Disclosure of Invention
The invention aims to at least solve one of the technical problems in the prior art, and provides a double-layer intelligent optimization method and device for a wind farm, which are influenced by environment.
In one aspect of the invention, a dual-layer intelligent optimization method for a wind farm with environmental influence is provided, which comprises the following specific steps:
Acquiring power of a wind turbine generator in a wind power plant;
setting a wake flow model of the wind turbine generator;
calculating the output power of the wind power plant;
setting a double-layer structure optimization model of the wind power plant;
optimizing the double-layer structure optimization model based on suburban wolf optimization algorithm.
Optionally, the obtaining the power of the wind turbine in the wind farm includes:
The Logistic function is used for representing the power curve of the wind turbine, and the method is as follows:
Wherein,Pip and vip are the power and wind speed at the inflection point of the wind power curve, Pwtr represents the rated active power of the wind turbine, and Pwt (v) represents the active power generated by the wind turbine at wind speed v.
Optionally, setting a wake model of the wind turbine generator, and specifically the following relation:
Vx,r is the wind speed at the distance x and radius r from the wind turbine generator, sigma represents the characteristic width of wake flow, K (x) is the undetermined coefficient of x, and CT is the thrust coefficient;
v0 denotes wind speed of the wind turbine, rr denotes rotor radius of the wind turbine, z0 denotes surface roughness length, and zh denotes hub height of the wind turbine.
Optionally, the calculating the output power of the wind farm is expressed as follows:
Wherein, Pwf represents the total output power of the wind farm, Pwt(vi) represents the output power of the ith wind turbine, and Nt represents the total number of wind turbines in the wind farm;
The total injected reactive power Qwf of the wind farm is calculated according to the following:
Qwf＝Pwf tanφ
where φ is the power factor angle of the wind farm.
Optionally, the setting a double-layer structure optimization model of the wind farm includes:
Setting a first layer of model to optimize capacity and local of the wind farm according to the wind energy comprehensive power generation cost;
a second layer model is set up to plan the generation plan of the generator in the grid.
Optionally, the first layer model expression is as follows:
Wherein,A matrix representing the coordinates of the wind turbine,F1 is the unit wind power generation cost, and consists of daily equivalent investment Cwt, power grid running cost Co and environment penalty cost Ce; AndThe upper limit and the lower limit of the x-axis coordinate of the wind turbine generator are represented,AndThe upper limit and the lower limit of the y-axis coordinate of the wind turbine generator are represented,AndRepresenting the coordinates of the x-axis and the y-axis of the ith wind turbine,AndRepresenting the x-axis and y-axis coordinates of the jth wind turbine, Cwf representing the capacity of the wind farm,AndRepresenting the upper and lower limits of the capacity of the wind farm, SPL representing the sound pressure level, SPLlim representing the lower limit of the boost level, Nt representing the number of wind turbines in the wind farm,Represents the mth (n) bus voltage at the t-th hour,AndThe upper and lower limits of the (n) th bus voltage at the t-th hour are shown,Represents the voltage phase angle difference between the mth and nth buses at t hours, Gmn represents the conductivity of line mn, Bmn represents the inductance of line mn,Representing the reactive output of the mth generator at time t,AndThe upper limit and the lower limit of the reactive output of the mth generator at the t hour are represented, and Qld,m(j) represents the reactive load value of the mth (j-th) bus at the t hour; Representing the reactive output of the wind farm at the t-th hour; representing the reactive power flow of the nth line at the t-th hour and the upper limit and the lower limit of the reactive power flow,Active power flow of the nth line at the t-th hour and upper and lower limits of the active power flow are represented;
minimum number of WTsAnd maximum number ofCalculated from the following
Wherein round is a rounding function; is rated active power of the wind turbine generator; Is the upper and lower limits of the output power of the wind farm.
Optionally, the constraint in the first layer model is formed by two parts, specifically including:
Boundary of wind farm planning area, minimum allowable distance between wind turbines, lower and upper limits of wind farm capacity and SPL limit of wind farm at each Nr receptor, and
Active and reactive power balance, bus voltage amplitude and phase angle constraints, active and reactive power output constraints of conventional generators, and bus active and reactive power flow constraints.
Optionally, the second layer model expression is as follows:
Wherein,Expressed as a quadratic function:
ao,i,a1,i,a2,i (co/MW) is the cost factor of the i-th generator;
Is the penalty cost for pollutant emissions at time t, expressed asWherein ρi (co/MW) is the unit penalty price;
is the amount of pollutant released by the ith generator at time t, and is estimated by the following formula:
where bo,i,b1,i,b2,i (lb/MW) is the emission coefficient of the ith generator.
Optionally, the optimizing the double-layer structure optimizing model based on suburban wolf optimizing algorithm includes:
randomly initializing suburban wolves and randomly grouping;
Growing suburban wolves in the group;
Producing and dying suburban wolves;
suburban wolves are driven off and received.
The invention further provides a double-layer intelligent optimization device for the wind power plant with environmental influence, which comprises an acquisition module, a wake flow model module, a calculation module, a double-layer structure optimization model module and an optimization module, wherein,
The acquisition module is used for acquiring the power of a wind turbine generator in a wind power plant;
The wake model module is used for setting a wake model of the wind turbine generator;
The calculation module is used for calculating the output power of the wind farm;
the double-layer structure optimization model is used for setting the double-layer structure optimization model of the wind power plant;
The optimizing module is used for optimizing the double-layer structure optimizing model based on suburban wolf optimizing algorithm.
The invention provides a double-layer intelligent optimization method for a wind power plant, which comprises the specific steps of obtaining power of a wind power plant in the wind power plant, setting a wake model of the wind power plant, calculating output power of the wind power plant, setting a double-layer structure optimization model of the wind power plant, and optimizing the double-layer structure optimization model based on suburban wolf optimization algorithm. Compared with a single-layer ratio optimization model, the double-layer optimization model provided by the invention can reduce the adverse effect of the wind power plant on the environment in the planning stage.
Detailed Description
The present invention will be described in further detail below with reference to the drawings and detailed description for the purpose of better understanding of the technical solution of the present invention to those skilled in the art. It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which can be made by a person skilled in the art without creative efforts, based on the described embodiments of the present invention belong to the protection scope of the present invention.
As shown in FIG. 1, in one aspect of the present invention, a method S100 for optimizing wind farm double-layer intelligence with environmental impact is provided, comprising the following specific steps S110 to S150:
S110, acquiring power of a wind turbine in the wind power plant.
Specifically, in this embodiment, a Logistic function is used to represent a power curve of a wind turbine, which is specifically as follows:
Wherein,Pip and vip are the power and wind speed at the inflection point of the wind power curve, Pwtr represents the rated active power of the wind turbine, and Pwt (v) represents the active power generated by the wind turbine at wind speed v.
S120, setting a wake flow model of the wind turbine generator, as shown in FIG. 2.
Gaussian wake model as shown in fig. 2, the mathematical model is expressed as follows:
Vx,r is the wind speed at the distance x and radius r from the wind turbine generator, sigma represents the characteristic width of wake flow, K (x) is the undetermined coefficient of x, and CT is the thrust coefficient;
v0 denotes wind speed of the wind turbine, rr denotes rotor radius of the wind turbine, z0 denotes surface roughness length, and zh denotes hub height of the wind turbine.
In this model, the radius of the wake is infinitely extended due to the nature of the exponential function. Thus, with respect to a given wind turbine, all wind turbines located downstream are in the wake, and no wind turbine will be partially in the wake of another wind turbine, regardless of the distance between them.
S130, calculating the output power of the wind power plant.
For a given inflow vector v= (v 0, θ), wind speed v0 (m/s), wind direction θ (°), the coordinates of the wind turbine should be transformed from the original layout [ X, Y ] to the new layout [ X ', Y' ] according to the following formula.
When determining the relative distance between wind turbines in the new coordinates, the wind speed vi of the i-th wind turbine (i=1, 2,..nt) can be calculated by:
Wherein, tij takes the value of 1. The speed deficit depends only on the location and direction of each wind turbine within the wind farm. Deltavij is the wake loss of the ith wind turbine caused by the jth wind turbine.
The total power output Pwf of the wind farm is represented as follows:
wherein, Pwf represents the total output power of the wind farm, Pwt(vi) represents the output power of the ith wind turbine, and Nt represents the total number of wind turbines in the wind farm.
The wind farm can be considered a PQ load bus when grid connected. The total injected active power Pwf and the reactive power Qwf of the wind farm are calculated according to the sum respectively.
Qwf＝Pwf tanφ (6)
Where φ is the power factor angle of the wind farm.
S140, setting a double-layer structure optimization model of the wind power plant.
When dealing with complex optimization problems of multi-level variables, double-layer optimization has two advantages over single-layer optimization. First, it significantly reduces the computational scale and reduces the complexity of the developed model. Furthermore, it greatly reduces the load on the communication network. The interaction mechanism between the first layer and the second layer model is as follows. The first layer establishes an overall optimization scheme of wind farm scale and wind farm microscopic site selection, and aims to minimize the comprehensive power generation cost of wind power and determine the optimal wind farm capacity, wind farm number and layout. The second tier is nested in the first tier, and by purchasing an optimal power generation plan for other generators in the grid, minimal operating and environmental penalty costs are generated for the first tier model. The power generation plan obtained from the second layer model is influenced by the result of the first layer model, and comprises the integration and arrangement scheme of the wind turbine generator. Furthermore, the optimization result of the first layer model is related to the power generation amount and the environmental penalty cost obtained from the second layer model.
First, a first layer model is set, and in order to optimize the capacity and layout of a wind farm with minimum wind energy comprehensive power generation cost, a first layer problem is expressed as follows:
Wherein,A matrix representing the coordinates of the wind turbine,F1 is the unit wind power generation cost, and consists of daily equivalent investment Cwt, power grid operation cost Co and environment penalty cost Ce.AndThe upper limit and the lower limit of the x-axis coordinate of the wind turbine generator are represented,AndThe upper limit and the lower limit of the y-axis coordinate of the wind turbine generator are represented,AndRepresenting the coordinates of the x-axis and the y-axis of the ith wind turbine,AndRepresenting the x-axis and y-axis coordinates of the jth wind turbine, Cwf representing the capacity of the wind farm,AndRepresenting the upper and lower limits of the capacity of the wind farm, SPL representing the sound pressure level, SPLlim representing the lower limit of the boost level, Nt representing the number of wind turbines in the wind farm,Represents the mth (n) bus voltage at the t-th hour,AndThe upper and lower limits of the (n) th bus voltage at the t-th hour are shown,Representing the voltage phase angle difference between the mth and nth buses at the t-th hour. Gmn denotes the conductivity of line mn, Bmn denotes the inductance of line mn,Representing the reactive output of the mth generator at time t,AndThe upper and lower limits of the reactive output of the mth generator at the t-th hour are indicated, and Qld,m(j) represents the reactive load value of the mth (j-th) bus at the t-th hour.Representing the reactive output of the wind farm at time t.Representing the reactive power flow of the nth line at the t-th hour and the upper limit and the lower limit of the reactive power flow,Active power flow of the nth line at the t-th hour and upper and lower limits of the active power flow are represented;
minimum number of WTsAnd maximum number ofCalculated from the following formula:
Wherein round is the rounding function,Is rated active power of the wind turbine generator.Is the upper and lower limits of the output power of the wind farm.
The constraints in the first layer model consist of two parts.
Cstr.1 relates to 1) boundaries of the wind farm planning area, 2) minimum allowable distances between wind turbines, 3) lower and upper limits of wind farm capacity, and 4) SPL limits of the wind farm at each Nr receptor, limited to less than 40dB (a).
SPL (T, Rk) is the equivalent continuous downwind octave band SPL of the whole wind farm, calculated as follows:
Wherein,Is an equivalent continuous downwind octave SPL at each receiver position calculated for each sound source of each of eight octaves j of nominal intermediate frequency (63 Hz-8 kHz).
The SPL for each octave (Lf) can be expressed as follows:
Lf＝Lw+Dc-Aw(f) (10)
where Lw is the octave band acoustic power from the noise source, Dc is the directivity correction for the non-omni-directional source (Dc is set to 0 in this study), and aw (f) is the octave band attenuation.
Cstr.2 is the set of power flow constraints at hour t, including 1) active and reactive power balance, 2) bus voltage magnitude and phase angle constraints, 3) active and reactive power output constraints of conventional generators, 4) bus active and reactive power flow constraints.
Cwt represents the daily equivalent investment cost of the wind turbine, and is expressed as follows:
Wherein, Co is the total running cost, including the daily running and maintenance costs of the wind farm, the power generation costs of other generators in the grid and the grid loss costs, expressed as follows:
Wherein the method comprises the steps ofIs the daily operational cost of the wind farm, which can be calculated by the following formula:
Where OPEXunit (co/MW/yr) is the annual operation and maintenance cost of a wind farm.Can be obtained from a second layer model.Is the network loss cost calculated by the following equation.
Wherein the method comprises the steps ofThe network loss at time t can be calculated by the following formula:
Wherein,Can be obtained from a second layer model.Can be obtained from a typical daily curve of wind speed and direction.
Ce is a pollutant emission penalty cost, which can be expressed by the following formula:
Co and Ce are both determined by the second tier model and are sensitive to the power output of the wind farm. The first and second layer models interact with each other through Co and Ce.
Ewf is the expected daily energy yield of the wind farm, which can be estimated by:
Wherein JPDF (v, θ) represents the joint probability distribution function of wind speed and direction. vin and vout are the cut-in and cut-out wind speeds of the wind turbines.
Second, second layer model
The purpose of the second layer model is to plan the generation plans of other generators in the grid. It is expressed as an optimal economic dispatch model:
Wherein,Expressed as a quadratic function:
ao,i,a1,i,a2,i (co/MW) is the cost factor of the i-th generator, which is determined by the properties of the generator.Is the penalty cost for pollutant emissions at hour t, expressed as:
wherein ρi (co/MW) is the unit penalty price;
is the amount of pollutant released by the ith generator at time t, and is estimated by the following formula:
where bo,i,b1,i,b2,i (lb/MW) is the emission coefficient of the ith generator.
These three constraints represent grid power balance, rotational reserve capacity limit, and generator output power limit, respectively. The optimal power generation schedule may be determined by the second tier model from which the components C0 and Ce in the first tier model F1 may be derived accordingly.
And S150, optimizing the double-layer structure optimization model based on suburban wolf optimization algorithm.
It should be noted that, suburban wolf optimization algorithm (coyote optimization algorithm, COA) was proposed in 2018, and is a novel intelligent optimization algorithm for simulating suburban wolf population living, growing, dying, being driven off and being accepted by groups, and the like. In the COA, each suburban wolf represents a candidate solution, each solution vector is composed of social state factors of the suburban wolf, the state factors comprise intrinsic factors and extrinsic factors of the suburban wolf, each state factor represents a decision variable, D state factors form a solution vector containing a decision variable, and each suburban wolf is measured by social adaptation capacity. The COA is mainly divided into four steps of randomly initializing suburban wolves and randomly grouping, growing suburban wolves in groups, generating and dying suburban wolves and driving suburban wolves away from groups and accepting suburban wolves. The optimization process of the algorithm is specifically as follows:
first, randomly initializing suburban wolves and randomly grouping
Setting parameters of suburban wolf group number Np, suburban wolf number Nc and maximum iteration number Ngen, randomly setting the following relation to the initial social state factors of each suburban wolf, calculating the social adaptability of each suburban wolf, and randomly grouping, wherein the specific relation is as follows:
socj＝lbj+rj*(ubj-lbj) (22)
Where lbj and ubj represent the lower and upper bounds, respectively, of the suburban wolf jth state factor, j=1, 2.
Second, suburban wolf growth
Determining optimal wolves alpha in a group, calculating a group cultural trend, randomly selecting suburban wolves, influencing the growth of suburban wolves by the four factors, and calculating the group cultural trend as follows:
cultj＝median(Aj) (23)
Wherein A is a matrix of Nc rows and D columns, representing Nc solution vectors, Aj represents the j-th column of the matrix A, and mean represents the fetch bits,
In the growth process of suburban wolves, firstly, calculating the difference delta1 between the optimal suburban wolves alpha in the group and suburban wolves randomly selected by one head in the group, wherein the difference delta2 between the cultural trend of the group and the other head of suburban wolves randomly selected by the other head in the group, and suburban wolves grow under the influence of delta1 and delta2, wherein the following relational expression is as follows:
δ1＝Lbest-socr1,δ2＝cult-socr2 (24)
new_socc＝socc+s1*δ1+s2*δ2 (25)
Wherein r1 and r2 represent different random suburban wolves, Lbest represents optimal suburban wolves alpha wolves in the group, s1 and s2 are random weights of delta1 and delta2 respectively, and s1 and s2 are random numbers uniformly distributed in [0,1 ].
After each suburban wolf in the group grows, the social adaptability is calculated by adopting the following relation, greedy selection is adopted, and the algorithm convergence speed is accelerated by keeping high-quality suburban wolves to participate in the growth of the rest suburban wolves in the group:
third, the suburban wolf is born and dead
It should be understood that two important evolutionary processes in nature are birth and death, and that the age of suburban wolves in COA is in years. After each group of suburban wolves grows, a new suburban wolf is produced. Birth and death of suburban wolves are shown in table 1 below.
TABLE 1 birth and death Process of suburban wolves
Specifically, the birth of a new suburban wolf is commonly affected by two randomly selected social conditions and social environments of the parent suburban wolf. The new suburban wolves are produced in the following manner:
Wherein cr1 and cr2 are two randomly different suburban wolves in the P-th group, j1 and j2 are two random dimensions of the newly-generated suburban wolves, Ps is a dispersion probability, Pa is a correlation probability, and the diversity of the newly-generated suburban wolves is affected by the dispersion and correlation probability. Rj is a random number of the decision variable in the value range, and rndj is a random number uniformly distributed in [0,1 ].
Ps=1/D
Pa＝(1-Ps)/2 (28)
Where ω represents suburban wolves with poorer adaptation than newborn wolves,The number of suburban wolves in the group and the age of the newly-grown suburban wolves was 0.
Fourth, suburban wolves are driven off and received
Suburban wolves were initially randomly assigned to groups, but suburban wolves sometimes left and added to other groups. The probability of suburban wolves being driven off and admitted by the group is denoted by Pe. This mechanism facilitates the exchange of information between groups of COAs, contributing to interactions between suburban wolves in the population.
After initialization and random grouping, the three steps of suburban wolf growth, suburban wolf death and suburban wolf driving and receiving are sequentially carried out. And outputting the optimal suburban wolf if the iteration termination condition is reached. Otherwise, jumping to the second step.
As shown in FIG. 3, in another aspect of the present invention, a dual-layer intelligent optimization device 200 for a wind farm with environmental impact is provided, which comprises an acquisition module 210, a wake model module 220, a calculation module 230, a dual-layer structure optimization model module 240 and an optimization module 250, wherein the acquisition module 210 is used for acquiring the power of a wind turbine in the wind farm, the wake model module 220 is used for setting the wake model of the wind turbine, the calculation module 230 is used for calculating the output power of the wind farm, the dual-layer structure optimization model 240 is used for setting the dual-layer structure optimization model of the wind farm, and the optimization module 250 is used for optimizing the dual-layer structure optimization model based on suburban wolf optimization algorithm.
It should be noted that, the method adopted by the apparatus of this embodiment is described with reference to the foregoing, and will not be described herein.
The wind farm double-layer intelligent optimization method for environmental influence is described in the following by using a specific embodiment:
1. Test system and data source
The wind farm to be optimized occupies a rectangular area, the basic information of which is given in table 2. The planned area of the wind farm is given in fig. 4, where red dots on the boundary represent noise receivers. During the wind farm layout planning phase, all homes in the vicinity are considered potential noise receivers. The receivers 1-8 marked with blue circles are located in eight typical directions (NW, N, NE, E, SE, S, SW, W).
TABLE 2 wind farm information
The parameters of the wind turbines installed in the wind farm are given in table 3. The power output and acoustic power emissivity curves of the wind turbine are shown in fig. 5.
TABLE 3 wind turbine parameters
The wind farm should be integrated in an IEEE 30 bus test system. The parameters of the generator in this system are given in table 4.
Table 4 genset parameters
Turbulent wind with an average wind speed of 10m/s and a turbulence intensity of 20% at zref = 62m was used. The parameters of WGA are set to the following formula:
Nwg＝120,Ng＝10,w＝0.7298,R＝1.5931,m＝0.85,Itmax=250 (30)
2. capacity and layout optimization of grid-connected wind farms with consideration of environmental impact
Firstly, the capacity and layout double-layer optimization model of the grid-connected wind power plant considering environmental influence is established, and the model is specifically expressed as follows:
1) First layer model
In order to optimize wind farm capacity and layout with minimal wind energy comprehensive generation costs, the first layer of problems is expressed as follows:
Wherein,A matrix representing the coordinates of the wind turbine,F1 is the unit wind power generation cost, and consists of daily equivalent investment Cwt, power grid operation cost Co and environment penalty cost Ce. The minimum and maximum number of WTs are calculated by
Where round is the rounding function.Is rated active power of the wind turbine generator.Is the upper and lower limits of the output power of the wind farm.
The constraints in the first layer model consist of two parts.
Cstr.1 relates to 1) boundaries of the wind farm planning area, 2) minimum allowable distances between wind turbines, 3) lower and upper limits of wind farm capacity, and 4) SPL limits of the wind farm at each Nr receptor, limited to less than 40dB (a).
SPL (T, Rk) is the equivalent continuous downwind octave band SPL of the whole wind farm, calculated as follows:
Wherein,Is an equivalent continuous downwind octave SPL at each receiver position calculated for each sound source of each of eight octaves j of nominal intermediate frequency (63 Hz-8 kHz).
The SPL for each octave (Lf) can be expressed as follows:
Lf＝Lw+Dc-Aw(f) (34)
where Lw is the octave band acoustic power from the noise source, Dc is the directivity correction for the non-omni-directional source (Dc is set to 0 in this study), and aw (f) is the octave band attenuation.
Cstr.2 is the set of power flow constraints at hour t, including 1) active and reactive power balance, 2) bus voltage magnitude and phase angle constraints, 3) active and reactive power output constraints of conventional generators, 4) bus active and reactive power flow constraints.
Cwt represents the daily equivalent investment cost of the wind turbine, and is expressed as follows:
Wherein, Co is the total running cost, including the daily running and maintenance costs of the wind farm, the power generation costs of other generators in the grid and the grid loss costs, expressed as follows:
Wherein the method comprises the steps ofIs the daily operational cost of the wind farm, which can be calculated by the following formula:
Where OPEXunit (co/MW/yr) is the annual operation and maintenance cost of a wind farm.Can be obtained from a second layer model.Is the network loss cost calculated by the following equation.
Wherein the method comprises the steps ofThe network loss at time t can be calculated by the following formula:
Wherein,Can be obtained from a second layer model.Can be obtained from a typical daily curve of wind speed and direction.
Ce is a pollutant emission penalty cost, which can be expressed by the following formula:
Co and Ce are both determined by the second tier model and are sensitive to the power output of the wind farm. The first and second layer models interact with each other through Co and Ce.
Ewf is the expected daily energy yield of the wind farm, which can be estimated by:
Wherein JPDF (v, θ) represents the joint probability distribution function of wind speed and direction. vin and vout are the cut-in and cut-out wind speeds of the wind turbines.
Second layer model
The purpose of the second layer model is to plan the generation plans of other generators in the grid. It is expressed as an optimal economic dispatch model:
Wherein,Expressed as a quadratic function:
ao,i,a1,i,a2,i (co/MW) is the cost factor of the i-th generator, which is determined by the properties of the generator.
Is the penalty cost for pollutant emissions at hour t, expressed as:
wherein ρi (co/MW) is the unit penalty price;
is the amount of pollutant released by the ith generator at time t, and is estimated by the following formula:
where bo,i,b1,i,b2,i (lb/MW) is the emission coefficient of the ith generator.
These three constraints represent grid power balance, rotational reserve capacity limit, and generator output power limit, respectively. The optimal power generation schedule may be determined by the second tier model from which the components C0 and Ce in the first tier model F1 may be derived accordingly.
Then optimizing the double-layer structure optimization model based on suburban wolf optimization algorithm, 1) initializing and randomly grouping
Setting parameters of suburban wolf group number Np, suburban wolf number Nc and maximum iteration number Ngen, randomly setting the following relation to the initial social state factors of each suburban wolf, calculating the social adaptability of each suburban wolf, and randomly grouping, wherein the specific relation is as follows:
socj＝lbj+rj*(ubj-lbj) (46)
Where lbj and ubj represent the lower and upper bounds, respectively, of the suburban wolf jth state factor, j=1, 2.
2) Suburban wolves grow in the group.
Determining optimal wolves alpha in a group, calculating a group cultural trend, randomly selecting suburban wolves, influencing the growth of suburban wolves by the four factors, and calculating the group cultural trend as follows:
cultj＝median(Aj) (47)
Wherein A is a matrix of Nc rows and D columns, representing Nc solution vectors, Aj represents the j-th column of the matrix A, and mean represents the fetch bits,
In the growth process of suburban wolves, firstly, calculating the difference delta1 between the optimal suburban wolves alpha in the group and suburban wolves randomly selected by one head in the group, wherein the difference delta2 between the cultural trend of the group and the other head of suburban wolves randomly selected by the other head in the group, and suburban wolves grow under the influence of delta1 and delta2, wherein the following relational expression is as follows:
δ1＝Lbest-socr1,δ2＝cult-socr2 (48)
new-socc＝socc+s1*δ1+s2*δ2 (49)
Wherein r1 and r2 represent different random suburban wolves, Lbest represents optimal suburban wolves alpha wolves in the group, s1 and s2 are random weights of delta1 and delta2 respectively, and s1 and s2 are random numbers uniformly distributed in [0,1 ];
After each suburban wolf in the group grows, the social adaptability is calculated by adopting the following relation, greedy selection is adopted, and the algorithm convergence speed is accelerated by keeping high-quality suburban wolves to participate in the growth of the rest suburban wolves in the group:
3) Suburb wolf life and death
Two important evolutionary processes in nature are birth and death, and the age of suburban wolves in COA is in years. After each group of suburban wolves grows, a new suburban wolf is produced. Birth and death of suburban wolves are shown in table 5 below:
Table 5 birth and death Process of suburban wolves
The birth of a new suburban wolf is commonly affected by the social condition and social environment of two randomly selected parent suburban wolves. The new suburban wolves are produced in the following manner:
Wherein cr1 and cr2 are two randomly different suburban wolves in the P-th group, j1 and j2 are two random dimensions of the newly-generated suburban wolves, Ps is a dispersion probability, Pa is a correlation probability, and the diversity of the newly-generated suburban wolves is affected by the dispersion and correlation probability. Rj is a random number of the decision variable in the value range, and rndj is a random number uniformly distributed in [0,1 ].
Ps=1/D
Pa＝(1-Ps)/2 (52)
Wherein ω represents suburban wolves with poorer adaptability than newly born wolves,The number of suburban wolves in the group and the age of the newly-grown suburban wolves was 0.
4) Suburban wolves are driven off and received
Suburban wolves were initially randomly assigned to groups, but suburban wolves sometimes left and added to other groups. The probability that suburban wolves are driven off and admitted by the group is denoted by Pe. This mechanism facilitates the exchange of information between groups of COAs, contributing to interactions between suburban wolves in the population.
After initialization and random grouping, the three steps of suburban wolf growth, suburban wolf death and suburban wolf driving and receiving are sequentially carried out. And outputting the optimal suburban wolf if the iteration termination condition is reached. Otherwise, jumping to the second step.
Finally, the suburban wolf optimization algorithm obtains the optimal layout and capacity of the wind farm.
Compared with the prior art, the method has the advantages that capacity and layout optimization of the grid-connected wind power plant with environmental influence are considered, and adverse effects of the wind power plant on the environment can be reduced in a planning stage based on a double-layer optimization model compared with a single-layer ratio optimization model.
It is to be understood that the above embodiments are merely illustrative of the application of the principles of the present invention, but not in limitation thereof. Various modifications and improvements may be made by those skilled in the art without departing from the spirit and substance of the invention, and are also considered to be within the scope of the invention.