Reactive power optimization system and method of power grid based on the double-fish-swarm algorithm

• US 10,439,398 B2
• Filed: 06/16/2017
• Issued: 10/08/2019
• Est. Priority Date: 12/07/2016
• Status: Active Grant

• Current Assignee: Northeastern University
• Original Assignee: Northeastern University
• Inventors: Zhang, Huaguang, Yang, Jun, Sun, Qiuye, Wu, Feiye, Liu, Xinrui, Yang, Dongsheng, Wang, Zhiliang, Feng, Jian, Huang, Bonan, Luo, Yanhong, Hui, Guotao, Liu, Zhenwei
• Primary Examiner(s): Karim, Ziaul
• Application Number:
  US15/577,127
• Publication Number:
  US 20180309294A1
• Time in Patent Office: 844 Days
• Field of Search: 700 89
• CPC Class Codes: G05F 1/14 using tap transformers or tap changing inductors as final control devices
• CPC Class Codes: H02J 2203/20 Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
• CPC Class Codes: H02J 3/16 by adjustment of reactive power
• CPC Class Codes: H02J 3/18 Arrangements for adjusting, eliminating or compensating reactive power in networks for adjustment of voltage H02J3/16
• CPC Class Codes: H02J 3/1878 using tap changing or phase shifting transformers
• CPC Class Codes: H02J 3/1885 using rotating means, e.g. synchronous generators
• CPC Class Codes: Y02E 40/30 Reactive power compensation
• CPC Class Codes: Y02E 40/70 Smart grids as climate change mitigation technology in the energy generation sector
• CPC Class Codes: Y02E 60/00 Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
• CPC Class Codes: Y04S 10/50 Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
• CPC Class Codes: Y04S 40/20 Information technology specific aspects, e.g. CAD, simulation, modelling, system security

Abstract

The invention provides a reactive power optimization system and a method of a power grid based on a double-fish-swarm algorithm. The system includes a power grid state data acquiring module, a reactive power regulating module and a reactive power executing module. The power grid state data acquiring module includes a power grid state data acquiring processor and a relay transmitter. The reactive power regulating module is a control terminal. The reactive power executing module includes generator terminal voltage regulators, transformer tap regulators and reactive power compensation regulators. The method is used for acquiring the initial data to be optimized in the current network; and optimizing the initial data to be optimized in the current network based on a double-fish-swarm algorithm so as to obtain optimal value of control variables in the power grid. According to the method, the distribution network to be optimized can realize reasonable reactive power flow distribution.

Claims

1. A reactive power optimization system of a power grid based on a double-fish-swarm algorithm, comprising a power grid state data acquiring module, a reactive power regulating module and a reactive power executing module, the power grid state data acquiring module connected to the reactive power regulating module, the reactive power regulating module connected to the reactive power executing module, wherein

the power grid state data acquiring module comprises a power grid state data acquiring processor and a relay transmitter;

the reactive power regulating module is a control terminal;

the reactive power executing module comprises generator terminal voltage regulators, transformer tap regulators and reactive power compensation regulators;

an input end of the power grid state data acquiring processor is connected with a power grid;

an output end of the power grid state data acquiring processor is connected with the input end of the relay transmitter;

the output end of the relay transmitter is connected with the input end of the control terminal;

the output end of the control terminal is connected with the input end of each of the generator terminal voltage regulators, the input end of each of the transformer tap regulators and the input end of each of the reactive power compensation regulators;

the output end of each of the generator terminal voltage regulators is connected with a corresponding generator in the power grid;

the output end of each of the transformer tap regulators is connected with a corresponding transformer in the power grid; and

the output end of each of the reactive power compensation regulators is connected with a corresponding reactive power compensation device in the power grid;

the power grid state data acquiring processor is used for acquiring the current network information of the power grid and judging whether the current network information meets the optimal state required by the power grid or not; and

if the current network information cannot meet the optimal state required by the power grid, the current network information is transmitted to the relay transmitter, wherein the current network information includes power grid node information, branch information, generator information, transformer information and reactive power compensation device information;

the relay transmitter is used for transmitting the power grid node information, the branch information, the generator information, the transformer information and the reactive power compensation device information which are acquired by the power grid state data acquiring processor and required for reactive power optimization in the network information to the control terminal;

the reactive power regulating module comprises a processor configured to function as a parameter acquiring unit, a double-fish-swarm algorithm-base reactive power optimization unit, and an optimization decision-making control unit, the parameter acquiring unit connected to the reactive power optimization unit, and the reactive power optimization unit connected to the optimization decision-making control unit;

the parameter acquiring unit is used for acquiring the power grid node information, the branch information, the generator information, the transformer information and the reactive power compensation device information transmitted by the relay transmitter and used as initial data to be optimized in the current network;

the double-fish-swarm algorithm-base reactive power optimization unit is used for establishing a mathematical model for reactive power optimization of a power system, and the acquired initial data to be optimized in the current network is optimized based on the double-fish-swarm algorithm, so that an optimal value of each of control variables in the power grid is obtained, wherein the control variables comprise generator terminal voltage amplitudes, transformer adjustable ratios, and the reactive power capacity of reactive power compensation devices;

the optimization decision-making control unit is used for transmitting optimal values of the control variables to the reactive power executing module;

the generator terminal voltage regulators are used for regulating generator terminal voltages according to the optimal values of the generator terminal voltage amplitudes, obtained by the reactive power regulating module;

the transformer tap regulators are used for regulating transformer taps according to the optimal values of the transformer adjustable ratios, obtained by the reactive power regulating module; and

the reactive power compensation regulators are used for regulating the compensation capacity of reactive power compensators according to the optimal values of the reactive power capacity of the reactive power compensation devices, obtained by the reactive power regulating module.

2. The reactive power optimization system of a power grid based on the double-fish-swarm algorithm according to claim 1, wherein the mathematical model for establishing the reactive power optimization of the power system is shown as follows:

3. The reactive power optimization system of a power grid based on the double-fish-swarm algorithm according to claim 2, wherein a specific process for optimizing the acquired initial data to be optimized in the current network based on the double-fish-swarm algorithm so as to obtain the optimal values of the control variables in the power grid comprises the steps of:

generating a small fish-swarm and a ferocious fish-swarm in a random manner according to the upper limit value and the lower limit value of the control variables to be optimized in the power grid;

initializing the optimal value of current food concentration, and displaying the initialized optimal value of current food concentration in a bulletin board;

by using the reciprocal 1/F of the objective function of the mathematical model for reactive power optimization of a power grid as a value FC of food concentration, calculating the value of food concentration in the current network and of small fishes and ferocious fishes under an initial data condition, taking the maximum value of food concentration as the optimal value of current food concentration, displaying the optimal value of current food concentration in the bulletin board, and saving the state and the current value FC;

enabling the small fish-swarm to act;

enabling small fish individuals to perform a foraging clustering behavior, a rear-end behavior and a protective clustering behavior according to the distance between each of the small fish individuals and ferocious fish individuals as well as the distance between the small fish individuals in the small fish-swarm, and updating the value in the bulletin board Y1;

enabling the ferocious fish-swarm to act;

enabling each of the ferocious fish individuals to perform a predation behavior, a tracking behavior and a clustering behavior according to the distance between each of the ferocious fishes and the small fish individuals as well as the distance between the ferocious fish individuals in the ferocious fish-swarm, and updating the value in the bulletin board Y2; and

when iteration times achieve the maximum iteration times, taking the greater values of food concentration displayed in the bulletin boards Y1 and Y2 as well as the state thereof as optimal results so as to obtain the optimal values of the control variables, namely the optimal values of the generator terminal voltage amplitudes, the transformer adjustable ratios, and the reactive power capacity of the reactive power compensation devices.

4. A method for reactive power optimization of a power grid by adopting the reactive power optimization system of a power grid based on the double-fish-swarm algorithm according to claim 3, comprising the steps of:

step 1;

acquiring the current network information of the power grid state data and judging whether the current network information meets the optimal state required by the power grid or not by the power grid state data acquisitor; and

if the current network information cannot meet the optimal state required by the power grid, transmitting the current network information to the relay transmitter;

step 2;

transmitting the power grid node information, the branch information, the generator information, the transformer information and the reactive power compensation device information which are acquired by the power grid state data acquiring processor and required for reactive power optimization in the network information to the control terminal, by the relay transmitter;

step 3;

acquiring the power grid node information, the branch information, the generator information, the transformer information and the reactive power compensation device information transmitted by the relay transmitter, by the control terminal as initial data to be optimized in the current network;

step 4;

establishing the mathematical model for the reactive power optimization of the power system through the control terminal, optimizing the acquired initial data to be optimized in the current network based on the double-fish-swarm algorithm so as to obtain the optimal values of the control variables in the power grid, and transmitting the optimal values of the control variables to the reactive power executing module; and

step 5;

regulating generator terminal voltages by using the generator terminal voltage regulators according to the optimal values of the generator terminal voltage amplitudes, regulating the transformer taps by using the transformer tap regulators according to optimal values of the transformer adjustable ratios, and regulating the compensation capacity of the reactive power compensators by using the reactive power compensation regulators according to the optimal values of the reactive power capacity of the reactive power compensation devices.

5. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 4, wherein the step 4 comprises the steps of:

step 4.1;

establishing a mathematical model for the reactive power optimization of the power system by using an alpha-method, with the minimum network loss and the optimal voltage level as the optimization objectives;

6. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 5, wherein the step 4.3 comprises the steps of:

step 4.3.1;

initializing the number t₁=0 of the generated small fishes;

step 4.3.2;

randomly generating the small fish individuals according to the value ranges of the control variables to be optimized in the power grid;

X=[U_G11,U_G12, . . . U_G1g,K_T11,K_T12, . . . ,K_T1t,Q_C11,Q_C12, . . . Q_C1c]=[x₁,x₂, . . . ,x_s, . . . ,x_S];

wherein, X is the value sequence of the control variables of the small fish individuals, x_s=x_{s min}+rand(g)×

(x_{s max}−

x_{s min}), rand(g) is a random number of an interval (0,

     1), x_{s min} and x_{s max} are respectively the lower limit value and the upper limit value of the corresponding control variables, U_G1g, K_T1t and Q_C1c are respectively the generator terminal voltage amplitude, the transformer adjustable ratio, and the reactive power capacity of the reactive power compensation devices in the randomly generated small and artificial fish individuals, x_s is the control variables of the randomly generated small fish individuals, 1≤

s≤

s, S=N_g+N+N_c, N_g is the number of the generators in the power grid, N_i is the number of the transformers in the power grid, and N_c is the number of the reactive power compensation devices in the power grid;

step 4.3.3;

performing power flow calculation on the control variables in the randomly generated small fish individuals in the step 4.3.2 by using a P−

Q decomposition method, if the power flow value of the current small fish individual is converged, retaining the small fish individual, generating a small fish number t₁=t₁+1 of the current small fish individual, and executing the step 4.3.4;

or else, not retaining the small fish individual, and returning to the step 4.3.2; and

step 4.3.4;

if t₁≥

N1, obtaining a current small fish-swarm, and executing the step 4.4;

or else, returning to the step 4.3.2.

7. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 5, wherein the step 4.4 comprises the steps of:

step 4.4.1;

initializing the number t₂=0 of the generated ferocious fishes;

step 4.4.2;

randomly generating the ferocious fish individuals according to the value ranges of the control variables to be optimized in the power grid;

W=[U_G21,U_G22, . . . U_G2g,K_T21,K_T22, . . . ,K_T2t,Q_C21,Q_C22, . . . ,Q_C2c]=[w₁,w₂, . . . ,w_s, . . . ,w_S];

wherein, W is the value sequence of the control variables of the ferocious fish individuals, w_s=x_{s max}+rand(g)×

(x_{s max}−

x_{s min}), U_G2g, K_T2t and Q_C2c are respectively the generator terminal voltage amplitude, the transformer adjustable ratio, and the reactive power capacity of the reactive power compensation devices in the randomly generated ferocious fish individuals, and w_s is the control variables of the randomly generated ferocious fish individuals;

step 4.4.3;

performing power flow calculation on the control variables in the randomly generated ferocious fish individuals in the step 4.4.2 by using a P−

Q decomposition method, if the power flow value of the current ferocious fish individual is converged, retaining the ferocious fish individual, generating a ferocious fish number t₂=t₂+1 of the current ferocious fish individual, and executing the step 4.4.4;

or else, not retaining the ferocious fish individual, and returning to the step 4.4.2; and

step 4.4.4;

if t₂≥

N2, obtaining a current ferocious fish-swarm, and executing the step 4.5;

or else, returning to the step 4.4.2.

8. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 5, wherein the step 4.6 comprises the steps of:

step 4.6.1;

determining the distance D={D_i1, D_i2, . . . , D_ij, . . . , D_iN2} between the current i small fish individual X_i and each of the ferocious fish individuals W_j, wherein, D_ij=∥

X_i−

W_j∥

, 1<

j<

N2, and N2 is the number of the ferocious fish individuals;

step 4.6.2;

if the distance D_ij≤

Visual1 between the current i th small fish individual X_i and each of the ferocious fish individuals W_j, executing the step 4.6.3;

or else, executing the step 4.6.5;

step 4.6.3;

determining the total number n of the small fish individuals within the perception range of the i th small fish individual X_i, to determine a protective clustering safe position X_safe, and enabling the i th small fish individual X_i to perform the protective clustering behavior so as to obtain an update value X_inext1 of the i th small fish individual performing the protective clustering behavior, wherein

the calculation formula for the protective clustering safe position X_safe is as follows;

X_safe=(X_i1+X_i2+ . . . +X_in)×

λ

/n,

wherein, λ

=(n+p)/n is an escape factor, p is the total number of the ferocious fish individuals within the perception range Visual1 of the small fish individual X_i, and X_in is the small fish individuals within the perception range of the small fish individual X_i;

the calculation formula for the update value X_inext1 of the i th small fish individual performing the protective clustering behavior is as follows;

X_inext1=X_i+rand(g)×

Step1×

δ

₁×

(X_safe−

X_i)/∥

X_safe−

X_i∥

;

step 4.6.4;

judging whether the update value X_inext1 of the current small fish individual performing the protective clustering behavior meets a constraint condition and the power flow is converged;

if yes, recording the state of the update value X_inext1, calculating the value of food concentration, updating the value in the bulletin board Y1, and executing the step 4.7;

or else, returning to the step 4.6.3;

step 4.6.5;

determining the small fish individual X_j′

with the maximum value FC_j′

of food concentration in the small fish-swarm within the perception range of the i th small fish individual X_i, and enabling the i th small fish individual X_i to perform the rear-end behavior so as to obtain an update value X_inext2 of the i th small fish individual X_i performing the rear-end behavior, wherein

the calculation formula for the update value X_inext2 of the i th small fish individual X_i performing the rear-end behavior is as follows;

X_inext2=X_i+rand(g)×

Step1×

δ

₁×

(X_j′

−

X_i)/∥

X_j′

−

X_i∥

;

step 4.6.6;

determining the total number n of the small fish individuals within the perception range of the i th small fish individual X_i to determine the central position X_c1 of the small fish individuals performing foraging clustering behavior, and enabling the i th small fish individual X_i to perform the foraging clustering behavior so as to obtain an update value X_inext3 of the i th small fish individual performing the foraging clustering behavior, wherein

the calculation formula for the central position X_c1 of the small fish individuals performing foraging clustering behavior is as follows;

X_c1=(X_i1+X_i2+ . . . +X_in)/n;

the calculation formula for the update value X_inext3 of the i th small fish individual performing the foraging clustering behavior is as follows;

X_inext3=X_i+rand(g)×

Step1×

δ

₁×

(X_c1−

X_i)/∥

X_c1−

X_i∥

; and

step 4.6.7;

judging whether the update value X_inext2 of the current small fish individual performing the rear-end behavior and the update value X_inext3 of the current small fish individual performing foraging clustering behavior meet constraint conditions and the power flow is converged;

if yes, recording the state of the update value X_inext2 and the state of the update value X_inext3, calculating the value of food concentration, updating the value in the bulletin board Y1, and executing the step 4.7;

or else, returning to the step 4.6.5.

9. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 5, wherein the step 4.7 comprises the steps of:

step 4.7.1;

determining the distance L={L<sub>i′

1</sub>, L<sub>i′

2</sub>, . . . , L<sub>i′

j</sub>, . . . , L<sub>i′

N1</sub>} between the current i′

th ferocious fish individual W_i′

and each of the small fish individuals X_j″

, wherein L<sub>i′

j″</sub>

=∥

W_i′

−

X_j′

∥

, 1<

j″

<

N1, and N1 is the number of the small fish individuals;

step 4.7.2;

if the distance L<sub>i′

j″</sub>

≤

Visual2 between the current i′

th ferocious and artificial fish individual W_i′

and each of the small and artificial fish individuals X_j″

, executing the step 4.7.3;

or else, executing the step 4.7.6;

step 4.7.3;

determining the total number m of the small fish individuals within the perception range of the current i′

th ferocious fish individual W_i′

, enabling the i′

th ferocious fish individual W_i′

to perform the predation behavior with a central position X_c2 of small fish-swarm within the perception range of the i′

th ferocious fish individual W_i′

as the maximum value of current food concentration, and updating the i′

th ferocious fish individual W_i′

as the central position X_c2 of the small fish-swarm so as to obtain an update value X<sub>1′

next1</sub>=X_c2 of the i′

th ferocious fish individual W_i′

performing the predation behavior;

the calculation formula for the central position X_c2 of small fish-swarms within the perception range of the i′

th ferocious fish individual W_i′

is as follows;

X_c2=(X<sub>i′

1</sub>+X<sub>i′

2</sub>+ . . . +X<sub>i′

m</sub>)/m;

wherein X<sub>i′

m</sub> is the small fishes within the perception range of the ferocious fish individual W_i′

;

step 4.7.4;

enabling the i′

th ferocious fish individual W_i′

to perform the tracking behavior with the central position X_c2 of small fish-swarms within the perception range of the current i′

th ferocious fish individual W_i′

as the maximum value of current food concentration so as to obtain an update value W<sub>i′

next2</sub> of the i′

th ferocious fish individual W_i′

performing the tracking behavior;

the calculation formula for the update value W<sub>i′

next2</sub> of the i′

th ferocious fish individual W_i′

performing the tracking behavior is as follows;

W<sub>i′

next2</sub>=W<sub>′

j</sub>+rand(g)×

Step2×

δ

₂×

(X_c2−

W_i′

)/∥

X_c2−

W_i′

∥

;

step 4.7.5;

judging whether the update value W<sub>i′

next1</sub> of the current ferocious fish individual W_i′

performing the predation behavior and the update value W<sub>i′

next2</sub> of the ferocious fish individual W_i′

preforming the tracking behavior meet the constraint conditions and the power flow is converged;

if yes, updating the value in the bulletin board Y2, and executing the step 4.7.6;

or else, returning to the step 4.7.3;

step 4.7.6;

determining the total number r of the ferocious fish individuals within the perception range of the i′

th ferocious fish individual W_i′

to determine the central position W_c of the ferocious fish individuals preforming the clustering behavior, and enabling the i′

ferocious fish individual W_i′

to perform the clustering behavior so as to obtain an update value W<sub>i′

next3</sub> of the i′

th ferocious fish individual W_i′

performing the clustering behavior, wherein

the calculation formula for the central position W_c of the fish individual performing the clustering behavior is as follows;

W_c=(W<sub>i′

1</sub>+W<sub>i′

2</sub>+ . . . +W<sub>i′

r</sub>)/r;

wherein, W<sub>i′

r</sub> is the ferocious fish individuals within the perception range of the i′

th ferocious fish individual W_i′

;

the calculation formula for the update value W<sub>i′

next3</sub> of the i′

th ferocious fish individual W_i′

performing the clustering behavior is as follows;

W<sub>i′

next1</sub>=W_i′

+rand(g)×

Step2×

δ

₂×

(W_c−

W_i′

)/∥

W_c−

W_i′

∥

; and

step 4.7.7;

judging whether the update value W<sub>i′

next3</sub> of the current ferocious fish individual W_i′

performing the clustering behavior meets the constraint condition and the power flow is converged;

if yes, recording the state of the update value W<sub>i′

next3</sub>, calculating the value of the food concentration, updating the value in the bulletin board Y2, and executing the step 4.8;

or else, returning to the step 4.7.6.

10. A method for reactive power optimization of a power grid by adopting the reactive power optimization system of a power grid based on the double-fish-swarm algorithm according to claim 2, comprising the steps of:

step 1;

acquiring the current network information of the power grid state data and judging whether the current network information meets the optimal state required by the power grid or not by the power grid state data acquiring processor; and

if the current network information cannot meet the optimal state required by the power grid, transmitting the current network information to the relay transmitter;

step 2;

transmitting the power grid node information, the branch information, the generator information, the transformer information and the reactive power compensation device information which are acquired by the power grid state data acquiring processor and required for reactive power optimization in the network information to the control terminal, by the relay transmitter;

step 3;

acquiring the power grid node information, the branch information, the generator information, the transformer information and the reactive power compensation device information transmitted by the relay transmitter, by the control terminal as initial data to be optimized in the current network;

step 4;

establishing the mathematical model for the reactive power optimization of the power system through the control terminal, optimizing the acquired initial data to be optimized in the current network based on the double-fish-swarm algorithm so as to obtain the optimal values of the control variables in the power grid, and transmitting the optimal values of the control variables to the reactive power executing module; and

step 5;

regulating generator terminal voltages by using the generator terminal voltage regulators according to the optimal values of the generator terminal voltage amplitudes, regulating the transformer taps by using the transformer tap regulators according to optimal values of the transformer adjustable ratios, and regulating the compensation capacity of the reactive power compensators by using the reactive power compensation regulators according to the optimal values of the reactive power capacity of the reactive power compensation devices.

11. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 10, wherein the step 4 comprises the steps of:

step 4.1;

establishing a mathematical model for the reactive power optimization of the power system by using an alpha-method, with the minimum network loss and the optimal voltage level as the optimization objectives;

12. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 11, wherein the step 4.3 comprises the steps of:

step 4.3.1;

initializing the number t₁=0 of the generated small fishes;

step 4.3.2;

randomly generating the small fish individuals according to the value ranges of the control variables to be optimized in the power grid;

X=[U_G11,U_G12, . . . U_G1g,K_T11,K_T12, . . . ,K_T1t,Q_C11,Q_C12, . . . Q_C1c]=[x₁,x₂, . . . ,x_s, . . . ,x_S];

wherein, X is the value sequence of the control variables of the small fish individuals, x_s=x_{s min}+rand(g)×

(x_{s max}−

x_{s min}), rand(g) is a random number of an interval (0,

     1), x_{s min} and x_{s max} are respectively the lower limit value and the upper limit value of the corresponding control variables, U_G1g, K_T1t and Q_C1c are respectively the generator terminal voltage amplitude, the transformer adjustable ratio, and the reactive power capacity of the reactive power compensation devices in the randomly generated small and artificial fish individuals, x_s is the control variables of the randomly generated small fish individuals, 1≤

s≤

s, S=N_g+N+N_c, N_g is the number of the generators in the power grid, N_i is the number of the transformers in the power grid, and N_c is the number of the reactive power compensation devices in the power grid;

step 4.3.3;

performing power flow calculation on the control variables in the randomly generated small fish individuals in the step 4.3.2 by using a P−

Q decomposition method, if the power flow value of the current small fish individual is converged, retaining the small fish individual, generating a small fish number t₁=t₁+1 of the current small fish individual, and executing the step 4.3.4;

or else, not retaining the small fish individual, and returning to the step 4.3.2; and

step 4.3.4;

if t₁≥

N1, obtaining a current small fish-swarm, and executing the step 4.4;

or else, returning to the step 4.3.2.

13. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 11, wherein the step 4.4 comprises the steps of:

step 4.4.1;

initializing the number t₂=0 of the generated ferocious fishes;

step 4.4.2;

randomly generating the ferocious fish individuals according to the value ranges of the control variables to be optimized in the power grid;

W=[U_G21,U_G22, . . . U_G2g,K_T21,K_T22, . . . ,K_T2t,Q_C21,Q_C22, . . . ,Q_C2c]=[w₁,w₂, . . . ,w_s, . . . ,w_S];

wherein, W is the value sequence of the control variables of the ferocious fish individuals, w_s=x_{s max}+rand(g)×

(x_{s max}−

x_{s min}), U_G2g, K_T2t and Q_C2c are respectively the generator terminal voltage amplitude, the transformer adjustable ratio, and the reactive power capacity of the reactive power compensation devices in the randomly generated ferocious fish individuals, and w_s is the control variables of the randomly generated ferocious fish individuals;

step 4.4.3;

performing power flow calculation on the control variables in the randomly generated ferocious fish individuals in the step 4.4.2 by using a P−

Q decomposition method, if the power flow value of the current ferocious fish individual is converged, retaining the ferocious fish individual, generating a ferocious fish number t₂=t₂+1 of the current ferocious fish individual, and executing the step 4.4.4;

or else, not retaining the ferocious fish individual, and returning to the step 4.4.2; and

step 4.4.4;

if t₂≥

N2, obtaining a current ferocious fish-swarm, and executing the step 4.5;

or else, returning to the step 4.4.2.

14. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 11, wherein the step 4.6 comprises the steps of:

step 4.6.1;

determining the distance D={D_i1, D_i2, . . . , D_ij, . . . , D_iN2} between the current i small fish individual X_i and each of the ferocious fish individuals W_j, wherein, D_ij=∥

X_i−

W_j∥

, 1<

j<

N2, and N2 is the number of the ferocious fish individuals;

step 4.6.2;

if the distance D_ij≤

Visual1 between the current i th small fish individual X_i and each of the ferocious fish individuals W_j, executing the step 4.6.3;

or else, executing the step 4.6.5;

step 4.6.3;

determining the total number n of the small fish individuals within the perception range of the i th small fish individual X_i, to determine a protective clustering safe position X_safe, and enabling the i th small fish individual X_i to perform the protective clustering behavior so as to obtain an update value X_inext1 of the i th small fish individual performing the protective clustering behavior, wherein

the calculation formula for the protective clustering safe position X_safe is as follows;

X_safe=(X_i1+X_i2+ . . . +X_in)×

λ

/n,

wherein, λ

=(n+p)/n is an escape factor, p is the total number of the ferocious fish individuals within the perception range Visual1 of the small fish individual X_i, and X_in is the small fish individuals within the perception range of the small fish individual X_i;

the calculation formula for the update value X_inext1 of the i th small fish individual performing the protective clustering behavior is as follows;

X_inext1=X_i+rand(g)×

Step1×

δ

₁×

(X_safe−

X_i)/∥

X_safe−

X_i∥

;

step 4.6.4;

judging whether the update value X_inext1 of the current small fish individual performing the protective clustering behavior meets a constraint condition and the power flow is converged;

if yes, recording the state of the update value X_inext1, calculating the value of food concentration, updating the value in the bulletin board Y1, and executing the step 4.7;

or else, returning to the step 4.6.3;

step 4.6.5;

determining the small fish individual X_j′

with the maximum value FC_j′

of food concentration in the small fish-swarm within the perception range of the i th small fish individual X_i, and enabling the i th small fish individual X_i to perform the rear-end behavior so as to obtain an update value X_inext2 of the i th small fish individual X_i performing the rear-end behavior, wherein

the calculation formula for the update value X_inext2 of the i th small fish individual X_i performing the rear-end behavior is as follows;

X_inext2=X_i+rand(g)×

Step1×

δ

₁×

(X_j′

−

X_i)/∥

X_j′

−

X_i∥

;

step 4.6.6;

determining the total number n of the small fish individuals within the perception range of the i th small fish individual X_i to determine the central position X_c1 of the small fish individuals performing foraging clustering behavior, and enabling the i th small fish individual X_i to perform the foraging clustering behavior so as to obtain an update value X_inext3 of the i th small fish individual performing the foraging clustering behavior, wherein

the calculation formula for the central position X_c1 of the small fish individuals performing foraging clustering behavior is as follows;

X_c1=(X_i1+X_i2+ . . . +X_in)/n;

the calculation formula for the update value X_inext3 of the i th small fish individual performing the foraging clustering behavior is as follows;

X_inext3=X_i+rand(g)×

Step1×

δ

₁×

(X_c1−

X_i)/∥

X_c1−

X_i∥

; and

step 4.6.7;

judging whether the update value X_inext2 of the current small fish individual performing the rear-end behavior and the update value X_inext3 of the current small fish individual performing foraging clustering behavior meet constraint conditions and the power flow is converged;

if yes, recording the state of the update value X_inext2 and the state of the update value X_inext3, calculating the value of food concentration, updating the value in the bulletin board Y1, and executing the step 4.7;

or else, returning to the step 4.6.5.

15. The method for reactive power optimization of a power grid based on the double-fish-swarm algorithm according to claim 11, wherein the step 4.7 comprises the steps of:

step 4.7.1;

determining the distance L={L<sub>i′

1</sub>, L<sub>i′

2</sub>, . . . , L<sub>i′

j</sub>, . . . , L<sub>i′

N1</sub>} between the current i′

th ferocious fish individual W_i′

and each of the small fish individuals X_j″

, wherein L<sub>i′

j″</sub>

=∥

W_i′

−

X_j′

∥

, 1<

j″

<

N1, and N1 is the number of the small fish individuals;

step 4.7.2;

if the distance L<sub>i′

j″</sub>

≤

Visual2 between the current i′

th ferocious and artificial fish individual W_i′

and each of the small and artificial fish individuals X_j″

, executing the step 4.7.3;

or else, executing the step 4.7.6;

step 4.7.3;

determining the total number m of the small fish individuals within the perception range of the current i′

th ferocious fish individual W_i′

, enabling the i′

th ferocious fish individual W_i′

to perform the predation behavior with a central position X_c2 of small fish-swarm within the perception range of the i′

th ferocious fish individual W_i′

as the maximum value of current food concentration, and updating the i′

th ferocious fish individual W_i′

as the central position X_c2 of the small fish-swarm so as to obtain an update value X<sub>1′

next1</sub>=X_c2 of the i′

th ferocious fish individual W_i′

performing the predation behavior;

the calculation formula for the central position X_c2 of small fish-swarms within the perception range of the i′

th ferocious fish individual W_i′

is as follows;

X_c2=(X<sub>i′

1</sub>+X<sub>i′

2</sub>+ . . . +X<sub>i′

m</sub>)/m;

wherein X<sub>i′

m</sub> is the small fishes within the perception range of the ferocious fish individual W_i′

;

step 4.7.4;

enabling the i′

th ferocious fish individual W_i′

to perform the tracking behavior with the central position X_c2 of small fish-swarms within the perception range of the current i′

th ferocious fish individual W_i′

as the maximum value of current food concentration so as to obtain an update value W<sub>i′

next2</sub> of the i′

th ferocious fish individual W_i′

performing the tracking behavior;

the calculation formula for the update value W<sub>i′

next2</sub> of the i′

th ferocious fish individual W_i′

performing the tracking behavior is as follows;

W<sub>i′

next2</sub>=W<sub>′

j</sub>+rand(g)×

Step2×

δ

₂×

(X_c2−

W_i′

)/∥

X_c2−

W_i′

∥

;

step 4.7.5;

judging whether the update value W<sub>i′

next1</sub> of the current ferocious fish individual W_i′

performing the predation behavior and the update value W<sub>i′

next2</sub> of the ferocious fish individual W_i′

preforming the tracking behavior meet the constraint conditions and the power flow is converged;

if yes, updating the value in the bulletin board Y2, and executing the step 4.7.6;

or else, returning to the step 4.7.3;

step 4.7.6;

determining the total number r of the ferocious fish individuals within the perception range of the i′

th ferocious fish individual W_i′

to determine the central position W_c of the ferocious fish individuals preforming the clustering behavior, and enabling the i′

ferocious fish individual W_i′

to perform the clustering behavior so as to obtain an update value W<sub>i′

next3</sub> of the i′

th ferocious fish individual W_i′

performing the clustering behavior, wherein

the calculation formula for the central position W_c of the fish individual performing the clustering behavior is as follows;

W_c=(W<sub>i′

1</sub>+W<sub>i′

2</sub>+ . . . +W<sub>i′

r</sub>)/r;

wherein, W<sub>i′

r</sub> is the ferocious fish individuals within the perception range of the i′

th ferocious fish individual W_i′

;

the calculation formula for the update value W<sub>i′

next3</sub> of the i′

th ferocious fish individual W_i′

performing the clustering behavior is as follows;

W<sub>i′

next1</sub>=W_i′

+rand(g)×

Step2×

δ

₂×

(W_c−

W_i′

)/∥

W_c−

W_i′

∥

; and

step 4.7.7;

judging whether the update value W<sub>i′

next3</sub> of the current ferocious fish individual W_i′

performing the clustering behavior meets the constraint condition and the power flow is converged;

if yes, recording the state of the update value W<sub>i′

next3</sub>, calculating the value of the food concentration, updating the value in the bulletin board Y2, and executing the step 4.8;

or else, returning to the step 4.7.6.

16. The reactive power optimization system of a power grid based on the double-fish-swarm algorithm according to claim 1, wherein the specific process for optimizing the acquired initial data to be optimized in the current network based on the double-fish-swarm algorithm so as to obtain the optimal values of the control variables in the power grid comprises the steps of:

generating a small fish-swarm and a ferocious fish-swarm in a random manner according to the upper limit value and the lower limit value of the control variables to be optimized in the power grid;

initializing the optimal value of current food concentration, and displaying the initialized optimal value of current food concentration in a bulletin board;

by using the reciprocal 1/F of the objective function of the mathematical model for reactive power optimization of a power grid as a value FC of food concentration, calculating the value of food concentration in the current network and of small fishes and ferocious fishes under an initial data condition, taking the maximum value of food concentration as the optimal value of current food concentration, displaying the optimal value of current food concentration in the bulletin board, and saving the state and the current value FC;

enabling the small fish-swarm to act;

enabling small fish individuals to perform a foraging clustering behavior, a rear-end behavior and a protective clustering behavior according to the distance between each of the small fish individuals and ferocious fish individuals as well as the distance between the small fish individuals in the small fish-swarm, and updating the value in the bulletin board Y1;

enabling the ferocious fish-swarm to act;

enabling each of the ferocious fish individuals to perform a predation behavior, a tracking behavior and a clustering behavior according to the distance between each of the ferocious fishes and the small fish individuals as well as the distance between the ferocious fish individuals in the ferocious fish-swarm, and updating the value in the bulletin board Y2; and

when iteration times achieve the maximum iteration times, taking the greater values of food concentration displayed in the bulletin boards Y1 and Y2 as well as the state thereof as optimal results so as to obtain the optimal values of the control variables, namely the optimal values of the generator terminal voltage amplitudes, the transformer adjustable ratios, and the reactive power capacity of the reactive power compensation devices.