A two-tier configuration method for distributed photovoltaic capacity considering both power generation and stability

Abstract

The present invention belongs to the field of distributed photovoltaic technology and relates to a two-tier distributed photovoltaic capacity configuration method that comprehensively considers both power generation and stability. The method constructs an upper-tier objective function with maximum power generation as the goal, a first lower-tier evaluation function with voltage stability as the goal, and a second lower-tier evaluation function with the year-on-year growth rate of network losses before and after the distributed photovoltaic connection. The third lower-tier evaluation function uses the average rate of change of static voltage before and after the distributed photovoltaic connection. The upper-tier objective function is solved to obtain the first c solutions. Starting from the optimal solution, each solution is subjected to stability evaluation using the three lower-tier evaluation functions, and the optimal solution that satisfies the stability evaluation is selected as the Class A and Class B distributed photovoltaic installed capacity configuration scheme. The present invention determines the optimal configuration capacity for Class A and Class B distributed photovoltaics using the upper-tier objective function and the three lower-tier evaluation functions, taking into account both power generation and stability requirements.

Description

Technical Field

[0001] The present invention belongs to the field of distributed photovoltaic technology and relates to a distributed photovoltaic capacity double-layer configuration method that comprehensively considers power generation and stability. Background Art

[0002] Distributed photovoltaic systems with a voltage of 10kV or below can be divided into two categories: PV systems connected to the public grid at 10kV are called Class A distributed photovoltaic systems and are intended for general commercial and industrial use, as well as for use by non-individual households. PV systems connected to the grid at a voltage of 380V are called Class B distributed photovoltaic systems, or household use, and include those used by individual households. The differences between Class A and Class B distributed photovoltaic systems are shown in Table 1.

[0003] Table 1

[0004]

[0005] Existing distributed photovoltaic capacity configuration methods have numerous shortcomings. For one thing, most fail to balance the power generation revenue of distributed photovoltaic service providers with the stability requirements of the grid. This can lead to suboptimal power generation in practical applications, or excessive pursuit of power generation can impact grid stability. Furthermore, when calculating and optimizing capacity, the algorithms employed may be inefficient or prone to falling into local optimal solutions, making it impossible to accurately identify the optimal configuration. For example, the standard particle swarm algorithm (PSO) is prone to falling into local optimal solutions when dealing with complex distributed photovoltaic capacity configuration problems, making it difficult to achieve an ideal configuration. These issues limit the rational application and development of distributed photovoltaics in power grids. Summary of the Invention

[0006] The purpose of this patent is to provide a two-tier configuration method for distributed photovoltaic capacity that comprehensively considers power generation and stability. The invention aims to determine the optimal configuration capacity of Class A and Class B distributed photovoltaics from the dual perspectives of distributed photovoltaic service providers (including self-use loads) and the grid side, under various constraints such as limited investment and grid stability, to maximize the power generation of distributed photovoltaic service providers while ensuring stable operation of the grid.

[0007] The present invention comprehensively considers the power generation and stability of distributed photovoltaic capacity dual-layer configuration method. The photovoltaic power generation system connected to the public grid through 10kV is called Class A distributed photovoltaic, and the photovoltaic power generation system connected to the grid through 380V voltage level is called Class B distributed photovoltaic. The upper objective function is constructed with the maximum power generation as the goal. :

[0008] ;

[0009] Where: is the maximum annual power generation hours of distributed photovoltaics, is the total power generation of distributed photovoltaic power generation of Class A and B, r represents the year, R is the project life; x is the discount rate, It is the Class A distributed photovoltaic installed capacity, It is the Class B distributed photovoltaic installed capacity;

[0010] Construct the first lower-level evaluation function with voltage stability as the goal:

[0011] ;

[0012] Where: represents the first lower-level evaluation function with voltage stability as the goal, , 、 、 are three weight coefficients, the sum of the three is 1; M is the number of nodes in the line, is the per-unit voltage value of node m, 、 They are the per-unit voltage values ​​of Class A and Class B distributed photovoltaic grid-connected points respectively; Respectively represent Class A and B distributed photovoltaic grid connection points;

[0013] The year-on-year growth rate of network loss before and after distributed photovoltaic access As the second lower evaluation function; the average change rate of static voltage before and after distributed photovoltaic access is the third lower layer evaluation function;

[0014] Solve the upper objective function , get the first c solutions, start from the optimal solution, use the three lower-level evaluation functions to evaluate the stability of each solution one by one, and select the best solution that meets the stability evaluation as the A and B distributed photovoltaic installed capacity configuration scheme.

[0015] Specifically, the second lower layer evaluation function is:

[0016] ;

[0017] Where: is the threshold value of the year-on-year growth rate of network loss rate; is the total network loss before distributed photovoltaic access; is the total network loss after distributed photovoltaic access.

[0018] Specifically, the third lower layer evaluation function is:

[0019] ;

[0020] ;

[0021] Where: is the per-unit voltage value of node m before the distributed photovoltaic is connected, is the per-unit voltage value of node m after distributed photovoltaic access, is the static voltage average change rate threshold.

[0022] More preferably, the upper objective function The constraints include: investment constraints, reverse load rate constraints, and line transmission capacity constraints.

[0023] Further optimization, using penalty function in the upper objective function Based on this, a new comprehensive objective function is constructed:

[0024] ;

[0025] Where: is the penalty factor, is the inequality constraint condition of the objective function, including investment constraint, reverse load rate constraint, and line transmission capacity constraint. The constructed penalty function is ;

[0026] The comprehensive objective function is solved by particle swarm optimization. During the solution process, the penalty factor is dynamically adjusted:

[0027] ;

[0028] Where: is the penalty factor for the g-th iteration, is the proportion of solutions that meet the constraints in the current population, is the target proportion of the current population that meets the constraint solution, is the adjustment factor, which is a constant greater than 1.

[0029] Further preferably, in the process of solving the comprehensive objective function by the particle swarm algorithm, the average position vector of the current particle is set for:

[0030] ;

[0031] Where: is the average position of the particle swarm in the dth dimension of the search space, is the number of particles in the population dimensional position, d={1,2,…,D}, N is the total number of particles;

[0032] Then the position variance of the particle swarm is for:

[0033] ;

[0034] Where, is the position variance of the particle swarm in the dth dimension of the search space, is a variable factor, and the value selection rules are as follows:

[0035] ;

[0036] For standard deviation exist For particles within the range, the velocity variation formula is calculated using the following formula:

[0037] ;

[0038] Where, is the number of particles in the population Dimensional speed, is the maximum velocity of all particles in the dth dimension, and rand represents the generation of a random number between 0 and 1;

[0039] For the standard deviation For particles within the range, the velocity variation formula is calculated using the following formula:

[0040] ;

[0041] Where, 、 are the maximum and minimum values ​​of all particle positions.

[0042] For the standard deviation For particles within the range, the velocity variation formula is calculated using the following formula:

[0043] ;

[0044] Where, is the minimum velocity of all particles in the dth dimension.

[0045] Further preferably, the three lower-level evaluation function constraints include:

[0046] Short-circuit current constraint: , where are the short-circuit current value and the maximum short-circuit current limit of node m respectively;

[0047] Node voltage constraints: , where is the per-unit voltage value of node m after distributed photovoltaic access, is the voltage rating;

[0048] Distributed photovoltaic output constraints: , where is a node The maximum output of the connected distributed photovoltaics, is a node The maximum capacity of the connected distributed photovoltaic access node.

[0049] Beneficial effects of the present invention:

[0050] By taking the maximum power generation as the optimization goal through the upper-level objective function, comprehensively considering the characteristics and cost-benefit model of Class A and B distributed photovoltaics, and determining the optimal configuration capacity under investment constraints, it helps distributed photovoltaic service providers to obtain maximum power generation benefits, improve the return on investment, and promote the economic sustainable development of the distributed photovoltaic industry.

[0051] The three lower-level evaluation functions evaluate the upper-level optimization results from multiple aspects, such as voltage stability, network loss change rate, and static voltage change rate, and set multiple constraints to ensure that the access of distributed photovoltaics will not have a negative impact on the stable operation of the power grid, maintain grid voltage stability, reduce network losses, and improve the reliability and safety of grid operation.

[0052] The improved particle swarm optimization algorithm effectively avoids falling into local optimality by introducing the concept of particle position variance and the corresponding velocity variation formula, improving the optimization accuracy and efficiency. It can more accurately find the optimal configuration capacity of distributed photovoltaics that meets multiple constraints.

[0053] This invention takes into account multiple factors such as investment, power generation efficiency, transmission loss, and grid stability from the perspectives of both distributed photovoltaic service providers and the grid side. Compared with traditional configuration methods, its configuration results are more scientific and reasonable, more in line with actual engineering needs, and are conducive to promoting the widespread application and efficient development of distributed photovoltaics in distribution networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Schematic diagram of a typical line model of a distribution network.

[0055] Figure 2 Flowchart of the present invention. DETAILED DESCRIPTION

[0056] The present invention is further described below with reference to the embodiments. It is necessary to point out that the following embodiments are only used to further illustrate the present invention and are not to be construed as limiting the scope of protection of the present invention. Non-essential improvements and adjustments made by persons skilled in the art based on the above-mentioned invention contents still fall within the scope of protection of the present invention.

[0057] A two-tiered distributed photovoltaic capacity configuration method comprehensively considers both power generation and stability. This method considers the configuration capacity of Class A and B distributed photovoltaic systems from both the perspectives of distributed photovoltaic service providers (which also include self-use loads) and the grid. The overall goal is to determine the optimal configuration capacity of Class A and B distributed photovoltaic systems within limited investment constraints, subject to various constraints such as grid stability, to maximize power generation for distributed photovoltaic service providers. The upper-tier objective function optimizes maximum power generation and uses an improved particle swarm optimization algorithm to find the optimal capacity configuration for Class A / B photovoltaic systems, with the goal of achieving maximum power generation, within certain investment constraints. The lower-tier objective optimizes grid stability and primarily evaluates whether the upper-tier optimization results meet grid-side stability control conditions such as voltage stability, network losses, voltage variation rate, and short-circuit current constraints, balancing the PV service provider's maximum power generation requirements with the grid-side stability control objectives.

[0058] The upper layer aims to maximize the power generation of distributed photovoltaic service providers, and Class A and Class B distributed photovoltaics differ in unit power investment cost, power generation efficiency, and transmission loss rate.

[0059] Class A distributed photovoltaic adopts the conventional full grid access mode to establish the annual power generation model of Class A distributed photovoltaic:

[0060] The annual power generation mainly considers the lighting resources and component attenuation, and can be calculated using the following formula:

[0061] ;

[0062] Where: is the power generation of Class A distributed photovoltaic in year r, 、 They are the installed capacity of Class A distributed photovoltaic power generation and the number of peak sunshine hours, r represents the year (r=0 is the initial investment year, r=1,2,…,R is the operation year), and R represents the project life; is the attenuation rate of Class A distributed photovoltaic, is the system efficiency, is the line loss rate of Class A distributed photovoltaic.

[0063] Total power generation of Class A distributed photovoltaics:

[0064] ;

[0065] Where: is the total power generation of Class A distributed photovoltaics, and x is the discount rate (usually the project capital cost or social discount rate).

[0066] Class B distributed photovoltaics adopt the conventional self-generation and self-consumption mode, that is, after deducting the self-consumption portion of the load, the surplus power is connected to the grid. The annual power generation model of Class B distributed photovoltaics is established:

[0067] The annual power generation mainly considers the lighting resources and component attenuation, and can be calculated using the following formula:

[0068] ;

[0069] Where: is the power generation of Class B distributed photovoltaic in year r, 、 They are Class B distributed photovoltaic installed capacity, peak sunshine hours, is the attenuation rate of Class B distributed photovoltaics, is the system efficiency, is the line loss rate of Class B distributed photovoltaic;

[0070] Total power generation of Class B distributed photovoltaics:

[0071] ;

[0072] Where: is the total power generation of Class B distributed photovoltaics.

[0073] The maximum power generation is defined as the total power generation during the entire life cycle of photovoltaics. Considering the combination of Class A and Class B distributed photovoltaics, the total power generation of Class A and Class B distributed photovoltaics is It can be expressed as:

[0074] ;

[0075] The larger it is, the greater the power generation capacity.

[0076] In order to facilitate the optimization, the objective function is normalized, and finally the upper objective function with the maximum power generation as the goal is obtained The following formula represents:

[0077] ;

[0078] Where: The maximum annual power generation hours of distributed photovoltaics;

[0079] Upper-level objective function The constraints include:

[0080] (1) Investment constraints:

[0081] ;

[0082] Where: It is the largest distributed photovoltaic service provider with the largest total investment. 、 They are the investment costs of Class A and Class B distributed photovoltaics respectively.

[0083] (2) Reverse load rate constraint:

[0084] Reverse load rate refers to the ratio of the reverse power flowing through the power transmission and transformation equipment (usually lines or transformers) to the rated capacity of the equipment. If the reverse load rate exceeds 80%, it is necessary to temporarily suspend the access of distributed photovoltaic projects. The expression is:

[0085] ;

[0086] ;

[0087] ;

[0088] Where: is the maximum grid-connected power of Class A distributed photovoltaics, It is the maximum grid-connected power of Class B distributed photovoltaics, usually a typical number. is the rated transmission capacity of line l; 、 They are the power generation efficiency of Class A and Class B distributed photovoltaics respectively.

[0089] (3) Line transmission capacity constraints

[0090] ;

[0091] is the transmission capacity of line l.

[0092] The above constraints are integrated into the upper objective function in the form of penalty function In the above example, an improved particle swarm algorithm is used to optimize the upper objective function:

[0093] Taking the maximum value of the upper objective function as an example, the objective function must meet the following conditions:

[0094] Maximum ;

[0095] st ;

[0096] When solving this type of optimization problem, a new objective function is constructed based on the original one using the penalty function, and then the objective function is optimized using a certain type of intelligent algorithm. The comprehensive objective function L is expressed as:

[0097] ;

[0098] Where: It is a penalty factor and its value is too large. is the inequality constraint condition of the objective function, which mainly includes investment constraint, reverse load rate constraint, and line transmission capacity constraint. The constructed penalty function is .

[0099] This embodiment adopts a dynamic adjustment of the penalty factor to avoid the problem of premature or unsolvable particle swarm optimization. The value of can be expressed as:

[0100] ;

[0101] Where: is the penalty factor for the g-th iteration, is the proportion of solutions that meet the constraints in the current population, is the target proportion of the current population that satisfies the constraint solution. is the adjustment factor, which is a constant greater than 1.

[0102] The principle of the standard particle swarm algorithm is as follows: Assume that the population size consists of N particles, and seek the best in a D-dimensional search space. Each particle has a D-dimensional vector of speed and position. The speed vector is represents the nth particle in the population speed, are the nth particle in the population. dimensional velocity, T represents transpose, and the position vector is represents the position of the nth particle in the population, are the nth particle in the population. dimensional position, which can describe the interaction between particles and also express the relationship between particles. The particle direct velocity update formula and position update formula can be calculated by the following two formulas:

[0103] ;

[0104] ;

[0105] Where: , K represents the number of iterations, Indicates the optimal solution found by the current nth particle, They represent the number of particles found by the current nth particle. dimensional optimal solution, represents the optimal solution found in the current population, Respectively represent the first Dimensional optimal solution. 、 is the learning factor, and its value is usually between (0, 2). 、 The step size is used to adjust the acceleration of the particle, and its value is between 0 and 1.

[0106] To solve the problem that the standard particle swarm algorithm is prone to fall into local optimality, this embodiment proposes the concept of particle position variance, setting the average position vector of the current particle for:

[0107] ;

[0108] Where: is the average position of the particle swarm in the dth dimension of the search space, is the number of particles in the population dimensional position, d={1,2,…,D}, and N is the total number of particles.

[0109] Then the position variance of the particle swarm is for:

[0110] ;

[0111] Where, is the position variance of the particle swarm in the dth dimension of the search space, is a variable factor, and the value selection rules are as follows:

[0112] ;

[0113] The larger the overall particle variance, the more uneven the particle population distribution, and the further away from the optimal solution. However, if the variance gradually stabilizes, there are two possibilities: local optimum and global optimum. To avoid falling into local optimum, this embodiment introduces particle velocity and position variation formulas based on the position variance calculation results above.

[0114] For standard deviation exist Particles within the range, this type of particle solution is close to the global prediction optimal solution, and its velocity variation formula is calculated using the following formula:

[0115] ;

[0116] Where, is the number of particles in the population Dimensional speed, is the maximum velocity of all particles in the dth dimension, and rand represents the generation of a random number between 0 and 1.

[0117] For the standard deviation Particles within the range, this type of particle solution moves near the global optimal solution. To prevent it from moving towards the local optimal solution, the particle speed and position must be mutated. The speed variation formula is calculated as follows:

[0118] ;

[0119] Where, 、 are the maximum and minimum values ​​of all particle positions.

[0120] For the standard deviation Particles within the range of this type are far away from the global optimal solution, and their velocity variation formula is calculated using the following formula:

[0121] ;

[0122] Where, is the minimum velocity of all particles in the dth dimension.

[0123] The current A and B type distributed photovoltaics are all connected to the 10kV public distribution network through the form of distribution network branches, and two nodes on the line are connected to the capacity of Distributed photovoltaics, The capacity is obtained by optimizing the upper objective function, and the actual maximum power on the Internet is , the power flow calculation method is used to solve the voltage of each node of the line.

[0124] Distributed photovoltaics will affect the voltage of the distribution network, and the voltage at the grid connection point increases with the increase of access capacity. Therefore, considering the stability of the grid connection point voltage and the voltage of other nodes, a first lower-level evaluation function with voltage stability as the goal is proposed:

[0125] ;

[0126] Where: represents the first lower-level evaluation function with voltage stability as the goal, , 、 、 are three weight coefficients, the sum of the three is 1; M is the number of nodes in the line, is the per-unit voltage value of node m, 、 They are the per-unit voltage values ​​of Class A and Class B distributed photovoltaic grid-connected points respectively; They represent Class A and Class B distributed photovoltaic grid-connected points respectively.

[0127] At the maximum online power When , the forward-backward power flow calculation method is used to solve the voltage of each node of the line; The constraints are:

[0128] ;

[0129] Where: The voltage threshold is set to determine whether the system meets the voltage-related standards or requirements.

[0130] The weight coefficient is calculated by using the analytic hierarchy process and entropy weight method to obtain a mixed weight:

[0131] ;

[0132] Where: 、 、 Represent the mixed weight of the qth weight index, the weight of the hierarchical analysis method and the weight obtained by the entropy weight method. 、 、 The weights of the analytic hierarchy process are obtained by the entropy weight method. 、 、 The entropy weight method weight is used, and then the mixed weight is obtained according to the above formula.

[0133] Before the distributed photovoltaic is connected, typical daily data is taken and the voltage of each node and the current of each branch are calculated using the power flow calculation method. Any line of the distribution network can be simplified to Figure 1 form, Figure 1 middle: and They represent the voltage vectors of line nodes i and j before distributed photovoltaic access, 、 is the impedance of the branch formed by nodes i and j, 、 and 、 Represents the active and reactive power of nodes i and j before distributed photovoltaic access.

[0134] Current vector of branch ij before distributed photovoltaic access for:

[0135] ;

[0136] in is the real part of the impedance of the branch formed by nodes i and j, is the imaginary part of the impedance of the branch formed by nodes i and j;

[0137] The power loss of branch ij before distributed photovoltaic access can be calculated using the following formula:

[0138] ;

[0139] in, is the power loss of branch ij before distributed photovoltaic access, is the current of branch ij before distributed photovoltaic access, is the imaginary part of the impedance of the branch formed by nodes i and j;

[0140] Decompose the above equation into real and imaginary parts until we get: the active power loss of branch ij before distributed photovoltaic access , Reactive power loss of branch ij before distributed photovoltaic access ;

[0141] but Valid values ​​are:

[0142] ;

[0143] Total network loss before distributed photovoltaic access for:

[0144] ;

[0145] in, is a collection of branches in the distribution network.

[0146] Use the power flow calculation method to calculate the voltage of each node and the current of each branch, and get and , and Respectively represent the voltage vectors of line nodes i and j after distributed photovoltaic access, 、 and 、 Represents the active and reactive power of nodes i and j after distributed photovoltaic access;

[0147] The current vector of branch ij after distributed photovoltaic access for:

[0148] ;

[0149] The power loss of branch ij after distributed photovoltaic access can be calculated using the following formula:

[0150] ;

[0151] in, is the current of branch ij after distributed photovoltaic access;

[0152] Decompose the real and imaginary parts of the above equation until: the active power loss of branch ij after distributed photovoltaic access is obtained , Reactive power loss of branch ij after distributed photovoltaic access ;

[0153] but Valid values ​​are:

[0154] ;

[0155] Total network loss after distributed photovoltaic access for:

[0156] ;

[0157] The year-on-year growth rate of network loss before and after distributed photovoltaic access As the second lower evaluation function:

[0158] ;

[0159] Where: The threshold for the year-on-year growth rate of the grid loss rate is used to determine whether the increase in the grid loss rate after access to distributed photovoltaics is within an acceptable range;

[0160] The average change rate of static voltage before and after distributed photovoltaic access It is the third lower evaluation function. If the average change rate of the static voltage is too large, it is not conducive to the regulation of the grid bus voltage, so it is used as the evaluation condition.

[0161] ;

[0162] ;

[0163] Where: M is the number of nodes on the line, is the per-unit voltage value of node m before the distributed photovoltaic is connected, is the per-unit voltage value of node m after distributed photovoltaic access, is the static voltage average change rate threshold.

[0164] The constraints of the lower layer stability evaluation function include:

[0165] (1) Short-circuit current constraint:

[0166] The short-circuit current of the source distribution network is calculated using the Norton and Thevenin equivalent network methods, which will not be described here.

[0167] ;

[0168] Where: are the short-circuit current value and the maximum short-circuit current limit of node m respectively.

[0169] (2) Node voltage constraints:

[0170] After distributed photovoltaics are connected to the grid, the voltage of the distribution network can be significantly increased. According to GB / T 12325, the range of node voltage change is -7% to +7%, which can be expressed by the following formula:

[0171] ;

[0172] Where: is the voltage rating.

[0173] (3) Distributed photovoltaic output constraints:

[0174] ;

[0175] Where: is a node The maximum output of the connected distributed photovoltaics, is a node The maximum capacity of the connected distributed photovoltaic access node is limited by the distribution transformer capacity of the substation.

[0176] The distributed photovoltaic capacity double-layer configuration method of the present invention comprehensively considers power generation and stability, such as Figure 2 As shown. The absolute value of the difference between the two optimization results must be less than a certain number to ensure that the global optimal solution is obtained. The convergence condition can be expressed as:

[0177] ;

[0178] Where: k is the number of iterations, is the grid-connected power of Class A distributed photovoltaics at the kth iteration, is the grid-connected power of Class A distributed photovoltaics in the k+1th iteration, is the grid-connected power of Class B distributed photovoltaics at the kth iteration, is the grid-connected power of Class B distributed photovoltaics in the k+1th iteration, is the convergence threshold.

[0179] Reference Figure 2 , the specific solution steps are as follows:

[0180] Step 1: Set various initial input parameters, including the investment scale of Class A and B distributed photovoltaics, project life, distribution network structure and impedance parameters, initialization parameters of the improved particle swarm algorithm used, and the maximum number of iterations.

[0181] Step 2: Based on the upper-level objective function and constraints, the improved particle swarm optimization algorithm is used to optimize the upper-level objective function. The position variance and standard deviation of the particles are calculated. The velocity variation formula is determined based on the range of the standard deviation, and the position and velocity of the particles are updated. The convergence condition is determined or the maximum number of iterations is reached. If not, the number of iterations is increased by 1. If yes, the top c solutions (i.e., the installed capacity of distributed photovoltaic power generation of Class A and Class B) that meet the constraints are output.

[0182] Step 3: Use the hierarchical analysis method and entropy weight method to comprehensively calculate the weight coefficients of the first lower-level evaluation function, and substitute the first c optimal solutions into the three lower-level evaluation functions one by one in the order of solution quality for stability evaluation, and judge whether the solution meets the judgment conditions and corresponding constraints of the three lower-level evaluation functions. If all are satisfied, go to step 4; if all are not satisfied or partially satisfied, judge whether the substituted solution is the cth one. If so, go to step 2 to re-optimize; if not, substitute the next solution into the three lower-level evaluation functions for re-evaluation.

[0183] Step 4: Save the Class A and Class B distributed photovoltaic installed capacity configuration schemes, and determine whether the convergence conditions are met or the maximum number of iterations is reached. If so, output the Class A and Class B distributed photovoltaic installed capacity configuration schemes and calculation results. If not, re-enter Step 2 for another iteration.

[0184] The above description merely represents preferred embodiments of the present invention and is not intended to limit the present invention in any other manner. Any person skilled in the art may utilize the above disclosure to modify or modify the present invention into equivalent embodiments. However, any simple modifications, equivalent variations, and modifications to the above embodiments that do not depart from the technical content of the present invention and are based on the technical essence of the present invention remain within the scope of protection of the present invention.

Claims

1. A two-tier configuration method for distributed photovoltaic capacity that comprehensively considers power generation and stability: a photovoltaic power generation system connected to the public grid via 10kV is called Class A distributed photovoltaic, and a photovoltaic power generation system connected to the grid via a 380V voltage level is called Class B distributed photovoltaic; its characteristics are: Construct the upper objective function with the maximum power generation as the goal : ； Where: is the maximum annual power generation hours of distributed photovoltaics, is the total power generation of distributed photovoltaic power generation of Class A and B, r represents the year, R represents the project life; x is the discount rate, It is the Class A distributed photovoltaic installed capacity, It is the Class B distributed photovoltaic installed capacity; The first lower-level evaluation function with voltage stability as the goal is: ； Where: represents the first lower-level evaluation function with voltage stability as the goal, , 、 、 are three weight coefficients, the sum of the three is 1; M is the number of nodes in the line, is the per-unit voltage value of node m, 、 They are the per-unit voltage values ​​of Class A and Class B distributed photovoltaic grid-connected points respectively; Respectively represent Class A and B distributed photovoltaic grid connection points; The year-on-year growth rate of network loss before and after distributed photovoltaic access As the second lower evaluation function; the average change rate of static voltage before and after distributed photovoltaic access is the third lower layer evaluation function; Solve the upper objective function , get the first c solutions, start from the optimal solution, use the three lower-level evaluation functions to evaluate the stability of each solution one by one, and select the best solution that meets the stability evaluation as the A and B distributed photovoltaic installed capacity configuration scheme.

2. The distributed photovoltaic capacity double-layer configuration method according to claim 1 is characterized in that: The second lower layer evaluation function is: ； Where: is the threshold value of the year-on-year growth rate of network loss rate; is the total network loss before distributed photovoltaic access; is the total network loss after distributed photovoltaic access.

3. The distributed photovoltaic capacity double-layer configuration method according to claim 1 is characterized in that: The third lower layer evaluation function is: ； ； Where: is the per-unit voltage value of node m before the distributed photovoltaic is connected, is the per-unit voltage value of node m after distributed photovoltaic access, is the static voltage average change rate threshold.

4. The distributed photovoltaic capacity double-layer configuration method according to claim 1 is characterized in that: The upper objective function The constraints include: investment constraints, reverse load rate constraints, and line transmission capacity constraints.

5. The distributed photovoltaic capacity double-layer configuration method according to claim 4 is characterized in that: Use the penalty function in the upper objective function Based on this, a new comprehensive objective function is constructed: ； Where: is the penalty factor, is the inequality constraint condition of the objective function, including investment constraint, reverse load rate constraint, and line transmission capacity constraint. The constructed penalty function is ; The comprehensive objective function is solved by particle swarm optimization. During the solution process, the penalty factor is dynamically adjusted: ； Where: is the penalty factor for the g-th iteration, is the proportion of solutions that meet the constraints in the current population, is the target proportion of the current population that meets the constraint solution, is the adjustment factor, which is a constant greater than 1.

6. The distributed photovoltaic capacity double-layer configuration method according to claim 5 is characterized in that: In the process of solving the comprehensive objective function through the particle swarm algorithm, the average position vector of the current particle is set for: ； Where: is the average position of the particle swarm in the dth dimension of the search space, is the number of particles in the population dimensional position, d={1,2,…,D}, D is the total dimension of the particle, and N is the total number of particles; Then the position variance of the particle swarm is for: ； Where, is the position variance of the particle swarm in the dth dimension of the search space, is a variable factor, and the value selection rules are as follows: ； For standard deviation exist For particles within the range, the velocity variation formula is calculated using the following formula: ； Where, is the number of particles in the population Dimensional speed, is the maximum velocity of all particles in the dth dimension, and rand represents the generation of a random number between 0 and 1; For the standard deviation For particles within the range, the velocity variation formula is calculated using the following formula: ； Where, 、 is the maximum and minimum value of all particle positions; For the standard deviation For particles within the range, the velocity variation formula is calculated using the following formula: ； Where, is the minimum velocity of all particles in the dth dimension.

7. The distributed photovoltaic capacity double-layer configuration method according to claim 1 is characterized in that: The three lower-level evaluation function constraints include: Short-circuit current constraint: , where are the short-circuit current value and the maximum short-circuit current limit of node m respectively; Node voltage constraints: , where is the per-unit voltage value of node m after distributed photovoltaic access, is the voltage rating; Distributed photovoltaic output constraints: , where is a node The maximum output of the connected distributed photovoltaics, is a node The maximum capacity of the connected distributed photovoltaic access node.

8. The distributed photovoltaic capacity double-layer configuration method according to claim 1 is characterized in that: The solution steps are as follows: Step 1: Set various initial input parameters, including the investment scale of Class A and B distributed photovoltaics, project life, distribution network structure and impedance parameters, particle swarm algorithm initialization parameters, and maximum number of iterations; Step 2: Based on the upper-level objective function and constraints, the particle swarm algorithm is used to optimize the upper-level objective function, calculate the position variance and standard deviation of the particles, determine the velocity variation formula based on the value range of the standard deviation, and update the position and velocity of the particles; determine whether the convergence condition is met or the maximum number of iterations is reached. If not, the number of iterations is increased by 1. If yes, the top c solutions that meet the constraints and maximize the power generation are output; Step 3: Use the analytic hierarchy process and entropy weight method to comprehensively calculate the weight coefficients of the first lower-level evaluation function, and substitute the first c optimal solutions into the three lower-level evaluation functions one by one in the order of their quality for stability evaluation, and judge whether they all meet the judgment conditions and corresponding constraints of the three lower-level evaluation functions. If they all meet, proceed to step 4; if they do not meet or partially meet, judge whether the substituted solution is the cth one. If so, proceed to step 2 to re-optimize; if not, substitute the next solution into the three lower-level evaluation functions for re-evaluation; Step 4: Save the Class A and Class B distributed photovoltaic installed capacity configuration schemes, and determine whether the convergence conditions are met or the maximum number of iterations is reached. If so, output the Class A and Class B distributed photovoltaic installed capacity configuration schemes and calculation results. If not, re-enter Step 2 for another iteration.

Images