Optimal configuration method of energy storage capacity for photovoltaic and energy storage combined system considering source-load uncertainty

Abstract

The present invention discloses a method for optimizing the configuration of energy storage capacity of a photovoltaic-storage combined system considering source-load uncertainty. First, the system structure and related data of a distribution station area to be analyzed are collected and obtained; then, a source-load uncertain output model is established and simulated to obtain an operation scenario with uncertain source-load output throughout the year and reduce it, and obtain a description of the source-load output under a typical load demand operation scenario and a photovoltaic operation scenario; secondly, a forward-backward power flow method is used to calculate the voltage and power loss of the typical load demand scenario after reduction, and the typical load demand operation scenario with the worst voltage is determined; a multi-objective mathematical optimization model is constructed, and the constraint conditions are determined with the objective functions of minimizing the abandonment rate, minimizing the investment and operation and maintenance cost, and minimizing the electric energy loss; finally, the optimization model is solved using an improved salp swarm algorithm to obtain an optimized configuration result under a typical photovoltaic operation scenario. The present invention improves the photovoltaic absorption capacity, reduces abandonment, and realizes green and sustainable development.

Description

Technical Field

[0001] The present invention relates to the field of new energy consumption and optimization technology, and in particular to a method for optimizing energy storage capacity configuration of a photovoltaic-storage combined system considering source-load uncertainty. Background Art

[0002] Despite the current "low-carbon and environmentally friendly" philosophy, the importance of distributed photovoltaics as a zero-carbon power source is growing year by year. However, the timing characteristics of source-load power make it difficult to fully absorb the output of both sources and loads during the same time period, leading to reverse power flow within the system and changes in the distribution patterns of voltage and line losses in the distribution network. Combining energy storage systems with photovoltaic systems leverages the characteristics of energy storage to smooth photovoltaic output, improve photovoltaic absorption, reduce curtailment, and ensure stable and safe operation of the distribution network.

[0003] Because the capacity configuration of energy storage and photovoltaic systems affects the operational characteristics of combined photovoltaic and storage systems, many scholars have conducted extensive research in recent years on the optimal capacity configuration of these systems. For example, some researchers have targeted peak shaving and valley filling as energy storage configuration objectives, achieving optimal energy storage capacity configuration by controlling the peak-valley difference, load variance, and load change rate. Other researchers have proposed a coordinated optimization configuration method for energy storage and renewable energy based on the "equal area principle." By establishing a grid power balance model, a schematic diagram of renewable energy consumption, and an "equal area principle" optimization model for energy storage and renewable energy consumption, energy storage and renewable energy capacity are configured based on the inherent characteristics of renewable energy and system balance principles, achieving decoupled configuration and coordinated planning. Other researchers have also considered the impact of demand response on capacity configuration, developing demand response models to mitigate the impact of energy storage on high-penetration distributed photovoltaic distribution systems. Other researchers have proposed a distributed photovoltaic and energy storage site selection and sizing planning method for distribution networks based on clustering. After clustering the distribution network according to the grid structure and node load characteristics, a two-tier coordinated site selection and sizing planning model for distributed photovoltaic and energy storage is established and solved using a hybrid particle swarm optimization algorithm.

[0004] However, the source-load uncertainty problem has not been fully reflected in the optimization configuration research. There is little research on distribution substations with dispersed load density. In addition, most of the established mathematical models are based on economic perspectives and ignore problems such as internal system losses. Therefore, there is a great need for a method that can effectively optimize the configuration of energy storage capacity of a photovoltaic-storage combined system considering source-load uncertainty. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method for optimizing the configuration of energy storage capacity of a photovoltaic-storage combined system considering source-load uncertainty. By establishing a source-load uncertain output model and performing simulation and reduction, a description of the source-load output under typical load demand operation scenarios and photovoltaic operation scenarios is obtained, and the severe load operation scenario is determined. By constructing a multi-objective mathematical optimization model, the objective function with the minimum abandonment rate, the lowest investment and operation and maintenance cost, and the minimum power loss is determined, and the constraint conditions are determined. The optimization model is solved using the improved Salp Unica swarm algorithm to obtain the optimized configuration results under different photovoltaic operation scenarios. The present invention improves the photovoltaic absorption capacity, reduces abandonment, and achieves green and sustainable development.

[0006] To solve the above technical problems, the technical solution adopted by the present invention is: a method for optimizing the energy storage capacity of a photovoltaic-storage combined system considering source-load uncertainty, specifically comprising the following steps:

[0007] Step S1: Collect and obtain the system structure and related data of the distribution area to be analyzed;

[0008] Step S2: Establish a source-load uncertain output model, simulate the source-load uncertain output model using a Markov Monte Carlo algorithm, and obtain a full-year source-load uncertain output operation scenario; reduce the full-year source-load uncertain operation scenario using a clustering algorithm to obtain a source-load output description under a typical load demand operation scenario and a photovoltaic operation scenario;

[0009] Step S3: Calculate the voltage and power loss of the typical load demand scenario after the reduction in step S2 using the forward-backward power flow method to determine the worst voltage typical load demand operation scenario;

[0010] Step S4: Construct a multi-objective mathematical optimization model to determine the objective functions of minimizing the curtailment rate, minimizing the investment and operation and maintenance costs, and minimizing the power loss. The constraints include node voltage constraints, energy storage power constraints, state of charge constraints, reverse power constraints, and power balance constraints.

[0011] Step S5: Use the improved salp swarm algorithm to solve the optimization model of step S4 to obtain the optimized configuration results under different photovoltaic operation scenarios.

[0012] A further improvement of the technical solution of the present invention is that the relevant data in step S1 includes obtaining the network topology diagram and line parameters of the distribution station area system, the number of system nodes, user load demand data at different nodes of the area system, and regional solar radiation intensity.

[0013] A further improvement of the technical solution of the present invention is that the step of establishing the source-load uncertain output model in step S2 is:

[0014] 21) Establishing a photovoltaic output uncertainty model: Using the mean light intensity μ gq and standard deviation σ gq The key parameters θ1 and θ2 of the Beta distribution are obtained, and the photovoltaic output uncertainty model is obtained:

[0015]

[0016] Where: σ gq is the standard deviation of light intensity; is the maximum photovoltaic output; N PV is the number of photovoltaic cells; S PV is the single photovoltaic area; η PV is the photovoltaic efficiency; is the predicted value of photovoltaic power at time t;

[0017] 22) Establishing a load output uncertainty model: Considering the large random differences in loads between different seasons, the load demand of users at different nodes is simulated using a time series load model to obtain a load output uncertainty model:

[0018]

[0019] Where: Load is the standard deviation of load forecast, ξ Load =P Load r%; r% is the percentage of standard deviation in load output; is the system load forecast value; μ Load is the mean value of load forecast.

[0020] A further improvement of the technical solution of the present invention is that: in step S2, the Markov Monte Carlo algorithm is combined to simulate the load output uncertainty model and photovoltaic output uncertainty model at different nodes in the source-load uncertainty output model to obtain the source-load output uncertainty operation scenario throughout the year; and the load demand data of users at different nodes are calculated to obtain the 24-hour total load demand output, thereby obtaining the matrix of the source-load uncertainty operation scenario throughout the year. and

[0021]

[0022] A further improvement of the technical solution of the present invention is that: in step S2, the source-load uncertainty operation scenarios throughout the year are reduced by a clustering algorithm to obtain source-load output descriptions under typical load demand operation scenarios and photovoltaic operation scenarios. The specific process is as follows;

[0023] Sure and The number of clusters C and fuzzy factor m are used to cluster the and To reduce the scenarios and obtain representative load demand operation scenarios, the steps are as follows:

[0024] k1, parameter setting: set the number of clusters C, fuzzy factor b and number of iterations;

[0025] k2, set the initial cluster center;

[0026] k3, calculate the membership degree between the sample and the cluster center according to formula (7);

[0027]

[0028] Where d() is the Euclidean distance from the sample data to the cluster center;

[0029] k4, update the cluster center according to formula (8);

[0030]

[0031] k5, update the membership matrix according to formula (9);

[0032]

[0033] k6. If the number of iterations reaches the maximum number of iterations or the difference between the two J values ​​is less than the maximum allowable error, the iteration ends. Otherwise, jump to step k3 to obtain C typical load demand operation scenarios and photovoltaic operation scenarios.

[0034] A further improvement of the technical solution of the present invention is that the specific steps of step S3 are as follows:

[0035] Based on the calculation of the 24-hour power flow of the distribution substation system under the C typical load demand operation scenario in each time period, the voltage change of each node in 24 hours is obtained, and the load demand operation scenario under the worst voltage conditions is determined.

[0036] A further improvement of the technical solution of the present invention is that the specific steps in step S4 are as follows:

[0037] 41) Establishing multi-objective functions of the optimization model:

[0038] F min ={F 1.min ,F 2.min ,F 3.min} (10)

[0039] Where: F 1.min The minimum light abandonment rate is F 2.minMinimize system investment and maintenance costs F 3.min The minimum power loss is ΔA;

[0040] 42) Calculate the abandoned light rate;

[0041]

[0042] Where: is the photovoltaic abandonment rate; N t is the total number of time periods; is the total power generation at time t after photovoltaic power output; The photovoltaic power consumed by the load at time t after the photovoltaic power is output; The photovoltaic power consumed by the energy storage system at time t after the photovoltaic power is output;

[0043] 43) Calculate investment and operation and maintenance costs;

[0044]

[0045] Where: The total investment and maintenance cost of the energy storage system; R ESS Planning capacity for energy storage systems; Plan the unit capacity investment cost for energy storage systems; The annual maintenance cost per unit capacity of the energy storage system is planned; r is the discount rate; N ESS The economic service life of the energy storage system;

[0046] 44) Calculate power loss;

[0047]

[0048] Where: ΔA t is the theoretical line loss of the distribution network in the tth period, kW·h; B is the total number of distribution network branches; Δt is the time of the tth period, h (in hours), I l is the current in branch 1, A; R l is the resistance of branch l, Ω;

[0049] 46) Establish constraints, including node voltage constraints, energy storage power constraints, state of charge constraints, reverse power constraints, and power balance constraints.

[0050] A further improvement of the technical solution of the present invention is that: in step 45)

[0051] The node voltage constraints are as follows:

[0052] U i,min ≤U i ≤U i,max (14)

[0053] Where: U i,max and U i,min They represent the upper and lower limits of the voltage at node i under the nominal voltage respectively;

[0054] The energy storage power constraints are as follows:

[0055]

[0056] Where: is the ESS charging power during period t, kW; is the ESS discharge power during period t, kW; η + and η - is the charge and discharge efficiency; is the maximum value of ESS charging power, kW; is the maximum discharge power of ESS, kW;

[0057] The state of charge constraints are as follows:

[0058] SOC min ≤SOC(t)≤SOC max (16)

[0059] Where: SOC min and SOC max Respectively represent the lower and upper limits of the energy storage system's state of charge, ranging from 20% to 90%;

[0060] The reverse power constraints are as follows:

[0061]

[0062] Where: is the reverse power of the photovoltaic and energy storage combined system, kW; The maximum value of the reverse power allowed in kW;

[0063] The system power balance constraints are as follows:

[0064]

[0065] Where: P si The active power injected into the system at node i; P pvi P is the active power injected into node i by distributed photovoltaic; loadi is the active power consumed by node i; Active power absorbed by the energy storage system; The active power generated by the energy storage system; Q si The reactive power injected into the system at node i; Q pvi The reactive power injected into node i by distributed photovoltaic; Q loadi is the reactive power consumed by node i; Absorbed reactive power of the energy storage system; The reactive power generated by the energy storage system; G ij B is the conductance of the branch to ground; ij is the branch circuit's admittance to ground.

[0066] A further improvement of the technical solution of the present invention is that the specific steps of step S5 are as follows:

[0067] 51) Improvements to the salp swarm algorithm include population initialization, leader and follower position update strategies;

[0068] 52) Based on the original population, a hybrid chaotic map is introduced to improve population initialization;

[0069] X D×N =rand(D,N)·(ub-lb)+lb (19)

[0070] +cos(π(4RX i (1-X i )+(1-R)sin(πX i )-0.5))

[0071] 53) Introducing the golden sine to improve the leader position update strategy;

[0072]

[0073] Where: R0 and R1 are random numbers, R0 determines the moving distance of the individual in the next iteration,

[0074] R0∈[0,2π]; R1 determines the position update direction of the i-th individual in the next iteration, R1∈[0,π];

[0075] X1 and X2 are the coefficients obtained by introducing the golden section. The specific formula is as follows:

[0076]

[0077] 54) Introducing inertia factor to improve follower position update strategy;

[0078]

[0079] Where: ω initial and ω final are the initial value and final value of the inertia factor, t is the current number of iterations, and t max is the maximum number of iterations;

[0080] 55) According to the typical operation scenario after photovoltaic curtailment, the improved Salp Swarm Algorithm is used to solve the optimization model of step S4. After multiple iterations, the optimal Pareto solution set is obtained, and then the optimal energy storage capacity configuration scheme under the typical photovoltaic operation scenario is obtained.

[0081] Due to the adoption of the above technical solution, the technical advancements achieved by the present invention are:

[0082] The present invention provides a method for optimizing the configuration of energy storage capacity of a photovoltaic and storage combined system considering source-load uncertainty. The Markov Monte Carlo algorithm is used to simulate the source-load situation of user load demands at different nodes in the substation throughout the year, and the fuzzy C-means clustering algorithm is used to obtain typical scenarios. A multi-objective function model is established to minimize the abandonment rate, investment and maintenance costs, and power loss. In addition, multiple constraints such as the charge and discharge balance and power flow backflow of energy storage are taken into account. The objective function is solved by an improved intelligent optimization algorithm to obtain the optimal configuration of energy storage, which can greatly reduce the abandonment rate, reduce investment and maintenance costs, and reduce power loss, improve the photovoltaic absorption capacity, reduce abandonment, and achieve green and sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 It is the overall flow chart of the present invention;

[0084] Figure 2 This is the structural diagram of a 33-node 0.4kV distribution network area;

[0085] Figure 3 Schematic diagram of the establishment process of a typical scenario;

[0086] Figure 4 Identify processes for severe operating scenarios;

[0087] Figure 5 Schematic diagram of the dynamic balance of power within the system. DETAILED DESCRIPTION

[0088] The present invention is described in further detail below in conjunction with the embodiments:

[0089] like Figure 1 As shown, the present invention provides a method for optimizing energy storage capacity configuration of a photovoltaic-storage combined system considering source-load uncertainty, specifically comprising the following steps:

[0090] Step S1: Collect and obtain the system structure and related data of the distribution area to be analyzed, including Figure 2 For example, obtain the network topology and line parameters of the distribution substation system, the number of system nodes, the user load demand data at different nodes of the substation system, and the regional solar intensity;

[0091] Step S2: Establish a source-load uncertain output model, simulate the source-load uncertain output model using the Markov Monte Carlo algorithm, and obtain the source-load uncertain output operation scenario throughout the year; use the clustering algorithm to reduce the source-load uncertain operation scenario throughout the year, and obtain the source-load output description under the typical load demand operation scenario and the photovoltaic operation scenario; the steps for establishing the source-load uncertain output model are as follows:

[0092] 21) Establishing a photovoltaic output uncertainty model: Using the mean light intensity μ gq and standard deviation σ gq The key parameters θ1 and θ2 of the Beta distribution are obtained, and the photovoltaic output uncertainty model is obtained:

[0093]

[0094]

[0095] Where: σ gq is the standard deviation of light intensity, is the maximum photovoltaic output; N PV is the number of photovoltaic cells; S PV is a single photovoltaic area; η PV is the photovoltaic efficiency; is the predicted value of photovoltaic power at time t.

[0096] 22) Establishing a load output uncertainty model: Considering the large random differences in loads between different seasons, the load demand of users at different nodes is simulated using a time series load model to obtain a load output uncertainty model:

[0097]

[0098] Where: Load is the standard deviation of load forecast, ξ Load =P Load r%; r% is the percentage of standard deviation in load output; is the system load forecast value; μ Load is the mean value of load forecast.

[0099] 23) Combined with the Markov Monte Carlo algorithm, the load output uncertainty model and photovoltaic output uncertainty model at different nodes in the source-load uncertainty output model are simulated to obtain the full energy-load output uncertainty operation scenario; and the user load demand data at different nodes are calculated to obtain the 24-hour total load demand output, thereby obtaining the matrix of the source-load uncertainty operation scenario throughout the year. and

[0100]

[0101] 24) Figure 3 For example, determine and The number of clusters C and fuzzy factor m are used to cluster the and Perform scenario reduction to obtain representative typical load demand operation scenarios and photovoltaic operation scenarios. The steps are as follows:

[0102] k1, parameter setting: set the number of clusters C, fuzzy factor b and number of iterations;

[0103] k2, set the initial cluster center;

[0104] k3, calculate the membership degree between the sample and the cluster center according to formula (7);

[0105]

[0106] Where d() is the Euclidean distance from the sample data to the cluster center;

[0107] k4, update the cluster center according to formula (8);

[0108]

[0109] k5, update the membership matrix according to formula (9);

[0110]

[0111] k6. If the number of iterations reaches the maximum number of iterations or the difference between the two J values ​​is less than the maximum allowable error, the iteration ends. Otherwise, jump to step k3 to obtain C typical load demand operation scenarios.

[0112] Step S3: Calculate the voltage and power loss of the typical load demand scenario after the reduction in step S2 using the forward-backward power flow method to determine the worst voltage typical load demand operation scenario;

[0113] The specific steps are as follows:

[0114] by Figure 4 Taking the load demand scenario as an example, the power flow of the 24-hour distribution substation system under the C typical load demand operation scenario is calculated based on each time period, the voltage change of each node in 24 hours is obtained, and the load demand operation scenario under the worst voltage conditions is determined.

[0115] Step S4: Construct a multi-objective mathematical optimization model to determine the objective functions of minimizing the curtailment rate, minimizing the investment and operation and maintenance costs, and minimizing the power loss. The constraints include node voltage constraints, energy storage power constraints, state of charge constraints, reverse power constraints, and power balance constraints.

[0116] The specific steps are as follows:

[0117] 41) Establishing multi-objective functions of the optimization model:

[0118] F min ={F 1.min ,F 2.min ,F 3.min} (10)

[0119] Where: F 1.min The minimum light abandonment rate is F 2.min Minimize system investment and maintenance costs F 3.min The minimum power loss is ΔA;

[0120] 42) Calculate the abandoned light rate;

[0121]

[0122] Where: is the photovoltaic abandonment rate; N t is the total number of time periods; is the total power generation at time t after photovoltaic power output; The photovoltaic power consumed by the load at time t after the photovoltaic power is output; The photovoltaic power consumed by the energy storage system at time t after the photovoltaic power is output;

[0123] 43) Calculate investment and operation and maintenance costs;

[0124]

[0125] Where: The total investment and maintenance cost of the energy storage system; R ESS Planning capacity for energy storage systems; Plan the unit capacity investment cost for energy storage systems; The annual maintenance cost per unit capacity of the energy storage system is planned; r is the discount rate; N ESS The economic service life of the energy storage system;

[0126] 44) Calculate power loss;

[0127]

[0128] Where: ΔA tis the theoretical line loss of the distribution network in the tth period, kW·h; B is the total number of distribution network branches; Δt is the time of the tth period, h (in hours). l is the current in branch 1, A; R l is the resistance of branch l, Ω;

[0129] 47) Establish constraints, including node voltage constraints, energy storage power constraints, state of charge constraints, reverse power constraints, and power balance constraints.

[0130] The node voltage constraints are as follows:

[0131] U i,min ≤U i ≤U i,max (14)

[0132] Where: U i,max and U i,min They represent the upper and lower limits of the voltage at node i under the nominal voltage respectively;

[0133] The energy storage power constraints are as follows:

[0134]

[0135] Where: is the ESS charging power during period t, kW; is the ESS discharge power during period t, kW; η + and η - is the charge and discharge efficiency; is the maximum value of ESS charging power, kW; is the maximum discharge power of ESS, kW;

[0136] The state of charge constraints are as follows:

[0137] SOC min ≤SOC(t)≤SOC max (16)

[0138] Where: SOC min and SOC max Respectively represent the lower and upper limits of the energy storage system's state of charge, ranging from 20% to 90%;

[0139] The reverse power constraints are as follows:

[0140]

[0141] Where: is the reverse power of the photovoltaic and energy storage combined system, kW; The maximum value of the reverse power allowed in kW;

[0142] The system power balance constraints are as follows:

[0143]

[0144] Where: P si The active power injected into the system at node i; P pvi P is the active power injected into node i by distributed photovoltaic; loadi is the active power consumed by node i; Active power absorbed by the energy storage system; The active power generated by the energy storage system; Q si The reactive power injected into the system at node i; Q pvi The reactive power injected into node i by distributed photovoltaic; Q loadi is the reactive power consumed by node i; Absorbed reactive power of the energy storage system; is the reactive power generated by the energy storage system, G ij is the branch conductance to ground, B ij is the branch circuit's admittance to ground.

[0145] Step S5: Use the improved salp swarm algorithm to solve the optimization model of step S4 to obtain the optimized configuration results under different photovoltaic operation scenarios.

[0146] The specific steps are as follows:

[0147] 51) Improvements to the salp swarm algorithm include population initialization, leader and follower position update strategies;

[0148] 52) Based on the original population, a hybrid chaotic map is introduced to improve population initialization;

[0149] X D×N =rand(D,N)·(ub-lb)+lb (19)

[0150] +cos(π(4RX i (1-X i )+(1-R)sin(πX i )-0.5))

[0151] 53) Introducing the golden sine to improve the leader position update strategy;

[0152]

[0153] Where: R0 and R1 are random numbers, R0 determines the moving distance of the individual in the next iteration, R0∈[0,2π]; R1 determines the position update direction of the i-th individual in the next iteration, R1∈[0,π];

[0154] X1 and X2 are the coefficients obtained by introducing the golden section. The specific formula is as follows:

[0155]

[0156] 54) Introducing inertia factor to improve follower position update strategy;

[0157]

[0158] Where: ω initial and ω final are the initial value and final value of the inertia factor, t is the current number of iterations, and t max is the maximum number of iterations;

[0159] 55) Based on the typical operation scenario after photovoltaic curtailment, the improved salp swarm algorithm is used to solve the optimization model of step S4, as follows:

[0160] 551) Set the population size with chaotic mapping, maximum number of iterations, and algorithm-related parameters of the algorithm;

[0161] 552) Calculate the fitness value of each individual and select the best individual as the food source;

[0162] 553) Update parameter C1, leader and follower positions;

[0163] 554) Compare and store the individual fitness values ​​found;

[0164] 555) Repeat 553)-554) until the convergence accuracy or the maximum number of iterations is reached, and the current food position is output as the best solution set, and then the optimal configuration scheme of energy storage capacity under the typical photovoltaic operation scenario is obtained. Figure 5 For example, the power flow within the distribution area system of the present invention is used to achieve dynamic balance of the system in the area.

[0165] The above embodiments are intended only to illustrate the technical concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand and implement the present invention accordingly. They are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the claims.

Claims

1. A method for optimizing energy storage capacity of a photovoltaic-storage combined system considering source-load uncertainty, characterized by: The specific steps include: Step S1, collecting and obtaining the system structure and related data of the distribution station area to be analyzed; Step S2, establish a source-load uncertain output model, simulate the source-load uncertain output model in combination with the Markov Monte Carlo algorithm, and obtain the source-load uncertain output operation scenario throughout the year; reduce the source-load uncertain operation scenario throughout the year through the clustering algorithm, and obtain the source-load output description under the typical load demand operation scenario and the photovoltaic operation scenario; Step S3, use the forward-backward power flow method to calculate the voltage and power loss of the typical load demand scenario after reduction in step S2, and determine the typical load demand operation scenario with the worst voltage; Step S4: construct a multi-objective mathematical optimization model, determine the objective function of minimizing the abandoned light rate, minimizing the investment and operation cost, and minimizing the power loss, and determine the constraints, including node voltage constraints, energy storage power constraints, state of charge constraints, reverse power constraints, and power balance constraints; Step S5, using the improved salp swarm algorithm to solve the optimization model of step S4, and obtain the optimization configuration results under different photovoltaic operation scenarios; The specific steps of step S5 are as follows: 51) The improvements of the salp swarm algorithm include population initialization, leader and follower position update strategies; 52) Based on the original population, the hybrid chaotic mapping is introduced to improve the population initialization; 53) Introducing the golden sine to improve the leader position update strategy; Where: R0 and R1 are random numbers, R0 determines the moving distance of the individual in the next iteration, R0∈[0,2π]; R1 determines the position update direction of the i-th individual in the next iteration, R1∈[0,π]; X1 and X2 are the coefficients obtained by introducing the golden section. The specific formula is as follows: 54) Introducing inertia factor to improve follower position update strategy; Where: initial and ω final are the initial value and final value of the inertia factor, t is the current iteration number, t max is the maximum number of iterations; 55) According to the typical operation scenario after photovoltaic curtailment, the optimization model of step S4 is solved using the improved Salp Swarm algorithm. After multiple iterations, the optimal Pareto solution set is obtained, and then the optimal configuration plan of energy storage capacity under the typical photovoltaic operation scenario is obtained.

2. The method for optimizing energy storage capacity of a photovoltaic-storage combined system considering source-load uncertainty according to claim 1 is characterized in that: The relevant data in step S1 include obtaining the network topology diagram and line parameters of the distribution station area system, the number of system nodes, user load demand data at different nodes of the area system, and regional solar radiation intensity.

3. The method for optimizing energy storage capacity of a photovoltaic-storage combined system considering source-load uncertainty according to claim 1 is characterized in that: The steps for establishing the source-load uncertain output model in step S2 are: 21) Establishing the uncertainty model of photovoltaic output: by using the mean value of light intensity μ gq and standard deviation σ gq The key parameters θ1 and θ2 of the Beta distribution are obtained, and the uncertainty model of photovoltaic output is obtained: Where: gq is the standard deviation of light intensity; is the maximum photovoltaic output; N PV is the number of photovoltaic cells; S PV is a single photovoltaic area; η PV is the photovoltaic efficiency; is the predicted value of photovoltaic power at time t; 22) Establish a load output uncertainty model: Considering the large random differences in loads between different seasons, the load demand of users at different nodes is simulated using a time series load model to obtain a load output uncertainty model: Where: Load is the standard deviation of load forecast, ξ Load =P Load ·r%; r% is the percentage of standard deviation in load output; is the system load forecast value; μ Load is the mean value of load forecast.

4. The method for optimizing energy storage capacity configuration of a photovoltaic-storage combined system considering source-load uncertainty according to claim 3 is characterized in that: In step S2, the Markov Monte Carlo algorithm is combined to simulate the load output uncertainty model and the photovoltaic output uncertainty model at different nodes in the source-load uncertain output model to obtain the source-load output uncertainty operation scenario throughout the year; and the load demand data of users at different nodes are calculated to obtain the 24-hour total load demand output, thereby obtaining the matrix of the source-load uncertain operation scenario throughout the year. and 5. The method for optimizing energy storage capacity configuration of a photovoltaic-storage combined system considering source-load uncertainty according to claim 4 is characterized in that: In step S2, the source-load uncertain operation scenarios throughout the year are reduced by a clustering algorithm to obtain a source-load output description under a typical load demand operation scenario and a photovoltaic operation scenario. The specific process is as follows; Sure and The number of clusters C and fuzzy factor m are used to cluster the and Perform scenario reduction to obtain representative typical load demand operation scenarios and photovoltaic operation scenarios. The steps are as follows: k1, parameter setting: set the number of clusters C, fuzzy factor b and number of iterations; k2, set the initial cluster center; k3, according to formula (7), calculate the membership degree between the sample and the cluster center; Where d() is the Euclidean distance from the sample data to the cluster center; k4, update the cluster center according to formula (8); k5, update the membership matrix according to formula (9); k6. If the number of iterations reaches the maximum number of iterations or the difference between the two Js is less than the maximum allowable error, the iteration ends. Otherwise, jump to step k3 to obtain C typical load demand operation scenarios and photovoltaic operation scenarios.

6. The method for optimizing energy storage capacity configuration of a photovoltaic-storage combined system considering source-load uncertainty according to claim 5 is characterized in that: The specific steps of step S3 are as follows: Based on the calculation of the power flow of the 24-hour distribution substation system under C typical load demand operation scenarios in each time period, the voltage change of each node in 24 hours is obtained, and the load demand operation scenario under the worst voltage conditions is determined.

7. The method for optimizing energy storage capacity configuration of a photovoltaic-storage combined system considering source-load uncertainty according to claim 1 is characterized in that: The specific steps in step S4 are as follows: 41) Establish multi-objective function of optimization model: F min ={F 1.min ,F 2.min ,F 3.min } (10) Where: F 1.min The minimum light abandonment rate is F 2.min Minimize system investment and maintenance costs F 3.min The minimum power loss is ΔA; 42) Calculate the abandoned light rate; Where: is the photovoltaic abandonment rate; N t is the total number of time periods; is the total power generation at time t after the photovoltaic power is output; The photovoltaic power consumed by the load at time t after the photovoltaic power is output; The photovoltaic power consumed by the energy storage system at time t after the photovoltaic power is output; 43) Calculate investment and operation and maintenance costs; Where: The total investment and maintenance cost of the energy storage system; R ESS Planning capacity for energy storage systems; Plan the unit capacity investment cost for energy storage systems; is the annual maintenance cost per unit capacity of the energy storage system; r is the discount rate; N ESS The economic service life of the energy storage system; 44) Calculate power loss; Where: ΔA t is the theoretical line loss of the distribution network in the tth period, kW·h; B is the total number of distribution network branches; Δt is the time of the tth period, h represents the number of hours, I l is the current in branch l, A; R l is the resistance of branch l, Ω; 45) Establish constraints, including node voltage constraints, energy storage power constraints, state of charge constraints, reverse power constraints and power balance constraints.

8. The method for optimizing energy storage capacity configuration of a photovoltaic-storage combined system considering source-load uncertainty according to claim 7 is characterized in that: In the step 45) The node voltage constraints are as follows: IN i,min ≤U i ≤U i,max (14) Where: U i,max and U i,min They represent the upper and lower limits of the voltage at node i under the nominal voltage respectively; The energy storage power constraints are as follows: Where: is the ESS charging power during period t, kW; is the ESS discharge power in period t, kW; η + and η - is the charge and discharge efficiency; is the maximum value of ESS charging power, kW; is the maximum value of ESS discharge power, kW; The state of charge constraints are as follows: SOC min ≤SOC(t)≤SOC max (16) Where: SOC min and SOC max They represent the lower and upper limits of the state of charge of the energy storage system, ranging from 20% to 90%; The reverse power constraints are as follows: Where: is the reverse power of the photovoltaic and energy storage combined system, kW; The maximum value of the reverse power allowed in kW; The system power balance constraints are as follows: Where: P si The active power injected into the system at node i; P pvi P is the active power injected into node i by distributed photovoltaic; loadi is the active power consumed by node i; Absorbed active power of the energy storage system; The active power generated by the energy storage system; Q si The reactive power injected into the system at node i; Q pvi The reactive power injected into node i by distributed photovoltaic; Q loadi is the reactive power consumed by node i; The absorbed reactive power of the energy storage system; The reactive power generated by the energy storage system; G ij B is the conductance of the branch to the ground; ij is the branch circuit's admittance to ground.

Images