CN111461478A - Large-scale water-light energy complementary scheduling method and system - Google Patents

Abstract

The invention discloses a large-scale water-light energy complementary scheduling method and system, belonging to the field of water-light complementary scheduling. First, the initial population is randomly generated in the search space, and the individual optimal position and the global optimal position in the current population are updated according to the fitness value of each individual, and then all individual positions of the population are updated; the local search strategy is used to improve the convergence speed of the population , use the adaptive mutation strategy to screen the population, and update the position of all individuals in the population by iterative calculation, and after reaching the maximum number of iterations, the global optimal position of the population is obtained as the optimal scheme of water-light energy complementary scheduling. The invention solves the technical problems existing in the existing GSA algorithm, such as difficulty in getting rid of local optimum and weak development ability, has the advantages of strong optimization ability, can reasonably handle the balance between exploration and development for the water-light cooperative scheduling problem, and has the advantages of Good engineering practicality.

Description

A large-scale water-light energy complementary scheduling method and system

technical field

The invention belongs to the field of water-light complementary scheduling, and more particularly relates to a large-scale water-light energy complementary scheduling method and system.

Background technique

其中f为目标函数值；ρ_s为第s个光伏电站发生输出概率；P_s,m,t为在s概率下发生后第m个光伏电站在t个时段的出力。L_t为第t个时段电力系统的负荷需求。P_n,t为第n个电站在第t个时段的出力；M是光伏电站的数目；N为水电站总数目；T为总时段数目。my country is rich in energy reserves, the main fossil energy is coal, the coal reserves ranks third in the world, and the theoretical reserves of hydraulic resources rank first in the world. Wind energy, solar energy and other new energy resources are abundant. However, my country has a large population, and the per capita share of energy is low, and the distribution is extremely uneven. With the improvement of the technical level, the direction of research scholars has gradually expanded to the new field of solar energy. There are usually many kinds of power loads participating in the operation of the power grid. For the coordinated dispatch of cascade reservoir groups and photovoltaic power stations, the variance of the participating loads of the power grid is usually used as the ultimate goal of optimization. The specific formula is:

where f is the objective function value; ρ _s is the output probability of the s-th photovoltaic power station; P _s,m,t is the output of the m-th photovoltaic power station in the t period after the occurrence of the s probability. L _t is the load demand of the power system in the t-th period. P _n,t is the output of the n-th power station in the t-th period; M is the number of photovoltaic power stations; N is the total number of hydropower stations; T is the total number of periods.

The constraints that need to be satisfied are as follows:

其中，NU_n为直接连接在第n个水电站的上游电站数目；V_n,t为第n个水电站在第t个时段的库容；q_n,t为第n个水电站在第t个时段的区间流量；I_n,t为第n个水电站在第t个时段的总入库流量；O_n,t为第n个水电站在第t个时段的总出库流量；Q_n,t为第n个水电站在第t个时段的发电流量。(1) Water balance constraints:

Among them, NU _n is the number of upstream power stations directly connected to the nth hydropower station; Vn _,t is the storage capacity of the nth hydropower station in the tth period; qn _,t is the interval of the nth hydropower station in the tth period Flow; I _n,t is the total inflow flow of the nth hydropower station in the tth period; O _n,t is the total outgoing flow of the nth hydropower station in the tth period; Qn _,t is the nth hydropower station The power generation flow of the hydropower station in the t-th period.

(2) Hydraulic head constraint: H _n,t =0.5(Zn _,t +Zn _,t-1 )-d _n,t . Among them, Z _n,t is the water level in front of the dam of the nth hydropower station in the tth period; Hn _, t is the water head of the nth hydropower station in the tth period; dn _,t is the nth hydropower station in the tth period downstream water level for a period of time.

其中，

为第n个水电站在第t个时段的坝前水位最小值；

为第n个水电站在第t个时段的坝前水位最大值；(3) Water level constraints in front of the dam:

in,

is the minimum water level before the dam of the nth hydropower station in the tth period;

is the maximum water level before the dam of the nth hydropower station in the tth period;

其中，

为第n个水电站在第t个时段的发电流量上限；

为第n个水电站在第t个时段的发电流量下限；(4) Power generation flow constraints:

in,

is the upper limit of the power generation flow of the nth hydropower station in the tth period;

is the lower limit of the power generation flow of the nth hydropower station in the tth period;

其中，

为第n个水电站在第t个时段的出库流量上限；

为第n个水电站在第t个时段的出库流量下限；(5) Reservoir outflow flow constraints:

in,

is the upper limit of the outbound flow of the nth hydropower station in the tth period;

is the lower limit of the outbound flow of the nth hydropower station in the tth period;

其中，

为第n个水电站在第t个时段的水电出力上限；

为第n个水电站在第t个时段的水电出力下限；(6) Output constraints of hydropower stations:

in,

is the upper limit of hydropower output of the nth hydropower station in the tth period;

is the lower limit of hydropower output of the nth hydropower station in the tth period;

其中，

为第n个水电站在第t个时段光能出力下限，

为第n个水电站在第t个时段光能出力上限。(7) Output constraints of photovoltaic power plants:

in,

is the lower limit of light energy output of the nth hydropower station in the tth period,

It is the upper limit of light energy output of the nth hydropower station in the tth period.

Typically, the optimization objective in a power system with cascaded hydropower stations is to determine the optimal dispatch scheme to maximize the benefits of the entire power system while satisfying a series of physical constraints (eg, water balance constraints, generation flow constraints). With the large-scale injection of solar photovoltaic power generation, the impact of the external environment on forecast uncertainty has become prominent, and it has become more difficult to formulate scientific decision-making programs. To effectively address this challenge, it is important to develop an appropriate optimization model considering the uncertainty of solar power generation. The Gravitational Search Algorithm (GSA) is a novel swarm-based evolutionary algorithm inspired by Newton's laws of gravity and motion. In GSA, each potential solution is treated as a planet in the universe, the quality of which can be measured by the fitness value associated with the target problem. However, because these factors tend to have individuals of the same quality in the later stages of evolution, it is difficult for GSA to get rid of the problems of local optima and weak development ability.

SUMMARY OF THE INVENTION

In view of the above defects or improvement needs of the prior art, the present invention provides a large-scale water-light energy complementary scheduling method and system, thereby solving the technical problems of the existing GSA algorithm, such as difficulty in getting rid of local optimum and weak development ability.

In order to achieve the above object, according to one aspect of the present invention, a large-scale water-light energy complementary scheduling method is provided, comprising the following steps:

(1) Determine the constraints between the power stations, the power stations include hydropower stations and photovoltaic power stations, take the outbound flow of each hydropower station at different times as a decision variable and encode it, and then randomly generate an initial population in the search space according to the decision variable, Taking the initial population as the current population, any individual in the population represents a complementary scheduling scheme of water-light energy;

(2) Obtain the fitness of all individuals in the current population, update the individual optimal position and the global optimal position in the current population by the fitness value of each individual, and then update the positions of all individuals in the population;

Wherein, in the first iteration of the population, the optimal position of the individual refers to the position of the individual in the first iteration of the population, and in the second and previous iterations, the optimal position of the individual refers to the position of the individual in the population The position in this iteration and the position in the previous iteration are the better positions; the global optimal position refers to the position of the optimal individual in this iteration of the population;

(3) Based on the population obtained after updating all the individual positions of the population, a local search strategy is used to perform a local search on the population to improve the population convergence speed;

(4) Use adaptive mutation strategy to screen populations to improve population diversity;

(5) Based on the population obtained by using the mutation strategy, return the individuals beyond the boundary to the boundary range to form the next generation population;

(6) Judging whether the algebra of the current population has reached the preset maximum number of iterations, if not, the next generation of the population will be used as the current population, and return to step (2). The optimal individual output is the optimal scheme for the complementary scheduling of water and light energy.

其中，N表示电站数目；T表示时段数目；

为X_i(k)中第n个水电站在第t个时段的出库流量，

为第n个水电站在第t个时段的发电流量下限，rand(0,1)为[0,1]区间均匀分布的随机数，k表示迭代次数，X_i(k)具有N*T个维度。Preferably, for any individual in the k-th generation population X _i (k) can be expressed as

Among them, N represents the number of power stations; T represents the number of time periods;

is the outbound flow of the nth hydropower station in the tth period in Xi ( _k ),

is the lower limit of the power generation flow of the nth hydropower station in the tth period, rand(0,1) is a random number uniformly distributed in the [0,1] interval, k represents the number of iterations, and X _i (k) has N*T dimensions .

Preferably, the fitness F[X _i (k)] of the i-th individual X _i (k) of the k-th generation is:

L_t为电力系统第t个时段的负荷需求；P_s,m,t为在概率ρ_s下第m个光伏电站在第t个时段的出力；P_n,t为第n个水电站在第t个时段的出力；M是光伏电站的总数目；N为水电站总数目；T为总时段数目；g_a[X_i(k)]和c_a分别为第a个不等式约束的约束违背值和惩罚系数；e_b[X_i(k)]和c_b分别为第b个等式的约束违背值和惩罚系数；A和B分别为不等式约束和等式约束的数目。Among them, ρ _s is the probability of occurrence of the s-th photovoltaic power station; S is the total probability number,

L _t is the load demand of the power system in the t-th period; P _s,m,t is the output of the m-th photovoltaic power station in the t-th period under the probability ρ _s ; P _n,t is the n-th hydropower station in the t-th period The output of each period; M is the total number of photovoltaic power stations; N is the total number of hydropower stations; T is the total number of periods; g _a [X _i (k)] and c _a are the constraint violation value and penalty of the a-th inequality constraint, respectively coefficients; e _b [X _i (k)] and c _b are the constraint violation value and penalty coefficient of the b-th equation, respectively; A and B are the number of inequality constraints and equality constraints, respectively.

更新当前种群中的全局最优位置，由

更新当前种群中的个体最优位置；其中，gBest(k)为种群第k次迭代的全局最优位置，gBest(k)＝{gBest^d(k),d＝1,2...,D}，gBest^d(k)为种群第k次迭代在第d维度的全局最优位置，gBest(k-1)种群第k-1次迭代的全局最优位置，

为第k次迭代第i个个体第d维度的个体最优位置，D为第i个个体的最大维度，pBest_i(k)为种群第k次迭代第i个个体的个体最优位置，pBest_i(k-1)为种群第k-1次迭代第i个个体的个体最优位置。Preferably, by

Update the global optimal position in the current population, given by

Update the individual optimal position in the current population; where, gBest(k) is the global optimal position of the k-th iteration of the population, gBest(k)={gBest ^d (k),d=1,2...,D }, gBest ^d (k) is the global optimal position of the k-th iteration of the population in the d-th dimension, and the global optimal position of the k-1 iteration of the gBest(k-1) population,

is the individual optimal position of the i-th individual in the k-th iteration of the d-th dimension, D is the maximum dimension of the i-th individual, pBest _i (k) is the individual optimal position of the i-th individual in the k-th iteration of the population, pBest _i (k-1) is the individual optimal position of the i-th individual in the k-1 iteration of the population.

更新种群所有个体位置，式中，Preferably, by

Update the positions of all individuals in the population, where,

R _ij (k)=||x _i (k)-x _j (k)||

和

分别为第k代种群和第k+1代种群第i个个体在第d维的位置，

为第k代种群第i个个体在第j维的位置；种群第k次迭代第i个个体的位置

种群第k+1次迭代第i个个体的位置

和

分别为第k次迭代和第k+1次迭代第i个个体在第d维的速度，

为第k次迭代第i个个体在第d维上的加速度，rand_j和rand_i是[0,1]之间均匀分布的随机数；Kbest为前K个具有更好适应度的个体；in,

and

are the positions of the i-th individual in the k-th generation population and the k+1-th generation population in the d-dimension, respectively,

is the position of the i-th individual in the k-th generation population in the j-th dimension; the position of the i-th individual in the k-th iteration of the population

The position of the ith individual in the k+1th iteration of the population

and

are the velocity of the i-th individual in the d-th dimension in the k-th iteration and the k+1-th iteration, respectively,

is the acceleration of the i-th individual in the k-th iteration on the d-th dimension, rand _j and rand _i are random numbers uniformly distributed between [0, 1]; Kbest is the first K individuals with better fitness;

为第k次迭代种群中所有个体最差适应度；

为第k次迭代种群中所有个体最优适应度；

为第i个个体对第j个个体在第d维上的作用力；M_pi(k)为第i个个体的被动质量，M_aj(k)为第j个个体的主动质量，R_ij(k)为第i个个体和第j个个体的欧氏距离，G₀为万有引力常数的初始值，G(k)是第k次迭代的万有引力常数值，α为衰减系数，ε为常数值。

is the worst fitness of all individuals in the k-th iteration population;

is the optimal fitness of all individuals in the k-th iteration population;

is the force of the i-th individual on the j-th individual on the d-th dimension; M _pi (k) is the passive mass of the i-th individual, M _aj (k) is the active mass of the j-th individual, and R _ij ( k) is the Euclidean distance between the i-th individual and the j-th individual, G ₀ is the initial value of the gravitational constant, G(k) is the gravitational constant value of the k-th iteration, α is the decay coefficient, and ε is the constant value.

Preferably, the population is locally searched using the following formula:

c ₂ =1-c ₁

为第k次迭代第i个个体在第d维度的局部搜索位置；r₃和r₄为[0,1]区间均匀分布的随机数；δ^d为搜索空间中第d维的中值；

为第k次迭代第i个个体在第d维度的对立因子；c₁和c₂为学习因子；

为搜索空间第d维度上限值；x ^d为搜索空间第d维度下限值；

为预设的最大迭代次数。In the formula,

is the local search position of the i-th individual in the d-th dimension of the k-th iteration; r ₃ and r ₄ are random numbers uniformly distributed in the [0,1] interval; δ ^d is the median of the d-th dimension in the search space;

is the opposition factor of the i-th individual in the d-th dimension in the k-th iteration; c ₁ and c ₂ are learning factors;

is the upper limit value of the d-th dimension of the search space; x ^d is the lower value of the d-th dimension of the search space;

is the preset maximum number of iterations.

Preferably, step (4) includes:

为第k次迭代第i个个体在第d维度的变异位置，α为种群中随机选择的个体下标，

表示第k次迭代第α个个体第d维度的位置；φ为[-0.5,0.5]区间均匀分布的随机数；Elite是从当前种群中得到的前三个最优个体位置集合，

为第k次迭代第β个个体第d维度的个体位置，β为Elite中随机选择的个体下标。The individual positions in the current population are sorted according to the fitness value, the first a (a < m) individuals directly enter the next iteration of the population, and the adaptive mutation operation is used for the remaining ma individuals to generate mutant individuals and the first a (a < m ) ) individuals as the population in the next iteration; among them, the adaptive mutation method is:

is the variation position of the i-th individual in the d-th dimension in the k-th iteration, α is the subscript of the randomly selected individual in the population,

Represents the position of the d-th dimension of the α-th individual in the k-th iteration; φ is a random number uniformly distributed in the interval [-0.5, 0.5]; Elite is the set of the first three optimal individual positions obtained from the current population,

is the individual position of the d-th dimension of the β-th individual in the k-th iteration, and β is the randomly selected individual subscript in Elite.

将超出边界的个体返回到边界范围内，其中，

r₁为[0,1]区间均匀分布的随机数。Preferably, by

Return individuals beyond the boundary to within the boundary, where,

r ₁ is a random number uniformly distributed in the interval [0,1].

According to another aspect of the present invention, a large-scale water-light energy complementary scheduling system is provided, comprising:

The initial population generation module is used to determine the constraints between the power stations, including hydropower stations and photovoltaic power stations. The outbound flow of each hydropower station at different times is used as a decision variable and coded, and then randomly selected in the search space according to the decision variable. Generate an initial population and take the initial population as the current population, where any individual in the population represents a complementary scheduling scheme of water-light energy;

The position update module is used to obtain the fitness of all individuals in the current population, update the individual optimal position and the global optimal position in the current population by the fitness value of each individual, and then update the positions of all individuals in the population; among them, In the first iteration of the population, the optimal position of the individual refers to the position of the individual in the first iteration of the population. In the second and previous iterations, the optimal position of the individual refers to the position of the individual in the current iteration of the population. The better position between the position in the iteration and the position in the previous iteration; the global optimal position refers to the position of the optimal individual in the current iteration of the population;

The local search module is used to perform a local search on the population by using a local search strategy based on the population obtained after updating the positions of all individuals in the population to improve the population convergence speed;

Adaptive mutation module, which is used to screen populations using adaptive mutation strategies to improve population diversity;

The next generation population generation module is used to return the individuals beyond the boundary to the boundary range based on the population obtained after mutation using the mutation strategy to form the next generation population;

The scheduling plan determination module is used to judge whether the algebra of the current population has reached the preset maximum number of iterations, if not, the next generation population will be regarded as the current population, and the operations from the location update module to the next generation population generation module will be repeated; , the calculation is stopped, and the optimal individual output corresponding to the global optimal position is used as the optimal scheme for the complementary scheduling of water-light energy.

In general, compared with the prior art, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

Using the invention to solve the problem of the combined system of the cascade hydropower station group and the photovoltaic power station, the principle is simple, the global search ability is enhanced, and the problem that the existing GSA algorithm is difficult to get rid of the local optimum is effectively avoided; the local search strategy is used to improve the convergence accuracy of the population, thereby enhancing the The algorithm development ability is improved; the adaptive mutation operation is used to improve the population diversity. In conclusion, the present invention has the advantages of strong optimization ability, not easy to converge prematurely, and easy to implement, etc., and can obtain satisfactory results in the coordinated scheduling of cascade hydropower and photovoltaic power station systems.

Description of drawings

1 is a schematic flowchart of a large-scale water-light energy complementary scheduling method provided by an embodiment of the present invention;

Fig. 2(a) is a box diagram of the coordinated dispatching result of the lower cascade power station and the photovoltaic power station in case 1-1 of the embodiment of the present invention;

Fig. 2(b) is a box diagram of the coordinated dispatching result of cascade power station and photovoltaic power station under Case 1-2 of the embodiment of the present invention;

Fig. 2(c) is a box diagram of the coordinated dispatching result of cascade power stations and photovoltaic power stations under Cases 1-3 of the embodiment of the present invention;

Fig. 2(d) is a box diagram of the coordinated dispatching result of the lower cascade power station and the photovoltaic power station in case 2-1 of the embodiment of the present invention;

Fig. 2(e) is a box diagram of the coordinated dispatching result of cascade power station and photovoltaic power station under Case 2-2 of the embodiment of the present invention;

Fig. 2(f) is a box diagram of the coordinated scheduling result of cascade power station and photovoltaic power station under Case 2-3 of the embodiment of the present invention;

Fig. 3 (a) is the scheduling process diagram of cascade power station and photovoltaic power station according to the method of the present invention under Case 1-1;

Fig. 3(b) is the scheduling process diagram of cascade power station and photovoltaic power station according to the method of the present invention under Case 1-2;

Fig. 3(c) is a diagram showing the scheduling process of cascade power plants and photovoltaic power plants according to the method of the present invention under Cases 1-3.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

In order to overcome that the standard GSA method is difficult to get rid of the local optimum and has weak development ability, the present invention proposes a large-scale water-light energy complementary scheduling method and system. On the basis of the standard GSA method, the local search strategy is used to improve the population convergence speed; the adaptive mutation operation is used to improve the population diversity. It has engineering practicality and feasibility.

1 is a schematic flowchart of a large-scale water-photovoltaic energy complementary scheduling method according to an embodiment of the present invention. The method involves the coordinated optimal scheduling of cascade reservoirs and photovoltaic power stations, and the specific steps include:

(1) Determine the constraints between the power stations, the power stations include hydropower stations and photovoltaic power stations. The parameters required for the calculation are selected and the outbound flow of each hydropower station at different time periods is used as a decision variable and coded. Let the number of iterations k=1, and initialize all individuals in the population in the search space.

可表示为

其中，N表示电站数目；T表示时段数目；For any individual in the k-th generation population

can be expressed as

Among them, N represents the number of power stations; T represents the number of time periods;

中第n个水电站在第t个时段的出库流量初始化方式为：

其中，

为

中第n个水电站在第t个时段的出库流量，

为第n个水电站在第t个时段的发电流量下限，rand(0,1)为[0,1]区间均匀分布的随机数，k表示迭代次数，

具有N*T个维度。where the k-th iteration is the i-th individual

The initialization method of the outbound flow of the nth hydropower station in the tth period is:

in,

for

The outbound flow of the nth hydropower station in the tth period,

is the lower limit of the power generation flow of the nth hydropower station in the tth period, rand(0,1) is a random number uniformly distributed in the interval [0,1], k represents the number of iterations,

Has N*T dimensions.

(2) Calculate the fitness of all individuals in the current population, then

The fitness F[X _i (k)] of the i-th individual X _i (k) in the k-th generation is:

L_t为电力系统第t个时段的负荷需求；P_s,m,t为在概率ρ_s下第m个光伏电站在第t个时段的出力；P_n,t为第n个水电站在第t个时段的出力；M是光伏电站的总数目；N为水电站总数目；T为总时段数目；g_a[X_i(k)]和c_a分别为第a个不等式约束的约束违背值和惩罚系数；e_b[X_i(k)]和c_b分别为第b个等式的约束违背值和惩罚系数；A和B分别为不等式约束和等式约束的数目。Among them, ρ _s is the probability of occurrence of the s-th photovoltaic power station; S is the total probability number,

L _t is the load demand of the power system in the t-th period; P _s,m,t is the output of the m-th photovoltaic power station in the t-th period under the probability ρ _s ; P _n,t is the n-th hydropower station in the t-th period The output of each period; M is the total number of photovoltaic power stations; N is the total number of hydropower stations; T is the total number of periods; g _a [X _i (k)] and c _a are the constraint violation value and penalty of the a-th inequality constraint, respectively coefficients; e _b [X _i (k)] and c _b are the constraint violation value and penalty coefficient of the b-th equation, respectively; A and B are the number of inequality constraints and equality constraints, respectively.

(3) Update the global optimal position of the population and the individual optimal position according to the fitness.

Wherein, in the first iteration of the population, the optimal position of the individual refers to the position of the individual in the first iteration of the population, and in the second and previous iterations, the optimal position of the individual refers to the position of the individual in the population The position in this iteration is a better position than the position in the previous iteration; the global optimal position refers to the position of the optimal individual of the population in this iteration.

更新当前种群中的全局最优位置，由

更新当前种群中的个体最优位置；其中，gBest(k)为种群第k次迭代的全局最优位置，gBest(k)＝{gBest^d(k),d＝1,2...,D}，gBest^d(k)为种群第k次迭代在第d维度的全局最优位置，gBest(k-1)种群第k-1次迭代的全局最优位置，

为第k次迭代第i个个体第d维度的个体最优位置，D为第i个个体的最大维度，pBest_i(k)为种群第k次迭代第i个个体的个体最优位置，pBest_i(k-1)为种群第k-1次迭代第i个个体的个体最优位置。Specifically, by

Update the global optimal position in the current population, given by

Update the individual optimal position in the current population; where, gBest(k) is the global optimal position of the k-th iteration of the population, gBest(k)={gBest ^d (k),d=1,2...,D }, gBest ^d (k) is the global optimal position of the k-th iteration of the population in the d-th dimension, and the global optimal position of the k-1 iteration of the gBest(k-1) population,

is the individual optimal position of the i-th individual in the k-th iteration of the d-th dimension, D is the maximum dimension of the i-th individual, pBest _i (k) is the individual optimal position of the i-th individual in the k-th iteration of the population, pBest _i (k-1) is the individual optimal position of the i-th individual in the k-1 iteration of the population.

(4)

更新种群所有个体位置，式中，Depend on

Update the positions of all individuals in the population, where,

R _ij (k)=||x _i (k)-x _j (k)||

和

分别为第k代种群和第k+1代种群第i个个体在第d维的位置，

为第k代种群第i个个体在第j维的位置；种群第k次迭代第i个个体的位置

种群第k+1次迭代第i个个体的位置

和

分别为第k次迭代和第k+1次迭代第i个个体在第d维的速度，

为第k次迭代第i个个体在第d维上的加速度，rand_j和rand_i是[0,1]之间均匀分布的随机数；Kbest为前K个具有更好适应度的个体；

为第k次迭代种群中所有个体最差适应度；

为第k次迭代种群中所有个体最优适应度；

为第i个个体对第j个个体在第d维上的作用力；M_pi(k)为第i个个体的被动质量，M_aj(k)为第j个个体的主动质量，R_ij(k)为第i个个体和第j个个体的欧氏距离，G₀为万有引力常数的初始值，G(k)是第k次迭代的万有引力常数值，α为衰减系数，ε为非常小的常数值。in,

and

are the positions of the i-th individual in the k-th generation population and the k+1-th generation population in the d-dimension, respectively,

is the position of the i-th individual in the k-th generation population in the j-th dimension; the position of the i-th individual in the k-th iteration of the population

The position of the ith individual in the k+1th iteration of the population

and

are the velocity of the i-th individual in the d-th dimension in the k-th iteration and the k+1-th iteration, respectively,

is the acceleration of the i-th individual in the k-th iteration on the d-th dimension, rand _j and rand _i are random numbers uniformly distributed between [0, 1]; Kbest is the first K individuals with better fitness;

is the worst fitness of all individuals in the k-th iteration population;

is the optimal fitness of all individuals in the k-th iteration population;

is the force of the i-th individual on the j-th individual on the d-th dimension; M _pi (k) is the passive mass of the i-th individual, M _aj (k) is the active mass of the j-th individual, and R _ij ( k) is the Euclidean distance between the i-th individual and the j-th individual, G ₀ is the initial value of the gravitational constant, G(k) is the gravitational constant value of the k-th iteration, α is the decay coefficient, and ε is very small constant value.

(5) Use the local search strategy to improve the population convergence speed, the expression is:

c ₂ =1-c ₁

为第k次迭代第i个个体在第d维度的局部搜索位置。δ^d为搜索空间中第d维的中值。

为第i个个体在第d维度的对立因子；c₁和c₂为学习因子；

为搜索空间第d维度上限值；x ^d为搜索空间第d维度下限值；

为预设的最大迭代次数。In the formula: r ₃ and r ₄ are random numbers uniformly distributed in the [0,1] interval.

is the local search position of the i-th individual in the d-th dimension for the k-th iteration. δ ^d is the median of the d-th dimension in the search space.

is the opposite factor of the i-th individual in the d-th dimension; c ₁ and c ₂ are learning factors;

is the upper limit value of the d-th dimension of the search space; x ^d is the lower value of the d-th dimension of the search space;

is the preset maximum number of iterations.

为第k次迭代第i个个体在第d维度的变异位置，α为种群中随机选择的个体下标，

表示第k次迭代第α个个体第d维度的位置；φ为[-0.5,0.5]区间均匀分布的随机数；Elite是从当前种群中得到的前三个最优个体位置集合，

为第k次迭代第β个个体第d维度的个体位置，β是从Elite随机选择的个体下标；(5) Use the adaptive mutation strategy to improve the diversity of the population, sort the positions of the individuals in the current population according to the fitness value, the first a (a < m) individuals directly enter the next iteration of the population, and the remaining ma individuals use the automatic The adaptive mutation operation generates mutant individuals and the first a (a<m) individuals as the population in the next iteration; among them, the adaptive mutation method is:

is the variation position of the i-th individual in the d-th dimension in the k-th iteration, α is the subscript of the randomly selected individual in the population,

Represents the position of the d-th dimension of the α-th individual in the k-th iteration; φ is a random number uniformly distributed in the interval [-0.5, 0.5]; Elite is the set of the first three optimal individual positions obtained from the current population,

is the individual position of the d-th dimension of the β-th individual in the k-th iteration, and β is the individual subscript randomly selected from Elite;

(6) Return the individuals beyond the boundary to the boundary range; the corresponding formula is:

In the formula: r ₁ is a random number uniformly distributed in the interval [0,1]. If the corrected individual is still outside the boundary, it is randomly generated in the search space of the extra dimension.

则返回步骤(2)，否则停止计算，将全局最优位置对应的个体输出作为梯级水库群与光伏电站系统协同调度的最优方案。

为预设的最大迭代次数。(7) Let k=k+1. like

Then go back to step (2), otherwise stop the calculation, and take the individual output corresponding to the global optimal position as the optimal solution for coordinated scheduling between cascade reservoir groups and photovoltaic power station systems.

is the preset maximum number of iterations.

The present invention will be further described below with reference to the accompanying drawings and embodiments.

各约束破坏惩罚系数均设定为1e3。为验证本发明实用性，将本发明与粒子群算法(PSO)、差分进化算法(DE)、正余弦算法(SCA)、灰狼优化算法(GWO)和引力搜索算法(GSA)进行对比。Taking the five power stations of Hongjiadu, Dongfeng, Suofengying, Wujiangdu and Goupitan and five photovoltaic power stations on the main stream of the Wujiang River as the implementation objects of the present invention, the corresponding parameters are set as: the population size is 30,

The penalty coefficient of each constraint destruction is set to 1e3. To verify the practicability of the present invention, the present invention is compared with particle swarm algorithm (PSO), differential evolution algorithm (DE), sine cosine algorithm (SCA), gray wolf optimization algorithm (GWO) and gravity search algorithm (GSA).

Four kinds of grid load requirements and three light energy inputs are selected as the implementation conditions. Table 1 shows the statistical results of 10 random runs under the six cases, and lists the maximum value, average value, worst value, standard deviation and range statistical results.

It can be seen from Table 1 that for all the listed index values, the performance of the method of the present invention is better than other methods. For example, compared to DE and PSO, the method of the present invention can achieve about 99.5% and 99.3% improvement in the target value range, respectively. Therefore, the method of the present invention can achieve a good trade-off between the development and exploration capabilities of the swarm.

Table 1

Figures 2(a) to 2(f) show the optimal solution distribution diagrams of the method of the present invention and the other five methods in six different cases. It can be seen from the figures that the method of the present invention has The best solution had a smaller proportional distribution and achieved the best performance among the five statistical metrics (maximum, mean, second or third quartile, median, and minimum) . Therefore, in this case, the robustness of the method of the present invention is fully demonstrated.

Figures 3(a) to 3(c) show the scheduling process diagrams of cascade power plants and photovoltaic power plants according to the method of the present invention in three cases. It can be seen from the figure that the method of the present invention can obtain a relatively stable residual load distribution curve. Therefore, it is explained that the solution provided by the method of the present invention is superior to the existing solution.

In another embodiment of the present invention, a large-scale water-light energy complementary scheduling system is also provided, including:

The initial population generation module is used to determine the constraints between the power stations, including hydropower stations and photovoltaic power stations. The outbound flow of each hydropower station at different times is used as a decision variable and coded, and then randomly selected in the search space according to the decision variable. Generate an initial population and take the initial population as the current population, where any individual in the population represents a complementary scheduling scheme of water-light energy;

The position update module is used to obtain the fitness of all individuals in the current population, update the individual optimal position and the global optimal position in the current population by the fitness value of each individual, and then update the positions of all individuals in the population; among them, In the first iteration of the population, the optimal position of the individual refers to the position of the individual in the first iteration of the population. In the second and previous iterations, the optimal position of the individual refers to the position of the individual in the current iteration of the population. The better position between the position in the iteration and the position in the previous iteration; the global optimal position refers to the position of the optimal individual in the current iteration of the population;

The local search module is used to perform a local search on the population by using a local search strategy based on the population obtained after updating the positions of all individuals in the population to improve the population convergence speed;

Adaptive mutation module, which is used to screen populations using adaptive mutation strategies to improve population diversity;

The next generation population generation module is used to return the individuals beyond the boundary to the boundary range based on the population obtained after mutation using the mutation strategy to form the next generation population;

The scheduling plan determination module is used to judge whether the algebra of the current population has reached the preset maximum number of iterations, if not, the next generation population will be regarded as the current population, and the operations from the location update module to the next generation population generation module will be repeated; , the calculation is stopped, and the optimal individual output corresponding to the global optimal position is used as the optimal scheme for the complementary scheduling of water-light energy.

For the specific implementation of each module, reference may be made to the description in the method embodiment, which will not be repeated in the embodiment of the present invention.

Those skilled in the art can easily understand that the above are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, etc., All should be included within the protection scope of the present invention.

Claims (9)

1. a large-scale water-light energy complementary scheduling method, is characterized in that, comprises the steps: (1) Determine the constraints between the power stations, the power stations include hydropower stations and photovoltaic power stations, take the outbound flow of each hydropower station at different times as a decision variable and encode it, and then randomly generate an initial population in the search space according to the decision variable, Taking the initial population as the current population, any individual in the population represents a complementary scheduling scheme of water-light energy; (2) Obtain the fitness of all individuals in the current population, update the individual optimal position and the global optimal position in the current population by the fitness value of each individual, and then update the positions of all individuals in the population; Wherein, in the first iteration of the population, the optimal position of the individual refers to the position of the individual in the first iteration of the population, and in the second and previous iterations, the optimal position of the individual refers to the position of the individual in the population The position in this iteration and the position in the previous iteration are the better positions; the global optimal position refers to the position of the optimal individual in this iteration of the population; (3) Based on the population obtained after updating all the individual positions of the population, a local search strategy is used to perform a local search on the population to improve the population convergence speed; (4) Use adaptive mutation strategy to screen populations to improve population diversity; (5) Based on the population obtained by using the mutation strategy, return the individuals beyond the boundary to the boundary range to form the next generation population; (6) Judging whether the algebra of the current population has reached the preset maximum number of iterations, if not, the next generation of the population will be used as the current population, and return to step (2). The optimal individual output is the optimal scheme for the complementary scheduling of water and light energy. 2. The method according to claim 1, characterized in that, for any individual X _i (k) in the k-th generation population, it can be expressed as

Among them, N represents the number of power stations; T represents the number of time periods;

is the outbound flow of the nth hydropower station in the tth period in Xi ( _k ),

is the lower limit of the power generation flow of the nth hydropower station in the tth period, rand(0,1) is a random number uniformly distributed in the [0,1] interval, k represents the number of iterations, and X _i (k) has N*T dimensions . 3. The method according to claim 2, wherein the fitness F[X _i (k)] of the i-th individual X _i (k) of the k-th generation is:

Among them, ρ _s is the probability of occurrence of the s-th photovoltaic power station; S is the total probability number,

L _t is the load demand of the power system in the t-th period; P _s,m,t is the output of the m-th photovoltaic power station in the t-th period under the probability ρ _s ; P _n,t is the n-th hydropower station in the t-th period The output of each period; M is the total number of photovoltaic power stations; N is the total number of hydropower stations; T is the total number of periods; g _a [X _i (k)] and c _a are the constraint violation value and penalty of the a-th inequality constraint, respectively coefficients; e _b [X _i (k)] and c _b are the constraint violation value and penalty coefficient of the b-th equation, respectively; A and B are the number of inequality constraints and equality constraints, respectively. 4. The method of claim 3, wherein the

Update the global optimal position in the current population, given by

Update the individual optimal position in the current population; where, gBest(k) is the global optimal position of the k-th iteration of the population, gBest(k)={gBest ^d (k),d=1,2...,D }, gBest ^d (k) is the global optimal position of the k-th iteration of the population in the d-th dimension, and the global optimal position of the k-1 iteration of the gBest(k-1) population,

is the individual optimal position of the i-th individual in the k-th iteration of the d-th dimension, D is the maximum dimension of the i-th individual, pBest _i (k) is the individual optimal position of the i-th individual in the k-th iteration of the population, pBest _i (k-1) is the individual optimal position of the i-th individual in the k-1 iteration of the population. 5. The method of claim 4, wherein the

Update the positions of all individuals in the population, where,

R _ij (k)=||x _i (k)-x _j (k)||

in,

and

are the positions of the i-th individual in the k-th generation population and the k+1-th generation population in the d-dimension, respectively,

is the position of the i-th individual in the k-th generation population in the j-th dimension; the position of the i-th individual in the k-th iteration of the population

The position of the ith individual in the k+1th iteration of the population

and

are the velocity of the i-th individual in the d-th dimension in the k-th iteration and the k+1-th iteration, respectively,

is the acceleration of the i-th individual in the k-th iteration on the d-th dimension, rand _j and rand _i are random numbers uniformly distributed between [0, 1]; Kbest is the first K individuals with better fitness;

is the worst fitness of all individuals in the k-th iteration population;

is the optimal fitness of all individuals in the k-th iteration population;

is the force of the i-th individual on the j-th individual on the d-th dimension; M _pi (k) is the passive mass of the i-th individual, M _aj (k) is the active mass of the j-th individual, and R _ij ( k) is the Euclidean distance between the i-th individual and the j-th individual, G ₀ is the initial value of the gravitational constant, G(k) is the gravitational constant value of the k-th iteration, α is the decay coefficient, and ε is the constant value. 6. The method of claim 5, wherein a local search is performed on the population using the following formula:

c ₂ =1-c ₁ In the formula,

is the local search position of the i-th individual in the d-th dimension of the k-th iteration; r ₃ and r ₄ are random numbers uniformly distributed in the [0,1] interval; δ ^d is the median of the d-th dimension in the search space;

is the opposition factor of the i-th individual in the d-th dimension in the k-th iteration; c ₁ and c ₂ are learning factors;

is the upper limit value of the d-th dimension of the search space; x ^d is the lower value of the d-th dimension of the search space;

is the preset maximum number of iterations. 7. The method of claim 6, wherein step (4) comprises: The individual positions in the current population are sorted according to the fitness value, the first a (a < m) individuals directly enter the next iteration of the population, and the adaptive mutation operation is used for the remaining ma individuals to generate mutant individuals and the first a (a < m ) ) individuals as the population in the next iteration; among them, the adaptive mutation method is:

is the variation position of the i-th individual in the d-th dimension in the k-th iteration, α is the subscript of the randomly selected individual in the population,

Represents the position of the d-th dimension of the α-th individual in the k-th iteration; φ is a random number uniformly distributed in the interval [-0.5, 0.5]; Elite is the set of the first three optimal individual positions obtained from the current population,

is the individual position of the d-th dimension of the β-th individual in the k-th iteration, and β is the randomly selected individual subscript in Elite. 8. The method of claim 7, wherein the

Return individuals beyond the boundary to within the boundary, where,

r ₁ is a random number uniformly distributed in the interval [0,1]. 9. A large-scale water-light energy complementary scheduling system, characterized in that, comprising: The initial population generation module is used to determine the constraints between the power stations, including hydropower stations and photovoltaic power stations. The outbound flow of each hydropower station at different times is used as a decision variable and coded, and then randomly selected in the search space according to the decision variable. Generate an initial population and take the initial population as the current population, where any individual in the population represents a complementary scheduling scheme of water-light energy; The position update module is used to obtain the fitness of all individuals in the current population, update the individual optimal position and the global optimal position in the current population by the fitness value of each individual, and then update the positions of all individuals in the population; among them, In the first iteration of the population, the optimal position of the individual refers to the position of the individual in the first iteration of the population. In the second and previous iterations, the optimal position of the individual refers to the position of the individual in the current iteration of the population. The better position between the position in the iteration and the position in the previous iteration; the global optimal position refers to the position of the optimal individual in the current iteration of the population; The local search module is used to perform a local search on the population by using a local search strategy based on the population obtained after updating the positions of all individuals in the population to improve the population convergence speed; Adaptive mutation module, which is used to screen populations using adaptive mutation strategies to improve population diversity; The next generation population generation module is used to return the individuals beyond the boundary to the boundary range based on the population obtained after mutation using the mutation strategy to form the next generation population; The scheduling plan determination module is used to judge whether the algebra of the current population has reached the preset maximum number of iterations, if not, the next generation population will be regarded as the current population, and the operations from the location update module to the next generation population generation module will be repeated; , the calculation is stopped, and the optimal individual output corresponding to the global optimal position is used as the optimal scheme for the complementary scheduling of water-light energy.

Images