New energy-containing multi-region power system economic dispatching method based on improved differential evolution algorithm

Abstract

The invention belongs to the technical field of power system dispatching, and particularly discloses a new energy-containing multi-region power system economic dispatching method based on an improved differential evolution algorithm, and the method comprises the steps: building a multi-region power system joint reserve capacity calculation model; the new energy uncertainty penalty cost of a certain region is calculated according to the multi-region power system joint reserve capacity obtained in the region, the comprehensive thermal power generation cost and the new energy uncertainty penalty cost are used as a target function, and constraint conditions for keeping stable operation of the power system are considered. Establishing a new energy-containing multi-region power system economic dispatching model; an improved differential evolution algorithm is adopted to solve the economic dispatching model of the new-energy-containing multi-region power system, and a solving result is obtained; analyzing an optimal strategy of economic dispatching of the new energy-containing multi-region power system according to a solving result; reasonable sharing and scheduling of the reserve capacity among the regions are realized, so that new energy consumption is promoted, and the operation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching, and in particular to an economic dispatching method for a multi-region power system containing new energy sources based on an improved differential evolution algorithm. Background Art

[0002] Spinning reserve capacity can be used to deal with renewable energy forecast errors. In a multi-regional power system, since different regions adapt to different renewable energy power generation technology solutions, it is a challenging task to adjust the spinning reserve capacity of each region according to the characteristics of renewable energy in different regions.

[0003] At present, for the multi-region interconnected economic dispatch containing new energy, the commonly used method for calculating rotating reserve capacity is the individual reserve capacity of a single region. However, the individual reserve capacity of a single region only considers the reserve capacity that can be provided by the region, and cannot comprehensively consider the reserve capacity that can be provided by all regions to each other, and cannot achieve reasonable allocation of reserve capacity according to regional energy characteristics. In view of the above problems, in order to flexibly respond to the uncertainty of new energy output in multi-regional power systems containing new energy and promote the consumption of new energy, it is of great prospect to establish a reasonable reserve capacity calculation theory.

[0004] When solving the economic dispatch problem of multi-regional power systems containing renewable energy, the differential evolution algorithm does not need to simplify the model, has high computational efficiency, and is highly adaptable to optimization problems. However, the algorithm's solution performance needs to be further improved. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an economic dispatch method for a multi-regional power system containing renewable energy based on an improved differential evolution algorithm.

[0006] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0007] A method for economic dispatching of a multi-regional power system containing renewable energy based on an improved differential evolution algorithm comprises the following steps:

[0008] S1. Construct a multi-regional power system consisting of new energy generators, thermal power generators and loads, and then establish a joint reserve capacity calculation model for the multi-regional power system;

[0009] S2. Calculate the uncertainty penalty cost of renewable energy in a region based on the joint reserve capacity of the multi-regional power system available in a certain region, take the comprehensive thermal power generation cost and the uncertainty penalty cost of renewable energy as the objective function, and consider the constraints of maintaining stable operation of the power system to establish an economic dispatch model of the multi-regional power system containing renewable energy;

[0010] S3. Use an improved differential evolution algorithm to solve the economic dispatch model of the multi-regional power system containing renewable energy to obtain a solution result; and analyze the optimal strategy for the economic dispatch of the multi-regional power system containing renewable energy based on the solution result.

[0011] Furthermore, in step S1, the specific steps of establishing a multi-regional power system joint reserve capacity calculation model are as follows:

[0012] According to the reserve capacity that each region can provide in the multi-region power system, on the basis of satisfying the tie line constraint and reserve capacity constraint, the calculation model of the negative rotation multi-region joint reserve capacity and the positive rotation multi-region joint reserve capacity is obtained as follows:

[0013]

[0014]

[0015] in, and are the negative rotation multi-region joint reserve capacity and the positive rotation multi-region joint reserve capacity of region i at time t; A k is the set of regions connected to region k; N is the number of regions; is the reserve capacity provided by region i at time t when calculating the joint reserve capacity of multiple regions; d i,l is a state variable used to indicate the direction of power transmission through the tie line; To calculate the power transfer value between area i and area l at time t when calculating the multi-area joint reserve capacity; T t k,l is the power transfer value between area k and area l at time t; is the maximum power flow limit from region l to region k; and are the maximum positive and negative spinning reserve capacities that region k can provide at time t, respectively, and are calculated using the following formula:

[0016]

[0017] Among them, M i is the number of thermal power generating units in region i; P i,t,j is the output power of the jth thermal power generating unit in region i at time t; and UR is the upper and lower limits of the output power of the jth thermal power generating unit in region i at time t; i,j and DR i,j are the up-ramp and down-ramp powers of the j-th thermal power generating unit in region i, respectively.

[0018] Furthermore, in step S2, taking the comprehensive thermal power generation cost and the new energy uncertainty penalty cost as the objective function specifically includes:

[0019] The objective function of comprehensive thermal power generation cost is as follows:

[0020]

[0021] Among them, f i,t,j (P i,t,j ) is the cost of the j-th thermal power generating unit in region i at time t; is the penalty cost of new energy uncertainty in region i at time t;

[0022]

[0023] Among them, α i,j ,β i,j ,γ i,j ,δ i,j and ε i,j The cost coefficient of the jth generator in region i;

[0024]

[0025] in, is the positive spinning reserve cost; is the negative spinning reserve cost; the specific calculation formula is as follows:

[0026]

[0027] in, is the planned output power of the new energy generator set in region i at time t; is the rated output power of the new energy power generation unit in region i; f new (w) is the probability distribution function of the output power of the renewable energy power generation unit, which obeys the Gaussian distribution; g is the penalty cost coefficient of the renewable energy output power;

[0028] The constraints for maintaining stable operation of the power system include: power balance constraints, generator set constraints and tie line constraints;

[0029] The power balance constraints are:

[0030]

[0031] in, is the load demand of area i at time t;

[0032] The constraints of the generator set are:

[0033]

[0034] The tie line constraints are:

[0035]

[0036] Furthermore, the step S3 specifically includes:

[0037] S31, processing the new energy and load data of the previous day, dividing the 24 hours of a day into T time periods, obtaining new energy output and load data curves, and establishing a problem model to be optimized according to the economic dispatch model of the multi-regional power system containing new energy;

[0038] S32, initializing algorithm parameters, including: maximum number of iterations G, population size NP, variable space dimension D of the economic dispatch problem of multi-regional power system containing renewable energy, upper bound ub and lower bound lb of the economic dispatch problem of multi-regional power system containing renewable energy;

[0039] S33, using a random number initialization method to initialize the population P, each individual in the population represents a scheduling result of the problem to be optimized;

[0040] S34, set the initial values ​​of the crossover operator memory archives MCr1, MCr2, MCr3, MCr4 and the mutation operator archives MF1, MF2, MF3, MF4, and the inferior archives The loop iteration variable is g=1, and the indexes k1=1, k2=1, k3=1, k4=1 indicating the different elements in the current archive are set;

[0041] S35. Set four groups of memory archives SCr1, SF1, SCr2, SF2, SCr3, SF3, SCr4, and SF4 as empty sets, and divide the population into four groups according to the fitness ranking of the population and the average ranking of the Euclidean distance between individuals in the population and their neighbors. The specific groups are: group P1 with higher convergence and diversity in the population, group P2 with higher convergence and lower diversity in the population, group P3 with lower convergence and higher diversity in the population, and group P4 with lower convergence and diversity in the population;

[0042] S36, according to the results of sorting and grouping, each individual in the four groups of populations is subjected to adaptive mutation based on the mutation operator archive library and adaptive crossover based on the crossover operator archive library, and then the individual vector in the population is updated according to the environment selection, and then the four groups of memory archives SCr1, SF1, SCr2, SF2, SCr3, SF3, SCr4, SF4 and the inferior archive library A after grouping are updated;

[0043] S37, taking the first set of memory archives SCr1 and SF1 as an example, the other three sets of memory archives are similar. When the memory archives SCr1 and SF1 are empty sets, the archives MCr1 and MF1 remain unchanged. When SCr1 and SF1 are not empty sets, MCr1 and MF1 are updated based on SCr1 and SF1. MCr1 and MF1 are updated according to the following formula:

[0044]

[0045] Where c is the learning factor of the archive library; mean WA (g) is the function for calculating the weighted arithmetic mean; mean WL (g) is the function for calculating the weighted Lehmer mean;

[0046]

[0047] Δf k =|f(U k )-f(X k )|;

[0048] k1=k1+1. If the index k1 is greater than the maximum number of storage elements H in the archive, set k1=1.

[0049] S38, g=g+1, determine whether the maximum number of iterations G is reached, if yes, jump out of the loop and execute step S39, otherwise return to step S35;

[0050] S39. Output the optimal solution and obtain the scheduling plan for the problem to be optimized.

[0051] Furthermore, in step S35, the method for sorting the population groups is:

[0052] S351, calculate the average Euclidean distance between each individual in the population and its neighbors;

[0053] S352. Under the condition that the constraint conditions are met, the objective function value of the economic dispatch model of the multi-regional power system containing new energy is calculated;

[0054] S353, arranging the objective function values ​​of the population in ascending order, and arranging the average values ​​of the Euclidean distances between individuals in the population and their neighbors in descending order;

[0055] S354. Individuals in the population that rank high in objective function value and average Euclidean distance between neighbors are divided into group P1 with higher convergence and diversity in the population; individuals in the population that rank high in objective function value and average Euclidean distance between neighbors are ranked low are divided into group P2 with higher convergence and lower diversity in the population; individuals in the population that rank low in objective function value and average Euclidean distance between neighbors are ranked high are divided into group P3 with lower convergence and higher diversity in the population; individuals in the population that rank low in objective function value and average Euclidean distance between neighbors are ranked low are divided into group P4 with lower convergence and diversity in the population.

[0056] Furthermore, in step S36, the four mutation strategies adopted for the four groups of populations are as follows:

[0057] For group P1 in the population:

[0058] V i =X i +F i ·(X gbest -X i )+F i ·(X r1 -X r2 );

[0059] Among them, X gbest represents the global optimal individual in the population; X r1 represents an individual randomly selected from the population P; X r2 represents an individual randomly selected from the archive A∪P;

[0060] For group P2 in the population:

[0061] V i =X i +F i ·(X P1 -X i )+F i ·(X r1 -X r2 );

[0062] Among them, X P1 represents an individual randomly selected from population P1; X r1 represents an individual randomly selected from the population P; X r2 represents an individual randomly selected from the archive A∪P;

[0063] For the P3 group in the population:

[0064] V i =X i +Fi ·(X P1∪P2 -X i )+F i ·(X r1 -X r2 );

[0065] Among them, X P1∪P2 represents an individual randomly selected from the population P1∪P2; X r1 and X r2 represents two individuals randomly selected from population P;

[0066] For the P4 group in the population:

[0067] V i =X i +F i ·(X P1∪P2∪P3 -X i )+F i ·(X r1 -X r2 );

[0068] Among them, X P1∪P2∪P3 represents an individual randomly selected from the population P1∪P2∪P3; X r1 and X r2 represents two individuals randomly selected from population P; F i is the variation factor;

[0069] The updating method of memory archive and inferior archive in the population is as follows: when the fitness of the generated test vector is better than that of the target vector, the SCr and SF of the group are updated, and the target vector is placed in the inferior archive A. When the archive A exceeds the maximum number of elements, the elements in the archive A are randomly selected and deleted until the number of elements in the archive A meets the conditions.

[0070] Compared with the prior art, the present invention establishes an economic dispatch model for a multi-regional power system containing new energy based on the joint reserve capacity of multiple regions. It can comprehensively consider the reserve capacity of all regions and realize the reasonable sharing and dispatch of reserve capacity between regions, thereby promoting the consumption of new energy and reducing operating costs. At the same time, by improving the differential evolution algorithm, the solution accuracy and solution stability are improved, the robustness of the algorithm in solving problems is enhanced, and a more accurate dispatch plan can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 The present invention is a flowchart of a method for optimizing economic dispatch of a multi-regional power system containing renewable energy based on an improved differential evolution algorithm according to an embodiment of the present invention.

[0072] Figure 2Schematic diagram of the system structure of the multi-regional power system joint reserve capacity according to an embodiment of the present invention.

[0073] Figure 3 The figure is a flow chart of an improved differential evolution algorithm according to an embodiment of the present invention.

[0074] Figure 4 It is an iteration diagram of the convergence process according to an embodiment of the present invention.

[0075] Figure 5 4 is a diagram showing the result of spare capacity scheduling according to an embodiment of the present invention. DETAILED DESCRIPTION

[0076] 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 embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the invention.

[0077] like Figure 1 As shown, this embodiment shows an economic dispatch method for a multi-regional power system containing new energy sources based on an improved differential evolution algorithm, and the specific steps are as follows:

[0078] S1. Construct a multi-regional power system consisting of new energy generators, thermal power generators and loads, and then establish a joint reserve capacity calculation model for the multi-regional power system;

[0079] The schematic diagram of the system structure for establishing the joint reserve capacity of the three-region power system with new energy is as follows Figure 2 As shown, region 1 includes 4 generator sets and 1 wind farm, region 2 includes 4 generator sets and 1 wind farm, region 3 includes 3 generator sets and 1 wind farm, and the maximum capacity of the interconnection line is 200MW. It should be noted that in this embodiment, the wind power forecast output of the wind farms in regions 1 and 3 is small, and the range forecast output values ​​are similar. The wind farm forecast output of region 2 is large, and all wind farms have the same rated power. Therefore, in this embodiment, both regions 1 and 3 require a large negative spinning reserve capacity. In addition, the uncertainty of wind power output in all regions cannot be fully covered; wherein, the dotted line represents the positive direction of power transmission between regions, the solid line represents the interconnection line between regions, T t k,l is the power transfer value between area k and area l at time t, To calculate the reserve capacity provided by region i at time t when calculating the joint reserve capacity of multiple regions, d i,l is a state variable, d i,l =1 means that area i transmits electric energy outward in the positive direction, d i,l =-1 means that region i absorbs electrical energy from other regions in the positive direction, To calculate the combined reserve capacity of a region, the power transfer value between region i and region l at time t is When the direction of power transmission is consistent with the specified positive direction, Is a positive number. When the direction of power transmission is opposite to the specified positive direction, Is a negative number.

[0080] The joint reserve capacity of the multi-regional power system, on the basis of satisfying the tie line constraint and the maximum reserve capacity constraint, the specific calculation model of the negative rotation multi-regional joint reserve capacity and the positive rotation multi-regional joint reserve capacity is:

[0081]

[0082]

[0083] in, and are the negative rotation multi-region joint reserve capacity and the positive rotation multi-region joint reserve capacity of region i at time t; A k is the set of regions connected to region k; N is the number of regions; is the maximum power flow limit from region l to region k; and are the maximum positive and negative spinning reserve capacities that region k can provide at time t, respectively, which can be calculated by the following formula:

[0084]

[0085] Among them, M i is the number of thermal power generating units in region i; P i,t,j is the output power of the jth thermal power generating unit in region i at time t; and UR is the upper and lower limits of the output power of the jth thermal power generating unit in region i at time t; i,j and DR i,j are the up-ramp and down-ramp powers of the j-th thermal power generating unit in region i, respectively.

[0086] S2. Calculate the uncertainty penalty cost of renewable energy in a region based on the joint reserve capacity of the multi-regional power system available in a certain region, take the comprehensive thermal power generation cost and the uncertainty penalty cost of renewable energy as the objective function, and consider the constraints of maintaining stable operation of the power system to establish an economic dispatch model of the multi-regional power system containing renewable energy;

[0087] S21. First, the operating cost of thermal power generating units and the penalty cost of new energy uncertainty should be constructed. Specifically, the operating cost of thermal power generating units should be constructed. i,t,j (P i,t,j) Penalty costs of overestimating renewable energy output Penalty costs for underestimating renewable energy output The total cost is as follows:

[0088]

[0089] Among them, f i,t,j (P i,t,j ) is the cost of the jth thermal power generating unit in region i at time t, is the penalty cost of new energy uncertainty in region i at time t, and express.

[0090] Exemplarily, the operating cost of a thermal power generating unit considering the valve point effect is:

[0091]

[0092] Among them, α i,j ,β i,j ,γ i,j ,δ i,j and ε i,j The cost coefficient of the jth generator in region i;

[0093] The specific new energy uncertainty penalty costs constructed are as follows:

[0094]

[0095] in, is the planned output power of the new energy generator set in region i at time t, is the rated output power of the renewable energy power generation unit in region i, f new (w) is the probability distribution function of the output power of the renewable energy power generation unit, which obeys the Gaussian distribution, and g is the penalty cost coefficient of the renewable energy output power.

[0096] S22. The proposed constraints include generator operation constraints, power balance constraints and inter-regional interconnection line capacity constraints.

[0097] The constraints of the generator set are:

[0098]

[0099] The tie line constraints are:

[0100]

[0101] The power balance constraints are:

[0102]

[0103] is the load demand of area i at time t.

[0104] S3, such as Figure 3 , using an improved differential evolution algorithm to solve the economic dispatch model of the multi-regional power system containing renewable energy, and obtaining a solution result; analyzing the optimal strategy for the economic dispatch of the multi-regional power system containing renewable energy according to the solution result;

[0105] It is proposed to improve the differential evolution algorithm and solve the economic dispatch problem of multi-regional power system containing renewable energy. The economic dispatch problem of multi-regional power system containing renewable energy can be divided into dynamic optimization problem and static optimization problem. Among them, the dynamic optimization problem requires the prediction of renewable energy output and load data in advance. In this process, due to the randomness of wind speed and light intensity, the load used by users cannot be accurately predicted. Therefore, when establishing the economic dispatch model of multi-regional power system containing renewable energy, the uncertainty of renewable energy needs to be considered. In the solution process, due to the complexity of the economic dispatch model of multi-regional power system containing renewable energy, it is difficult for the differential evolution algorithm to obtain a more accurate solution. The use of the improved differential evolution algorithm can effectively improve the accuracy of the algorithm solution. The specific algorithm steps are as follows:

[0106] S31. Process the predicted new energy output and load data of the day before, divide the 24 hours of a day into 24 time periods, obtain new energy output and load data curves, and establish a problem model to be optimized based on the economic dispatch model of the multi-regional power system containing new energy.

[0107] S32. Initialize algorithm parameters, including: maximum number of iterations G, population size NP, variable space dimension D of the economic dispatch problem of multi-regional power system containing renewable energy, upper bound ub and lower bound lb of the economic dispatch problem of multi-regional power system containing renewable energy.

[0108] S33, using the random number initialization method to initialize the population P, the specific method for generating a single individual in the population is as follows:

[0109] X i =lb+(ub-lb)·rand(0,1);

[0110] When initializing the population, the operating constraints of the generator sets, power balance constraints and inter-regional interconnection line capacity constraints should be met. Then, based on the objective function of the economic dispatch model of the multi-regional power system containing new energy, the cost of each individual in the population is obtained. Each individual in the population may represent a dispatch result of the problem to be optimized.

[0111] S34, set the initial values ​​of the crossover operator memory archives MCr1, MCr2, MCr3, MCr4 and the mutation operator archives MF1, MF2, MF3, MF4 to 0.5, and the inferior archives The loop iteration variable is g=1, and the indexes k1=1, k2=1, k3=1, k4=1 indicating the different elements in the current archive are set;

[0112] S35. Set four groups of memory archives SCr1, SF1, SCr2, SF2, SCr3, SF3, SCr4, and SF4 as empty sets. According to the fitness ranking of the population and the average Euclidean distance ranking between individuals and neighbors in the population, the population is divided into four groups. The specific groups are: group P1 with higher convergence and diversity in the population, group P2 with higher convergence and lower diversity in the population, group P3 with lower convergence and higher diversity in the population, and group P4 with lower convergence and diversity in the population. The method for grouping and sorting the population is:

[0113] S351. Calculate the average Euclidean distance between each individual in the population and its neighbors.

[0114] S352. Under the condition that the constraints are met, calculate the objective function value of the economic dispatch model of the multi-regional power system containing new energy.

[0115] S353. Arrange the objective function values ​​of the population in ascending order, and arrange the average Euclidean distances between individuals and their neighbors in the population in descending order.

[0116] S354. The individuals in the population that rank in the top 1 / 2 in terms of objective function value and the average value of the Euclidean distance between the population and the neighbors are divided into group P1 with higher convergence and diversity in the population; the individuals in the population that rank in the top 1 / 2 in terms of objective function value and the bottom 1 / 2 in terms of the average value of the Euclidean distance between the population and the neighbors are divided into group P2 with higher convergence and lower diversity in the population; the individuals in the population that rank in the bottom 1 / 2 in terms of objective function value and the top 1 / 2 in terms of the average value of the Euclidean distance between the population and the neighbors are divided into group P3 with lower convergence and higher diversity; the individuals in the population that rank in the bottom 1 / 2 in terms of objective function value and the average value of the Euclidean distance between the population and the neighbors are divided into group P4 with lower convergence and diversity in the population.

[0117] S36. According to the results of sorting and grouping, each individual in the four groups of populations is subjected to adaptive mutation based on the mutation operator archive library, adaptive crossover based on the crossover operator archive library, and then the individual vector in the population is updated according to the environment selection, and then the four groups of memory archives SCr1, SF1, SCr2, SF2, SCr3, SF3, SCr4, SF4 and the inferior archive library A after grouping are updated. The four mutation strategies adopted for the four groups of populations are as follows:

[0118] For group P1 in the population:

[0119] V i =X i+F i ·(X gbest -X i )+F i ·(X r1 -X r2 );

[0120] Among them, X gbest represents the global optimal individual in the population, X r1 represents an individual randomly selected from the population P, X r2 represents an individual randomly selected from the archive A∪P;

[0121] For group P2 in the population:

[0122] V i =X i +F i ·(X P1 -X i )+F i ·(X r1 -X r2 );

[0123] Among them, X P1 represents an individual randomly selected from population P1, X r1 represents an individual randomly selected from the population P, X r2 represents an individual randomly selected from the archive A∪P;

[0124] For the P3 group in the population:

[0125] V i =X i +F i ·(X P1∪P2 -X i )+F i ·(X r1 -X r2 );

[0126] Among them, X P1∪P2 represents an individual randomly selected from the population P1∪P2, X r1 and X r2 represents two individuals randomly selected from population P;

[0127] For the P4 group in the population:

[0128] V i =X i +F i ·(X P1∪P2∪P3 -X i )+F i ·(X r1 -X r2);

[0129] Among them, X P1∪P2∪P3 represents an individual randomly selected from the population P1∪P2∪P3, X r1 and X r2 Represent two individuals randomly selected from the population P.

[0130] In the mutation strategy, the mutation factor is obtained by the following formula:

[0131] F i =randc i (MF(ri),0.1);

[0132] Among them, F i should be limited to [0,1]. i If F i =1, when the F generated by the above mutation factor generation method i If it is less than 0, the above formula is repeated until a feasible value is obtained. ri is an index randomly selected from [1,H], and randc i (μ,σ 2 ) is the mean μ and variance σ 2 A value chosen randomly from a Cauchy distribution.

[0133] After the individuals in the population undergo mutation operations, they then undergo crossover operations as shown below:

[0134]

[0135] Among them, Cr i is the cross factor, j rand is a random integer in [1,D], u i,j It is a decision variable of the newly generated individual in the jth dimension.

[0136] In the above crossover operation, the crossover factor is obtained by the following formula:

[0137] Cr i =randn i (MCr(ri),0.1);

[0138] Among them, Cr i Should be limited to [0,1], when Cr generated by the above mutation factor generation method i If Cr is greater than 1, i =1, when Cr generated by the above mutation factor generation method i Less than, let Cr i =0, ri is an index randomly selected from [1,H], randni (μ,σ 2 ) is the mean μ and variance σ 2 A value randomly selected from the normal distribution.

[0139] Next, the individual environment selection after mutation and crossover is as follows:

[0140]

[0141] Among them, f(U i ) and f(X i ) are the objective function values ​​of the experimental vector and the target vector respectively.

[0142] The updating method of memory archive and inferior archive library in the population: when the fitness of the generated test vector is better than the target vector, Cr i and F i They are added to the archive libraries SCr and SF of their respective groups, and the target vector is placed in the inferior archive library A. When the archive library A exceeds the maximum number of elements, the elements in the archive library A are randomly selected and deleted until the number of elements in the archive library A meets the conditions.

[0143] S37. Taking the first set of memory archive libraries SCr1 and SF1 as an example, the other three sets of memory archive libraries are similar. When the memory archive libraries SCr1 and SF1 are empty sets, the archive libraries MCr1 and MF1 remain unchanged. When SCr1 and SF1 are not empty sets, MCr1 and MF1 are updated based on SCr1 and SF1. The specific method for updating the archive library is as follows:

[0144] S371, update MCr1 and MF1 according to the following formula

[0145]

[0146] Among them, c is the learning factor of the archive library, mean WA (g) is the function for calculating the weighted arithmetic mean, mean WL (g) is the function for calculating the weighted Lehmer mean.

[0147]

[0148] Where Δf k =|f(U k )-f(X k )|.

[0149] S372, k1=k1+1, if the index k1 is greater than the maximum number of storage elements H in the archive library, set k1=1;

[0150] S38, g=g+1, determine whether the maximum number of iterations G is reached, if yes, jump out of the loop and execute step S3.9, otherwise return to step S3.5;

[0151] S39. Output the optimal solution and obtain the scheduling plan for the problem to be optimized.

[0152] In order to verify the effectiveness of this method, two simulation experiments are designed to verify:

[0153] Simulation experiment 1

[0154] This experiment used Figure 2 The simulation analysis of the three-region power system with new energy is shown, with 24h set as an optimization time interval, the optimization cycle is 1h, region 1 includes 4 generators and 1 wind farm, region 2 includes 4 generators, region 3 includes 3 generators, the maximum capacity of the tie line is 200MW, and the rated power of the wind farm is 400MW. In this embodiment, the uncertainty of wind power output can be fully covered.

[0155] like Figure 4 As shown in the figure, the blue curve is the proposed improved differential evolution algorithm. The improved differential evolution algorithm has the highest solution accuracy and continuous convergence in the iterative solution process, indicating that the algorithm has a good effect in balancing convergence and diversity.

[0156] As shown in Table 1, the improved differential evolution algorithm proposed in the present invention is used to solve the economic dispatch problem of multi-regional power system containing new energy based on traditional single-region reserve capacity and the economic dispatch problem of multi-regional power system containing new energy based on multi-regional joint reserve capacity, respectively. The results show that the economic dispatch model of multi-regional power system containing new energy based on multi-regional joint reserve capacity proposed in the present invention has lower cost, indicating that when the uncertainty of new energy output can be fully covered, the economic dispatch model of multi-regional power system containing new energy based on multi-regional joint reserve capacity has better flexibility.

[0157] Table 1 Comparison of the two models

[0158]

[0159] Simulation experiment 2

[0160] This experiment still uses Figure 2The three-region power system containing new energy is simulated and analyzed, with 24h set as an optimization time interval and an optimization cycle of 1h. Region 1 includes 4 generators and 1 wind farm, region 2 includes 4 generators and 1 wind farm, region 3 includes 3 generators and 1 wind farm, and the maximum capacity of the interconnection line is 200MW. It should be noted that, in this embodiment, the predicted wind power output of the wind farms in regions 1 and 3 is relatively small, and the range predicted output values ​​are similar. The predicted output of the wind farms in region 2 is relatively large, and all wind farms have the same rated power. Therefore, in this embodiment, both regions 1 and 3 require a larger negative rotating reserve capacity. In addition, the uncertainty of wind power output in all regions cannot be fully covered.

[0161] like Figure 5 It can be seen that the model based on the individual reserve capacity of a single region has a larger negative spinning reserve capacity in region 1, and a smaller negative spinning reserve capacity in regions 2 and 3. The model based on the reserve capacity of multiple regions has a smaller negative spinning reserve capacity in region 1, and a larger negative spinning reserve capacity in regions 2 and 3. Since the predicted wind power output values ​​in regions 1 and 3 are similar and have the same rated power, in order to ensure the lowest penalty cost for uncertainty of new energy, the interval with a higher probability density should be covered as much as possible, that is, the negative spinning reserve in regions 1 and 3 should be relatively even. Figure 5 It can be seen that the model based on multi-regional reserve capacity is significantly better than the model based on single-region individual reserve capacity.

[0162] It can be seen from the above simulation that the present invention can effectively solve the uncertainty problem caused by the introduction of new energy into multi-regional power systems. The experimental results also show that the improved differential evolution algorithm has a higher solution accuracy.

[0163] The technical solution of the present invention is not limited to the above-mentioned specific embodiments. All technical variations made according to the technical solution of the present invention fall within the protection scope of the present invention.

Claims

1. An economic dispatch method for a multi-regional power system containing new energy based on an improved differential evolution algorithm, characterized in that: The following steps are involved: S1. Construct a multi-regional power system consisting of new energy generators, thermal power generators and loads, and then establish a joint reserve capacity calculation model for the multi-regional power system; S2. Calculate the uncertainty penalty cost of renewable energy in a region based on the joint reserve capacity of the multi-regional power system available in a certain region, take the comprehensive thermal power generation cost and the uncertainty penalty cost of renewable energy as the objective function, and consider the constraints of maintaining stable operation of the power system to establish an economic dispatch model of the multi-regional power system containing renewable energy; S3. Use an improved differential evolution algorithm to solve the economic dispatch model of the multi-regional power system containing renewable energy to obtain a solution result; and analyze the optimal strategy for the economic dispatch of the multi-regional power system containing renewable energy based on the solution result.

2. The economic dispatch method for a multi-regional power system containing new energy based on an improved differential evolution algorithm according to claim 1 is characterized in that: In step S1, the specific steps of establishing a multi-regional power system joint reserve capacity calculation model are as follows: According to the reserve capacity that each region can provide in the multi-region power system, on the basis of satisfying the tie line constraint and reserve capacity constraint, the calculation model of the negative rotation multi-region joint reserve capacity and the positive rotation multi-region joint reserve capacity is obtained as follows: in, and are the negative rotation multi-region joint reserve capacity and the positive rotation multi-region joint reserve capacity of region i at time t; A k is the set of regions connected to region k; N is the number of regions; is the reserve capacity provided by region i at time t when calculating the joint reserve capacity of multiple regions; d i,l is a state variable used to indicate the direction of power transmission through the tie line; To calculate the power transfer value between area i and area l at time t when calculating the multi-area joint reserve capacity; T t k,l is the power transfer value between area k and area l at time t; is the maximum power flow limit from region l to region k; and are the maximum positive and negative spinning reserve capacities that region k can provide at time t, respectively, and are calculated using the following formula: Among them, M i is the number of thermal power generating units in region i; P i,t,j is the output power of the jth thermal power generating unit in region i at time t; and UR is the upper and lower limits of the output power of the jth thermal power generating unit in region i at time t; i,j and DR i,j are the up-ramp and down-ramp powers of the j-th thermal power generating unit in region i, respectively.

3. The economic dispatch method for a multi-regional power system containing new energy based on an improved differential evolution algorithm according to claim 1 is characterized in that: In step S2, taking the comprehensive thermal power generation cost and the new energy uncertainty penalty cost as the objective function specifically includes: The objective function of comprehensive thermal power generation cost is as follows: Among them, f i,t,j (P i,t,j ) is the cost of the j-th thermal power generating unit in region i at time t; is the penalty cost of new energy uncertainty in region i at time t; Among them, α i,j ,β i,j ,γ i,j ,δ i,j and ε i,j The cost coefficient of the jth generator in region i; in, is the positive spinning reserve cost; is the negative spinning reserve cost; the specific calculation formula is as follows: in, is the planned output power of the new energy generator set in region i at time t; is the rated output power of the new energy power generation unit in region i; f new (w) is the probability distribution function of the output power of the renewable energy power generation unit, which obeys the Gaussian distribution; g is the penalty cost coefficient of the renewable energy output power; The constraints for maintaining stable operation of the power system include: power balance constraints, generator set constraints and tie line constraints; The power balance constraints are: in, is the load demand of area i at time t; The constraints of the generator set are: The tie line constraints are:

4. The economic dispatch method for a multi-regional power system containing new energy based on an improved differential evolution algorithm according to claim 1 is characterized in that: The step S3 specifically includes: S31, processing the new energy and load data of the previous day, dividing the 24 hours of a day into T time periods, obtaining new energy output and load data curves, and establishing a problem model to be optimized according to the economic dispatch model of the multi-regional power system containing new energy; S32, initializing algorithm parameters, including: maximum number of iterations G, population size NP, variable space dimension D of the economic dispatch problem of multi-regional power system containing renewable energy, upper bound ub and lower bound lb of the economic dispatch problem of multi-regional power system containing renewable energy; S33, using a random number initialization method to initialize the population P, where each individual in the population represents a scheduling result of the problem to be optimized; S34, set the initial values ​​of the crossover operator memory archives MCr1, MCr2, MCr3, MCr4 and the mutation operator archives MF1, MF2, MF3, MF4, and the inferior archives The loop iteration variable is g=1, and the indexes k1=1, k2=1, k3=1, k4=1 indicating the different elements in the current archive are set; S35. Set four groups of memory archives SCr1, SF1, SCr2, SF2, SCr3, SF3, SCr4, and SF4 as empty sets, and divide the population into four groups according to the fitness ranking of the population and the average ranking of the Euclidean distance between individuals in the population and their neighbors. The specific groups are: group P1 with higher convergence and diversity in the population, group P2 with higher convergence and lower diversity in the population, group P3 with lower convergence and higher diversity in the population, and group P4 with lower convergence and diversity in the population; S36, according to the results of sorting and grouping, each individual in the four groups of populations is subjected to adaptive mutation based on the mutation operator archive library and adaptive crossover based on the crossover operator archive library, and then the individual vector in the population is updated according to the environment selection, and then the four groups of memory archives SCr1, SF1, SCr2, SF2, SCr3, SF3, SCr4, SF4 and the inferior archive library A after grouping are updated; S37, taking the first set of memory archives SCr1 and SF1 as an example, the other three sets of memory archives are similar. When the memory archives SCr1 and SF1 are empty sets, the archives MCr1 and MF1 remain unchanged. When SCr1 and SF1 are not empty sets, MCr1 and MF1 are updated based on SCr1 and SF1. MCr1 and MF1 are updated according to the following formula: Where c is the learning factor of the archive library; mean WA (g) is the function for calculating the weighted arithmetic mean; mean WL (g) is the function for calculating the weighted Lehmer mean; Δf k =|f(U k )-f(X k )|; k1=k1+1. If the index k1 is greater than the maximum number of storage elements H in the archive, set k1=1. S38, g=g+1, determine whether the maximum number of iterations G is reached, if yes, jump out of the loop and execute step S39, otherwise return to step S35; S39. Output the optimal solution and obtain the scheduling plan for the problem to be optimized.

5. The economic dispatch method for a multi-regional power system containing new energy based on an improved differential evolution algorithm according to claim 4 is characterized in that: In step S35, the method for sorting the population groups is: S351, calculate the average Euclidean distance between each individual in the population and its neighbors; S352. Under the condition that the constraint conditions are met, the objective function value of the economic dispatch model of the multi-regional power system containing new energy is calculated; S353, arranging the objective function values ​​of the population in ascending order, and arranging the average values ​​of the Euclidean distances between individuals in the population and their neighbors in descending order; S354. Individuals in the population that rank high in objective function value and average Euclidean distance between neighbors are divided into group P1 with higher convergence and diversity in the population; individuals in the population that rank high in objective function value and average Euclidean distance between neighbors are ranked low are divided into group P2 with higher convergence and lower diversity in the population; individuals in the population that rank low in objective function value and average Euclidean distance between neighbors are ranked high are divided into group P3 with lower convergence and higher diversity in the population; individuals in the population that rank low in objective function value and average Euclidean distance between neighbors are ranked low are divided into group P4 with lower convergence and diversity in the population.

6. The economic dispatch method for a multi-regional power system containing new energy based on an improved differential evolution algorithm according to claim 1 is characterized in that: In step S36, the four mutation strategies adopted for the four groups of populations are as follows: For group P1 in the population: V i =X i +F i ·(X gbest -X i )+F i ·(X r1 -X r2 ); Among them, X gbest represents the global optimal individual in the population; X r1 represents an individual randomly selected from the population P; X r2 represents an individual randomly selected from the archive A∪P; For group P2 in the population: V i =X i +F i ·(X P1 -X i )+F i ·(X r1 -X r2 ); Among them, X P1 represents an individual randomly selected from population P1; X r1 represents an individual randomly selected from the population P; X r2 represents an individual randomly selected from the archive A∪P; For the P3 group in the population: V i =X i +F i ·(X P1∪P2 -X i )+F i ·(X r1 -X r2 ); Among them, X P1∪P2 represents an individual randomly selected from the population P1∪P2; X r1 and X r2 represents two individuals randomly selected from population P; For the P4 group in the population: V i =X i +F i ·(X P1∪P2∪P3 -X i )+F i ·(X r1 -X r2 ); Among them, X P1∪P2∪P3 represents an individual randomly selected from the population P1∪P2∪P3; X r1 and X r2 represents two individuals randomly selected from population P; F i is the variation factor; The updating method of memory archive and inferior archive in the population is as follows: when the fitness of the generated test vector is better than that of the target vector, the SCr and SF of the group are updated, and the target vector is placed in the inferior archive A. When the archive A exceeds the maximum number of elements, the elements in the archive A are randomly selected and deleted until the number of elements in the archive A meets the conditions.

Images