Active power distribution network distributed power supply optimal configuration method

Abstract

The invention relates to an active power distribution network distributed power supply optimization configuration method, and belongs to the field of distributed power supply (DG) optimization configuration of a power distribution network. According to the method, an improved leapfrog algorithm is utilized to solve an upper-layer planning model and a lower-layer planning model, a DG access position and installation capacity are obtained through an upper-layer planning result and transmitted to a lower-layer function, a DG output scheme is obtained through a lower-layer planning result and reversely transmitted to an upper layer, and finally a DG overall configuration scheme is obtained through calculation. According to the method, the fluctuation and uncertainty of the output of the DG can be more accurately considered, the access position and the installation capacity of the DG in the power distribution network are optimized, and the operation efficiency and the economical efficiency of the power distribution network are improved.

Description

Technical Field

[0001] The present invention belongs to the field of distributed power (DG) optimization configuration of distribution networks, and in particular relates to a method for optimizing the configuration of distributed power in active distribution networks. Background Art

[0002] Existing DG optimization configuration research is mostly based on the assumption that DG output and load are constant values, ignoring the timing of DG output, especially for uncontrollable factors such as photovoltaic and wind power generation. This assumption is quite different from the actual operation situation. Summary of the invention

[0003] The purpose of the present invention is to overcome the defects of the prior art and provide a method for optimizing the configuration of distributed power sources in an active power distribution network so as to improve the operating efficiency and economy of the power distribution network.

[0004] To achieve the above purpose, the technical solution of the present invention is: a method for optimizing the configuration of distributed power sources in an active distribution network, which uses an improved frog leaping algorithm to solve the upper and lower planning models, and the upper planning result obtains a DG access position and installation capacity, and passes it to the lower function, and the lower planning result obtains a DG output plan, and passes it back to the upper layer, and finally calculates an optimal DG overall configuration plan, that is, the access position and installation capacity of the DG in the distribution network.

[0005] In one embodiment of the present invention, the method comprises the following steps:

[0006] (1) Analyze the time series characteristics of DG output and load, divide the average value of load and DG output in each period by the maximum value in the whole period, and use the ratio curve as its time series characteristic;

[0007] (2) Establish a two-layer optimization configuration model that considers timing characteristics and active management strategies;

[0008] (3) An improved frog leaping algorithm that combines genetic algorithm and frog leaping algorithm is used to optimize and solve the two-layer optimization configuration model and output the optimal DG overall configuration plan.

[0009] In one embodiment of the present invention, step (2) is specifically implemented as follows:

[0010] A two-layer optimal configuration model of the distribution network is established considering the DG access location and installation capacity and the DG output at different time periods:

[0011] min[C total +(1-U average )]

[0012] The upper layer is the planning problem of DG access location and installation capacity in the distribution network, and the lower layer is the optimization problem of DG output size in the distribution network under different climatic conditions and at different time periods;

[0013] The upper model is converted to the annual comprehensive cost of distributed power generation C total The minimum is the goal, and the objective function is:

[0014] minC total =C invest +C loss +C purchase +C co2

[0015] In the formula, C invest is the annual comprehensive investment cost of distributed power generation; C loss is the loss cost of the distribution network; C purchase The cost of purchasing electricity from the distribution network; C co2 Carbon emission fees for distribution networks;

[0016] The lower model controls the DG output with the average value of the system node voltage closest to the per-unit value as the target, and maintains the voltage quality stability of the system on the basis of fully absorbing DG, thereby ensuring the power safety of the user end. The objective function is:

[0017] min(1-U average )

[0018] Where U average It is the average node voltage in each period under each climate condition.

[0019] In one embodiment of the present invention, the specific calculation formulas for the comprehensive investment cost of distributed power sources, the loss cost of the distribution network, the power purchase cost of the distribution network, and the carbon emission cost of the distribution network are as follows:

[0020] 1) Comprehensive investment cost of distributed power generation

[0021]

[0022] In the formula, C DGi is the DG unit capacity investment cost converted to annual; r is the fixed annual interest rate; N DG is the number of DG; Y DG C is the service life of DG; ri is the annual operation and maintenance cost per unit capacity of DG; im (n) is the active power output of the ith DG at the nth hour under the mth weather condition;

[0023] 2) Distribution network loss costs

[0024] C loss =Ploss-DG τ max C p

[0025] Where P loss-DG is the active power loss of the distribution network after connecting to DG; τ max is the maximum load loss hours; C p is the unit electricity price;

[0026] 3) Power purchase costs from distribution network

[0027]

[0028] Where P L is the line loss of the distribution network in the nth hour under the mth climate condition; im (n) is the active power output of the ith DG at the nth hour under the mth climate condition;

[0029] 4) Carbon emission costs of distribution networks

[0030]

[0031] In the formula, C c Penalized for carbon emissions.

[0032] In one embodiment of the present invention, the constraints of the upper model are as follows:

[0033] 1) Node power balance constraints

[0034]

[0035] Where N is the number of system nodes; P i , Q i are the power values ​​of node i respectively; e i 、f i are the real and imaginary parts of the voltage at node i, respectively; G ij , B ij is the conductance and susceptance of the line between nodes i and j;

[0036] 2) Node voltage constraints

[0037] U imin ≤U i ≤U imax

[0038] Where U imax , U imin are the upper and lower limits of the voltage at node i respectively;

[0039] 3) DG access capacity upper limit constraint

[0040]

[0041] In the formula, α is the upper limit of DG access capacity penetration rate;

[0042] 4) Branch power constraints

[0043] S i ≤S i.max

[0044] In the formula, S i is the apparent power passing through the i-th line; S i.max is the maximum apparent power allowed to pass through the i-th line.

[0045] In one embodiment of the present invention, the constraints of the lower model are as follows:

[0046] 1) Node power balance constraints

[0047]

[0048] Where N is the number of system nodes; P i , Q i are the power values ​​of node i respectively; e i 、f i are the real and imaginary parts of the voltage at node i, respectively; G ij , B ij is the conductance and susceptance of the line between nodes i and j;

[0049] 2) Node voltage constraints

[0050] U imin ≤U i ≤U imax

[0051] Where U imax , U imin are the upper and lower limits of the voltage at node i respectively;

[0052] 3) Branch power constraints

[0053] S i ≤S i.max

[0054] In the formula, S i is the apparent power passing through the i-th line; S i.max is the maximum apparent power allowed to pass through the i-th line;

[0055] 4) OLTC regulation constraints

[0056] K min ≤K m (n)≤K max

[0057] In the formula, K m (n) is the OLTC tap position in the nth period under the mth climate condition; K min , K max are the minimum and maximum values ​​of the OLTC tap position, respectively;

[0058] 5) Reactive power compensation constraints

[0059] Q Cmin ≤Q Cm (n)≤Q Cmax

[0060] In the formula, Q Cm (n) is the grid-connected capacity of the reactive power compensation device in the nth period under the mth climate condition; Q Cmin , Q Cmax They are the minimum and maximum values ​​of the grid-connected capacity of reactive power compensation equipment respectively;

[0061] 6) DG active output constraints

[0062] P min ≤P m (n)≤P max

[0063] Where P m (n) is the active power output of DG in the nth period under the mth climate condition; P min , P max They are the minimum and maximum values ​​of DG active output respectively.

[0064] In one embodiment of the present invention, the improved frog leaping algorithm is implemented as follows:

[0065] 1) Integrate the particle update strategy in the genetic algorithm into the hybrid frog leaping algorithm;

[0066] 2) Update the individual with the lowest fitness ranking in the subgroup;

[0067] 3) If the global optimal value does not improve, a new individual is generated to replace the original one.

[0068] In one embodiment of the present invention, the specific implementation steps of the improved frog leaping algorithm are as follows:

[0069] Step 1, initialization: according to the original parameters of the example node system, generate a total of F (F = m × n) candidate solutions as frog individuals in the definition domain; set the worst number K in the group and the number of consecutive unimproved fitness values ​​of the optimal solution of the population T; the number of global iterations S and the number of local iterations G;

[0070] Step 2: Calculate the fitness of each frog according to the objective function, and sort the frogs according to the fitness size as required. The frog individual ranked first is the global optimal individual X g ; And group all frog individuals;

[0071] Step 3: Update the K individuals with the lowest fitness ranking in each subgroup. If the fitness of the updated individual is better, replace the original individual, otherwise randomly generate a new individual to replace the original individual; if the number of times the global optimal value has not been improved is less than T, continue to update the remaining n-2-K individuals in the subgroup. If the fitness of the updated individual is better, replace the original individual; when the number of times the global optimal value has not been improved reaches T, randomly generate n-2-K individuals to replace the original individual; repeat this update process until the number of local iterations is reached;

[0072] Step 4: Determine whether the number of global iterations has reached S. If so, stop the algorithm and output the optimal solution; otherwise, go to step 3.

[0073] The present invention also provides an active distribution network distributed power optimization configuration system, including a memory, a processor, and computer program instructions stored in the memory and capable of being executed by the processor. When the processor executes the computer program instructions, the method steps described above can be implemented.

[0074] The present invention also provides a computer-readable storage medium, on which computer program instructions that can be executed by a processor are stored. When the processor executes the computer program instructions, the method steps described above can be implemented.

[0075] Compared with the prior art, the present invention has the following beneficial effects: the method of the present invention can more accurately consider the volatility and uncertainty of DG output, optimize the access location and installation capacity of DG in the distribution network, and improve the operating efficiency and economy of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 Schematic diagram of the two-layer model for optimal configuration of distributed power sources.

[0077] Figure 2 This is the flow chart of the hybrid frog leaping algorithm.

[0078] Figure 3 Schematic diagram of chromosome crossover operation.

[0079] Figure 4 Schematic diagram of chromosome mutation crossover operation

[0080] Figure 5 To improve the optimization flow chart of the frog leaping algorithm.

[0081] Figure 6Flowchart for optimizing the two-layer configuration model. DETAILED DESCRIPTION

[0082] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings.

[0083] The present invention provides a method for optimizing the configuration of distributed power sources in an active distribution network. The improved frog leaping algorithm is used to solve the upper and lower planning models. The upper planning result obtains a DG access position and installation capacity, and passes it to the lower function. The lower planning result obtains a DG output plan, and passes it back to the upper layer. Finally, an optimal DG overall configuration plan, that is, the access position and installation capacity of the DG in the distribution network, is calculated.

[0084] The following is the specific implementation process of the present invention.

[0085] The present invention provides a method for optimizing the configuration of distributed power sources in an active power distribution network, and the specific implementation steps are as follows:

[0086] 1. Analysis of distributed power output and load time characteristics:

[0087] Based on the hour, the ratio of the average value to the maximum value of the load and DG output in each period is calculated to obtain the time series characteristic ratio.

[0088] 2. Establishment of a two-layer optimization configuration model:

[0089] 2.1. Distributed power optimization configuration The upper model aims to minimize the annual comprehensive cost of distributed power, and the objective function is:

[0090] min C total =C invest +C loss +C purchase +C co2 (1)

[0091] In the formula, C invest is the annual comprehensive investment cost of distributed power generation; C loss is the loss cost of the distribution network; C purchase The cost of purchasing electricity from the distribution network; C co2 is the carbon emission fee of the distribution network. The specific calculation formulas for each fee are as follows:

[0092] DG's comprehensive investment cost:

[0093]

[0094] In the formula, C DGi is the DG unit capacity investment cost converted to annual; r is the fixed annual interest rate; N DG is the number of DG; Y DGC is the service life of DG; ri is the annual operation and maintenance cost per unit capacity of DG; im (n) is the active power output of the ith DG in the nth hour under the mth weather condition.

[0095] Distribution network loss costs:

[0096] C loss =P loss-DG τ max C p (3)

[0097] Where P loss-DG is the active power loss of the distribution network after connecting to DG; τ max is the maximum load loss hours; C p The unit electricity price.

[0098] Electricity purchase cost from distribution network:

[0099]

[0100] Where P L is the line loss of the distribution network in the nth hour under the mth climate condition; im (n) is the active power output of the ith DG at the nth hour under the mth climate condition.

[0101] Carbon emission charges for distribution network:

[0102]

[0103] In the formula, C c Penalty for carbon emissions. When using traditional fossil energy to generate electricity, 0.344 tons of coal are needed to generate 1MW·h. Burning one ton of coal produces 2620kg of carbon dioxide. The national penalty for carbon emissions is 9.75 yuan / t.

[0104] Constraints:

[0105] (1) Node power balance constraints

[0106]

[0107] Where N is the number of system nodes; P i , Q i are the power values ​​of node i respectively; e i 、f i are the real and imaginary parts of the voltage at node i, respectively; G ij , B ij are the conductance and susceptance of the line between nodes i and j.

[0108] (2) Node voltage constraints

[0109] U imin ≤U i ≤U imax (7)

[0110] Where U imax , U imin are the upper and lower limits of the voltage at node i respectively.

[0111] (3) DG access capacity upper limit constraint

[0112]

[0113] Where α is the upper limit of DG access capacity penetration.

[0114] (4) Branch power constraints

[0115] S i ≤S i.max (9)

[0116] In the formula, S i is the apparent power passing through the i-th line; S i.max is the maximum apparent power allowed to pass through the i-th line.

[0117] 2.2. Distributed power optimization configuration The lower model controls the DG output with the average value of the system node voltage closest to the per-unit value as the target, and maintains the voltage quality stability of the system on the basis of fully absorbing DG, thereby ensuring the power safety of the user end. The objective function is:

[0118] min(1-U average )(10)

[0119] Where U average It is the average node voltage in each period under each climate condition.

[0120] Constraints:

[0121] (1) Node power balance constraints

[0122]

[0123] Where N is the number of system nodes; P i , Q i are the power values ​​of node i respectively; e i 、f i are the real and imaginary parts of the voltage at node i, respectively; G ij , B ij are the conductance and susceptance of the line between nodes i and j.

[0124] (2) Node voltage constraints

[0125] U imin ≤U i ≤U imax (12)

[0126] Where U imax , U imin are the upper and lower limits of the voltage at node i respectively.

[0127] (3) Branch power constraints

[0128] S i ≤S i.max (13)

[0129] In the formula, S i is the apparent power passing through the i-th line; S i.max is the maximum apparent power allowed to pass through the i-th line.

[0130] (4) OLTC Regulation Constraints

[0131] K min ≤K m (n)≤K max (14)

[0132] In the formula, K m (n) is the OLTC tap position in the nth period under the mth climate condition; K min , K max are the minimum and maximum values ​​of the OLTC tap position respectively.

[0133] (5) Reactive power compensation constraints

[0134] Q Cmin ≤Q Cm (n)≤Q Cmax (15)

[0135] In the formula, Q Cm (n) is the grid-connected capacity of the reactive power compensation device in the nth period under the mth climate condition; Q Cmin , Q Cmax They are respectively the minimum and maximum values ​​of the grid-connected capacity of the reactive power compensation equipment.

[0136] (6) DG active output constraints

[0137] P min ≤P m (n)≤P max (16)

[0138] Where P m (n) is the active power output of DG in the nth period under the mth climate condition; P min, P max They are the minimum and maximum values ​​of DG active output respectively.

[0139] 2.3. By integrating the upper and lower models of distributed generation optimization configuration proposed above, we can obtain a two-layer planning model of the distribution network that takes into account the DG access location and installation capacity and the DG output at different time periods:

[0140] min[C total +(1-U average )] (17)

[0141] The upper layer is the planning problem of the DG access location and installation capacity in the distribution network, and the lower layer is the optimization problem of the DG output size in the distribution network under different climatic conditions and different time periods. The upper layer passes the DG construction plan in the distribution network to the lower layer. The lower layer optimizes the DG output based on the upper layer and passes the DG output plan to the upper layer. The upper layer then uses the DG output plan passed from the lower layer to optimize the DG construction plan, and so on. After many exchanges and communications between the upper and lower planning models, the optimal DG construction plan is finally obtained, that is, the access location and installation capacity of DG in the distribution network. The optimization process of the overall model of the two-layer planning of the distribution network is as follows: Figure 1 shown.

[0142] 3. Improved frog leaping algorithm

[0143] According to the established two-layer optimization configuration model of DG in active distribution network, the particle update strategy in genetic algorithm is integrated into the hybrid frog leaping algorithm according to the characteristics of model optimization, aiming to obtain an optimization scheme with high performance, low complexity, easy implementation and fast convergence speed.

[0144] 3.1. Hybrid Frog Leaping Algorithm SFLA simulates the optimization process as when frogs go out to hunt, they are divided into multiple subgroups for internal search, and all frogs are mixed regularly to exchange information and sort the frog individuals, and then the subgroups are searched again, so that the information between different frog individuals can be globally exchanged. This process continues until the set convergence condition is met or the maximum number of iterations is reached.

[0145] The parameters of the hybrid frog leaping algorithm are: population size P (P = m × n), number of subpopulations m, number of individuals in a subpopulation n, maximum jump step length S max The basic optimization steps of SFLA are as follows:

[0146] Step 1: Initialization, define the number of subgroups m, the number of individuals in each subgroup n, and thus determine the total number of individuals P in the initial frog population.

[0147] Step 2: Randomly generate P frogs (representing candidate solutions to the problem) in the feasible domain and make them the initial frog population, denoted as X i =(xi1 ,x i2 ,…x iD ), where 0≤i≤n, and D is the dimension of the candidate solution. Calculate the fitness value of each frog.

[0148] Step 3: Sort the P frogs in descending order of fitness value. Store them as a sequence X = {U(i), f(i), i = 1, ... P}, when i = 1, it means that this is the frog with the best fitness value, record the position P of the best frog in the frog group g (P g =U(1)).

[0149] Step 4: Divide the sequence X into m subgroups, each containing n frogs. The frog ranked first is divided into the first subgroup, the frog ranked second is divided into the second subgroup, ..., the frog ranked m is divided into the mth subgroup, the frog ranked m+1 is divided into the first subgroup, and so on, until all frogs are divided into their respective subgroups. The subgroup sequence is Y = (Y1, Y2, ... Y m ) [i] .

[0150] Step 5: Assume that the initial values ​​of im and in are both 0, and im=im+1, in=in+1. im is used to calculate the number of subgroups and compare it with the total number of subgroups m, and in is used to calculate the number of evolutions in each subgroup and compare it with the given number of evolutions N. Since each frog approaches the optimal goal by improving their memes, and each frog can obtain ideas from the best frog individual in the subgroup, this may cause these individuals to fall into a local optimal solution. Therefore, it is necessary to adopt a sub-meme subgroup. The frog selection strategy of the sub-meme is based on the weight of the frog. The higher the weight of the frog, the more likely it is to be selected, and the lower the weight of the frog, the smaller the probability of being selected. The weight is obtained through the triangular distribution probability, and the formula is p j =2(n+1-j) / n(n+1), j=1,...,n. In this way, the frog individual with the highest fitness value in each subgroup has the highest probability p1=2 / (n+1), and the frog individual with the lowest fitness value has the lowest probability p1=2 / (n+1). n =2 / n(n+1). In this way, q frogs can be randomly selected from each subgroup and combined into a sub-memetic composite sequence Z. After the frogs are grouped in descending order of fitness, an internal search is performed in each subgroup. Let the individual with the best internal fitness in the sub-memetic composite sequence be X b , the individual with the worst fitness is X w , the individual with the best fitness in the entire population is X g . For the worst individual X in the subgroup w To update:

[0151] Xnew =X w +r×(X b -X w ) (18)

[0152] In the formula, X new is the new individual generated after the update, r is a random number between 0 and 1. new The fitness is better than X w , then X w =X new Otherwise, replace X b Replace with X g Re-update, if no X with better fitness is generated new , then a new individual is randomly generated to replace X w .

[0153] Step 6: After completing the update evolution in each subgroup, remix all frogs to generate a new sequence X, arrange the sequence in descending order of fitness value, and update the best position in the population

[0154] Step 7: If the convergence condition is met, stop. Otherwise, return to step 3. The convergence condition here can be determined by two conditions. One is that when the number of consecutive cycles in which at least one frog in the evolutionary frog group has not changed reaches the set value, stop. The other is that when the number of evolutionary cycles of the entire population reaches the set value, stop.

[0155] Figure 2 It is the optimization flowchart of the hybrid frog leaping algorithm.

[0156] 3.2. The performance of genetic algorithms is mainly determined by the following three points: the selection of encoding methods, the formulation of fitness functions, and the selection, crossover and mutation operators. The basic optimization process of genetic algorithms is as follows:

[0157] (1) Coding:

[0158] (2) Fitness function: For example, to find the maximum value:

[0159]

[0160] In the formula, C min is the minimum estimate of f(X). When solving the minimum problem:

[0161]

[0162] In the formula, C max Is an arbitrarily large number, which is the maximum value estimate of f(X). If the objective function value f(X) is always positive, the fitness function can also select the inverse of the objective function:

[0163]

[0164] (3) Selection: After evaluating the individuals in the initial population according to the fitness function, the process of generating the next generation of individuals and forming subpopulations begins. Suppose there are M individuals in the population, F i is the fitness value of the ith individual, then the probability of the ith individual being selected is:

[0165]

[0166] The fitness value of an individual is proportional to its probability of being selected and passed on to the next generation. This cycle is repeated n times, and n child individuals are obtained.

[0167] (4) Crossover: The crossover operation is to exchange some genes of two individuals according to the initially set crossover probability to generate two new individual genes. The fitness value of the new individuals may be better than that of the previous generation, so as to achieve the purpose of optimization. The basic schematic diagram of the crossover operation is as follows: Figure 3 shown.

[0168] (5) Mutation: The mutation operation in the genetic algorithm refers to the conversion of some genes in the individual chromosomes into their alleles according to the initially set mutation probability, thereby generating new individuals of the next generation. The basic diagram of the mutation operation is as follows: Figure 4 As shown:

[0169] (6) Algorithm termination condition: A termination algebra can be set as the termination condition. When the genetic algorithm optimizes to the set algebra, the loop stops and the optimal solution is output. A certain rule can also be set as the termination condition of the genetic algorithm, that is, when it is judged that all candidate solutions have reached a mature state and no longer have an evolutionary trend, the loop is terminated and the optimal solution is output.

[0170] 3.3. SFLA is prone to fall into local optimum and has a slow search speed. To improve this defect and prevent premature convergence, the present invention combines the idea of ​​genetic algorithm with frog leaping algorithm. The information of frog individuals with higher fitness values ​​is compiled and merged into the K frog individuals with lower rankings:

[0171]

[0172] In the formula, is the frog individual ranked t in the i-th subgroup; r1 and r2 are both random numbers between 0 and 1.

[0173] To further improve the performance of the algorithm, the update strategy for the remaining frog individuals in the subgroup is improved. The top two frog individuals in the subgroup are regarded as excellent reference individuals, and the remaining n-2-K frog individuals in the subgroup are updated. If the algorithm falls into a local optimum and the fitness value of the global optimal individual has not been improved for many times, the n-2-K frog individuals in the subgroup are updated again to jump out of the local optimum and improve the reliability of the algorithm:

[0174]

[0175] In the formula, is the frog individual ranked t in the i-th subgroup; r is a random number between 0 and 1.

[0176] If the updated If the fitness value of the frog individual is not better than that of the frog individual before the update, a new individual will be generated to replace the original one.

[0177]

[0178] In the formula, X b is the best individual in the subgroup; r is a random number between 0 and 1; X g is the optimal individual in the population.

[0179] The optimization steps of the improved frog leaping algorithm for planning DG in active distribution network are as follows:

[0180] Step 1: Initialization. According to the original parameters of the example node system, a total of F (F = m × n) candidate solutions are generated in the definition domain as frog individuals; the worst number K in the group and the number of consecutive unimproved fitness values ​​of the optimal solution of the population T are set; the number of global iterations S and the number of local iterations G are set.

[0181] Step 2: Calculate the fitness of each frog according to the objective function, and sort the frogs according to the fitness size as required. The frog individual ranked first is the global optimal individual X g . And divide all frog individuals into groups.

[0182] Step 3: Update the K individuals with the lowest fitness ranking in each subgroup. If the fitness of the updated individual is better, replace the original individual. Otherwise, randomly generate a new individual to replace the original individual. If the number of times the global optimal value has not been improved is less than T, continue to update the remaining n-2-K individuals in the subgroup. If the fitness of the updated individual is better, replace the original individual. When the number of times the global optimal value has not been improved reaches T, randomly generate n-2-K individuals to replace the original individual. Repeat this update process until the number of local iterations is reached.

[0183] Step 4: Determine whether the number of global iterations has reached S. If so, stop the algorithm and output the optimal solution; otherwise, go to step 3.

[0184] Improved frog leaping algorithm optimization flowchart reference Figure 5 .

[0185] The present invention uses the improved frog leaping algorithm to solve the upper and lower planning models. The upper planning result obtains a DG access location and installation capacity, and passes it to the lower function. The lower planning result obtains the DG output plan, and passes it back to the upper layer, and finally calculates a DG overall configuration plan. The specific flow chart is as follows Figure 6 shown.

[0186] The present invention also provides an active distribution network distributed power optimization configuration system, including a memory, a processor, and computer program instructions stored in the memory and capable of being executed by the processor. When the processor executes the computer program instructions, the method steps described above can be implemented.

[0187] The present invention also provides a computer-readable storage medium, on which computer program instructions that can be executed by a processor are stored. When the processor executes the computer program instructions, the method steps described above can be implemented.

[0188] The above are preferred embodiments of the present invention. Any changes made according to the technical solution of the present invention, as long as the resulting functions do not exceed the scope of the technical solution of the present invention, belong to the protection scope of the present invention.

Claims

1. A method for optimizing the configuration of distributed power sources in an active power distribution network, characterized in that: The improved frog leaping algorithm is used to solve the upper and lower planning models. The upper planning result obtains a DG access location and installation capacity, which are passed to the lower function. The lower planning result obtains the DG output plan and is passed back to the upper layer. Finally, an optimal DG overall configuration plan is calculated, that is, the access location and installation capacity of DG in the distribution network.

2. The method for optimizing the configuration of distributed power sources in an active power distribution network according to claim 1, characterized in that: The steps include: (1) Analyze the time series characteristics of DG output and load, divide the average value of load and DG output in each period by the maximum value in the whole period, and use the ratio curve as its time series characteristic; (2) Establish a two-layer optimization configuration model that considers timing characteristics and active management strategies; (3) An improved frog leaping algorithm that combines genetic algorithm and frog leaping algorithm is used to optimize and solve the two-layer optimization configuration model and output the optimal DG overall configuration plan.

3. The method for optimizing the configuration of distributed power sources in an active power distribution network according to claim 2, characterized in that: Step (2) is specifically implemented as follows: A two-layer optimal configuration model of the distribution network is established considering the DG access location and installation capacity and the DG output at different time periods: min[C total +(1-U average )] The upper layer is the planning problem of DG access location and installation capacity in the distribution network, and the lower layer is the optimization problem of DG output size in the distribution network under different climatic conditions and at different time periods; The upper model is converted to the annual comprehensive cost of distributed power generation C total The minimum is the goal, and the objective function is: minC total =C invest +C loss +C purchase +C co2 In the formula, C invest is the annual comprehensive investment cost of distributed power generation; C loss is the loss cost of the distribution network; C purchase The cost of purchasing electricity from the distribution network; C co2 Carbon emission fees for distribution networks; The lower model controls the DG output with the average value of the system node voltage closest to the per-unit value as the target, and maintains the voltage quality stability of the system on the basis of fully absorbing DG, thereby ensuring the power safety of the user end. The objective function is: min(1-U average ) Where U average It is the average node voltage in each period under each climate condition.

4. The method for optimizing the configuration of distributed power sources in an active power distribution network according to claim 2, characterized in that: The specific calculation formulas for the comprehensive investment cost of distributed power generation, distribution network loss cost, distribution network power purchase cost, and distribution network carbon emission cost are as follows: 1) Comprehensive investment cost of distributed power generation In the formula, C DGi is the DG unit capacity investment cost converted to annual; r is the fixed annual interest rate; N DG is the number of DG; Y DG C is the service life of DG; ri is the annual operation and maintenance cost per unit capacity of DG; im (n) is the active power output of the ith DG at the nth hour under the mth weather condition; 2) Distribution network loss costs C loss =P loss-DG t max C p Where P loss-DG is the active power loss of the distribution network after connecting to DG; τ max is the maximum load loss hours; C p is the unit electricity price; 3) Power purchase costs from distribution network Where P L is the line loss of the distribution network in the nth hour under the mth climate condition; im (n) is the active power output of the ith DG at the nth hour under the mth climate condition; 4) Carbon emission costs of distribution networks In the formula, C c Penalized for carbon emissions.

5. The method for optimizing the configuration of distributed power sources in an active power distribution network according to claim 3, characterized in that: The constraints of the upper model are as follows: 1) Node power balance constraints Where N is the number of system nodes; P i , Q i are the power values ​​of node i respectively; e i 、f i are the real and imaginary parts of the voltage at node i, respectively; G ij , B ij is the conductance and susceptance of the line between nodes i and j; 2) Node voltage constraints IN imin ≤U i ≤U imax Where U imax , U imin are the upper and lower limits of the voltage at node i respectively; 3) DG access capacity upper limit constraint In the formula, α is the upper limit of DG access capacity penetration rate; 4) Branch power constraints S i ≤S i.max In the formula, S i is the apparent power passing through the i-th line; S i.max is the maximum apparent power allowed to pass through the i-th line.

6. The method for optimizing the configuration of distributed power sources in an active power distribution network according to claim 3, characterized in that: The constraints of the lower model are as follows: 1) Node power balance constraints Where N is the number of system nodes; P i , Q i are the power values ​​of node i respectively; e i 、f i are the real and imaginary parts of the voltage at node i, respectively; G ij , B ij is the conductance and susceptance of the line between nodes i and j; 2) Node voltage constraints IN imin ≤U i ≤U imax Where U imax , U imin are the upper and lower limits of the voltage at node i respectively; 3) Branch power constraints S i ≤S i.max In the formula, S i is the apparent power passing through the i-th line; S i.max is the maximum apparent power allowed to pass through the i-th line; 4) OLTC regulation constraints K min ≤K m (n)≤K max In the formula, K m (n) is the OLTC tap position in the nth period under the mth climate condition; K min , K max are the minimum and maximum values ​​of the OLTC tap position, respectively; 5) Reactive power compensation constraints Q Cmin ≤Q Cm (n)≤Q Cmax In the formula, Q Cm (n) is the grid-connected capacity of the reactive power compensation device in the nth period under the mth climate condition; Q Cmin , Q Cmax They are the minimum and maximum values ​​of the grid-connected capacity of reactive power compensation equipment respectively; 6) DG active output constraints P min ≤P m (n)≤P max Where P m (n) is the active power output of DG in the nth period under the mth climate condition; P min , P max They are the minimum and maximum values ​​of DG active output respectively.

7. The method for optimizing the configuration of distributed power sources in an active power distribution network according to claim 1, characterized in that: The steps to implement the improved frog leaping algorithm are as follows: 1) Integrate the particle update strategy in the genetic algorithm into the hybrid frog leaping algorithm; 2) Update the individual with the lowest fitness ranking in the subgroup; 3) If the global optimal value does not improve, a new individual is generated to replace the original one.

8. The method for optimizing the configuration of distributed power sources in an active power distribution network according to claim 1, characterized in that: The specific implementation steps of the improved frog leaping algorithm are as follows: Step 1, initialization: according to the original parameters of the example node system, generate a total of F (F = m × n) candidate solutions as frog individuals in the definition domain; set the worst number K in the group and the number of consecutive unimproved fitness values ​​of the optimal solution of the population T; the number of global iterations S and the number of local iterations G; Step 2: Calculate the fitness of each frog according to the objective function, and sort the frogs according to the fitness size as required. The frog individual ranked first is the global optimal individual X g ; And group all frog individuals; Step 3: Update the K individuals with the lowest fitness ranking in each subgroup. If the fitness of the updated individual is better, replace the original individual, otherwise randomly generate a new individual to replace the original individual; if the number of times the global optimal value has not been improved is less than T, continue to update the remaining n-2-K individuals in the subgroup. If the fitness of the updated individual is better, replace the original individual; when the number of times the global optimal value has not been improved reaches T, randomly generate n-2-K individuals to replace the original individual; repeat this update process until the number of local iterations is reached; Step 4: Determine whether the number of global iterations has reached S. If so, stop the algorithm and output the optimal solution; otherwise, go to step 3.

9. An active distribution network distributed power optimization configuration system, characterized in that: The method comprises a memory, a processor and computer program instructions stored in the memory and capable of being executed by the processor. When the processor executes the computer program instructions, the method steps as claimed in any one of claims 1 to 8 can be implemented.

10. A computer-readable storage medium having stored thereon computer program instructions that can be executed by a processor, and when the processor executes the computer program instructions, the method steps according to any one of claims 1 to 8 can be implemented.

Images