A power distribution network energy storage dynamic programming method and system

Abstract

The application discloses a kind of power distribution network energy storage dynamic programming method and system, it is related to power distribution network control technical field.The application includes: when power distribution network is not installed energy storage, construct basic system value income model;After installing energy storage in power distribution network, construct dynamic programming configuration model;The difference between the present value of value income of each stage of power distribution network obtained by basic system value income model and the present value of value income of each stage of power distribution network obtained by dynamic programming configuration model is used as energy storage system value model;Improved particle swarm optimization algorithm is used to solve energy storage system value model, and the optimal energy storage configuration scheme is obtained, and according to the optimal energy storage configuration scheme, energy storage system is configured in power distribution network in stage.The application realizes the overall planning quantification and capture of energy storage multidimensional value in full time scale, makes energy storage configuration from cost benefit accounting upgrade to system value creation, and provides scientific decision basis for power distribution network source network load storage collaborative planning.

Description

Technical Field

[0001] This invention relates to the field of power distribution network control technology, and in particular to a dynamic planning method and system for power distribution network energy storage. Background Technology

[0002] In recent years, as the global energy structure has shifted towards a green and low-carbon direction, the distribution network, as the link between the main grid and users in the power system, has assumed increasingly important responsibilities. Directly facing users, the distribution network's development has always aimed to meet user needs, playing a crucial role in ensuring energy supply and guiding the transformation of the energy system. Driven by low-carbon goals and user demands, the distribution network faces the challenges of integrating a high proportion of distributed renewable energy and continuously growing load, increasing the difficulty of ensuring the safe and stable operation of the grid and guaranteeing power quality. Energy storage has advantages such as rapid power regulation and flexible energy management. Its application in the distribution network can not only effectively promote the consumption of renewable energy but also respond promptly to the continuously growing load demand on the user side. However, the relatively high investment cost of energy storage limits its development. Therefore, it is of great significance to plan and configure energy storage according to the load demand of the distribution network at each stage and the size of the installed capacity of new energy sources to maximize the value and benefits of energy storage and obtain higher economic returns.

[0003] When planning energy storage configurations, existing methods are typically based on pre-defined typical operating scenarios or simplified rules. First, based on historical load and electricity price data, candidate nodes are initially selected through experience or sensitivity analysis, with the single objective of minimizing investment costs or maximizing annual net revenue. Then, a single-objective optimization model centered on economics is established and solved using standard optimization algorithms, ultimately outputting an economically optimal energy storage configuration scheme under specific assumptions.

[0004] However, existing methods, based on pre-defined typical operating scenarios or simplified rules, narrowly equate the value of energy storage with its direct economic benefits in limited scenarios. This results in planning outcomes that are limited to short-term financial gains, underestimate the multi-dimensional comprehensive value of energy storage, and fail to maximize the system value of the distribution network throughout its entire life cycle. Summary of the Invention

[0005] Therefore, it is necessary to provide a dynamic planning method and system for energy storage in power distribution networks to address the aforementioned technical problems.

[0006] This invention provides a dynamic planning method for energy storage in distribution networks, comprising: When energy storage is not installed in the distribution network, a basic system value-benefit model is constructed with the goal of maximizing the present value of the distribution network at each stage; after energy storage is installed in the distribution network, a dynamic programming configuration model is constructed with the goal of maximizing the present value of the distribution network at each stage after the installation of energy storage. The difference between the present value of distribution network value at each stage obtained from the basic system value benefit model and the present value of distribution network value at each stage obtained from the dynamic programming configuration model is used as the value model of the energy storage system. An improved particle swarm optimization algorithm is used to solve the value model of the energy storage system to obtain the optimal energy storage configuration scheme. Based on the optimal energy storage configuration scheme, the energy storage system is configured in stages in the distribution network. The improved particle swarm optimization algorithm is obtained by introducing adaptive inertia weights and crossover mutation operations in the search phase of the particle swarm optimization algorithm.

[0007] Optionally, based on the following formula, when energy storage is not installed in the distribution network, a basic system value-benefit model is constructed with the objective of maximizing the present value of the value benefits of the distribution network at each stage: ; ; ; ; P d,t = P v,t - P w,t - P pv,t + P loss,t ; in, H 1 represents the present value of the basic system's value and benefits during the planning phase. Y For the planning period, C v,y Indicates the distribution network number y Annual electricity sales revenue C p,y Indicates the distribution network number y The annual cost of purchasing electricity from the main grid. C e,y Indicates the distribution network number y The annual carbon emission costs borne by the company ρ Indicates the discount rate. e u,t This indicates the time-of-use electricity price for electricity sold through the distribution network. P v,t Indicates the distribution network in t Total load for the time period, Δ t Indicates the duration of a time period. P d,t Indicates that the distribution network is connected to the main grid. t Total power purchased during the time period P w,t , P pv,t The distribution network is respectivelyt Total wind / solar power absorbed during the time period P loss,t For the distribution network in t Total network loss during the time period e g,t Indicates the distribution network in t The time-of-use electricity price for electricity purchased from the main grid during certain periods.

[0008] Optionally, the allocation of carbon emission credits in carbon emission costs can be determined based on the following formula: ; in, E r The annual carbon emission allowances allocated to the power distribution network free of charge. ς The carbon emission allocation coefficient per unit of electricity; The system carbon emissions in carbon emission costs are determined based on the following formula: ; in, E c This represents the system's total annual carbon emissions. λ t for t Carbon emission factors over a given period of time.

[0009] Optionally, it also includes adding node power balance constraints, line current carrying constraints, and node voltage constraints when constructing the basic system value benefit model: Add node power balance constraints based on the following formula: ; ; in, C To remove the line ij External and Node j A set of connected lines, P ij,t For the line ij exist t The active power flowing through during a given time period. Q ij,t For the line ij exist t The reactive power flowing through the time period r ij For the line ij The resistor on x ij For the line ij The reactance on, f ij,t For the line ij exist t The square of the current value flowing through the time period. Pload,j,t For nodes j exist t Active load during a given time period Q load,j,t For nodes j exist t Reactive load during a given period For nodes j superior t Energy storage during a given period of time provides active power output. For nodes j superior t Reactive power output of energy storage during certain periods. P w,j,t For nodes j superior t Wind power output during a given period Q w,j,t For nodes j superior t Wind power reactive power output during certain periods P pv,j,t For nodes j superior t Solar power output during certain periods Q pv,j,t For nodes j superior t Solar reactive power output during certain periods; Add line current-carrying constraints based on the following formula: ; in, v i,t For nodes i In the t The square of the voltage value over a time period For the line ij The upper limit of the square of the current; Add node voltage constraints based on the following formula: ; ; in, v max and v min These represent the upper and lower limits of the square of the voltage, respectively.

[0010] Optionally, based on the following formula, after installing energy storage in the distribution network, a dynamic programming configuration model is established using the difference method to maximize the present value of the value benefits of the distribution network at each stage after the installation of energy storage: ; in, H 2 represents the present value of the distribution network's revenue after the installation of energy storage devices. For the first time after installing energy storage devicesy The annual electricity purchase cost from the main grid by the distribution network operator. For the first time after installing energy storage devices y The annual carbon emission costs borne by the power distribution network C ess This refers to the various investment costs of the energy storage device during the planning period.

[0011] Optionally, the power purchase cost of a distribution network including energy storage can be determined based on the following formula: ; ; ; in, To facilitate the installation of energy storage devices in the distribution network from the main grid t Total power purchased during the time period for t Charging power of time-of-use energy storage for t Discharge power of time-limited energy storage After installing energy storage devices, the distribution network will be in t Total network loss during the time period; The carbon emission cost of a system including energy storage is determined based on the following formula: ; ; in, This represents the total annual carbon emissions of the system after the installation of energy storage devices. The energy storage investment cost is determined based on the following formula: ; ; ; ; in, C s For energy storage installation costs, k For energy storage lifespan, N For the number of energy storage installations, For energy storage in the first y Annual unit power cost For energy storage in the first y Annual unit capacity cost For the system in the first y The first year installed n The rated power of the energy storage unit, For the system in the first y The first year installed n The rated capacity of the energy storage, The annual operating and maintenance cost per unit capacity of energy storage. For the first y The total energy storage capacity installed in the system annually. C p The replacement cost of energy storage within the planning cycle, M ess For installation at the y The number of times annual energy storage needs to be replaced within the planning cycle. For installation at the y The unit power cost of replacing energy storage devices annually. For installation at the y The unit capacity cost of replacing energy storage devices annually INT () is the floor function.

[0012] Optionally, it also includes adding energy storage operation constraints and candidate node installation constraints when constructing the dynamic programming configuration model: Energy storage operation constraints are added based on the following formula: ; ; ; ; Add installation constraints for candidate nodes based on the following formula: ; in, I These are the nodes to be installed in the system. X i For binary decision variables, N ess This represents the number of energy storage devices connected.

[0013] Optionally, an improved particle swarm optimization algorithm is used to solve the value model of the energy storage system to obtain the optimal energy storage configuration scheme, specifically including: A set of particles is randomly generated based on the energy storage system value model, with each particle representing an energy storage configuration scheme for a distribution network. The fitness value of each particle is determined by a fitness function, which characterizes the system value of the corresponding energy storage configuration scheme. The particle with the largest fitness value is taken as the global optimal solution. For each particle other than the global optimum, the inertia weight of the particle is dynamically adjusted through adaptive inertia weight to update the velocity and position of each particle; it is determined whether the difference between each updated particle and the global optimum is less than a set threshold; if so, a crossover mutation operation is performed on the particle, and the individual optimum and the global optimum are updated through the fitness value of the particle. The process continues until the maximum number of iterations is reached or the global optimal solution shows no significant change in multiple consecutive iterations. The final global optimal solution is then obtained, and the energy storage configuration scheme corresponding to the final global optimal solution is taken as the optimal energy storage configuration scheme.

[0014] Optionally, the particle's inertial weight can be dynamically adjusted based on the following formula using adaptive inertial weights: ; ; in, X i ( k ) is a particle i In the k The difference distance between the particle and the best particle in the population at the next iteration. x max The maximum value of the particle's position. x min This represents the minimum value of the particle's position. D To solve for spatial dimensions, w i ( k ) is a particle i In the k Inertia weights in the next iteration a and b For auxiliary values; The crossover and mutation operation of particles is performed based on the following formula: ; ; in, r A random number between 0 and 1.

[0015] This invention provides a dynamic planning system for energy storage in a distribution network, comprising: The model building module is used to construct a basic system value-benefit model when the distribution network has no energy storage installed, with the goal of maximizing the present value of the value-benefit of the distribution network at each stage; and to construct a dynamic programming configuration model after the distribution network has energy storage installed, with the goal of maximizing the present value of the value-benefit of the distribution network at each stage after the installation of energy storage. The value determination module is used to take the difference between the present value of the distribution network at each stage obtained from the basic system value revenue model and the present value of the distribution network at each stage obtained from the dynamic programming configuration model as the value model of the energy storage system. The planning module uses an improved particle swarm optimization algorithm to solve the value model of the energy storage system, obtains the optimal energy storage configuration scheme, and configures the energy storage system in the distribution network in stages according to the optimal energy storage configuration scheme. The improved particle swarm optimization algorithm is obtained by introducing adaptive inertia weights and crossover mutation operations in the search phase of the particle swarm optimization algorithm.

[0016] The dynamic planning method and system for energy storage in power distribution networks provided in this invention have the following advantages compared with the prior art: This invention introduces adaptive inertia weights into the improved particle swarm optimization algorithm to balance global and local search capabilities, and integrates crossover and mutation operations to avoid premature convergence. This ensures that the energy storage configuration sequence that truly maximizes the system's benefits throughout its entire lifecycle is found in the complex dynamic programming solution space. It realizes the comprehensive quantification and capture of the multidimensional value of energy storage across the entire time scale, elevating energy storage configuration from cost-benefit accounting to system value creation, and providing a scientific decision-making basis for the coordinated planning of power grid, source, grid, load and storage. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the dynamic planning configuration of energy storage in a distribution network according to an embodiment of a dynamic planning method for energy storage in a distribution network. Figure 2 A graph illustrating the carbon trading mechanism of a dynamic planning method for energy storage in a distribution network provided in one embodiment; Figure 3 An adaptive inertia weight curve is provided for a dynamic planning method for energy storage in a distribution network in one embodiment. Figure 4 A flowchart of an improved particle swarm optimization algorithm for a power distribution network energy storage dynamic planning method is provided in one embodiment; Figure 5 This is an IEEE 33-node distribution network structure diagram of a distribution network energy storage dynamic planning method provided in one embodiment; Figure 6 A typical daily wind power output curve is shown in one embodiment of a dynamic planning method for energy storage in a distribution network. Figure 7 A typical daily photovoltaic output curve is shown in one embodiment of a dynamic planning method for energy storage in a distribution network. Figure 8 This is a graph showing the changes in energy storage capacity and present value of value returns for a dynamic planning method for energy storage in a power distribution network provided in one embodiment. Figure 9 This is a schematic diagram of CO2 emission factor data and typical daily load curves for a power distribution network energy storage dynamic planning method provided in one embodiment. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] Based on the difference method in system value assessment, this paper first establishes a basic system value and revenue model, corresponding to the "original system" part of the distribution network without energy storage in the difference method, and calculates the present value of the distribution network revenue without energy storage during the planning period. Then, with the objective function of maximizing the present value of the distribution network value and revenue after energy storage installation, a dynamic planning configuration model for energy storage is established, corresponding to the "new system" part of the distribution network after energy storage installation in the difference method. The optimal configuration of energy storage is carried out according to the load demand of the system and the installed capacity of new energy sources during the planning stage. Finally, a simulation example is selected from the IEEE 33-node distribution network to obtain the energy storage planning configuration scheme and analyze the change process of energy storage value and revenue at each stage, verifying the effectiveness of the proposed method and model.

[0020] This invention provides a dynamic planning method for energy storage in distribution networks, the basic framework of which is as follows: Figure 1 As shown, the method includes: When energy storage is not installed in the distribution network, a basic system value-benefit model is constructed with the objective of maximizing the present value of the distribution network's value benefits at each stage. After energy storage is installed in the distribution network, a dynamic programming configuration model is constructed with the objective of maximizing the present value of the distribution network's value benefits at each stage after energy storage installation.

[0021] The difference between the present value of distribution network value at each stage obtained from the basic system value benefit model and the present value of distribution network value at each stage obtained from the dynamic programming configuration model is used as the value model of the energy storage system.

[0022] An improved particle swarm optimization (PSO) algorithm is used to solve the value model of the energy storage system, obtaining the optimal energy storage configuration scheme. Based on this optimal scheme, the energy storage system is then configured in stages within the distribution network. The improved PSO algorithm is achieved by introducing adaptive inertia weights and crossover / mutation operations during the search phase of the PSO algorithm.

[0023] Preferably, an improved particle swarm optimization algorithm is used to solve the value model of the energy storage system to obtain the optimal energy storage configuration scheme, specifically including: A set of particles is randomly generated based on the energy storage system value model, with each particle representing an energy storage configuration scheme for a distribution network. The fitness value of each particle is determined by a fitness function, which characterizes the system value of the corresponding energy storage configuration scheme. The particle with the largest fitness value is then selected as the global optimal solution.

[0024] For each particle other than the global optimum, its inertia weight is dynamically adjusted using adaptive inertia weighting to update its velocity and position. The difference between the updated particle and the global optimum is then checked against a set threshold. If so, a crossover mutation operation is performed on the particle, and its fitness value is used to update both the individual optimum and the global optimum.

[0025] The process continues until the maximum number of iterations is reached or the global optimal solution shows no significant change in multiple consecutive iterations. The final global optimal solution is then obtained, and the energy storage configuration scheme corresponding to the final global optimal solution is taken as the optimal energy storage configuration scheme.

[0026] A specific embodiment of the present invention is provided: 1. Basic system value and revenue model.

[0027] 1.1 Objective function.

[0028] First, with the goal of maximizing the present value of distribution network revenue during the planning phase, a basic system value-revenue model is established, the objective function of which can be expressed as: (1) in, H 1 represents the present value of the basic system's value and benefits during the planning phase; Y For the planning period; C v,y Indicates the distribution network number y Annual electricity sales revenue; C p,y Indicates the distribution network number y The cost of purchasing electricity from the main grid annually; C e,y Indicates the distribution network number y The annual carbon emission costs borne by the organization; ρ This represents the discount rate.

[0029] 1) Annual electricity sales revenue.

[0030] (2) in, e u,t This indicates the time-of-use electricity price for electricity sold through the distribution network; P v,t Indicates the distribution network in t Total load for the time period; Δ t Indicates the duration of a time period.

[0031] 2) Annual electricity purchase cost.

[0032] (3) in, P d,t = P v,t - P w,t - P pv,t + P loss,t This indicates that the distribution network is connected to the main grid. t Total power purchased during the time period; Pw,t , P pv,t The distribution network is respectively t Total wind / solar power absorbed during the time period; P loss,t For the distribution network in t Total network loss during the time period; e g,t Indicates the distribution network in t The time-of-use electricity price for electricity purchased from the main grid during certain periods.

[0033] 3) Carbon emission costs.

[0034] To achieve low-carbon electricity, carbon emissions need to be considered in the planning and operation of the power distribution network. To further demonstrate the effectiveness of emission reduction efforts and achieve carbon reduction targets, a tiered carbon trading mechanism should be implemented to limit the purchase of emission rights within the system. The trading mechanism would be as follows: Figure 2 As shown.

[0035] ① Allocation of carbon emission allowances.

[0036] Traditional thermal power plant carbon emission allowances are typically allocated using a composite method, primarily through free allocation, supplemented by paid allocation. Referring to this calculation method, the system's free carbon emission allowances for the corresponding planning period are calculated as follows:

[0037] (4) in, E r This represents the annual carbon emission allowance allocated to the distribution network free of charge. ς The carbon emission allocation coefficient represents the amount of carbon emissions per unit of electricity.

[0038] ②System carbon emissions.

[0039] (5) in, E c This indicates the system's total annual carbon emissions; λ t for t Carbon emission factors over a given period of time.

[0040] ③ Tiered carbon trading mechanism.

[0041] (6) in, γ The initial carbon trading price; μ , φμ These represent the price increases for the second / third tiers of carbon trading after the system's carbon emissions exceed the free carbon emission allowance. φ ≥2; ηThe range length for carbon emissions.

[0042] 1.2 Constraints of the Energy Storage Dynamic Programming Configuration Model.

[0043] 1) Node power balance constraints.

[0044] (7) in, C Indicates except for the line ij External and Node j A set of connected lines; P ij,t and Q ij,t The lines are respectively ij exist t Active / reactive power flowing through a given time period; r ij , x ij They represent the lines respectively. ij Resistance and reactance; f ij,t For the line ij exist t The square of the current flowing through the time period; P load,j,t and Q load,j,t They are nodes j exist t Active / reactive load during a given time period; and They are nodes j superior t Active / reactive power output of energy storage during a given time period; P w,j,t and Q w,j,t They are nodes j superior t Active / reactive power output of wind power during a given time period; P pv,j,t and Q pv,j,t They are nodes j superior t Photovoltaic active / reactive power output during a given time period.

[0045] The regulations stipulate that power in the distribution network is not allowed to be fed back to the upstream power grid, that is: (8) 2) Line current carrying capacity constraints.

[0046] (9) in, v i,t For nodesi In the t The square of the voltage value over a time period and The lines are respectively ij Active power and reactive power on For the line ij The square of the current, For the line ij The upper limit of the square of the current. There is a quadratic term on the right side of equation (9), which can be converted into a second-order cone form, as shown in equation (10).

[0047] (10) 3) Node voltage constraints.

[0048] (11) Considering that the quadratic term in equation (11) is much smaller than the other two terms, it can be ignored. To ensure that the node voltage is within the feasible region, the following constraints should also be added:

[0049] (12) in, v max and v min These represent the upper and lower limits of the square of the voltage, respectively.

[0050] 2. Dynamic programming configuration model for energy storage.

[0051] 2.1 Objective function.

[0052] To maximize the present value of the distribution network's revenue including newly planned energy storage, a dynamic programming configuration model for energy storage is established. Its objective function can be expressed as: (13) in, H 2 represents the present value of the distribution network's value after the installation of energy storage devices; Indicates the first day after the installation of the energy storage device y Annual electricity purchase fees from the main grid by the distribution network operator; Indicates the first day after the installation of the energy storage device y The annual carbon emission costs borne by the power distribution network; C ess This indicates the various investment costs of the energy storage device during the planning period.

[0053] 1) Electricity purchase cost of distribution network including energy storage.

[0054] (14) in, This indicates that after the energy storage device is installed, the distribution network will be connected to the main grid. t Total power purchased during the time period; ,in, and They represent in t The charging / discharging power of time-of-use energy storage After installing energy storage devices, the distribution network will be in t Total network loss during the time period.

[0055] 2) Carbon emission costs of systems including energy storage.

[0056] ① Calculation of system emissions.

[0057] (15) ② Calculation of carbon trading costs.

[0058] (16) in, This indicates the total annual carbon emissions of the system after the installation of energy storage devices.

[0059] 3) Energy storage investment costs.

[0060] Considering that the initial investment cost of installing energy storage devices in each planning stage is relatively large, the installation cost is converted into the annual equivalent investment cost according to the service life of the energy storage, as shown in equations (17), (18), (19) and (20).

[0061] (17) (18) (19) (20) in, C s This indicates the cost of energy storage installation; k This refers to the lifespan of energy storage. N The number of energy storage installations; , Energy storage in the first y Annual cost per unit power / capacity; , The system at the 1st y The first year installed n The rated power / capacity of each energy storage unit; Annual operating and maintenance costs per unit capacity of energy storage; For the first y Total energy storage capacity installed in the system annually; C pThis indicates the replacement cost of energy storage within the planning cycle; M ess Indicates installation at the y The number of times annual energy storage needs to be replaced within the planning cycle; , These respectively indicate the installation location at the... y Annual cost per unit power / capacity when replacing energy storage devices; INT () is the floor function.

[0062] 2.2 Constraints.

[0063] The dynamic programming configuration model for energy storage includes two parts: distribution network operation constraints and energy storage-related constraints. The distribution network operation constraints are consistent with the constraints in the basic system value and benefit model.

[0064] 1) Constraints on energy storage operation.

[0065] (twenty one) (twenty two) (twenty three) (twenty four) 2) Installation constraints for candidate nodes.

[0066] (25) in, I This indicates the nodes to be installed in the system; X i For binary decision variables, a value of 1 represents a node. i Connect the energy storage device; a value of 0 indicates that it is not connected. N ess This indicates the number of energy storage devices connected.

[0067] 3. Evaluation indicators.

[0068] Net Present Value (NPV) is used. NPV ) and Internal Rate of Return (IRR) IRR The results of energy storage configuration are analyzed as an economic analysis indicator.

[0069] 1) Net Present Value NPV .

[0070] NPV This refers to the sum of the present values ​​of the net cash flows of a project in each year, discounted to the initial stage using a certain discount rate. The calculation formula is: (26) in,( CI - CO ) y Indicates the first y Net cash flow for the year; CI Cash inflow is equivalent to energy storage revenue; CO This refers to cash outflows, i.e., various cost expenditures.

[0071] 3) Internal Rate of Return IRR .

[0072] IRR It is an important economic evaluation indicator that reflects the degree of acceptance of currency devaluation during project operation and can be used to assess the feasibility of project investment. IRR It is the discount rate at which the total present value of cash inflows equals the total present value of cash outflows when the net present value is zero. Generally, IRR A return rate greater than or equal to the benchmark rate of return indicates that the project is feasible. The calculation formula can be expressed as:

[0073] (27) 4. Solution algorithm.

[0074] To address the aforementioned energy storage optimization configuration problem, the Particle Swarm Optimization (PSO) algorithm can be used to solve it. As a swarm intelligence optimization algorithm, PSO initializes with a swarm of randomly generated particles. Each particle iteratively evolves based on its current position, historical position, and population information to search for the global optimum. The particle velocity and position update formulas are shown below:

[0075] (28) (29) in, v id ( k )and v id ( k +1) represent particles respectively i In the k Second and third k The speed in +1 iterations; x id ( k )and x id ( k +1) represent particles respectively i In the k Second and third k Position in +1 iteration; w This is the inertia weighting coefficient;c 1. c 2 represents the particle learning factor; r 1. r 2 represents a random value distributed between the intervals (0,1); p id ( k ) is a particle i In the d The optimal value of the dimension; g d ( k ) for the population in the 1st d The optimal value of dimension.

[0076] In the conventional PSO algorithm, the inertia weight coefficient is used during the problem-solving process. w The selection of values ​​lacks guidance, and when particles search for a local optimum, population diversity is lost too quickly, easily leading to local optima. Therefore, as... Figure 4 The improved PSO algorithm is shown in the diagram:

[0077] (1) Adaptive adjustment of inertia weight: Conventional PSO algorithms do not consider particle characteristics during iteration and mainly use linear or nonlinear decreasing methods to adjust inertia weights. w Therefore, calculations are performed. w The selection of values ​​lacks guidance. The difference between the particle's current position and the optimal particle position in the population is chosen as the guiding factor, and the periodicity and "convex-concave" property of trigonometric functions are utilized to... w The values ​​are dynamically adjusted to improve the algorithm's search capability. The calculation formula is shown below:

[0078] (30) (31) in, X i ( k ) is a particle i In the k The difference distance between the particle and the best particle in the population at the next iteration; x max , x min These represent the maximum and minimum values ​​of the particle's position, respectively. D To solve for the spatial dimension; w i ( k ) is a particle i In the k Inertia weights in the next iteration; a and b Adjust as an auxiliary value w It is within a reasonable range.

[0079] when a =1, b When =1, the adaptive inertia weight curve is as follows: Figure 3 As shown. By Figure 3 It can be seen that when the difference is small, the inertia weight w When the value of is small, the optimization is performed at a slower speed, giving the particle better local search capabilities; when the difference is large, the inertial weight [is used]. w The value of is also relatively large. At this time, the particle can have a better global search capability by searching at a faster speed.

[0080] (2) Crossover variation The crossover and mutation mechanism can effectively solve the problems of premature convergence and easy getting trapped in local optima in the PSO algorithm during iterative optimization. The difference between the particle position and the global optimum of the population is used as the criterion for mutation, and the specific operation can be divided into the following five steps:

[0081] 1) Calculate the difference X minimum value X min Cross rate P c Variation rate P m .

[0082] 2) Determine the particle X i The size. If X i < X min At this point, it is easy to get trapped in a local solution, requiring analysis of the particles. i Perform crossover and mutation operations; otherwise, proceed to step 5.

[0083] 3) For particles i Perform mutation operations. For particles i Each positional component is selected from a random number in the range [0,1]. r id ,like r id < p m Then for particles i The d The dimensional components are initialized as shown in equation (32).

[0084] 4) Perform a crossover operation on the position vectors of the mutated particles. If r id < p c Then for particles i The position vector of the first dCrossover operations are performed on the dimensions, with the global optimal solution of the population as the crossover target. The operation method is shown in equation (4.33).

[0085] 5) End the crossover and mutation operation.

[0086] (32) (33) in, r A random number between 0 and 1.

[0087] Based on the conventional PSO algorithm, inertia weights w The adaptive adjustment enhances the particle's optimization search capability, improving convergence speed while maintaining algorithm accuracy. Simultaneously, the crossover and mutation operations on particles expand population diversity, effectively addressing issues such as premature convergence and susceptibility to local optima during computation, thus improving algorithm accuracy.

[0088] 5. Case Analysis.

[0089] 5.1 Basic Data.

[0090] To verify the effectiveness of the established energy storage planning and configuration model, a case study analysis was conducted using the IEEE 33-node distribution network test system. The IEEE 33-node distribution network comprises 32 branches, with a system base capacity of 10 MVA and a base voltage of 12.66 kV. The planning period for the distribution network in the scheme is 16 years, with each phase lasting 4 years. Taking 2022 as the base year, the load growth rate for the first 8 years is 2.3%, and the annual growth rate from the 9th year onwards is 4.6%, with an annual growth rate of 6.7% for wind power and photovoltaic installed capacity.

[0091] The system structure of the distribution network in the initial planning stage, i.e., the base year, is as follows: Figure 5 As shown, nodes 9, 20, and 24 are connected to distributed wind power with a rated capacity of 1.5MW, and nodes 5 and 15 are connected to distributed photovoltaic power with a rated capacity of 1.0MW. The year is divided into four typical scenarios: spring, summer, autumn, and winter. The typical daily wind and photovoltaic power output curves for the baseline year are shown below. Figure 6 and Figure 7 Carbon emission allocation factor per unit of electricity ς The value is 0.65 t / (MWh), the initial carbon trading price. γ The price is 254.85 yuan / t, with a range length of [missing information]. η =1000t, the increase in tiered carbon trading prices μ =25%, ϕ The value is 2. The time-of-use electricity price for the distribution network purchasing electricity from the main grid is shown in Table 1, and the electricity price for the distribution network selling electricity is 1.25e. g,tThe CO2 emission factor data for each period and the typical daily load curve for the baseline year are shown below. Figure 9 As shown ( Figure 9 (In the middle: green solid line represents spring, red solid line represents summer, purple solid line represents autumn, and blue solid line represents winter).

[0092] Table 1. Time-of-use electricity pricing table Lithium iron phosphate batteries were selected as the energy storage device. The unit capacity cost of energy storage in the four planning phases was RMB 1500 / kWh, RMB 1300 / kWh, RMB 1100 / kWh, and RMB 900 / kWh, respectively; the annual operation and maintenance cost per unit capacity was RMB 0.47 / kWh; and the unit power cost in each phase was RMB 500 / kW, RMB 450 / kW, RMB 400 / kW, and RMB 350 / kW, respectively. The number of energy storage devices connected was [not specified]. N ess ≤5, Energy storage lifespan k =10 years, discount rate ρ =3%, charge / discharge efficiency β c = β d =90%, with a state of charge range of 10% to 90%.

[0093] 5.2 Analysis of Energy Storage Planning and Configuration Results.

[0094] By solving the basic system value-benefit model and the dynamic programming configuration model of energy storage, the present value of system value-benefit under different planning stages is shown in Table 2, and the dynamic programming configuration results of energy storage and the present value of system revenue under different planning stages are shown in Table 3.

[0095] Table 2 Present Value of System Revenues at Different Planning Stages (excluding energy storage) Table 3. Dynamic energy storage configuration results and present value of system revenue at different planning stages Comparing Tables 2 and 3, it can be seen that planning and configuring energy storage at different planning stages reduces both system electricity purchase costs and carbon emission costs to varying degrees. In the initial planning stage, energy storage purchase costs are relatively high, the configured capacity is small, and its value is not yet apparent, resulting in high system electricity purchase costs and carbon emission costs. As the load on the distribution network and the installed capacity of wind power gradually increase, the configured energy storage capacity in the system continuously increases, and its value becomes increasingly prominent. The present value of the reduced electricity purchase revenue brought by energy storage in Stage 1 increases from RMB 689,900 / stage to RMB 4,807,400 / stage in Stage 4, and the present value of emission reduction benefits increases from RMB 75,200 / stage to RMB 1,932,800 / stage. Therefore, the system value of energy storage shows a continuously increasing trend during the planning stage.

[0096] To compare with the dynamic planning configuration scheme for energy storage, and referring to the 2021 requirements for new energy storage allocation, energy storage is configured in the form of "new energy + energy storage," with energy storage devices allocated according to a standard of 20% of the installed capacity, a backup power duration of 2 hours, and the energy storage site selection scheme is consistent with the above scheme. The results of the conventional energy storage configuration are shown in Table 4.

[0097] Table 4. Conventional configuration results and present value of system revenue under different planning stages As can be seen from Table 4, the energy storage capacity configured in the first stage of the conventional configuration scheme is relatively large. As the energy storage capacity of each stage continues to increase, its value returns also show a continuous upward trend during the planning stage.

[0098] 5.3 Economic analysis of energy storage.

[0099] Table 5 compares the inputs and outputs of the two configuration schemes during the planning phase. Considering the issues of partial energy storage replacement and the long remaining service life of energy storage during the planning phase, the recovery value of energy storage at the end of the planning period is calculated based on the loss cost (initial investment cost × 9% / year).

[0100] Table 5. Economic Comparison of Energy Storage Investment Value As shown in Table 5, the conventional energy storage configuration scheme requires relatively little investment throughout the planning period, with a total present value of energy storage revenue of RMB 13.3063 million and a net present value of only RMB 2.7464 million, resulting in an internal rate of return (IRR) of 6.55%. In the dynamic energy storage configuration scheme, the total present value of revenue reaches RMB 16.5339 million, RMB 3.2276 million higher than the conventional configuration scheme, and the net present value of energy storage reaches RMB 3.7902 million, RMB 1.0438 million higher than the conventional configuration scheme. Furthermore, its IRR is higher than both the benchmark discount rate and the discount rate of the conventional configuration scheme, demonstrating superior economic efficiency.

[0101] Depend on Figure 8 It can be seen that the conventional energy storage configuration scheme has a relatively high capacity in the first phase of the planning cycle, while the capacity increase is relatively low in subsequent phases, showing a trend of rapid initial growth followed by slower growth in value returns. The dynamic energy storage configuration scheme, considering the relatively low installed capacity of new energy sources and high energy storage costs in the initial planning stage, only invests a small amount of energy storage in the first phase to achieve low-charge, high-discharge goals, resulting in significantly lower value returns compared to the conventional scheme in this phase. However, as the load and installed capacity of new energy sources in the distribution network continue to grow, the system's demand for energy storage increases, thus requiring the investment of more energy storage capacity in subsequent phases, leading to higher value returns than the conventional scheme. When the distribution network is in the fourth phase of the planning cycle, the installed capacity of new energy sources in the system is already at a high level, and the original distribution network can no longer effectively absorb the output of new energy sources. Therefore, the energy storage installed in the system can significantly improve its absorption capacity, thereby greatly reducing electricity purchase costs and carbon emission costs, effectively improving the value returns of energy storage.

[0102] 6. Summary.

[0103] Based on the difference method in system value assessment, a basic system value benefit model is first established. Then, a dynamic programming configuration model for energy storage is established with the objective function of maximizing the present value of the distribution network value benefit after energy storage configuration. Finally, the proposed model is verified using an IEEE 33-node distribution network in a numerical example, and compared with conventional planning and configuration schemes, leading to the following conclusions: 1) Dynamic planning configurations incorporating energy storage offer significant economic advantages compared to conventional energy storage configurations. Within the planning period, energy storage, through its ability to charge and discharge at low speeds and flexibly absorb renewable energy sources, can effectively reduce electricity purchase costs and carbon emission costs.

[0104] 2) The value and benefits of energy storage show a year-on-year increasing trend during the planning period. In the early stage of planning, the load demand of the distribution network is high and the installed capacity of new energy is relatively low. Energy storage mainly plays a role of low charging and high discharging in the system, resulting in low value and benefits. However, the annual growth rate of new energy in the distribution network is faster than the annual growth rate of load. By the end of the planning stage, the proportion of new energy in the system is high, and energy storage can effectively realize the absorption of new energy by leveraging its flexible adjustment characteristics, thus significantly improving the value and benefits.

[0105] Based on the same inventive concept, embodiments of the present invention also provide a dynamic planning system for energy storage in a distribution network, the system comprising: The model building module is used to construct a basic system value-benefit model when the distribution network has no energy storage installed, with the objective of maximizing the present value of the distribution network's value-benefit at each stage. After energy storage is installed in the distribution network, a dynamic programming configuration model is constructed with the objective of maximizing the present value of the distribution network's value-benefit at each stage after energy storage installation.

[0106] The value determination module is used to take the difference between the present value of the distribution network at each stage obtained from the basic system value benefit model and the present value of the distribution network at each stage obtained from the dynamic programming configuration model as the value model of the energy storage system.

[0107] The planning module employs an improved particle swarm optimization algorithm to solve the value model of the energy storage system, obtaining the optimal energy storage configuration scheme. Based on this optimal scheme, the energy storage system is then configured in stages within the distribution network. The improved particle swarm optimization algorithm is achieved by introducing adaptive inertia weights and crossover / mutation operations during the search phase of the algorithm.

[0108] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A dynamic programming method for energy storage in a distribution network, characterized in that, include: When energy storage is not installed in the distribution network, a basic system value-benefit model is constructed with the goal of maximizing the present value of the value benefits of the distribution network at each stage. After energy storage is installed in the distribution network, a dynamic programming configuration model is constructed with the goal of maximizing the present value of the value returns of the distribution network at each stage after the installation of energy storage. The difference between the present value of distribution network value at each stage obtained from the basic system value benefit model and the present value of distribution network value at each stage obtained from the dynamic programming configuration model is used as the energy storage system value model. An improved particle swarm optimization algorithm is used to solve the value model of the energy storage system to obtain the optimal energy storage configuration scheme. The energy storage system is then configured in stages in the distribution network according to the optimal energy storage configuration scheme. The improved particle swarm optimization algorithm is obtained by introducing adaptive inertia weights and crossover mutation operations in the search phase of the particle swarm optimization algorithm.

2. The dynamic planning method for energy storage in a distribution network as described in claim 1, characterized in that, Based on the following formula, when energy storage is not installed in the distribution network, a basic system value-benefit model is constructed with the objective of maximizing the present value of the value benefits of the distribution network at each stage: ； ； ； ； P d,t = P v,t − P w,t − P pv,t + P loss,t ； in, H 1 represents the present value of the basic system's value and benefits during the planning phase. Y For the planning period, C v,y Indicates the distribution network number y Annual electricity sales revenue C p,y Indicates the distribution network number y The annual cost of purchasing electricity from the main grid. C e,y Indicates the distribution network number y The annual carbon emission costs borne by the company ρ Indicates the discount rate. e u,t This indicates the time-of-use electricity price for electricity sold through the distribution network. P v,t Indicates the distribution network in t Total load for the time period, Δ t Indicates the duration of a time period. P d,t Indicates that the distribution network is connected to the main grid. t Total power purchased during the time period P w,t , P pv,t The distribution network is respectively t Total wind / solar power absorbed during the time period P loss,t For the distribution network in t Total network loss during the time period e g,t Indicates the distribution network in t The time-of-use electricity price for electricity purchased from the main grid during certain periods.

3. The dynamic planning method for energy storage in a distribution network as described in claim 2, characterized in that, The allocation of carbon emission credits in carbon emission costs is determined based on the following formula: ； in, E r The annual carbon emission allowances allocated to the power distribution network free of charge. ς The carbon emission allocation coefficient per unit of electricity; The system carbon emissions in carbon emission costs are determined based on the following formula: ； in, E c This represents the system's total annual carbon emissions. λ t for t Carbon emission factors over a given period of time.

4. The dynamic planning method for energy storage in a distribution network as described in claim 1, characterized in that, This also includes adding node power balance constraints, line current carrying constraints, and node voltage constraints when constructing the basic system value benefit model: Add node power balance constraints based on the following formula: ； ； in, C To remove the line ij External and Node j A set of connected lines, P ij,t For the line ij exist t The active power flowing through during a given time period. Q ij,t For the line ij exist t The reactive power flowing through the time period r ij For the line ij The resistor on x ij For the line ij The reactance on, f ij,t For the line ij exist t The square of the current value flowing through the time period. P load,j,t For nodes j exist t Active load during a given time period Q load,j,t For nodes j exist t Reactive load during a given period For nodes j superior t Energy storage during a given period of time provides active power output. For nodes j superior t Reactive power output of energy storage during certain periods. P w,j,t For nodes j superior t Wind power output during a given period Q w,j,t For nodes j superior t Wind power reactive power output during certain periods P pv,j,t For nodes j superior t Solar power output during certain periods Q pv,j,t For nodes j superior t Solar reactive power output during certain periods; Add line current-carrying constraints based on the following formula: ； in, v i,t For nodes i In the t The square of the voltage value over a time period For the line ij The upper limit of the square of the current; Add node voltage constraints based on the following formula: ； ； in, v max and v min These represent the upper and lower limits of the square of the voltage, respectively.

5. The dynamic planning method for energy storage in a distribution network as described in claim 1, characterized in that, Based on the following formula, after installing energy storage in the distribution network, a dynamic programming configuration model is established with the objective of maximizing the present value of the distribution network's value returns at each stage after energy storage installation: ； in, H 2 represents the present value of the distribution network's revenue after the installation of energy storage devices. For the first time after installing energy storage devices y The annual electricity purchase cost from the main grid by the distribution network operator. For the first time after installing energy storage devices y The annual carbon emission costs borne by the power distribution network C ess This refers to the various investment costs of the energy storage device during the planning period.

6. The dynamic planning method for energy storage in a distribution network as described in claim 5, characterized in that, The power purchase cost of a distribution network including energy storage is determined based on the following formula: ； ； ； in, To facilitate the installation of energy storage devices in the distribution network from the main grid t Total power purchased during the time period for t Charging power of time-of-use energy storage for t Discharge power of time-limited energy storage After installing energy storage devices, the distribution network will be in t Total network loss during the time period; The carbon emission cost of a system including energy storage is determined based on the following formula: ； ； in, This represents the total annual carbon emissions of the system after the installation of energy storage devices. The energy storage investment cost is determined based on the following formula: ； ； ； ； in, C s For energy storage installation costs, k For energy storage lifespan, N For the number of energy storage installations, For energy storage in the first y Annual unit power cost For energy storage in the first y Annual unit capacity cost For the system in the first y The first year installed n The rated power of the energy storage unit, For the system in the first y The first year installed n The rated capacity of the energy storage, The annual operating and maintenance cost per unit capacity of energy storage. For the first y The total energy storage capacity installed in the system annually. C p The replacement cost of energy storage within the planning cycle, M ess For installation at the y The number of times annual energy storage needs to be replaced within the planning cycle. For installation at the y The unit power cost of replacing energy storage devices annually. For installation at the y The unit capacity cost of replacing energy storage devices annually INT () is the floor function.

7. The dynamic planning method for energy storage in a distribution network as described in claim 1, characterized in that, This also includes adding energy storage operation constraints and candidate node installation constraints when building the dynamic programming configuration model: Energy storage operation constraints are added based on the following formula: ； ； ； ； Add installation constraints for candidate nodes based on the following formula: ； in, I These are the nodes to be installed in the system. X i For binary decision variables, N ess This represents the number of energy storage devices connected.

8. The dynamic planning method for energy storage in a distribution network as described in claim 1, characterized in that, An improved particle swarm optimization algorithm is used to solve the value model of the energy storage system to obtain the optimal energy storage configuration scheme, which includes: A set of particles is randomly generated based on the energy storage system value model, with each particle representing an energy storage configuration scheme for a distribution network. The fitness value of each particle is determined by a fitness function, whereby the fitness value characterizes the system value of the corresponding energy storage configuration scheme, and the particle with the largest fitness value is taken as the global optimal solution. For each particle other than the global optimum, the inertia weight of the particle is dynamically adjusted through adaptive inertia weight to update the velocity and position of each particle; it is determined whether the difference between each updated particle and the global optimum is less than a set threshold; if so, a crossover mutation operation is performed on the particle, and the individual optimum and the global optimum are updated through the fitness value of the particle. The process continues until the maximum number of iterations is reached or the global optimal solution shows no significant change in multiple consecutive iterations. The final global optimal solution is then obtained, and the energy storage configuration scheme corresponding to the final global optimal solution is taken as the optimal energy storage configuration scheme.

9. The dynamic planning method for energy storage in a distribution network as described in claim 8, characterized in that, The particle's inertial weight is dynamically adjusted based on the following formula using adaptive inertial weighting: ； ； in, X i ( k ) is a particle i In the k The difference distance between the particle and the best particle in the population at the next iteration. x max The maximum value of the particle's position. x min This represents the minimum value of the particle's position. D To solve for spatial dimensions, w i ( k ) is a particle i In the k Inertia weights in the next iteration a and b For auxiliary values; The crossover and mutation operation of particles is performed based on the following formula: ； ； in, r A random number between 0 and 1.

10. A dynamic planning system for energy storage in a power distribution network, characterized in that, include: The model building module is used to construct a basic system value-benefit model with the goal of maximizing the present value of the value benefits of each stage of the distribution network when energy storage is not installed in the distribution network. After energy storage is installed in the distribution network, a dynamic programming configuration model is constructed with the goal of maximizing the present value of the value returns of the distribution network at each stage after the installation of energy storage. The value determination module is used to take the difference between the present value of the distribution network at each stage obtained from the basic system value benefit model and the present value of the distribution network at each stage obtained from the dynamic programming configuration model as the value model of the energy storage system. The planning module uses an improved particle swarm optimization algorithm to solve the value model of the energy storage system, obtains the optimal energy storage configuration scheme, and configures the energy storage system in the distribution network in stages according to the optimal energy storage configuration scheme. The improved particle swarm optimization algorithm is obtained by introducing adaptive inertia weights and crossover mutation operations in the search phase of the particle swarm optimization algorithm.

Images