Remote mountainous area power distribution network light storage combined planning method

By establishing a photovoltaic and load uncertainty processing model and a double-layer planning model in the remote mountain distribution network, combining ICOA and NSGA-II algorithms to optimize the location and operation of photovoltaic and energy storage, the voltage problem and economics in remote mountainous areas are solved. It has achieved efficient joint planning of photoreservation and voltage quality improvement.

CN119067460APending Publication Date: 2024-12-03国网陕西省电力有限公司安康供电公司

Patent Information

Application Number
CN202410890873.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-04
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing joint planning technology for optical storage in distribution networks is mainly concentrated in urban areas. Remote mountainous areas have prominent voltage problems due to geographical and economic factors. The existing technology is difficult to meet the demand for high-quality power supply, and the solution methods are mostly classic optimization algorithms and parameter settings Complexity affects the accuracy of the result.

Method used

A joint planning method for photovoltaic storage in remote mountainous power distribution networks is proposed. By establishing a photovoltaic and load uncertainty processing model, a two-layer planning model is adopted, and combined with ICOA and NSGA-II algorithms are used to optimize the location and location of photovoltaic and energy storage. Capacity selection and voltage scheduling is performed in the lower operating layer to reduce network loss and light abandonment rate.

Benefits of technology

Effectively improve the voltage quality in remote mountainous areas, reduce network losses, improve new energy consumption rate, improve economic benefits and power supply safety of the distribution network, and improve the solution algorithm improves accuracy and reduces local optimal risks. .

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067460A_ABST
    Figure CN119067460A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of power distribution network optical storage combined planning, and particularly relates to a remote mountainous area power distribution network optical storage combined planning method. The remote mountainous area power distribution network optical storage joint optimization model comprises a remote mountainous area PV and load uncertainty processing model, a bilevel planning model and a remote mountainous area power distribution network optical storage collaborative optimization bilevel model solving algorithm; the remote mountainous area PV and load uncertainty processing model comprises a photovoltaic probability model, a load probability model and load flow calculation considering uncertainty processing; the double-layer planning model comprises an upper-layer planning model and a lower-layer operation model; the remote mountainous area power distribution network optical storage collaborative optimization double-layer model solving algorithm comprises upper layer model solving based on an ICOA optimization algorithm and lower layer model solving based on an NSGA-II algorithm. The invention provides an efficient model solving method, and a double-layer optimization model not only can effectively reduce the network loss, but also has higher voltage quality, can also improve the new energy consumption rate, and gives consideration to the economic benefits of the power distribution network.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of photovoltaic and storage joint planning of distribution networks, and specifically relates to a photovoltaic and storage joint planning method for distribution networks in remote mountainous areas. Background Art

[0002] It is of great significance to ensure the economic development of remote areas, promote the equalization of urban and rural power services, and promote the construction of a new rural energy system. Remote mountainous areas are rich in light resources, but the load is dispersed and the power supply distance is long, resulting in prominent voltage problems. Photovoltaic and energy storage are flexible to install and are less restricted by geographical factors. With their rapid development, their application in remote mountainous areas can not only improve the quality of electricity for residents, but also help save energy and reduce emissions, which is of great significance to promoting the development of remote mountainous areas.

[0003] At present, there are many research results on the joint planning of photovoltaic and storage in distribution networks. The research mainly starts from three perspectives: source-load uncertainty processing, model construction, and solution method. In view of the uncertainty of source and load in the system, photovoltaic power is mostly processed by Beta probability model, interval model, and three-dimensional Markov chain model; load is mostly processed by normal distribution probability model, robust optimization, and fuzzy model. The existing source-load uncertainty processing methods are very mature. For model construction, from the perspective of economy, the existing technology considers demand response, reduces energy storage configuration, improves economic benefits, and establishes an optimal configuration model for photovoltaic and storage microgrids for high penetration photovoltaic distribution networks; the existing technology considers cost and benefit for power supply planning in photovoltaic and storage microgrids, integrates time-of-use electricity prices, and establishes a photovoltaic power supply planning model with the greatest benefit; the existing technology considers the single configuration of energy storage and the coordinated configuration mode of photovoltaic and storage for the problem of energy storage output volatility, analyzes the investment costs of different types of energy storage, and establishes an energy storage configuration optimization model; the existing technology considers the problem of wind and solar uncertainty and the limited output of energy storage, in order to reduce active power loss, enhance voltage stability, and improve economic benefits, and establishes a single-layer model for the dispatch of wind, solar, and storage multi-objective systems. From the perspective of safety, reliability and economy, the existing technology establishes a two-stage photovoltaic and storage collaborative operation scheduling model. The first stage determines the photovoltaic and storage regulation nodes, and the second stage optimizes the photovoltaic and storage scheduling from an economic perspective, so as to control voltage over-limit and achieve economic operation at the same time. For the problem of voltage over-limit after photovoltaic access, a single-layer model of photovoltaic and storage collaborative voltage regulation is established considering the equipment loss cost, which has good economy. Based on multi-scenario analysis, a two-layer model of photovoltaic and storage joint planning is established to increase revenue and absorb photovoltaics. For the problem of low voltage crossing, a storage and reactive compensation collaborative regulation model is established to achieve system power balance and safe operation and improve the stability of photovoltaic grid-connected points. For isolated photovoltaic and storage microgrids, the existing technology studies the capacity configuration of photovoltaic and storage access at different load points, and analyzes the impact of different photovoltaic and storage configurations on the reliability and economy of the distribution network. However, most of the existing research objects are cities, and less attention is paid to remote mountainous areas. Although the external conditions and resources of remote mountainous areas are rich, the voltage quality is poor due to restrictions such as economy, terrain, and load distribution. In addition, most of the research is based on economics. Although economic operation is an important factor that cannot be ignored, the use of electricity is far from enough. Therefore, considering the quality of power supply and combining it with the objective problem of backward economy in remote mountainous areas to achieve economic investment is an important idea, and it is also a concrete manifestation of the distribution network company's social responsibility.

[0004] Regarding the solution methods, they are mainly divided into two categories: intelligent optimization algorithms and commercial solvers such as Yalmip+Cplex. The existing technology proposes to improve the second-order oscillating particle swarm algorithm and the adaptive ε-dominated multi-objective particle swarm algorithm to solve the two-layer model, which has a faster convergence speed; the existing technology uses the particle swarm algorithm for solution; the existing technology proposes to improve the multi-objective particle swarm algorithm and combine it with the TOPSIS method to improve the solution accuracy; the existing technology uses the co-evolution genetic algorithm for solution; the existing technology proposes a parallel double quantum differential evolution algorithm to speed up the solution speed; the existing technology proposes a simulated annealing particle swarm algorithm; the existing technology proposes an improved ant colony algorithm to improve the solution speed and accuracy. The existing technology uses second-order cone relaxation, linearization processing and other methods, based on the Yalmip platform, and uses mature commercial solvers such as Cplex and Mosek to solve the established mixed integer nonlinear programming model, but this method is more difficult to handle multi-objective optimization problems. The above solutions are mostly classic optimization algorithms. The effectiveness of traditional solution methods cannot be denied, but they have many parameter settings, which have a certain impact on the results. Therefore, it is necessary to adopt appropriate intelligent optimization algorithms with better performance.

[0005] In view of the above problems, a photovoltaic and energy storage joint planning method for distribution networks in remote mountainous areas was proposed, taking quality as the starting point and economy as the end point. Summary of the invention

[0006] The purpose of the present invention is to address the problems existing in the prior art and provide a method for joint planning of photovoltaic and energy storage for distribution networks in remote mountainous areas. First, the uncertainty of photovoltaic and load is processed. Secondly, the upper planning layer uses the location and capacity of photovoltaic and energy storage as decision variables, and selects the site and capacity of PV and ESS with the goal of minimizing investment cost and operation and maintenance cost; the lower operation layer uses the actual output of photovoltaic in each season and time period and the charging and discharging power of energy storage in each season and time period as decision variables, and performs scheduling with the comprehensive optimization of small voltage offset, small total network loss and low abandonment rate as the goal.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for joint planning of photovoltaic and storage distribution networks in remote mountainous areas, characterized by comprising the following steps:

[0009] Step 1: Establish a PV and load uncertainty processing model for remote mountainous areas; the PV and load uncertainty processing model for remote mountainous areas includes a photovoltaic probability model, a load probability model, and a power flow calculation considering uncertainty processing;

[0010] Step 2: Establish a two-layer planning model; the two-layer planning model includes an upper-layer planning model and a lower-layer operation model;

[0011] Step 3. An algorithm for solving the double-layer model of photovoltaic and storage collaborative optimization for distribution networks in remote mountainous areas; the algorithm for solving the double-layer model of photovoltaic and storage collaborative optimization for distribution networks in remote mountainous areas includes an upper-layer model solution based on the ICOA optimization algorithm and a lower-layer model solution based on the NSGA-Ⅱ algorithm.

[0012] Preferably, in step 1, in the photovoltaic probability model, the light intensity satisfies the Beta distribution, and its probability density function f(S) is:

[0013]

[0014] Where: S represents the actual value of light intensity; S max Represents the maximum light intensity; α and β represent the two parameters of the probability density function respectively;

[0015] Given the probability density function of light intensity, the probability density function of photovoltaic output is:

[0016]

[0017] Where: P PV is the actual output of photovoltaic power; P max is the maximum photovoltaic output, and P max =AηS max , where A is the effective area of ​​the photovoltaic array and η is the photoelectric conversion efficiency.

[0018] Preferably, in step 1, in the probability model of the load, the load in remote mountainous areas is mainly the load of residents and obeys the normal distribution:

[0019]

[0020] Where: P L , Q L are active load and reactive load respectively; are the variances of active load and reactive load respectively; μ P , μ Q are the mean values ​​of active load and reactive load respectively.

[0021] Preferably, in step one, in the flow calculation considering uncertainty processing, based on the historical hourly light intensity and load data in remote mountainous areas, Monte Carlo simulation is used to obtain light intensity and demand load samples with uncertainty, and then K-means clustering is used, with the four seasons of spring, summer, autumn and winter as typical scenarios, and the day is divided into 24 time periods by the time period division method, and the photovoltaic output and load output are clustered to obtain the typical daily output curves of photovoltaic and load in each scenario; in the flow calculation, the steady-state operation of photovoltaic adopts a constant power factor control strategy and is processed as a PQ node, and the load adopts a constant power model and is processed as a PQ node.

[0022] Preferably, in step 2, in the two-layer planning model, in order to solve the low voltage problem in remote mountainous areas, a photovoltaic-storage synergy model is established with reference to the establishment method of the traditional planning model, as shown in formula (5).

[0023]

[0024] Where: G(·) and H(·) are upper-level constraints, g(·) and h(·) are lower-level constraints; x I 、x O are upper and lower level decision variables respectively, and the model is two-level and bidirectional;

[0025] Among them, the upper planning layer, in view of the underdeveloped economy in remote mountainous areas, selects the objective function based on economy, reduces investment losses, and takes the minimum comprehensive cost as the goal. The decision variables are the location capacity of PV and ESS, and the constraints are the investment and site selection constraints of each equipment; the lower operation layer, in view of the special conditions of long power supply distance, low terminal voltage, small wire diameter, large active power loss and abundant light resources in remote mountainous areas, in order to improve voltage quality, improve safety, power supply economy and greenness, reduce network loss, and reduce abandoned light, the voltage offset ΔU is used. all Small, total network loss P loss Less, abandoned light rate PV The overall optimal goal is to consider the scheduling of PV and ESS, and the constraints are system constraints and output constraints of each device; the planning decision variable x in the two-layer model I and run the decision variable x o is separable.

[0026] Preferably, in step 2, the upper-level planning model includes an upper-level model objective function and an upper-level model constraint condition;

[0027] In the upper model objective function, the upper planning layer takes the minimum comprehensive cost as the goal, as shown in formula (6); C I Including annual PV investment costs ESS investment cost The sum of O Including the annual PV operation and maintenance costs ESS operation and maintenance costs Cost of purchasing electricity from the upper grid Abandoned light cost sum;

[0028] The objective function F of the upper model is:

[0029] minF=C I +C O (6)

[0030] The cost calculation of each fee is as follows:

[0031] (1) The investment cost for photovoltaic is:

[0032]

[0033] Where: b is the discount rate; y PV is the service life of photovoltaic; Ω PV is the set of photovoltaic candidate node installation locations, c PV The investment cost of a single photovoltaic unit; is the number of photovoltaic configurations at node i;

[0034] (2) Cost of investment for ESS:

[0035]

[0036] Where: y ESS is the service life of ESS; Ω ESS is the set of ESS candidate node installation locations, c ESS The investment cost of a single ESS; is the number of ESSs configured at node i;

[0037] (3) PV operation and maintenance costs:

[0038]

[0039] Where: D s is the number of days in the season; PV is a collection of photovoltaic installation node locations; It is the operation and maintenance cost of photovoltaic power generation per unit; is the active power injected by photovoltaic power in the i-node in the t-period of the season; Δt is 1h;

[0040] (4) ESS operation and maintenance costs:

[0041]

[0042] Where: ESS A set of ESS installation node locations; It is the operation and maintenance cost of ESS unit charging and discharging power; are the ESS charging and discharging power of node i, season s, period t;

[0043] (5) Cost of purchasing electricity from the upper power grid:

[0044]

[0045] Where: A L (1,:) is the set of terminal nodes of the AC branch with the generator node as the head node; P 1j,s,t is the active power flowing into the local power grid from the upper power grid during period t in season s;

[0046] (6) Cost of abandoned light:

[0047]

[0048] Where: The cost of abandoned solar power per unit of electricity; The maximum active power output of the photovoltaic installed at node i in period t during season s;

[0049] The upper model constraints are:

[0050] (1) Installation capacity and power constraints of candidate PV and ESS installation nodes:

[0051]

[0052] Where: The maximum power of the ESS installed at node i in period t during season s; The capacity of the ESS system installed for node i in period t during season s; are the lower and upper limits of the capacity of the ESS installed at node i in period t during season s;

[0053] (2) Installation quantity constraints for PV and ESS candidate installation nodes:

[0054]

[0055] Preferably, the lower layer operation model includes a lower layer model objective function and a lower layer model constraint condition;

[0056] In the objective function of the lower model, the lower operating layer is based on the voltage offset ΔU all Small, total network loss P loss Less, abandoned light rate PV The goal is low overall optimum;

[0057] The objective function f of the lower model is:

[0058] minf=(ΔU all ,P loss ,l PV ) (15)

[0059] (1)ΔU all is the voltage offset:

[0060]

[0061] Where: is the per unit value of the voltage at node i in season s, period t; Ω AN is a collection of communication nodes;

[0062] (2)P LOSS is the total network loss:

[0063]

[0064] Where: P LOSS Represents the active network loss of the system; V i represents the voltage amplitude of node i; G ij , B ij , δ ij Respectively represent the conductance, susceptance and phase angle difference between nodes i and j; n is the total number of nodes; Γ represents the set of nodes connected to node j;

[0065] (3)l PV The light abandonment rate is:

[0066]

[0067] Where: is the maximum output value of the i-th photovoltaic at time t;

[0068] The lower model constraints are:

[0069] (1) DistFlow constraints of AC distribution network:

[0070]

[0071] Where: A L (j,:) is the set of end nodes of the AC branch with j as the head node; A L (:,j) is the set of the head-end nodes of the AC branch with j as the terminal node; The reactive power output of the PV installed at node i in period t during season s; P ij,s,t , P jk,s,t and Q ij,s,t , Q jk,s,t are the active and reactive power on lines ij and jk in time period t during season s; X ij is the reactance of line ij; and are respectively the active and reactive power of the original load of node j in time period t in season s; U i,s,t The voltage of node i;

[0072] (2) System security constraints:

[0073]

[0074] Where: U i,max and U i,min are the upper and lower limits of the voltage at node i; I ij,max is the maximum safe current allowed to flow through line ij;

[0075] (3) Power constraints injected into the power grid by the generator node:

[0076]

[0077] Where: Q 1j,s,t P is the reactive power flowing from the upper power grid to the local power grid; 1 max and P 1 min They are the upper and lower limits of active power flowing from the upper power grid to the local power grid; and They are the upper and lower limits of reactive power flowing from the upper power grid to the local power grid;

[0078] (4) PV operation constraints:

[0079]

[0080] Where: is the power factor angle of PV, which is taken as 0.98;

[0081] (5) ESS operation constraints

[0082] ESS can play the role of "peak shaving and valley filling" and improve PV absorption capacity. It can also improve economic efficiency when combined with time-of-use electricity prices.

[0083]

[0084] Where: and Respectively represent the charging and discharging status of ESS, which are 0-1 variables; P i ESS,max is the maximum value of ESS charging and discharging power of node i; is the upper limit of ESS power of node i; η + and η - are the charge and discharge efficiency respectively.

[0085] Preferably, in step three, in the two-layer model solution algorithm for the photovoltaic and storage collaborative optimization of the distribution network in remote mountainous areas, since the upper layer is a single-objective problem and the lower layer is a multi-objective problem, according to the characteristics of the problem, the upper layer adopts ICOA and the lower layer adopts NGSA-Ⅱ combined with fuzzy membership function to solve the problem and obtain the distribution network planning-operation plan.

[0086] Preferably, in step three, in solving the upper model based on the ICOA optimization algorithm, the Coati Optimization Algorithm (COA) is improved by using the Tent chaos mapping strategy and the elite retention strategy;

[0087] In the Tent chaotic mapping strategy, since the Tent mapping can generate initial values ​​uniformly distributed in [0,1], it has better ergodicity and uniformity. Therefore, the Tent mapping is used to generate a [0,1] uniform random variable y. The mathematical expression is:

[0088]

[0089] Where: t represents the current iteration number; u represents a random number in [0,1];

[0090] Then transform it to the individual solution space to increase the initial population diversity. The mathematical expression of the individual initial solution is:

[0091] X i =lb+y·(ub-lb), i=1,2,3...,N (25)

[0092] Where: X i Represents the decision variables of the upper-level planning scheme, which consists of PV location, PV planning capacity, energy storage location, and energy storage capacity; ub and lb represent the upper and lower limits of each decision variable, respectively; N represents the population size;

[0093] In the elite retention strategy, the parent and child populations are first merged, then sorted according to the fitness value, and finally the first N individuals are selected as the new population for the next iteration.

[0094] Preferably, in step 3, in solving the lower model based on the NSGA-Ⅱ algorithm, since the dimension of the objective function of the lower model is 3, the NSGA-Ⅱ algorithm not only ensures the convergence of the solution, but also takes into account the diversity of the solution, thereby obtaining a Pareto frontier solution set that meets the conditions; in order to select the most ideal compromise solution so as to pass it to the upper planning layer, a fuzzy membership function is used, as shown in formula (26):

[0095]

[0096] Where: ui is the value of the i-th objective function; u imax and u imin are the upper and lower limits of the i-th objective function respectively;

[0097] Among all solutions, the one with the highest satisfaction value is the best compromise:

[0098]

[0099] Compared with the prior art, the present invention has the following beneficial effects:

[0100] 1. The photovoltaic and energy storage joint optimization model of the distribution network of the present invention first processes the uncertainty of photovoltaic and load. Secondly, the upper planning layer uses the location and capacity of photovoltaic and energy storage as decision variables, and selects the site and determines the capacity of PV and ESS with the goal of minimizing investment cost and operation and maintenance cost; the lower operation layer uses the actual output of photovoltaic in each season and time period and the charging and discharging power of energy storage in each season and time period as decision variables, and dispatches with the comprehensive optimization of small voltage offset, small total network loss and low abandonment rate as the goal; according to the model characteristics, the upper layer uses ICOA to obtain the optimal planning scheme, and the lower layer combines NSGA-Ⅱ with fuzzy membership function to obtain the optimal operation scheme of the distribution network; finally, based on the actual data of a remote mountainous area in the central and western regions, combined with the improved IEEE33 node system, the verification is carried out. The results show that the proposed model effectively improves the voltage quality in remote mountainous areas, makes the power supply of distribution network companies safer, greener and more economical, and the proposed improved solution algorithm has higher solution accuracy.

[0101] 2. The present invention proposes a distribution network photovoltaic storage joint planning model that comprehensively considers power supply in remote mountainous areas, and provides an efficient solution method for the model. The two-layer optimization model can not only effectively reduce network losses, but also has high voltage quality, and can improve the new energy consumption rate, while taking into account the economic benefits of the distribution network.

[0102] 3. The COA algorithm of the present invention does not require basic parameter settings. Compared with traditional algorithms, it can reduce interference conditions in calculations, but it is easy to fall into local optimality. Compared with GA, PSO, and COA algorithms, the improved ICOA using chaos strategy and elite retention strategy can improve the accuracy of model solving and has certain advantages.

[0103] 4. The model established by the present invention comprehensively considers the characteristics of remote mountainous areas, can effectively consider the electricity shortage of the mountainous population, change the quality of electricity consumption of users, and at the same time can effectively smooth load fluctuations, reduce peaks and fill valleys, and can utilize renewable energy in mountainous areas, adapt to local conditions, and provide a reference for power supply in remote mountainous areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Figure 1 The photovoltaic and storage combined double-layer optimization model for the remote mountain distribution network of the present invention;

[0105] Figure 2 The solution process of the photovoltaic energy storage joint optimization model of the distribution network of the present invention;

[0106] Figure 3 The ICOA algorithm of the present invention solves the upper model flow chart;

[0107] Figure 4 The improved IEEE33 node system topology diagram of the present invention;

[0108] Figure 5 The typical photovoltaic daily output curve in Example 4 of the present invention;

[0109] Figure 6 is a typical daily load output curve in Example 4 of the present invention;

[0110] Figure 7 This is a diagram of the 13-node photovoltaic power abandonment in four seasons in Example 4 of the present invention;

[0111] Figure 8 This is a diagram of the four-season charging and discharging conditions of the four-node energy storage in Example 4 of the present invention;

[0112] Fig. 9 This is a diagram showing the peak shaving and valley filling of the total load in four seasons in Example 4 of the present invention;

[0113] Fig.10 Graph showing voltage conditions of nodes in four seasons in Embodiment 4 of the present invention;

[0114] Fig.11 This is a diagram of the four seasons load and the output of each device in Example 4 of the present invention. DETAILED DESCRIPTION

[0115] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0116] Example 1

[0117] 1. PV and load uncertainty processing in remote mountainous areas

[0118] PV output is constrained by light intensity and has obvious uncertainty, but it still has a certain regularity. Its temporal output characteristics conform to a certain probability distribution and have obvious seasonal characteristics.

[0119] The load in remote mountainous areas is generally residential load. The number of people returning home in winter increases, the electricity load increases, and the demand load shows an obvious "winter high" situation. In view of the situation in remote mountainous areas, based on the multi-scenario analysis method, the photovoltaic output and load output are classified into four typical scenarios: spring, summer, autumn and winter.

[0120] After obtaining the uncertainty description of PV output and load, power flow calculations are performed in each scenario to evaluate the operating performance of the model under different conditions.

[0121] 1.1 Probabilistic Model of Photovoltaics

[0122] The light intensity satisfies the Beta distribution, and its probability density function f(S) is:

[0123]

[0124] Where: S represents the actual value of light intensity; S max Represents the maximum light intensity; α and β represent the two parameters of the probability density function respectively.

[0125] Given the probability density function of light intensity, the probability density function of photovoltaic output is:

[0126]

[0127] Where: P PV is the actual output of photovoltaic power; P max is the maximum photovoltaic output, and P max =AηS max , where A is the effective area of ​​the photovoltaic array and η is the photoelectric conversion efficiency.

[0128] 1.2 Probabilistic model of load

[0129] The load in remote mountainous areas is mainly residential load, which obeys the normal distribution:

[0130]

[0131] Where: P L , Q L are active load and reactive load respectively; are the variances of active load and reactive load respectively; μ P , μ Q are the mean values ​​of active load and reactive load respectively.

[0132] 1.3 Power flow calculation considering uncertainty processing

[0133] Based on the historical hourly light intensity and load data in remote mountainous areas, Monte Carlo simulation was used to obtain uncertain light intensity and demand load samples. Secondly, K-means clustering was adopted. Taking spring, summer, autumn and winter as typical scenarios, the day was divided into 24 time periods by the time period division method. The photovoltaic output and load output were clustered to obtain the typical daily output curves of photovoltaic and load in each scenario.

[0134] In the power flow calculation, the steady-state operation of photovoltaic adopts a constant power factor control strategy and is treated as a PQ node, and the load adopts a constant power model and is treated as a PQ node.

[0135] Example 2

[0136] 2. Two-level planning model

[0137] In order to solve the low voltage problem in remote mountainous areas, a photovoltaic-storage synergy model is established by referring to the establishment method of the traditional planning model, as shown in formula (5).

[0138]

[0139] Where: G(·) and H(·) are upper-level constraints, and g(·) and h(·) are lower-level constraints. I 、x O are the upper and lower decision variables respectively. This model is a two-layer bidirectional model.

[0140] The photovoltaic and energy storage joint optimization model for remote mountain distribution network is as follows: Figure 1 As shown. In the upper planning layer, the objective function is selected based on economic efficiency to reduce investment losses and minimize the comprehensive cost in view of the underdeveloped economy in remote mountainous areas. The decision variables are the location capacity of PV and ESS, and the constraints are the investment and site selection constraints of each device. In the lower operation layer, in view of the special conditions of long power supply distance, low terminal voltage, small wire diameter, large active power loss and abundant light resources in remote mountainous areas, in order to improve voltage quality, safety, power supply economy and greenness, reduce network loss and reduce abandoned light, the voltage offset ΔU is used. all Small, total network loss P loss Less, abandoned light rate PV The overall optimality is the goal, and the scheduling of PV and ESS is considered. The constraints are system constraints and the output constraints of each device. I and run the decision variable x o is separable.

[0141] 2.1 Upper-level planning model

[0142] The upper planning layer aims to minimize the comprehensive cost, as shown in formula (6). I Including annual PV investment costs ESS investment cost The sum of O Including the annual PV operation and maintenance costs ESS operation and maintenance costs Cost of purchasing electricity from the upper grid Abandoned light cost sum.

[0143] 2.1.1 Upper-level model objective function

[0144] The objective function F of the upper model is:

[0145] minF=C I +C O (6)

[0146] The cost calculation of each fee is as follows:

[0147] (1) The investment cost for photovoltaic is:

[0148]

[0149] Where: b is the discount rate; y PV is the service life of photovoltaic; Ω PV is the set of photovoltaic candidate node installation locations, c PV The investment cost of a single photovoltaic unit; is the number of PV configurations at node i.

[0150] (2) Cost of investment for ESS:

[0151]

[0152] Where: y ESS is the service life of ESS; Ω ESS is the set of ESS candidate node installation locations, c ESS The investment cost of a single ESS; is the number of ESSs configured at node i.

[0153] (3) PV operation and maintenance costs:

[0154]

[0155] Where: D s is the number of days in the season; PV is a collection of photovoltaic installation node locations; It is the operation and maintenance cost of photovoltaic power generation per unit; is the active power injected by photovoltaic power at node i in season s and period t; Δt is 1h.

[0156] (4) ESS operation and maintenance costs:

[0157]

[0158] Where: ESS A set of ESS installation node locations; It is the operation and maintenance cost of ESS unit charging and discharging power; are the charging and discharging power of ESS at node i, season s, and period t respectively.

[0159] (5) Cost of purchasing electricity from the upper power grid:

[0160]

[0161] Where: A L (1,:) is the set of terminal nodes of the AC branch with the generator node as the head node; P 1j,s,t It is the active power flowing into the local power grid from the upper power grid during period t in season s.

[0162] (6) For abandoned light cost:

[0163]

[0164] Where: The cost of abandoned solar power per unit of electricity; The maximum active power output of the PV installed at node i in period t during season s.

[0165] 2.1.2 Upper-level model constraints

[0166] (1) Installation capacity and power constraints of candidate PV and ESS installation nodes:

[0167]

[0168] Where: The maximum power of the ESS installed at node i in period t during season s; The capacity of the ESS system installed for node i in period t during season s; are the lower and upper limits of the capacity of the ESS installed at node i in period t during season s, respectively.

[0169] (2) Installation quantity constraints for PV and ESS candidate installation nodes:

[0170]

[0171] 2.2 Lower layer operation model

[0172] The lower operating layer, with a voltage offset ΔU all Small, total network loss P loss Less, abandoned light rate PV The goal is low overall optimum.

[0173] 2.2.1 Lower-level model objective function

[0174] The objective function f of the lower model is:

[0175] minf=(ΔU all ,P loss ,l PV ) (15)

[0176] (1)ΔU all is the voltage offset:

[0177]

[0178] Where: is the per unit value of the voltage at node i in season s, period t; Ω AN A collection of communication nodes.

[0179] (2)P LOSS is the total network loss:

[0180]

[0181] Where: P LOSS Represents the active network loss of the system; V i represents the voltage amplitude of node i; G ij , B ij ,δ ij Respectively represent the conductance, susceptance and phase difference between nodes i and j; n is the total number of nodes; Γ represents the set of nodes connected to node j.

[0182] (3)l PV The light abandonment rate is:

[0183]

[0184] Where: is the maximum output value of the i-th photovoltaic at time t.

[0185] 2.2.2 Lower-level model constraints

[0186] (1) DistFlow constraints of AC distribution network

[25] :

[0187]

[0188] Where: A L(j,:) is the set of end nodes of the AC branch with j as the head node; A L (:,j) is the set of the head-end nodes of the AC branch with j as the terminal node; The reactive power output of the PV installed at node i in period t during season s; P ij,s,t , P jk,s,t and Q ij,s,t , Q jk,s,t are the active and reactive power on lines ij and jk in time period t during season s; X ij is the reactance of line ij; and are respectively the active and reactive power of the original load of node j in time period t in season s; U i,s,t The voltage of node i;

[0189] (2) System security constraints:

[0190]

[0191] Where: U i,max and U i,min are the upper and lower limits of the voltage at node i; I ij,max It is the maximum safe current allowed to flow through line ij.

[0192] (4) Power constraints injected into the power grid by the generator node:

[0193]

[0194] Where: Q 1j,s,t P is the reactive power flowing from the upper power grid to the local power grid; 1 max and P 1 min They are the upper and lower limits of active power flowing from the upper power grid to the local power grid; and They are respectively the upper and lower limits of the reactive power flowing into the local power grid from the superior power grid.

[0195] (4) PV operation constraints:

[0196]

[0197] Where: is the power factor angle of PV, which is taken as 0.98.

[0198] (5) ESS operation constraints

[0199] ESS can play the role of "peak shaving and valley filling" and improve PV absorption capacity. It can improve economic efficiency when combined with time-of-use electricity prices.

[0200]

[0201] Where: and Respectively represent the charging and discharging status of ESS, which are 0-1 variables; P i ESS,max is the maximum value of ESS charging and discharging power of node i; is the upper limit of ESS power of node i; η + and η - are the charge and discharge efficiency respectively.

[0202] Example 3

[0203] 3. Algorithm for solving the two-layer model of photovoltaic and storage collaborative optimization of distribution networks in remote mountainous areas

[0204] Since the upper layer is a single-objective problem and the lower layer is a multi-objective problem, according to the characteristics of the problem, the upper layer uses ICOA and the lower layer uses NGSA-Ⅱ combined with fuzzy membership function to solve the problem and obtain the distribution network planning-operation plan. The solution process is as follows: Figure 2 The process of ICOA solving the upper model is as follows Figure 3 shown.

[0205] 3.1 Solution of upper model based on ICOA optimization algorithm

[0206] Coati Optimization Algorithm (COA) is a new meta-heuristic optimization algorithm proposed by Dehghani Mohammad et al. in 2023.

[26] Compared with classic algorithms such as genetic algorithm (GA) and particle swarm algorithm (PSO), COA has faster convergence speed and higher optimization accuracy, and there is no basic parameter setting. In this model, the use of COA can reduce the risk of performance degradation due to improper parameter setting and ensure the accuracy of the decision variables transmitted by the upper model. However, since the algorithm uses a random initial population, it is difficult to spread the population and cannot reflect the diversity of the population. It is easy to produce local optimality, and the greedy selection strategy cannot save high-quality parent populations. Therefore, the chaotic strategy and elite retention strategy are used to improve the algorithm.

[0207] 3.1.1 Tent Chaos Mapping

[0208] Since the Tent map can generate initial values ​​uniformly distributed in [0,1], it has better ergodicity and uniformity. Therefore, the Tent map is used to generate a [0,1] uniform random variable y. The mathematical expression is:

[0209]

[0210] Where: t represents the current iteration number; u represents a random number in [0,1].

[0211] Then transform it to the individual solution space to increase the initial population diversity. The mathematical expression of the individual initial solution is:

[0212] X i =lb+y·(ub-lb), i=1,2,3...,N (25)

[0213] Where: X i It represents the decision variables of the upper-level planning scheme, which consists of PV location, PV planning capacity, energy storage location and energy storage capacity; ub and lb represent the upper and lower limits of each decision variable respectively; N represents the population size.

[0214] 3.1.2 Elite Retention Strategy

[0215] In order to retain high-quality individuals before the next iteration and improve convergence, an elite retention strategy is proposed. First, the parent and child populations are merged, then sorted according to the fitness value, and finally the first N individuals are selected as the new population for the next iteration.

[0216] 3.2 Solution of the lower model based on NSGA-Ⅱ algorithm

[0217] The NSGA-Ⅱ algorithm optimizes the calculation process and is particularly effective in dealing with complex optimization problems. Since the dimension of the objective function of the lower model is 3, the NSGA-Ⅱ algorithm not only ensures the convergence of the solution, but also takes into account the diversity of solutions, thereby obtaining a Pareto frontier solution set that meets the conditions. In order to select the most ideal compromise solution so that it can be passed to the upper planning layer, a fuzzy membership function is used, as shown in formula (26):

[0218]

[0219] Where: u i is the value of the i-th objective function; u imax and u imin are the upper and lower limits of the i-th objective function respectively.

[0220] Among all solutions, the one with the highest satisfaction value is the best compromise.

[0221]

[0222] Example 4

[0223] 4. Example Analysis

[0224] 4.1 Parameter settings

[0225] In order to verify the effectiveness of the proposed model, the improved IEEE33-node distribution network is taken as the research object, and its topology diagram is shown in Figure 4 As shown in Figure 2, the distance and impedance data of the conductor in a remote mountainous area in the central and western regions are used as line parameters, and the 1h-level light intensity and load in 2023 are used as basic data for processing to obtain the typical daily photovoltaic output curve and load output curve of the remote mountainous area in four seasons, as shown in Figure 2. Figure 5 , Figure 6 The specific parameters of photovoltaic and energy storage are shown in Table 1. The penalty cost for abandoning light is 0.6 yuan / kWh, the initial state of charge of energy storage is 0.5, the upper and lower limits are 0.9 and 0.1 respectively, and the charging and discharging efficiency is 0.9. Due to the geographical factors of the mountainous area, the location and capacity suitable for installing photovoltaic and energy storage are shown in Table 2, and the electricity prices at different time periods in remote mountainous areas are shown in Table 3.

[0226] Table 1 Photovoltaic and energy storage parameters

[0227]

[0228] Table 2 Restrictions on photovoltaic and energy storage in remote mountainous areas

[0229]

[0230] Table 3 Electricity prices at different times in remote mountainous areas

[0231]

[0232] 4.2 Results Analysis

[0233] In order to verify the effectiveness of the proposed model, the following scheme is set for comparison.

[0234] Model 1: The model proposed in this paper;

[0235] Model 2: Based on Model 1, energy storage is not considered when planning, only photovoltaics are considered;

[0236] Model 3: Based on Model 1, PV is not considered during planning, only energy storage is considered;

[0237] Model 4: Based on Model 1, the lower-level objective function is replaced by only considering the operation and maintenance cost during the simulation operation;

[0238] Models 1, 2, and 3 compare and analyze the planning results under single-agent and multi-agent conditions from the perspective of participating entities to verify the rationality of the proposed models. Models 1 and 4 compare and analyze the improvement of voltage quality in remote mountainous areas under different goal orientations from the perspective of objective functions to verify the necessity of setting the objective function of the proposed models.

[0239] The planning results of each model are shown in Table 4, and the planning objective function results are shown in Table 5.

[0240] 4.2.1 Analysis of the results of the constructed model

[0241] The specific costs of the planning scheme of the constructed model are shown in Table 6.

[0242] Table 4 Results of planning schemes for each model

[0243]

[0244] Table 5 Objective function results of each model planning scheme

[0245]

[0246] Table 6 Costs of the planning scheme of the model

[0247]

[0248] From the analysis of Table 5 and Table 6, we can see that the comprehensive cost of the planning scheme of the model in this paper is 9.7664 million yuan, of which the annual investment cost is 917,600 yuan and the annual operating cost is 8.8488 million yuan; the abandonment rate is only 5.68%, which makes the power supply of the distribution network company low-carbon; the active power loss is 377.56MW, which is 24.05% lower than that of the original distribution network, making the power supply of the distribution network company more economical; the voltage offset is 5894.33, which is 20.25% lower than that of the original distribution network; making the power supply of the distribution network company safer and of higher quality.

[0249] The total installed capacity of photovoltaic power reached 1.54MW, the penetration rate was 41.45%, and the total absorption rate reached 94.32%. The absorption rates in the four seasons reached 98.40%, 97.75%, 99.11%, and 99.05% respectively, indicating that the model proposed in this paper has transformed the distribution network in remote mountainous areas into a "green distribution network". Taking the photovoltaic power planned with 13 nodes as an example, the abandoned light situation in the four seasons is as follows: Figure 7 The situation of curtailment of solar power mostly occurs during the peak period of photovoltaic power output. This is because the photovoltaic power output is high during this period, while the load demand in the mountainous area is low. However, some of the excess photovoltaic power output will charge the energy storage, so the overall curtailment rate is low.

[0250] Taking the energy storage planned for 4 nodes as an example, its charging and discharging conditions in four seasons are as follows: Figure 8As shown, positive values ​​indicate discharge, and negative values ​​indicate charge. In general, energy storage is basically charged during the low load period, acting as a "load", discharged during the peak period, acting as a "power source", and is in a static state during normal periods. The spring state of energy storage at this location is analyzed in detail in combination with the spring load output curve and electricity price: 1:00-7:00, the energy storage is in a charging state. This is because this period is in the low load period in the mountainous area, the photovoltaic output is almost 0, but the electricity price is low, and the energy storage is charged by purchasing electricity from the upper power grid. Although 8:00-9:00 is also in the low period, the energy storage is in a static state because the energy storage SOC state reaches 0.9. In order to extend the life of the energy storage device, charging is stopped, but the load level is low at this time, so the energy storage does not discharge. From 13:00 to 18:00, although the electricity price is in the normal period, the photovoltaic output is high, so the energy storage is in a charging state to absorb part of the photovoltaic output. The energy storage is in a discharging state from 10:00 to 12:00 and 19:00 to 24:00. This is because it is the peak period in the mountainous area, with a high load level demand and low light intensity, resulting in less photovoltaic output but a higher electricity price. Therefore, the energy storage is discharged to reduce electricity purchases from higher levels and improve the economy of the distribution network company.

[0251] The node voltages all meet the requirements, among which the voltage level is lower in winter, ranging from 0.932 to 1.0pu. This is because people in remote mountainous areas return home in winter and the load demand level is greatly increased, resulting in the terminal node voltage level being lower than in other seasons, but still meeting the voltage quality requirements.

[0252] The output of each device in four seasons is as follows Fig.11 As shown in the analysis, it can be seen that in general, during the peak period of photovoltaic output, the amount of electricity purchased from the upper power grid is greatly reduced; energy storage charging is mostly during the low load period, and discharge is mostly during the peak load period; active power loss is higher in winter because the load demand is larger in winter and the photovoltaic output in winter is less than in other seasons.

[0253] In order to reflect the peak-shaving and valley-filling effect of energy storage, the total load before and after access to energy storage in each period of the four seasons is as follows: Fig. 9 Analysis shows that the original load has a large peak-to-valley difference overall, with the maximum being 2664.93kW in winter. After energy storage is connected, the peak-to-valley difference is reduced to 2455.16kW, a decrease of 7.98%, indicating that energy storage can improve the load distribution level, has a more obvious "peak shaving and valley filling" effect, and reduces the impact on the distribution network.

[0254] The voltage conditions of each node in four seasons are as follows Fig.10After planning, the voltage of each node meets the requirements. The voltage level in winter is relatively low, ranging from 0.932 to 1.0 pu. This is because people in remote mountainous areas return home in winter, and the load demand level is greatly increased, resulting in a lower voltage level at the terminal node than in other seasons, but it still meets the voltage quality requirements.

[0255] The output of each device in four seasons is as follows Fig.11 As shown in the analysis, it can be seen that in general, during the peak period of photovoltaic output, the amount of electricity purchased from the upper power grid is greatly reduced; energy storage charging is mostly during the low load period, and discharge is mostly during the peak load period; active power loss is higher in winter because the load demand is larger in winter and the photovoltaic output in winter is less than in other seasons.

[0256] 4.2.2 Model comparison results analysis

[0257] By analyzing Table 4 and Table 5 and comparing the results of Model 1 and 2, it can be seen that Model 2 does not consider energy storage, but the comprehensive cost increases by 51,600 yuan. This is because although the investment cost of energy storage is reduced, due to the lack of charging and discharging function of energy storage, when the load demand is high at night, electricity is purchased from the superior power grid, resulting in increased operating costs and ultimately increased comprehensive costs; the number of planned photovoltaics decreased, the penetration rate was 37.68%, a decrease of 3.77%, and the annual abandonment rate increased by 0.96%; the active power loss increased by 40.38MW, and the voltage offset increased by 115.25.

[0258] Comparing the results of models 1 and 3, it can be seen that although model 3 does not consider photovoltaics, the comprehensive cost has increased by 62,900 yuan. This is because although the photovoltaic investment cost has been reduced, the operating cost has increased due to the high electricity purchase price from the upper power grid during the peak load period at noon, resulting in an increase in comprehensive costs; the photovoltaic penetration rate is 0%, which is not conducive to the low-carbon transformation of distribution network companies; the active power loss has increased by 107.01MW, and the voltage offset has increased by 1428.22. Models 2 and 3 are not conducive to the power supply security, quality and economy of distribution network companies.

[0259] By comparing the results of models 1 and 4, it can be seen that when model 4 is simulated and operated, power supply security, low carbon and high quality are not considered, and only economic efficiency is considered. The comprehensive cost is reduced by 22,300 yuan, and the reduced costs are all reflected in the operating costs; but the abandonment rate in terms of low carbon efficiency has increased by 0.45%, the voltage offset in terms of safety and quality has increased by 69.31, and the active power loss has increased by 12.86MW, which is equivalent to an increased cost of about 9,800 yuan. Therefore, when only economic efficiency is considered, the effect is general and does not meet the model planning requirements.

[0260] 4.2.3 Analysis of solution results

[0261] In order to verify the effectiveness of the proposed solution method and the superiority of the improved upper-layer algorithm in terms of accuracy, the upper layer selects GA, PSO, and COA algorithms for solution, and the lower layer solution method remains unchanged for comparison. The comparison results are shown in Table 7. Analysis shows that the solution accuracy of the proposed solution method in the upper layer is improved by 1.72%, 1.31%, and 0.61% respectively compared with other methods, and the convergence speed is improved by 51.22%, 42.86%, and 25.93% respectively. This is because the COA algorithm itself does not need to set parameters compared with GA and PSO, and has better performance.

[26] However, since the random initialization strategy is still used when initializing the population, it is difficult to spread the population and cannot reflect the diversity of the population. There is still a risk of falling into the local optimal solution. The initialization population obtained after the improvement by the chaotic mapping strategy is better and not easy to fall into the local optimal solution. Therefore, the convergence times are accelerated and the solution accuracy is improved. However, due to the large comprehensive cost benchmark value, the improvement percentage is low. In terms of solution time, although the algorithm divides the population into two parts compared with GA and PSO, the total number of objective function calculations is the same as that of GA and PSO, so there is almost no change compared with GA and PSO. The improvement strategy proposed in this paper has no improvement in the time scale, but it will not increase the time further. In summary, the proposed algorithm has certain advantages in solution accuracy.

[0262] Table 7 Comparison of solution algorithm results

[0263]

[0264] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for joint planning of photovoltaic and storage distribution networks in remote mountainous areas. It is characterized in that The steps include: Step 1: Establish a PV and load uncertainty processing model for remote mountainous areas; the PV and load uncertainty processing model for remote mountainous areas includes a photovoltaic probability model, a load probability model, and a power flow calculation considering uncertainty processing; Step 2: Establish a two-layer planning model; the two-layer planning model includes an upper-layer planning model and a lower-layer operation model; Step 3. An algorithm for solving the double-layer model of photovoltaic and storage collaborative optimization for distribution networks in remote mountainous areas; the algorithm for solving the double-layer model of photovoltaic and storage collaborative optimization for distribution networks in remote mountainous areas includes an upper-layer model solution based on the ICOA optimization algorithm and a lower-layer model solution based on the NSGA-Ⅱ algorithm.

2. According to claim 1, a method for joint planning of photovoltaic and storage distribution networks in remote mountainous areas, Features: In step 1, in the photovoltaic probability model, the light intensity satisfies the Beta distribution, and its probability density function f(S) is: Where: S represents the actual value of light intensity; S max represents the maximum light intensity; α and β represent the two parameters of the probability density function respectively; Given the probability density function of light intensity, the probability density function of photovoltaic output is: Where: P PV is the actual output of photovoltaic power; P max is the maximum photovoltaic output, and P max =AηS max , where A is the effective area of ​​the photovoltaic array and η is the photoelectric conversion efficiency.

3. According to claim 2, a method for joint planning of photovoltaic and storage distribution networks in remote mountainous areas, Features: In step 1, in the probability model of the load, the load in remote mountainous areas is mainly residential load, which obeys the normal distribution: Where: P L , Q L They are active load and reactive load respectively; are the variances of active load and reactive load respectively; μ P , μ Q are the mean values ​​of active load and reactive load respectively.

4. According to claim 3, a method for joint planning of photovoltaic and storage distribution networks in remote mountainous areas, Features: In step one, in the flow calculation considering uncertainty processing, based on the historical hourly light intensity and load data in remote mountainous areas, Monte Carlo simulation is used to obtain light intensity and demand load samples with uncertainty. Secondly, K-means clustering is used, with spring, summer, autumn and winter as typical scenes, and a day is divided into 24 time periods by time division method. The photovoltaic output and load output are clustered to obtain the typical daily output curves of photovoltaic and load in each scene. In the flow calculation, the steady-state operation of photovoltaic adopts a constant power factor control strategy and is processed as a PQ node, and the load adopts a constant power model and is processed as a PQ node.

5. According to claim 4, a method for joint planning of photovoltaic and storage in remote mountainous distribution networks, Features: In step 2, in the two-layer planning model, including the upper planning layer and the lower operation layer, in order to solve the low voltage problem in remote mountainous areas, the photovoltaic storage synergy model is established with reference to the establishment method of the traditional planning model, as shown in formula (5): Where: G(·) and H(·) are upper-level constraints, g(·) and h(·) are lower-level constraints; x I 、x O are upper and lower level decision variables respectively, and the model is two-level and bidirectional; Among them, the upper planning layer, in view of the underdeveloped economy in remote mountainous areas, selects the objective function based on economy, reduces investment losses, and takes the minimum comprehensive cost as the goal. The decision variables are the location capacity of PV and ESS, and the constraints are the investment and site selection constraints of each equipment; the lower operation layer, in view of the special conditions of long power supply distance, low terminal voltage, small wire diameter, large active power loss and abundant light resources in remote mountainous areas, in order to improve voltage quality, improve safety, power supply economy and greenness, reduce network loss, and reduce abandoned light, the voltage offset ΔU is used. all Small, total network loss P loss Less, abandoned light rate PV The overall optimal goal is to consider the scheduling of PV and ESS, and the constraints are system constraints and output constraints of each device; the planning decision variable x in the two-layer model I and run the decision variable x o is separable.

6. According to claim 5, a method for joint planning of photovoltaic and storage in remote mountainous distribution networks, Features: The upper-level planning model includes an upper-level model objective function and an upper-level model constraint condition; In the upper model objective function, the upper planning layer takes the minimum comprehensive cost as the goal, as shown in formula (6); C I Including annual PV investment costs ESS investment cost The sum of O Including the annual PV operation and maintenance costs ESS operation and maintenance costs Cost of purchasing electricity from the upper grid Abandoned light cost sum; The objective function F of the upper model is: min F=C I +C O (6) The cost calculation of each fee is as follows: (1) The investment cost for photovoltaic is: Where: b is the discount rate; y PV is the service life of photovoltaic; Ω PV is the set of photovoltaic candidate node installation locations, c PV The investment cost of a single photovoltaic unit; is the number of photovoltaic configurations at node i; (2) Cost of investment for ESS: Where: y ESS is the service life of ESS; Ω ESS is the set of ESS candidate node installation locations, c ESS The investment cost of a single ESS; is the number of ESSs configured at node i; (3) PV operation and maintenance costs: Where: D s is the number of days in the season; PV is a collection of photovoltaic installation node locations; It is the operation and maintenance cost of photovoltaic power generation per unit; is the active power injected by photovoltaic power in the i-node in the t-period of the season; Δt is 1h; (4) ESS operation and maintenance costs: Where: ESS A set of ESS installation node locations; It is the operation and maintenance cost of ESS unit charging and discharging power; are the ESS charging and discharging power of node i, season s, period t; (5) Cost of purchasing electricity from the upper power grid: Where: A L (1,:) is the set of terminal nodes of the AC branch with the generator node as the head node; P 1j,s,t is the active power flowing into the local power grid from the upper power grid during period t in season s; (6) Cost of abandoned light: Where: The cost of abandoned solar power per unit of electricity; The maximum active power output of the photovoltaic installed at node i in period t during season s; The upper model constraints are: (1) Installation capacity and power constraints of candidate PV and ESS installation nodes: Where: The maximum power of the ESS installed at node i in period t during season s; The capacity of the ESS system installed for node i in period t during season s; are the lower and upper limits of the capacity of the ESS installed at node i in period t during season s; (2) Installation quantity constraints for PV and ESS candidate installation nodes:

7. According to claim 6, a method for joint planning of photovoltaic and storage in remote mountainous distribution networks, Features: In step 2, the lower layer operation model includes the lower layer model objective function and the lower layer model constraint conditions; In the objective function of the lower model, the lower operating layer is based on the voltage offset ΔU all Small, total network loss P loss Less, abandoned light rate PV The goal is low overall optimum; The objective function f of the lower model is: min f=(ΔU all ,P loss ,l PV ) (15) (1)ΔU all is the voltage offset: Where: is the per unit value of the voltage at node i in season s, period t; Ω AN is a collection of communication nodes; (2)P LOSS is the total network loss: Where: P LOSS Represents the active network loss of the system; V i represents the voltage amplitude of node i; G ij , B ij ,δ ij Respectively represent the conductance, susceptance and phase angle difference between nodes i and j; n is the total number of nodes; Γ represents the set of nodes connected to node j; (3)l PV The light abandonment rate is: Where: is the maximum output value of the i-th photovoltaic at time t; The lower model constraints are: (1) DistFlow constraints of AC distribution network: Where: A L (j,:) is the set of end nodes of the AC branch with j as the head node; A L (:,j) is the set of the head-end nodes of the AC branch with j as the terminal node; The reactive power output of the PV installed at node i in period t during season s; P ij,s,t , P jk,s,t and Q ij,s,t , Q jk,s,t are the active and reactive power on lines ij and jk in time period t during season s; X ij is the reactance of line ij; and are respectively the active and reactive power of the original load of node j in time period t in season s; U i,s,t The voltage of node i; (2) System security constraints: Where: U i,max and U i,min are the upper and lower limits of the voltage at node i; I ij,max is the maximum safe current allowed to flow through line ij; (3) Power constraints injected into the power grid by the generator node: Where: Q 1j,s,t P is the reactive power flowing from the upper power grid to the local power grid; 1 max and P 1 min They are the upper and lower limits of the active power flowing from the upper power grid to the local power grid; Q 1 max and Q 1 min They are the upper and lower limits of reactive power flowing from the upper power grid to the local power grid; (4) PV operation constraints: Where: is the power factor angle of PV, which is taken as 0.98; (5) ESS operation constraints ESS can play the role of "peak shaving and valley filling" and improve PV absorption capacity. It can also improve economic efficiency when combined with time-of-use electricity prices. Where: and Respectively represent the charging and discharging status of ESS, which are 0-1 variables; P i ESS,max is the maximum value of ESS charging and discharging power of node i; is the upper limit of ESS power of node i; η + and η - are the charge and discharge efficiency respectively.

8. According to claim 7, a method for joint planning of photovoltaic and storage in remote mountainous area distribution network, Features: In step 3, in solving the upper model based on the ICOA optimization algorithm, the Coati Optimization Algorithm (COA) is improved by using the Tent chaos mapping strategy and the elite retention strategy; In the Tent chaotic mapping strategy, since the Tent mapping can generate initial values ​​uniformly distributed in [0,1], it has better ergodicity and uniformity. Therefore, the Tent mapping is used to generate a [0,1] uniform random variable y. The mathematical expression is: Where: t represents the current iteration number; u represents a random number in [0,1]; Then transform it to the individual solution space to increase the initial population diversity. The mathematical expression of the individual initial solution is: X i =lb+y·(ub-lb),i=1,2,3...,N (25) Where: X i Represents the decision variables of the upper-level planning scheme, which consists of PV location, PV planning capacity, energy storage location, and energy storage capacity; ub and lb represent the upper and lower limits of each decision variable, respectively; N represents the population size; In the elite retention strategy, the parent and child populations are first merged, then sorted according to the fitness value, and finally the first N individuals are selected as the new population for the next iteration.

9. A method for joint planning of photovoltaic and storage in remote mountainous distribution networks according to claim 8, Features: In step 3, in the solution of the lower model based on the NSGA-Ⅱ algorithm, since the dimension of the objective function of the lower model is 3, the NSGA-Ⅱ algorithm not only ensures the convergence of the solution, but also takes into account the diversity of the solution, thereby obtaining a Pareto frontier solution set that meets the conditions; in order to select the most ideal compromise solution so as to pass it to the upper planning layer, a fuzzy membership function is used, as shown in formula (26): Where: u i is the i-th objective function value; u imax and u imin are the upper and lower limits of the i-th objective function respectively; Among all solutions, the one with the highest satisfaction value is the best compromise:

Citation Information

Patent Citations

  • Whale optimized damaged terracotta figure fragment registration method based on chaos reverse learning

    CN113034554A

  • Power distribution network double-layer optimization scheduling method considering wind and light uncertainty

    CN114825348A

  • Low-carbon power distribution network double-layer planning method considering OLTC fuzzy control

    CN116565838A

  • Source load coordinated peak regulation method and system considering demand side response

    CN117439196A

  • Multi-scene operation-oriented optical storage optimization configuration method and terminal

    CN117996796A

Cited By

  • Rural energy system double-layer planning method considering flexible load cooperation

    CN120414637A

  • Power distribution network multi-electric energy router planning method based on time sequence region division

    CN120657766A

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

    CN120657766B