is the discount rate; k DPV 、k ESS 、k line are the service life of DPV, ESS and the line to be built respectively; c DPV 、c ESS are the unit capacity investment costs of DPV and ESS respectively; c line is the investment cost per unit length of line; j∈Ω DPV 、i∈Ω ESS are the sets of DPV and ESS installation nodes respectively; ij∈Ω line For construction line collection; E ESS,i 、P DPV,j are the capacities of ESS and DPV installed at nodes i and j respectively; l ij is the length of the line to be built.

[0020] The constraints of the upper-level planning model in step 2.1 include DPV, ESS, line load margin constraints and grid connectivity constraints; among them, ①DPV capacity constraint is: 0≤P DPV,i ≤P DPV,max (11), where: P DPV,i is the DPV installed capacity at node i; P DPV,max is the maximum installed capacity of DPV at node i;② ESS capacity constraint is: Where, are the upper and lower limits of the ESS installation capacity at node i; ③ The line load margin constraint is: F ij,t ≥0(13); ④ The grid connectivity constraint is: when building lines for new load nodes, no ring network is formed, the distribution network is kept in a radial form, and there are no isolated islands or points after DPV is connected.

[0021] The lower-level planning model in step 2.2 is specifically as follows: the lower-level operation model takes the upper-level planning results as input, takes the optimal operation economy as the goal, and the objective function is the sum of the network loss cost of each scenario and the cost of purchasing electricity from the upper level. The decision variables are the capacity of DPV and ESS, and the objective function is: Where f2 is the objective function of the lower layer operation model; Closs (s) is the network loss cost under scenario s; C sub (s) is the cost of electricity purchased from the upper power grid under scenario s; c t is the time-of-use electricity price during period t; P loss,t (s) is the network loss power in time period t under scenario s; P sub,t (s) is the power transmitted to the upper grid during period t under scenario s.

[0022] The constraints of the lower-level planning model in step 2.2 include power flow equality constraints, node voltage inequality constraints, ESS operation inequality constraints, and power balance constraints. ① The power flow constraint is: Where, P inj,i , Q inj,i Respectively represent the active power and reactive power injected at node i; U i 、U j are the voltage amplitudes of nodes i and j respectively; n is the total number of nodes; G ij 、B ij are the conductance and susceptance in the node admittance matrix respectively; δ ij is the voltage phase angle difference between nodes i and j;② The node voltage constraint is: U min ≤U i ≤U max (16), where U max 、U min are the upper and lower limits of voltage amplitude respectively; ③ Line current carrying capacity constraint is: I ij ≤η ij |I ij,max |(17), where I ij is the current flowing through the branch; η ij is the state variable of branch ij. When the value is 1, it means that node i is connected to node j. When the value is 0, it means that node i is not connected to node j. ij,max is the upper limit of the current amplitude of branch ij; ④ESS operation constraints are: Where, are the charging and discharging power of ESS at node j during period t; y j,t is a 0-1 variable, which represents the charge and discharge status of the ESS during period t, where 1 indicates discharge and 0 indicates charge; is the charge of ESS at node j during period t; are the maximum and minimum charges of ESS at node j respectively; are the ESS charging and discharging efficiency, respectively; are the charges at the initial and final periods of the ESS operation cycle at node j respectively; ⑤The power balance constraint is: Where, P sub,tis the transmission power of the upper power grid in period t; P ESS,i,t is the power provided by the ESS at node i during period t, with discharge being positive and charge being negative; P DPV,j,t The power provided by DPV in time period t; P load,t is the total load demand of the distribution system during period t.

[0023] The solution of the distribution network source-grid-storage two-layer planning model in step 3 is specifically as follows: the medium-voltage distribution network source-grid-storage two-layer planning model considering flexibility is essentially a combined planning of the distribution network grid structure and two flexible resources, DPV and ESS. The model contains both 0-1 integer decision variables and continuous decision variables such as DPV and ESS output. To reduce the difficulty of solving the mixed integer model, the initial grid structure is first determined by the minimum spanning tree method; then, the fast non-dominated genetic algorithm (NSGA-II) is used to solve the multi-objective planning problem of optimal upper-level economy and flexibility, and then the particle swarm algorithm is used to solve the optimal model of lower-level operation economy; finally, the grid structure is determined to achieve the site selection and sizing of DPV and ESS; the steps include:

[0024] Step 3.1: Scenario clustering: Affected by the dual uncertainty of DPV output and load fluctuation, the complexity of the operating scenarios faced by the distribution network has been greatly increased. The planning and optimization operation of the distribution network need to meet the optimal comprehensive operation level in all scenarios. However, the changeable historical scenarios and solution efficiency make it difficult to traverse every scenario in the planning and optimization solution process. In order to take into account the overall operation of the distribution network throughout the year and improve the solution efficiency, an uncertainty planning method based on multi-scenario analysis is adopted; the correlation between DPV output and load power time series matching problems affects the change of the net load curve, which in turn affects the accuracy of the distribution network source network storage coordination planning and optimization operation; therefore, by selecting the daily characteristic indicators of DPV output and load power, and using the K-means clustering and hierarchical agglomerative clustering composite clustering algorithm to achieve scenario reduction, and using the clustering effectiveness index CH (+) Determine the optimal number of clusters and obtain typical scenarios and their probabilities; the selected daily characteristic indicators are: Where, X DPV,mean 、X load,mean are the average values ​​of DPV daily power and load daily power respectively; X DPV,std 、X load,std are the variances of DPV daily power and load daily power respectively; P DPV,t 、P load,t is the DPV output and load power during period t;

[0025] Step 3.2: Overall solution process.

[0026] The overall solution process in step 3.2 includes the following steps:

[0027] Step 3.21: Input the original load, the load to be planned, and the grid structure to be expanded;

[0028] Step 3.22: Generate the initial grid structure based on the minimum spanning tree algorithm;

[0029] Step 3.23: Use the NSGA-II algorithm to solve the upper-level multi-objective planning model and obtain the location and capacity of the DPV and ESS under the planning scheme with the best economy and flexibility.

[0030] Step 3.24: Input the upper-level planning scheme as the initial value into the lower-level optimization model, use the particle swarm optimization algorithm to solve the lower-level operation model, and determine whether it converges. If it converges, the optimal ESS charging and discharging power strategy under the planning scheme is obtained. If not, return to step 3.23;

[0031] Step 3.25: Determine whether the convergence condition is met. If not, return to step 3.23; if so, end and obtain the optimal distribution network source-grid-storage coordinated planning scheme.

[0032] Beneficial effects of the invention: The present invention is a medium-voltage distribution network source-grid-storage two-layer planning method considering flexibility. Aiming at the problem of insufficient flexibility after high-penetration DPV and large-scale power loads are connected to the medium-voltage distribution network, the present invention establishes a distribution network flexibility level evaluation index from the two aspects of flexibility supply and demand matching and network transmission capacity, and based on the index, establishes a medium-voltage distribution network source-grid-storage two-layer planning method considering flexibility, aiming to improve the economy and flexibility of distribution network expansion planning and power system operation; in use, the present invention proposes a two-layer planning model for the joint coordinated planning of source, grid and storage of medium-voltage distribution network considering flexibility; wherein, the upper objective function is a multi-objective function of optimal flexibility and economy, and the optimal economy is the traditional planning Objective: flexibility is an innovative goal added to adapt to the new power system containing DPV and load surge. The two are closely linked and together constitute the upper-level planning goal of this article. The planning scheme is determined by the site selection and sizing of the lines to be built, DPV and ESS, and the upper-level planning scheme is used as the input of the lower-level optimization model; the lower-level objective function is to optimize the operating economy. Under the planning scheme determined by the upper layer, a multi-scenario operation simulation of the source and load is carried out, and the relevant parameters of the optimal economic operation strategy are returned to the upper layer. The upper and lower layers are continuously iterated, and finally a planning scheme that takes flexibility and economy into consideration is obtained; the present invention has the advantages of establishing flexibility level evaluation indicators from flexibility supply and demand matching and network transmission capacity, and improving the economy and flexibility of expansion planning and system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] FIG1 is a framework diagram of a medium voltage distribution network planning model of the present invention.

[0034] FIG2 is a flow chart of solving the double-layer planning of the distribution network according to the present invention.

[0035] FIG3 is a network topology diagram of the distribution network in the planning area of ​​the present invention.

[0036] Figure 4 shows the clustering effectiveness index CH of the method of the present invention. (+) curve chart.

[0037] FIG5 is a schematic diagram of a typical scenario of DPV and load output according to the present invention.

[0038] FIG6 is a graph showing the power and SOC status of ESS1 in Solution 1 of the present invention.

[0039] FIG7 is a graph showing the power and SOC status of ESS2 in solution 1 of the present invention.

[0040] FIG8 is a graph showing the power and SOC status of ESS3 in Solution 1 of the present invention.

[0041] FIG9 is a graph showing the power and SOC status of ESS4 in solution 1 of the present invention.

[0042] FIG10 is a graph showing the power and SOC status of ESS1 in solution 2 of the present invention.

[0043] FIG11 is a graph showing the power and SOC status of ESS2 in solution 2 of the present invention.

[0044] FIG12 is a graph showing the power and SOC status of ESS3 in solution 2 of the present invention.

[0045] FIG13 is a graph showing the power and SOC status of ESS4 in solution 2 of the present invention.

[0046] FIG14 is a schematic diagram of the photovoltaic power output history data for the entire year according to the present invention.

[0047] FIG15 is a diagram showing historical data of load output throughout the year according to the present invention. DETAILED DESCRIPTION

[0048] The present invention will be further described below with reference to the accompanying drawings.

[0049] Example 1

[0050] As shown in Figure 1-15, a two-tier planning method for medium-voltage distribution networks, including source, network, and storage, takes flexibility into consideration. The method includes the following steps:

[0051] Step 1: A two-layer planning framework of source, grid and storage for distribution network based on flexibility evaluation index;

[0052] Step 2: Distribution network source-grid-storage two-layer planning model;

[0053] Step 3: Solve the source-grid-storage two-level programming model of the distribution network;

[0054] Step 4: Demonstration application analysis.

[0055] The distribution network source-grid-storage two-layer planning framework based on flexibility evaluation index in step 1 includes the following steps:

[0056] Step 1.1: Construction of flexibility indicators;

[0057] Step 1.2: Distribution network source-grid-storage two-layer planning model framework.

[0058] In this embodiment, at this stage, the flexibility of the power system mostly refers to the ability of the system to quickly respond to uncertainties on both the supply and demand sides while taking into account economic efficiency. Node flexibility resources refer to node controllable devices that participate in responding to power changes; network flexibility resources refer to the transmission network that provides transmission channel capabilities for node flexibility resources within different time scales while meeting relevant physical constraints; the flexibility of the distribution network mainly comes from the flexibility resources of the physical layer such as the transmission grid and ESS, among which ESS provides flexibility support and supply to the external system on a node basis, which can effectively suppress net load fluctuations; although the grid transmission channel cannot directly participate in power balance regulation, it can provide an important carrier for the distribution network node-type flexibility resources to exert their responsiveness; based on the power system flexibility requirements and combined with the characteristics of the distribution system flexibility resources, the present invention defines the distribution network flexibility level evaluation index from two aspects: net load fluctuation rate and line load margin balance.

[0059] The construction of the flexibility index in step 1.1 includes the following steps:

[0060] Step 1.11: Net load fluctuation index: The net load fluctuation curve is affected by both load demand and DPV penetration. The net load fluctuation index refers to the system's ability to withstand uncertainty and volatility under large-scale loads and high-penetration DPV access. It can clearly characterize the system's flexibility requirements from a time scale. The smaller the index, the flatter the overall net load fluctuation curve, indicating that the system is more adaptable to large loads and DPV access, the higher the overall source-load matching of the system, the better the overall flexibility of the distribution network, and the less flexibility resources need to be added or modified in the planning scheme. Conversely, the worse the system flexibility, the more flexibility resources need to be supplied in the planning scheme. The net load fluctuation index F in the sth scenario need (s) can be expressed as: Where, P NL,t (s) is the net load size of period t under scenario s; P NL,ave (s) is the average net load of the system in scenario s; T is the total number of time periods in the scheduling cycle. In this paper, the scheduling cycle is taken as 24 hours and the scheduling time interval is 1 hour;

[0061] Step 1.12: Line load margin balance index: To balance the supply and demand of distribution network flexibility, in addition to the net load fluctuation rate index, reasonable requirements for grid flexibility must also be made. Although grid flexibility resources cannot directly provide flexibility to the distribution network, they can provide access for flexibility resources and avoid the occurrence of "flexibility resource unavailability". Therefore, it is necessary to propose line load margin balance to characterize the balance of distribution system grid power flow. The load margin of a single line is described by the real-time load margin of the line transmission channel, reflecting the degree of line transmission congestion. The calculation formula is: Where, F ij,t I is the load margin of a single line between nodes i and j in period t; ij,t is the line transmission current between nodes i and j during period t; I ij,max is the upper limit of the line transmission current between nodes i and j; F ij,t The larger the index value, the lower the possibility of line blockage, so the line load margin can be used to evaluate the grid transmission capacity; when F ij,t ≥0, indicating that the line has sufficient load margin and can adapt to load power fluctuations; when F ij,t When ≤0, it indicates that the line load margin is insufficient and line congestion will occur; therefore, the load margin balance index F of the power supply line in the s scenario is net (s) can be expressed as: Where, F ave (s) is the mean value of the load margin of the power supply line; L is the set of system lines.

[0062] In this embodiment, the framework of the distribution network source-grid-storage two-layer planning model is as follows: the grid connection of high-penetration DPV urgently requires the integration of various flexibility resources for joint planning, and the increasingly blurred source-load boundary makes the distribution network planning and operation problems highly coupled. In the planning stage, it is necessary to consider the impact of the operation strategy at the same time, which further increases the difficulty of solving the planning problem; because the present invention simultaneously considers the optimal configuration of multiple flexibility resources and the optimal operation of the distribution network with improved flexibility, based on the idea of ​​decomposition and coordination, it is necessary to decouple the optimal configuration of flexibility resources and the optimal operation of the distribution network;

[0063] In summary, the present invention proposes a two-layer planning model of source, network and storage for medium-voltage distribution network considering flexibility, as shown in Figure 1, wherein the upper-layer planning model is a multi-objective problem composed of the annual comprehensive cost of the distribution network, the net load fluctuation rate index and the line load margin balance index, and the decision variables are whether the line to be built is invested, the location and capacity of DPV and ESS; the lower-layer operation model takes the upper-layer planning results as input, and takes the optimal operation economy as the goal. The objective function is the sum of the network loss cost of each scenario and the cost of purchasing electricity from the superior, and the decision variable is the charging and discharging power of the ESS. The current distribution network loss power, ESS charging and discharging power, the cost of purchasing electricity from the superior power grid and the line transmission current are transmitted to the upper layer; the upper and lower layers continuously iterate and optimize each other, and finally obtain a distribution network expansion planning scheme that takes into account both economy and flexibility.

[0064] The distribution network source-grid-storage two-layer planning model in step 2 includes the following steps:

[0065] Step 2.1: Upper level planning model;

[0066] Step 2.2: Lower-level planning model.

[0067] The upper-level planning model in step 2.1 is specifically as follows: the upper-level planning model objective function is composed of the annual comprehensive cost of the distribution network, the net load fluctuation index and the line load margin balance index, and the decision variables are whether the line is newly built, the DPV location and capacity, and the ESS location and capacity; the objective function is: f1 = {minC total ,minF net ,minF need}(6), C total =C inv +C mai +C ope (7), Where f1 is the objective function of the upper-level planning model; C total is the annual comprehensive cost of the distribution network; C inv is the annual investment cost of the equipment; C mai is the annual operation and maintenance cost of the equipment; C ope is the annual operation and maintenance cost of the distribution system, including the annual network loss cost of the distribution system and the annual power purchase cost of the upper power grid; λ is the equipment maintenance cost coefficient; Λ is the lower-level operation scenario set; is the probability of scene s appearing; C ope (s) is the annual operation and maintenance cost of the distribution system under scenario s; α is the discount rate; k DPV 、k ESS 、k line are the service life of DPV, ESS and the line to be built respectively; c DPV 、c ESSare the unit capacity investment costs of DPV and ESS respectively; c line is the investment cost per unit length of line; j∈Ω DPV 、i∈Ω ESS are the sets of DPV and ESS installation nodes respectively; ij∈Ω line For construction line collection; E ESS,i 、P DPV,j are the capacities of ESS and DPV installed at nodes i and j respectively; l ij is the length of the line to be built.

[0068] The constraints of the upper-level planning model in step 2.1 include DPV, ESS, line load margin constraints and grid connectivity constraints; among them, ①DPV capacity constraint is: 0≤P DPV,i ≤P DPV,max (11), where: P DPV,i is the DPV installed capacity at node i; P DPV,max is the maximum installed capacity of DPV at node i;② ESS capacity constraint is: Where, are the upper and lower limits of the ESS installation capacity at node i; ③ The line load margin constraint is: F ij,t ≥0(13); ④ The grid connectivity constraint is: when building lines for new load nodes, no ring network is formed, the distribution network is kept in a radial form, and there are no isolated islands or points after DPV is connected.

[0069] The lower-level planning model in step 2.2 is specifically as follows: the lower-level operation model takes the upper-level planning results as input, takes the optimal operation economy as the goal, and the objective function is the sum of the network loss cost of each scenario and the cost of purchasing electricity from the upper level. The decision variables are the capacity of DPV and ESS, and the objective function is: Where f2 is the objective function of the lower layer operation model; C loss (s) is the network loss cost under scenario s; C sub (s) is the cost of electricity purchased from the upper power grid under scenario s; c t is the time-of-use electricity price during period t; P loss,t (s) is the network loss power in time period t under scenario s; P sub,t (s) is the power transmitted to the upper grid during period t under scenario s.

[0070] The constraints of the lower-level planning model in step 2.2 include power flow equality constraints, node voltage inequality constraints, ESS operation inequality constraints, and power balance constraints. ① The power flow constraint is: Where, P inj,i , Q inj,i Respectively represent the active power and reactive power injected at node i; Ui 、U j are the voltage amplitudes of nodes i and j respectively; n is the total number of nodes; G ij 、B ij are the conductance and susceptance in the node admittance matrix respectively; δ ij is the voltage phase angle difference between nodes i and j;② The node voltage constraint is: U min ≤U i ≤U max (16), where U max 、U min are the upper and lower limits of voltage amplitude respectively; ③ Line current carrying capacity constraint is: I ij ≤η ij |I ij,max |(17), where I ij is the current flowing through the branch; η ij is the state variable of branch ij. When the value is 1, it means that node i is connected to node j. When the value is 0, it means that node i is not connected to node j. ij,max is the upper limit of the current amplitude of branch ij; ④ESS operation constraints are: Where, are the charging and discharging power of ESS at node j during period t; y j,t is a 0-1 variable, which represents the charge and discharge status of the ESS during period t, where 1 indicates discharge and 0 indicates charge; is the charge of ESS at node j during period t; are the maximum and minimum charges of ESS at node j respectively; are the ESS charging and discharging efficiency, respectively; are the charges at the initial and final periods of the ESS operation cycle at node j respectively; ⑤The power balance constraint is: Where, P sub,t is the transmission power of the upper power grid in period t; P ESS,i,t is the power provided by the ESS at node i during period t, with discharge being positive and charge being negative; P DPV,j,t The power provided by DPV in time period t; P load,t is the total load demand of the distribution system during period t.

[0071] The solution of the distribution network source-grid-storage two-layer planning model in step 3 is specifically as follows: the medium-voltage distribution network source-grid-storage two-layer planning model considering flexibility is essentially a combined planning of the distribution network grid structure and two flexible resources, DPV and ESS. The model contains both 0-1 integer decision variables and continuous decision variables such as DPV and ESS output. To reduce the difficulty of solving the mixed integer model, the initial grid structure is first determined by the minimum spanning tree method; then, the fast non-dominated genetic algorithm (NSGA-II) is used to solve the multi-objective planning problem of optimal upper-level economy and flexibility, and then the particle swarm algorithm is used to solve the optimal model of lower-level operation economy; finally, the grid structure is determined to achieve the site selection and sizing of DPV and ESS; the steps include:

[0072] Step 3.1: Scenario clustering: Affected by the dual uncertainty of DPV output and load fluctuation, the complexity of the operating scenarios faced by the distribution network has been greatly increased. The planning and optimization operation of the distribution network need to meet the optimal comprehensive operation level in all scenarios. However, the changeable historical scenarios and solution efficiency make it difficult to traverse every scenario in the planning and optimization solution process. In order to take into account the overall operation of the distribution network throughout the year and improve the solution efficiency, an uncertainty planning method based on multi-scenario analysis is adopted; the correlation between DPV output and load power time series matching problems affects the change of the net load curve, which in turn affects the accuracy of the distribution network source network storage coordination planning and optimization operation; therefore, by selecting the daily characteristic indicators of DPV output and load power, and using the K-means clustering and hierarchical agglomerative clustering composite clustering algorithm to achieve scenario reduction, and using the clustering effectiveness index CH (+) Determine the optimal number of clusters and obtain typical scenarios and their probabilities; the selected daily characteristic indicators are: Where, X DPV,mean 、X load,mean are the average values ​​of DPV daily power and load daily power respectively; X DPV,std 、X load,std are the variances of DPV daily power and load daily power respectively; P DPV,t 、P load,t is the DPV output and load power during period t;

[0073] Step 3.2: Overall solution process.

[0074] The overall solution process in step 3.2 includes the following steps:

[0075] Step 3.21: Input the original load, the load to be planned, and the grid structure to be expanded;

[0076] Step 3.22: Generate the initial grid structure based on the minimum spanning tree algorithm;

[0077] Step 3.23: Use the NSGA-II algorithm to solve the upper-level multi-objective planning model and obtain the location and capacity of the DPV and ESS under the planning scheme with the best economy and flexibility.

[0078] Step 3.24: Input the upper-level planning scheme as the initial value into the lower-level optimization model, use the particle swarm optimization algorithm to solve the lower-level operation model, and determine whether it converges. If it converges, the optimal ESS charging and discharging power strategy under the planning scheme is obtained. If not, return to step 3.23;

[0079] Step 3.25: Determine whether the convergence condition is met. If not, return to step 3.23; if so, end and obtain the optimal distribution network source-grid-storage coordinated planning scheme.

[0080] The present invention is a medium-voltage distribution network source-grid-storage two-layer planning method considering flexibility. Aiming at the problem of insufficient flexibility after high-penetration DPV and large-scale power loads are connected to the medium-voltage distribution network, the present invention establishes a distribution network flexibility level evaluation index from the two aspects of flexibility supply and demand matching and network transmission capacity, and establishes a medium-voltage distribution network source-grid-storage two-layer planning method considering flexibility based on the index, aiming to improve the economy and flexibility of distribution network expansion planning and power system operation. In use, the present invention proposes a two-layer planning model for the joint coordinated planning of source, grid and storage of medium-voltage distribution network considering flexibility. Among them, the upper objective function is a multi-objective function of optimal flexibility and economy, optimal economy is the traditional goal in planning, and optimal flexibility is the goal of optimal flexibility. Activity is an innovative goal added to adapt to the new power system containing DPV and a sharp increase in load. The two are closely linked and together constitute the upper-level planning goal of this article. The planning scheme is determined by the site selection and sizing of the lines to be built, DPV and ESS, and the upper-level planning scheme is used as the input of the lower-level optimization model; the lower-level objective function is to optimize the operating economy. Under the planning scheme determined by the upper layer, a multi-scenario operation simulation of source and load is carried out, and the relevant parameters of the optimal economic operation strategy are returned to the upper layer. The upper and lower layers are continuously iterated, and finally a planning scheme that takes flexibility and economy into consideration is obtained; the present invention has the advantages of establishing flexibility level evaluation indicators from flexibility supply and demand matching and network transmission capacity, and improving the economy and flexibility of expansion planning and system operation.

[0081] Example 2

[0082] As shown in Figure 1-15, a two-tier planning method for medium-voltage distribution networks, including source, network, and storage, takes flexibility into consideration. The method includes the following steps:

[0083] Step 1: A two-layer planning framework of source, grid and storage for distribution network based on flexibility evaluation index;

[0084] Step 2: Distribution network source-grid-storage two-layer planning model;

[0085] Step 3: Solve the source-grid-storage two-level programming model of the distribution network;

[0086] Step 4: Demonstration application analysis.

[0087] In this embodiment, the application analysis of the demonstration example is as follows: ① Demonstration example overview: The improved IEEE33-node system is used to verify the source-grid-storage two-layer planning model of the medium-voltage distribution network considering flexibility. The structural topology is shown in Figure 3; Nodes 34-43 in Figure 3 are newly added load nodes, and the dotted lines represent the distribution lines to be built. The specific parameters of the lines to be built and the size of the newly added loads are shown in Appendix Tables A1 and A2. The total load peak in the current system is 6134.26+j3147.32kVA; assuming that each node is allowed to access DPV and ESS except the balancing node and the maximum number of DPV / ESS per node is 1, the total number of ESS planned configurations in the system is 4; the maximum output of DPV is 500kW, and the rated capacity of ESS is 0 -500MVA, the ESS power upper limit is 40kW, and the ESS charge state fluctuation range is allowed to be 0.2-0.9; to simplify the calculation example, the present invention unifies the line model to be built in the distribution system, selects the line model LGJ240, the unit impedance is 0.132+0.357jΩ / km, the transmission current upper limit of lines 1-2, 2-3, 3-4, and 4-5 is 500A, and the transmission current upper limit of the remaining lines is 300A; the load demand and DPV output historical data of the whole year in the present invention are shown in Figures 14 and 15 respectively; the time-of-use electricity price policy of Henan Province is adopted, and the equipment and related electricity prices are shown in Appendix Table A3; the photovoltaic daily output curve and daily load curve data of the whole year are clustered to obtain CH (+) The changing curve of is shown in Figure 4; As can be seen from Figure 4, the clustering effectiveness index CH (+) The probability of each typical scenario is shown in Table 1.

[0088] Table 1 Typical scenarios and their probabilities

[0089] ② Impact of flexibility index on distribution network planning scheme: To verify the impact of the flexibility index proposed in this invention on the planning scheme, the following two schemes are designed and compared based on the premise of installing four ESSs in the system:

[0090] Solution 1: Considering the impact of flexibility indicators on planning results, which is the solution proposed in this paper;

[0091] Solution 2: Do not consider the impact of flexibility index on the planning scheme, that is, remove the two indicators related to system flexibility in the objective function, and let F need 、F net All are 0, only economic performance is retained;

[0092] The planning results are shown in Table 2, and the comparison of various indicators is shown in Table 3. The ESS operating power results and SOC state curves for Scheme 1 are shown in Figures 6-9; the ESS operating power results and SOC state curves for Scheme 2 are shown in Figures 10-13.

[0093] Table 2 Comparison of configurations of Schemes 1 and 2

[0094] Table 3 Comparison of indicators and costs between Schemes 1 and 2

[0095] Comparing Tables 2 and 3, we can draw the following conclusions:

[0096] 1) In terms of total cost, Scheme 1, which takes into account the impact of flexibility, has a larger DPV installation capacity. Although the corresponding investment cost also increases, the combined effect of flexibility resources and the grid structure significantly reduces line loss costs and power purchase costs from the upper power grid, thereby significantly reducing the total system cost. Scheme 2 only considers economic efficiency in its objective function, and the equipment investment is smaller, but the total cost of Scheme 1 is 217,000 yuan lower than that of Scheme 2. In other words, considering the impact of flexibility in planning helps save total system cost.

[0097] 2) In terms of DPV construction, the PV penetration rate in Scheme 1 is 32.3%, 6.6% higher than that in Scheme 2. Data results show that considering the impact of flexibility indicators in planning can increase the system's DPV access capacity. Analysis shows that the lack of flexibility indicators in the Scheme 2 model leads to a smaller margin for dealing with uncertain events in order to achieve optimal economic performance, resulting in a corresponding reduction in DPV penetration.

[0098] 3) In terms of ESS distribution, the ESS in Scheme 2 is mostly concentrated near the DPV, and the PV storage is mostly installed at the end of the feeder. The reason for this is that the end line is far from the main transformer, resulting in high line losses along the way. Connecting the DPV and ESS at the end can significantly alleviate the problem of the end line voltage exceeding the lower limit. In contrast, the ESS installation in Scheme 1 is relatively dispersed, providing more loads with local charging and discharging opportunities compared to Scheme 2. Analysis of the ESS curve shows that the ESS relies on peak-valley electricity prices for arbitrage, absorbing electricity from the grid for charging during low load periods and discharging to the grid during peak load periods, further improving system economic efficiency.

[0099] appendix

[0100] Table A1 New load parameter table

[0101] Table A2 Parameters of the lines to be built in the distribution network

[0102] Table A3 Cost parameters related to the example

[0103] The present invention is a medium-voltage distribution network source-grid-storage two-layer planning method considering flexibility. Aiming at a series of problems brought to the distribution network by the large-scale access of DPV and loads, the present invention proposes a medium-voltage distribution network source-grid-storage two-layer planning model considering flexibility; the upper layer is the planning layer, which determines the planning scheme by selecting the site and sizing the lines to be built, DPV and ESS, and uses the upper planning scheme as the input of the lower optimization model; the lower layer is the operation layer, which performs multi-scenario operation simulation on the planning scheme determined by the upper layer to obtain the operation strategy with the best economic efficiency of the lower layer, and returns the operation parameters to the upper planning layer, and obtains the optimal planning scheme through continuous iteration between the upper and lower layers. The two-layer planning model is solved by fast non-dominated genetic algorithm and particle swarm algorithm respectively; the demonstration example analysis and comparison illustrate that the consideration proposed by the present invention The source-grid-storage coordinated planning model for medium-voltage distribution networks that considers flexible supply and demand has two characteristics and advantages: ① The influence of flexibility indicators is fully considered during the planning process. Although the investment in flexibility resources ESS is increased, it can reduce the overall annual comprehensive cost of the distribution network and improve the overall economy and flexibility of the distribution system; ② The coordinated planning of sources, grids and storage can fully utilize the peak-shaving and valley-filling effect of flexibility resources ESS. On the one hand, it can increase the penetration rate of DPV, and on the other hand, it can meet the needs of most loads for local power supply and consumption, thereby reducing the purchase of electricity by the upper power grid and better adapting to the development of new power systems with large-scale source-load access in the future; the present invention has the advantages of establishing flexibility level evaluation indicators from the matching of flexibility supply and demand and network transmission capacity, and improving the economy and flexibility of expansion planning and system operation.

Claims

1. A medium voltage distribution network source-grid-storage two-layer planning method considering flexibility, characterized by: The method comprises the following steps: Step 1: A two-level planning framework of distribution network source, grid and storage based on flexibility evaluation index; Step 2: Distribution network source-grid-storage two-layer planning model; Step 3: Solve the source-grid-storage two-level planning model of the distribution network; Step 4: Demonstration application analysis.

2. A medium voltage distribution network source-grid-storage two-layer planning method considering flexibility as claimed in claim 1, characterized in that: The distribution network source-grid-storage two-layer planning framework based on flexibility evaluation index in step 1 includes the following steps: Step 1.1: Construction of flexibility index; Step 1.2: The framework of the two-level planning model of distribution network source, network and storage.

3. A medium voltage distribution network source-grid-storage two-layer planning method considering flexibility as claimed in claim 2, characterized in that: The construction of the flexibility index in step 1.1 includes the following steps: Step 1.11: Net load fluctuation index: The net load fluctuation curve is affected by both load demand and DPV penetration rate. The net load fluctuation index refers to the uncertainty and volatility of the system under large-scale load and high penetration DPV access, which can clearly characterize the system flexibility demand from a time scale. The smaller the index is, the flatter the net load fluctuation curve is as a whole, indicating that the system is more adaptable to large loads and DPV access, the higher the overall source-load matching of the system, the better the overall flexibility of the distribution network, and the less flexibility resources need to be added or modified in the planning scheme. On the contrary, the system flexibility is worse, and more flexibility resources are needed in the planning scheme. Net load fluctuation index F in the sth scenario need (s) can be expressed as: Where P NL,t (s) is the net load size in time period t under scenario s; P NL,ave (s) is the average net load of the system in scenario s; T is the total number of time periods in the scheduling cycle. In this paper, the scheduling cycle is taken as 24h and the scheduling time interval is 1h; Step 1.12: Line load margin balance index: In order to balance the supply and demand of distribution network flexibility, in addition to the net load fluctuation rate index, reasonable requirements for grid flexibility are also required; although grid flexibility resources cannot directly provide flexibility for distribution networks, they can provide channels for flexibility resources to avoid the occurrence of "flexibility resources are unavailable" phenomenon. For this reason, it is necessary to propose line load margin balance to characterize the balance of distribution system grid flow; the load margin of a single line is described by the real-time load margin of the line transmission channel, reflecting the degree of line transmission congestion, and the calculation formula is: In the formula, F ij,t I is the load margin of a single line between nodes i and j in period t; ij,t I is the line transmission current between nodes i and j during period t; ij,max is the upper limit of the line transmission current between nodes i and j; F ij,t The larger the index value, the lower the possibility of line blockage. Therefore, the line load margin can be used to evaluate the grid transmission capacity. ij,t ≥0, indicating that the line load margin is sufficient and can adapt to load power fluctuations; when F ij,t When ≤0, it indicates that the line load margin is insufficient and line congestion will occur; therefore, the load margin balance index F of the power supply line in scenario s is net (s) can be expressed as: In the formula, F ave (s) is the mean value of the load margin of the power supply line; L is the set of system lines.

4. A medium voltage distribution network source-grid-storage two-layer planning method considering flexibility as claimed in claim 1, characterized in that: The distribution network source-network-storage dual-layer planning model in step 2 includes the following steps: Step 2.1: Upper level planning model; Step 2.2: Lower-level planning model.

5. A medium voltage distribution network source-grid-storage two-layer planning method considering flexibility as claimed in claim 4, characterized in that: The upper-level planning model in step 2.1 is specifically as follows: the upper-level planning model objective function is composed of the annual comprehensive cost of the distribution network, the net load fluctuation index and the line load margin balance index, and the decision variables are whether the line is newly built, the DPV location and capacity, and the ESS location and capacity; The objective function is: f1 = {min C total ,min F net ,min F need }(6), C total =C inv +C mai +C ope (7) Where f1 is the objective function of the upper-level planning model; C total is the annual comprehensive cost of the distribution network; C inv is the annual investment cost of the equipment; C mai is the annual operation and maintenance cost of the equipment; C ope is the annual operation and maintenance cost of the distribution system, including the annual network loss cost of the distribution system and the annual power purchase cost of the upper power grid; λ is the equipment maintenance cost coefficient; Λ is the lower-level operation scenario set; is the probability of scene s appearing; C ope (s) is the annual operation and maintenance cost of the distribution system under scenario s; α is the discount rate; k DPV , k ESS , k line are the service life of DPV, ESS and the line to be built respectively; c DPV 、c ESS are the investment costs per unit capacity of DPV and ESS respectively; c line is the investment cost per unit length of line; j∈Ω DPV , i∈Ω ESS are the sets of DPV and ESS installation nodes respectively; ij∈Ω line To build a line collection; E ESS,i , P DPV,j are the installed capacities of ESS and DPV at nodes i and j respectively; l ij is the length of the line to be built.

6. A medium voltage distribution network source-grid-storage dual-layer planning method considering flexibility as claimed in claim 5, characterized in that: The constraints of the upper-level planning model in step 2.1 include DPV, ESS, line load margin constraints and grid connectivity constraints; among them, ① DPV capacity constraint is: 0≤P DPV,i ≤P DPV,max (11), where: P DPV,i is the DPV installed capacity at node i; P DPV,max is the maximum installed capacity of DPV at node i;②ESS capacity constraint is: In the formula, are the upper and lower limits of the ESS installation capacity at node i respectively; ③ The line load margin constraint is: F ij,t ≥0(13); ④ The grid connectivity constraint is: when building lines for new load nodes, no ring network is formed, the distribution network is kept radial, and there are no isolated islands or points after DPV is connected.

7. A medium voltage distribution network source-grid-storage dual-layer planning method considering flexibility as claimed in claim 4, characterized in that: The lower-level planning model in step 2.2 is specifically as follows: the lower-level operation model takes the upper-level planning results as input, takes the optimal operation economy as the goal, and the objective function is the sum of the network loss cost of each scenario and the cost of purchasing electricity from the upper level. The decision variables are the capacity of DPV and ESS, and the objective function is: Where f2 is the objective function of the lower layer operation model; C loss (s) is the network loss cost under scenario s; C sub (s) is the cost of purchasing electricity from the upper power grid under scenario s; c t is the time-of-use electricity price during period t; P loss,t (s) is the network loss power in time period t under scenario s; P sub,t (s) is the power transmitted to the upper grid during period t under scenario s.

8. A medium voltage distribution network source-grid-storage double-layer planning method considering flexibility as claimed in claim 7, characterized in that: The constraints of the lower-level planning model in step 2.2 include power flow equality constraints, node voltage inequality constraints, ESS operation inequality constraints and power balance constraints; wherein, ① the power flow constraint is: Where P inj,i , Q inj,i Respectively represent the active power and reactive power injected at node i; U i , U j are the voltage amplitudes of nodes i and j respectively; n is the total number of nodes; G ij , B ij are the conductance and susceptance in the node admittance matrix respectively; δ ij is the voltage phase difference between nodes i and j; ② The node voltage constraint is: U min ≤U i ≤U max (16), where U max , U min are the upper and lower limits of the voltage amplitude respectively; ③The line current carrying capacity constraint is: I ij ≤η ij |I ij,max |(17), where I ij is the current flowing through the branch; η ij is the state variable of branch ij. When the value is 1, it means that node i is connected to node j. When the value is 0, it means that node i is not connected to node j. ij,max is the upper limit of the current amplitude of branch ij; ④ESS operation constraints are: In the formula, are the charging and discharging power of ESS at node j during period t; y j,t is a 0-1 variable, which is the charge and discharge status of the ESS during period t, where 1 indicates discharge and 0 indicates charge; is the charge of ESS at node j during period t; are the maximum and minimum charges of ESS at node j respectively; are the ESS charging and discharging efficiency respectively; are the charges at the initial and final periods of the ESS operation cycle at node j respectively; ⑤The power balance constraint is: Where P sub,t is the transmission power of the upper power grid in period t; P ESS,i,t is the power provided by the ESS at node i in period t, with positive for discharge and negative for charge; P DPV,j,t The power provided by DPV in time period t; P load,t is the total load demand of the distribution system during period t.

9. A medium voltage distribution network source-grid-storage dual-layer planning method considering flexibility as claimed in claim 1, characterized in that: The solution of the distribution network source-grid-storage two-layer planning model in step 3 is specifically as follows: the medium-voltage distribution network source-grid-storage two-layer planning model considering flexibility is essentially a combination planning of the distribution network grid structure and two flexible resources, DPV and ESS. The model simultaneously contains decision variables in the form of 0-1 integers and continuous decision variables such as DPV and ESS output. In order to reduce the difficulty of solving the mixed integer model, the initial grid structure is first determined by the minimum spanning tree method; then the fast non-dominated genetic algorithm, i.e., NSGA-II algorithm, is used to solve the multi-objective planning problem of optimal upper-level economy and flexibility, and then the particle swarm algorithm is used to solve the optimal model of lower-level operation economy; finally, the grid structure is determined to realize the site selection and sizing of DPV and ESS; the following steps are included: Step 3.1: Scenario clustering: Affected by the dual uncertainty of DPV output and load fluctuation, the complexity of the operating scenarios faced by the distribution network has been greatly increased. The planning and optimization operation of the distribution network need to meet the optimal comprehensive operation level in all scenarios. However, the changeable historical scenarios and solution efficiency make it difficult for the planning and optimization solution process to traverse every scenario. In order to take into account the overall operation of the distribution network throughout the year and improve the solution efficiency, an uncertainty planning method based on multi-scenario analysis is adopted; the correlation between the DPV output and the load power timing matching problem affects the change of the net load curve, which in turn affects the accuracy of the distribution network source network storage coordination planning and optimization operation; therefore, by selecting the daily characteristic indicators of DPV output and load power, and using the K-means clustering and hierarchical agglomerative clustering composite clustering algorithm to achieve scenario reduction, and using the clustering effectiveness index CH (+) Determine the optimal number of clusters and obtain typical scenarios and their probabilities; the selected daily characteristic indicators are: Where, X DPV,mean , X load,mean are the average values ​​of DPV daily power and load daily power respectively; X DPV,std , X load,std are the variances of DPV daily power and load daily power respectively; P DPV,t , P load,t is the DPV output and load power in period t; Step 3.2: Overall solution process.

10. A medium voltage distribution network source-grid-storage dual-layer planning method considering flexibility as claimed in claim 9, characterized in that: The overall solution process in step 3.2 includes the following steps: Step 3.21: Input the original load, the load to be planned and the grid structure to be expanded; Step 3.22: Generate an initial grid structure based on the minimum spanning tree algorithm; Step 3.23: Use the NSGA-II algorithm to solve the upper multi-objective planning model and obtain the location and capacity of DPV and ESS under the planning scheme with the best economy and flexibility; Step 3.24: Input the upper-level planning scheme as the initial value into the lower-level optimization model, use the particle swarm algorithm to solve the lower-level operation model, and determine whether it converges. If it converges, the optimal ESS charging and discharging power strategy under the planning scheme is obtained. If it does not converge, return to step 3.23; Step 3.25: Determine whether the convergence condition is met. If not, return to step 3.23; if satisfied, end and obtain the optimal distribution network source-grid-storage coordination planning scheme.

Images