CN104112237A - WAMS-based genetic algorithm-improved power grid reactive capacity optimization configuration method - Google Patents

Abstract

The invention relates to a genetic algorithm-improved power grid reactive capacity optimization configuration method based on real-time information of a Wide Area Measurement System (WAMS) and an Energy Management System (EMS) of a power grid. The method includes: step S1. obtaining power grid measurement information from the power grid WAMS in real time; step S2. obtaining from the EMS the power grid structure, parameters and running section information in the EMS of a current time section, and forming an admittance matrix of a whole network containing all load node equivalent admittance; step S3. performing hybrid coding on a control variable of power grid reactive optimization; step S4. establishing a fitness function; step S5. executing a selection operator; step S6. executing a cross and mutation operator; and step S7. determining the rate of convergence and convergence termination criteria. The method in the invention makes improvements on the basis of a traditional genetic algorithm, makes improvements in multiple aspects to improve defects of the traditional genetic algorithm in reactive optimization, improves the rate of convergence and quality of a solution, and prevents the problem of occurrence of a local optimal solution.

Description

A kind of electric network reactive-load capacity configuration optimizing method of the improved genetic algorithms method based on WAMS

Technical field

The present invention relates to the collocation method in a kind of electric network intelligent scheduling back-up system operation of power networks state estimation and early warning field, specifically relate to a kind of electric network reactive-load capacity configuration optimizing method of the improved genetic algorithms method based on WAMS.

Background technology

Along with the fast development of society, economy and power industry, electrical network develops on a large scale gradually, at a distance, extra-high voltage alternating current-direct current is interconnected and the access ratio increasing of new forms of energy etc., has increased the uncertainty of operation of power networks.And receiving end electrical network is mainly centered by load area of concentration, by periphery interconnection and the remote broad sense power supply of making a start, is connected, and then realizes the equilibrium of supply and demand of electric energy.Because the energy and load center areal distribution are not mated, and the restriction of the factor such as environment, receiving-end system internal electric source underbraced, a large amount of electric energy need carry out from a distant place long-distance transmissions, and receiving-end system scale is swift and violent to be increased and its complexity is got over complicated.

Since the eighties in 20th century, in succession there is lasting on the low side, the collapse of voltage event of a lot of voltage in a plurality of large-scale power systems in the world, cause huge economic loss and social influence, voltage stabilization becomes the focus that International Power educational circles pays close attention to gradually, and the on-line monitoring of the receiving end Network Voltage Stability under running environment is had higher requirement.At present, for the numerical algorithm comparative maturity of power system steady state voltage stability and dynamic electric voltage stability simulation, cause that reason that simulation result and real system are misfitted is mainly the inaccurate of component models and parameter in system.In addition, according to the analytical approach of mathematical modeling and emulation, find, be subject to the restriction of the factors such as electric network model, parameter and numerical evaluation, at aspects such as application scale, speed and reliabilities, be difficult to adapt to the requirement of the online real-time assessment of voltage stabilization.

WAMS (Wide Area Measurement System, WAMS) can realize the synchronous real-time measurement of the true running status of the Wide Area Power, for the receiving end Network Voltage Stability online evaluation based on WAMS measuring track information provides new opportunity.For the dynamic recovery characteristics of physics and external presentation, Voltage-stabilizing Problems mainly focuses on receiving end electrical network, has certain local feature.Voltage Instability (or collapse) may occur in a certain microvariations of normal operating condition, or occur in large disturbance transient process, how to effectively utilize WAMS measuring track information, the Voltage Stability Evaluation of realizing receiving end electrical network normal condition and large state of disturbance is the focus of current power engineering circles research.

Reactive power optimization of power system is exactly that by the throwing of regulator generator terminal voltage, on-load transformer tap changer and reactive-load compensation equipment, to move back object be to guarantee under the prerequisite of power system safety and stability operation, realizes electric system active loss minimum and keep good voltage.For this reason, idle work optimization is the non-linear optimizing problem of mixing [1-2] of a multiple goal, multivariate, multiple constraint.In recent years, numerous scholars have carried out a large amount of research to this problem.Proposed a lot of algorithms, and intelligent algorithm becomes study hotspot in this field, mainly contain genetic algorithm, immune algorithm, particle swarm optimization algorithm, hybrid algorithm etc.

Genetic algorithm is because its principle is simple, easy to operate, also, without the mathematical operation such as carry out that differentiate is inverted, is particularly suitable for processing Nonlinear Multiobjective optimization problem.For this reason, in reactive power optimization of power system, be widely used.Yet traditional genetic algorithm still comes with some shortcomings in reactive power optimization of power system process, as slow in speed of convergence, there is precocious phenomenon and form the problems such as locally optimal solution.

Summary of the invention

For the deficiencies in the prior art, the electric network reactive-load capacity configuration optimizing method that the object of this invention is to provide a kind of improved genetic algorithms method based on WAMS, the present invention improves traditional genetic algorithm, comprises the aspects such as coded system, colony's selection, fitness function, crossover and mutation and convergence criterion.Improve from many aspects to improve the deficiency of traditional genetic algorithm aspect idle work optimization, improve the quality that speed of convergence is conciliate, avoid occurring locally optimal solution problem.

The object of the invention is to adopt following technical proposals to realize:

The invention provides a kind of electric network reactive-load capacity configuration optimizing method of the improved genetic algorithms method based on WAMS, its improvements are, the real-time information of described method based on electrical network wide area monitoring system WAMS and energy management system EMS, described method comprises the steps:

Step S1: Real-time Obtaining electrical network measurement information from electrical network wide area monitoring system WAMS;

Step S2: obtain electric network composition, parameter and operation section information in the energy management system EMS of current time section from energy management system EMS, and form the whole network admittance matrix that comprises all load bus equivalent admittances;

Step S3: the control variable of reactive power optimization is carried out to hybrid coding;

Step S4: set up fitness function;

Step S5: carry out and select operator;

Step S6: carry out crossover and mutation operator;

Step S7: determine that speed of convergence and convergence stop criterion.

Further, in described step S1, (PMU is that the English of vector measurement unit is called for short to each vector measurement unit of Real-time Obtaining electrical network PMU from electrical network wide area monitoring system WAMS, can be arranged on generating plant and transformer station, be only the PMU data that are taken at Installation in Plant point in this patent) dynamic measurement information that website is installed.

Further, in described step S2, from energy management system EMS, obtain the QS data file that simultaneously comprises topological structure of electric, parameter and operation trend section, carry out trend calculating, to obtain the complete network admittance matrix of taking into account load equivalent admittance, for subsequent calculations voltage, fall the Dai Weinan equivalent parameters of out-of-limit node and do data preparation;

It is target that electric network reactive-load capacity optimizes to reduce the whole network network loss, realizes the economical operation of the stable and electric system of voltage levvl, and objective function mathematical model expression formula is as follows:

$P_L=\underset{k=i,j}{\underset{k\∈\mathrm{Ng}}{\mathrm{\Σ}}}{g}_k(U_i^2+U_j^2-2U_i{U}_j\cos {\mathrm{\θ}}_{\mathrm{ij}})---\left(1\right);$

In formula, P _lactive power loss for system; Ng is a way; g _kfor the electricity of branch road k is led; U _ivoltage magnitude for load bus i; θ _ijfor the phasing degree between load bus i and j;

Choose generator voltage, reactive compensation capacity and on-load transformer tap changer gear as control variable, choose generator reactive and exert oneself and economize on electricity voltage magnitude as state variable; All state variables all must meet variable bound condition, and the introducing by penalty function when state variable exceeds upper lower limit value is processed objective function, and its expression formula is:

$\begin{array}{c}F=\min (P_L+{\mathrm{\λ}}_V\mathrm{\Σ}\left(\frac{\mathrm{\Δ}{V}_i}{V_{i\max }-V_{i\min }}\right)+\\ {\mathrm{\λ}}_Q\mathrm{\Σ}{\left(\frac{\mathrm{\Δ}{Q}_{\mathrm{Gi}}}{Q_{\mathrm{Gi}\max }-Q_{\mathrm{Gi}\min }}\right)}^2)\end{array}---\left(2\right);$

In formula: λ _vfor the node voltage penalty factor that crosses the border; λ _qthe penalty factor that crosses the border for generator output reactive power; for the penalty function of crossing the border of node voltage, the penalty function of crossing the border for generator output reactive power; Δ V _i, V _imax, V _iminbe illustrated respectively in given " generation " deviate, maximum node voltage, the minimum node voltage of i node voltage and average; Δ Q _gi, Q _gimax, Q _giminbe illustrated respectively in given " generation ", the deviate of i generator output reactive power and average is, the minimum reactive power of the maximum reactive power of generator output, generator output;

Equation of constraint, the constraint such as comprises and does not wait equation of constraint; Wherein, waiting equation of constraint is the trend equilibrium equation of node, and expression formula is as follows:

$\left\{\begin{array}{c}{P}_i=V_i\underset{j\∈i}{\mathrm{\Σ}}{V}_j(G_{\mathrm{ij}}\cos {\mathrm{\δ}}_{\mathrm{ij}}+B_{\mathrm{ij}}\sin {\mathrm{\δ}}_{\mathrm{ij}})\\ Q_i=V_i\underset{j\∈i}{\mathrm{\Σ}}{V}_j(G_{\mathrm{ij}}\sin {\mathrm{\δ}}_{\mathrm{ij}}+B_{\mathrm{ij}}\cos {\mathrm{\δ}}_{\mathrm{ij}})\end{array}\right.---\left(3\right);$

In formula, P _i, Q _ibe respectively injection active power and the reactive power of node i; V _i, V _jbe respectively the voltage magnitude of node i and j; B _ij, G _ijand δ _ijbe respectively that electricity between node i and j is led, susceptance and voltage-phase angular difference; J ∈ i represents the node that all nodes are connected with node i;

Do not wait equation of constraint to comprise the equation of constraint such as or not does not wait equation of constraint and control variable of state variable;

The equation of constraint that do not wait of state variable is:

$\left\{\begin{array}{c}{Q}_{\mathrm{Gi}\min }\≤Q_{\mathrm{Gi}}\≤Q_{\mathrm{Gi}\max }\\ V_{\mathrm{di}\min }\≤V_{\mathrm{di}}\≤V_{\mathrm{di}\max }\end{array}\right.---\left(4\right);$

In formula, Q _gifor generator reactive is exerted oneself; Q _giminand Q _gimaxrepresent respectively the upper and lower bound that generator reactive is exerted oneself; V _difor load bus voltage magnitude; V _diminand V _dimaxthe upper and lower bound that represents respectively load bus voltage magnitude;

The equation of constraint that do not wait of control variable is:

$\left\{\begin{array}{c}{V}_{\mathrm{Gi}\min }\≤V_{\mathrm{Gi}}\≤V_{\mathrm{Gi}\max }\\ Q_{\mathrm{Ci}\min }\≤Q_{\mathrm{Ci}}\≤Q_{\mathrm{Ci}\max }\\ T_{\mathrm{ti}\min }\≤T_{\mathrm{ti}}\≤T_{\mathrm{ti}\max }\end{array}\right.---\left(5\right);$

In formula, V _gifor generator voltage; Q _cifor the reactive compensation capacity in node i; T _tion-load transformer tap changer gear;

Adopt P-Q decomposition method to carry out trend calculating.

Further, in described step S3, take real number and integer hybrid coding mode, be in the control variable of reactive power optimization, the set end voltage of generator is continuous variable, adopt real coding, node reactive-load compensation amount and on-load transformer tap changer are discrete variable, adopt integer coding.

Further, in described step S4, adopt method reciprocal to set up fitness function, that is:

$F\left(x\right)=\frac{K}{f\left(x\right)-z+1}---\left(6\right);$

Wherein, F (x) is fitness value, and f (x) is objective function, and z is that experience is estimated minimum value, and K is enlargement factor, and 1 for preventing that denominator from being 0.

Further, in described step S5, the mixing selection strategy that adopts tournament method and optimum preservation method to combine; First, by championship, in N random individual, optimum individuality is selected, then carried out the preservation of optimum individual, the individuality of fitness maximum in parent is replaced to the poorest individuality of fitness in filial generation, the advantage genes of individuals of making is continued.

Further, in described step S6, adopt new crossover and mutation mode, i.e. crossover probability Pc and variation probability P m, the change of crossover probability and variation probability is along with fitness changes automatically, and expression formula is as follows:

$P_c=\left\{\begin{array}{cc}{P}_{c1}-\frac{(P_{c1}-P_{c2})(f_c-f_p)}{f_{\max }-f_p},& (f_c\&GreaterEqual;f_p)\\ P_{c1}& (f_c<f_p)\end{array}\right.---\left(7\right);$

$P_m=\left\{\begin{array}{cc}{P}_{m1}-\frac{(P_{m1}-P_{m2})(f_m-f_p)}{f_{\max }-f_p},& (f_m\&GreaterEqual;f_p)\\ P_{m1}& (f_m<f_p)\end{array}\right.---\left(8\right);$

In formula: f _maxfor the maximum adaptation degree value in population; f _pfor the average fitness value in population; f _cfor intersecting individual fitness value; f _mfor the individual fitness value that makes a variation; P _c1for produced first crossover probability value, P _c2for produced second crossover probability value, P _m1for produced the 1st variation probable value, P _m2for the 2nd produced variation probable value.

Further, in described step S7, described speed of convergence, according to the mode of taboo question blank, is set up chromosome question blank, to per generation population carry out chromosome inquiry, storage, prevent that phase homologous chromosomes from carrying out repeated power flow calculating, for improving speed of convergence; Adopt average fitness value determining method as stop technology criterion, guarantee just to stop finishing after output global optimum, that is:

$|{\stackrel{\&OverBar;}{F}}^{\left(n\right)}-{\stackrel{\&OverBar;}{F}}^{(n-1)}|<\mathrm{\&epsiv;}---\left(9\right);$

Wherein, ε determines according to the fitness value of practical problems; with the average fitness value that represents respectively n generation and n-1 generation, n is natural number.

Compared with the prior art, the beneficial effect that the present invention reaches is:

1. for the mathematical model of idle work optimization, the present invention improves traditional genetic algorithm.In objective function, adopt penalty function, effectively avoided the out-of-limit of control variable.

2. when encoding, for the feature of idle work optimization control variable, take the hybrid mode of real number and integer coding, more closing to reality situation.

3. in order to prevent, in operating process, may lose optimum solution, adopt mode reciprocal to set up fitness function, and the mixing selection strategy that has adopted tournament method and optimum preservation method to combine on population is selected, on crossover and mutation, also taked self-adaptation crossover probability and variation probability, counting yield is significantly improved.

4. the present invention, completely by the existing information of WAMS and EMS system, adapts to the real-time situation ability of electrical network strong, meets the requirement of online evaluation.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the electric network reactive-load capacity configuration optimizing method of the improved genetic algorithms method based on WAMS provided by the invention.

Embodiment

Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described in further detail.

The electric network reactive-load capacity configuration optimizing method that the invention provides a kind of improved genetic algorithms method based on WAMS and EMS real-time information, its process flow diagram as shown in Figure 1, comprises the steps:

Step S1: each vector measurement unit of Real-time Obtaining electrical network PMU installs the dynamic measurement information of website from electrical network wide area monitoring system WAMS, using this measurement data as power plant's end (comprise power plant's merit angle, send meritorious, idle);

Step S2: obtain the QS data file that simultaneously comprises topological structure of electric, parameter and operation trend section from energy management system EMS, carry out trend calculating, to obtain the complete network admittance matrix of taking into account load equivalent admittance, for subsequent calculations voltage, fall the Dai Weinan equivalent parameters of out-of-limit node and do data preparation;

It is target that electric network reactive-load capacity optimizes to reduce the whole network network loss, realizes the economical operation of the stable and electric system of voltage levvl, and objective function mathematical model expression formula is as follows:

$P_L=\underset{k=i,j}{\underset{k\&Element;\mathrm{Ng}}{\mathrm{\&Sigma;}}}{g}_k(U_i^2+U_j^2-2U_i{U}_j\cos {\mathrm{\&theta;}}_{\mathrm{ij}})---\left(1\right);$

In formula, P _lactive power loss for system; Ng is a way; g _kfor the electricity of branch road k is led; U _ivoltage magnitude for load bus i; θ _ijfor the phasing degree between load bus i and j;

Choose generator voltage, reactive compensation capacity and on-load transformer tap changer gear as control variable, choose generator reactive and exert oneself and economize on electricity voltage magnitude as state variable; All state variables all must meet variable bound condition, and the introducing by penalty function when state variable exceeds upper lower limit value is processed objective function, and its expression formula is:

$\begin{array}{c}F=\min (P_L+{\mathrm{\&lambda;}}_V\mathrm{\&Sigma;}\left(\frac{\mathrm{\&Delta;}{V}_i}{V_{i\max }-V_{i\min }}\right)+\\ {\mathrm{\&lambda;}}_Q\mathrm{\&Sigma;}{\left(\frac{\mathrm{\&Delta;}{Q}_{\mathrm{Gi}}}{Q_{\mathrm{Gi}\max }-Q_{\mathrm{Gi}\min }}\right)}^2)\end{array}---\left(2\right);$

In formula: λ _vfor the node voltage penalty factor that crosses the border; λ _qthe penalty factor that crosses the border for generator output reactive power; for the penalty function of crossing the border of node voltage, the penalty function of crossing the border for generator output reactive power; △ V _i, V _imax, V _iminbe illustrated respectively in given " generation " deviate, maximum node voltage, the minimum node voltage of i node voltage and average; △ Q _gi, Q _gimax, Q _giminbe illustrated respectively in given " generation ", the deviate of i generator output reactive power and average is, the minimum reactive power of the maximum reactive power of generator output, generator output;

Equation of constraint, the constraint such as comprises and does not wait equation of constraint; Wherein, waiting equation of constraint is the trend equilibrium equation of node, and expression formula is as follows:

$\left\{\begin{array}{c}{P}_i=V_i\underset{j\&Element;i}{\mathrm{\&Sigma;}}{V}_j(G_{\mathrm{ij}}\cos {\mathrm{\&delta;}}_{\mathrm{ij}}+B_{\mathrm{ij}}\sin {\mathrm{\&delta;}}_{\mathrm{ij}})\\ Q_i=V_i\underset{j\&Element;i}{\mathrm{\&Sigma;}}{V}_j(G_{\mathrm{ij}}\sin {\mathrm{\&delta;}}_{\mathrm{ij}}+B_{\mathrm{ij}}\cos {\mathrm{\&delta;}}_{\mathrm{ij}})\end{array}\right.---\left(3\right);$

In formula, P _i, Q _ibe respectively injection active power and the reactive power of node i; V _i, V _jbe respectively the voltage magnitude of node i and j; B _ij, G _ijand δ _ijbe respectively that electricity between node i and j is led, susceptance and voltage-phase angular difference; J ∈ i represents the node that all nodes are connected with node i;

Do not wait equation of constraint to comprise the equation of constraint such as or not does not wait equation of constraint and control variable of state variable;

The equation of constraint that do not wait of state variable is:

$\left\{\begin{array}{c}{Q}_{\mathrm{Gi}\min }\&le;Q_{\mathrm{Gi}}\&le;Q_{\mathrm{Gi}\max }\\ V_{\mathrm{di}\min }\&le;V_{\mathrm{di}}\&le;V_{\mathrm{di}\max }\end{array}\right.---\left(4\right);$

In formula, Q _gifor generator reactive is exerted oneself; Q _giminand Q _gimaxrepresent respectively the upper and lower bound that generator reactive is exerted oneself; V _difor load bus voltage magnitude; V _diminand V _dimaxthe upper and lower bound that represents respectively load bus voltage magnitude;

The equation of constraint that do not wait of control variable is:

$\left\{\begin{array}{c}{V}_{\mathrm{Gi}\min }\&le;V_{\mathrm{Gi}}\&le;V_{\mathrm{Gi}\max }\\ Q_{\mathrm{Ci}\min }\&le;Q_{\mathrm{Ci}}\&le;Q_{\mathrm{Ci}\max }\\ T_{\mathrm{ti}\min }\&le;T_{\mathrm{ti}}\&le;T_{\mathrm{ti}\max }\end{array}\right.---\left(5\right);$

In formula, V _gifor generator voltage; Q _cifor the reactive compensation capacity in node i; T _tion-load transformer tap changer gear;

In the process of idle work optimization, need calculate the initial value of electric system active power loss and the preliminary examination state of control variable and state variable by trend, these preliminary examination values can be used as the raw data based on genetic algorithm idle work optimization.In addition, while adopting genetic algorithm to carry out idle work optimization, every iteration once, need be carried out trend calculating to new individuality, and for this reason, the present invention adopts P-Q decomposition method to carry out trend calculating, and then improves the bulk velocity of algorithm.

Step S3: the control variable of reactive power optimization is carried out to hybrid coding; Improvement to coded system.Coding is the first step of genetic algorithm, takes real number and integer hybrid coding mode herein.Be in the control variable of idle work optimization, the set end voltage of generator is continuous variable, adopt real coding, and node reactive-load compensation amount and on-load transformer tap changer is discrete variable, adopts integer coding, and closing to reality encodes more.

Step S4: set up fitness function; In order to guarantee that optimizing maximizes, fitness function can not take simply to add and subtract shift method, can in operating process, may lose optimum solution like this, and for this reason, the present invention adopts method reciprocal to set up fitness function, that is:

$F\left(x\right)=\frac{K}{f\left(x\right)-z+1}---\left(6\right);$

Wherein, F (x) is fitness value, and f (x) is objective function, and z is that experience is estimated minimum value, and K is enlargement factor, and 1 for preventing that denominator from being 0, can improve significantly optimizing and maximize.

Step S5: carry out and select operator: to selecting the improvement of operator.Select to select the individuality that vitality is stronger to produce new population processes in Shi colony the mixing selection strategy that the present invention adopts tournament method and optimum preservation method to combine.First, by championship, in N random individual, optimum individuality is selected, then carry out the preservation of optimum individual, the individuality of fitness maximum in parent is replaced to the poorest individuality of fitness in filial generation, the advantage genes of individuals of making is continued, and can when effectively preventing being absorbed in locally optimal solution, improve search speed.

Step S6: carry out crossover and mutation operator: adopt new crossover and mutation mode, i.e. crossover probability Pc and variation probability P m, the change of crossover probability and variation probability is along with fitness changes automatically, and expression formula is as follows:

$P_c=\left\{\begin{array}{cc}{P}_{c1}-\frac{(P_{c1}-P_{c2})(f_c-f_p)}{f_{\max }-f_p},& (f_c\&GreaterEqual;f_p)\\ P_{c1}& (f_c<f_p)\end{array}\right.---\left(7\right);$

$P_m=\left\{\begin{array}{cc}{P}_{m1}-\frac{(P_{m1}-P_{m2})(f_m-f_p)}{f_{\max }-f_p},& (f_m\&GreaterEqual;f_p)\\ P_{m1}& (f_m<f_p)\end{array}\right.---\left(8\right);$

In formula: f _maxfor the maximum adaptation degree value in population; f _pfor the average fitness value in population; f _cfor intersecting individual fitness value; f _mfor the individual fitness value that makes a variation; P _c1for produced first crossover probability value, P _c2for produced second crossover probability value, P _m1for produced the 1st variation probable value, P _m2for the 2nd produced variation probable value.The speed of convergence that can realize in searching process is controlled.

P _c1, P _c2p _m1, P _m2all random numbers that produce between a 0-1.

Step S7: determine that speed of convergence and convergence stop criterion: speed of convergence is herein according to the mode of taboo question blank, set up chromosome question blank, to per generation population carry out chromosome inquiry, storage, prevent that phase homologous chromosomes from carrying out repeated power flow calculating, improve speed of convergence; For convergence, stop criterion problem, the present invention adopts average fitness determining method as stop technology criterion, guarantees just to stop finishing after output global optimum, that is:

$|{\stackrel{\&OverBar;}{F}}^{\left(n\right)}-{\stackrel{\&OverBar;}{F}}^{(n-1)}|<\mathrm{\&epsiv;}---\left(9\right);$

Wherein, ε determines according to the fitness value of practical problems; with the average fitness value that represents respectively n generation and n-1 generation, n is natural number.

Finally should be noted that: above embodiment is only in order to illustrate that technical scheme of the present invention is not intended to limit, although the present invention is had been described in detail with reference to above-described embodiment, those of ordinary skill in the field are to be understood that: still can modify or be equal to replacement the specific embodiment of the present invention, and do not depart from any modification of spirit and scope of the invention or be equal to replacement, it all should be encompassed in the middle of claim scope of the present invention.

Claims (8)

1. an electric network reactive-load capacity configuration optimizing method for the improved genetic algorithms method based on WAMS, is characterized in that, the real-time information of described method based on electrical network wide area monitoring system WAMS and energy management system EMS, and described method comprises the steps:

Step S1: Real-time Obtaining electrical network measurement information from electrical network wide area monitoring system WAMS;

Step S2: obtain electric network composition, parameter and operation section information in the energy management system EMS of current time section from energy management system EMS, and form the whole network admittance matrix that comprises all load bus equivalent admittances;

Step S3: the control variable of reactive power optimization is carried out to hybrid coding;

Step S4: set up fitness function;

Step S5: carry out and select operator;

Step S6: carry out crossover and mutation operator;

Step S7: determine that speed of convergence and convergence stop criterion.

2. electric network reactive-load capacity configuration optimizing method as claimed in claim 1, is characterized in that, in described step S1, from electrical network wide area monitoring system WAMS, each vector measurement unit of Real-time Obtaining electrical network PMU installs the dynamic measurement information of website.

3. electric network reactive-load capacity configuration optimizing method as claimed in claim 1, it is characterized in that, in described step S2, from energy management system EMS, obtain the QS data file that comprises topological structure of electric, parameter and operation trend section, carry out trend calculating, to obtain the complete network admittance matrix of taking into account load equivalent admittance, for subsequent calculations voltage, fall the Dai Weinan equivalent parameters of out-of-limit node and do data preparation;

Electric network reactive-load capacity optimizes that to take the reduction the whole network network loss shown in following formula (1) objective function mathematical model expression formula be target, realizes the economical operation of the stable and electric system of voltage levvl:

$P_L=\underset{k=i,j}{\underset{k\&Element;\mathrm{Ng}}{\mathrm{\&Sigma;}}}{g}_k(U_i^2+U_j^2-2U_i{U}_j\cos {\mathrm{\&theta;}}_{\mathrm{ij}})---\left(1\right);$

In formula, P _lactive power loss for system; Ng is a way; g _kfor the electricity of branch road k is led; U _ivoltage magnitude for load bus i; θ _ijfor the phasing degree between load bus i and j;

Choose generator voltage, reactive compensation capacity and on-load transformer tap changer gear as control variable, choose generator reactive and exert oneself and economize on electricity voltage magnitude as state variable; All state variables all must meet variable bound condition, and the introducing by penalty function when state variable exceeds upper lower limit value is processed objective function, and its expression formula is:

$\begin{array}{c}F=\min (P_L+{\mathrm{\&lambda;}}_V\mathrm{\&Sigma;}\left(\frac{\mathrm{\&Delta;}{V}_i}{V_{i\max }-V_{i\min }}\right)+\\ {\mathrm{\&lambda;}}_Q\mathrm{\&Sigma;}{\left(\frac{\mathrm{\&Delta;}{Q}_{\mathrm{Gi}}}{Q_{\mathrm{Gi}\max }-Q_{\mathrm{Gi}\min }}\right)}^2)\end{array}---\left(2\right);$

In formula: λ _vfor the node voltage penalty factor that crosses the border; λ _qthe penalty factor that crosses the border for generator output reactive power; for the penalty function of crossing the border of node voltage, the penalty function of crossing the border for generator output reactive power; Δ V _i, V _imax, V _iminbe illustrated respectively in given " generation " deviate, maximum node voltage, the minimum node voltage of i node voltage and average; Δ Q _gi, Q _gimax, Q _giminbe illustrated respectively in given " generation ", the deviate of i generator output reactive power and average is, the minimum reactive power of the maximum reactive power of generator output, generator output;

Equation of constraint, the constraint such as comprises and does not wait equation of constraint; Wherein, waiting equation of constraint is the trend equilibrium equation of node, and expression formula is as follows:

$\left\{\begin{array}{c}{P}_i=V_i\underset{j\&Element;i}{\mathrm{\&Sigma;}}{V}_j(G_{\mathrm{ij}}\cos {\mathrm{\&delta;}}_{\mathrm{ij}}+B_{\mathrm{ij}}\sin {\mathrm{\&delta;}}_{\mathrm{ij}})\\ Q_i=V_i\underset{j\&Element;i}{\mathrm{\&Sigma;}}{V}_j(G_{\mathrm{ij}}\sin {\mathrm{\&delta;}}_{\mathrm{ij}}+B_{\mathrm{ij}}\cos {\mathrm{\&delta;}}_{\mathrm{ij}})\end{array}\right.---\left(3\right);$

In formula, P _i, Q _ibe respectively injection active power and the reactive power of node i; V _i, V _jbe respectively the voltage magnitude of node i and j; B _ij, G _ijand δ _ijbe respectively that electricity between node i and j is led, susceptance and voltage-phase angular difference; J ∈ i represents the node that all nodes are connected with node i;

Do not wait equation of constraint to comprise the equation of constraint such as or not does not wait equation of constraint and control variable of state variable;

The equation of constraint that do not wait of state variable is:

$\left\{\begin{array}{c}{Q}_{\mathrm{Gi}\min }\&le;Q_{\mathrm{Gi}}\&le;Q_{\mathrm{Gi}\max }\\ V_{\mathrm{di}\min }\&le;V_{\mathrm{di}}\&le;V_{\mathrm{di}\max }\end{array}\right.---\left(4\right);$

In formula, Q _gifor generator reactive is exerted oneself; Q _giminand Q _gimaxrepresent respectively the upper and lower bound that generator reactive is exerted oneself; V _difor load bus voltage magnitude; V _diminand V _dimaxthe upper and lower bound that represents respectively load bus voltage magnitude;

The equation of constraint that do not wait of control variable is:

$\left\{\begin{array}{c}{V}_{\mathrm{Gi}\min }\&le;V_{\mathrm{Gi}}\&le;V_{\mathrm{Gi}\max }\\ Q_{\mathrm{Ci}\min }\&le;Q_{\mathrm{Ci}}\&le;Q_{\mathrm{Ci}\max }\\ T_{\mathrm{ti}\min }\&le;T_{\mathrm{ti}}\&le;T_{\mathrm{ti}\max }\end{array}\right.---\left(5\right);$

In formula, V _gifor generator voltage; Q _cifor the reactive compensation capacity in node i; T _tion-load transformer tap changer gear;

Adopt P-Q decomposition method to carry out trend calculating.

4. electric network reactive-load capacity configuration optimizing method as claimed in claim 1, it is characterized in that, in described step S3, take real number and integer hybrid coding mode, be in the control variable of reactive power optimization, the set end voltage of generator is continuous variable, adopts real coding, node reactive-load compensation amount and on-load transformer tap changer are discrete variable, adopt integer coding.

5. electric network reactive-load capacity configuration optimizing method as claimed in claim 1, is characterized in that, in described step S4, adopts method reciprocal to set up fitness function, that is:

$F\left(x\right)=\frac{K}{f\left(x\right)-z+1}---\left(6\right);$

Wherein, F (x) is fitness value, and f (x) is objective function, and z is that experience is estimated minimum value, and K is enlargement factor, and 1 for preventing that denominator from being 0.

6. electric network reactive-load capacity configuration optimizing method as claimed in claim 1, is characterized in that, in described step S5, and the mixing selection strategy that adopts tournament method and optimum preservation method to combine; First, by championship, in N random individual, optimum individuality is selected, then carried out the preservation of optimum individual, the individuality of fitness maximum in parent is replaced to the poorest individuality of fitness in filial generation, the advantage genes of individuals of making is continued.

7. electric network reactive-load capacity configuration optimizing method as claimed in claim 1, is characterized in that, in described step S6, adopt new crossover and mutation mode, be crossover probability Pc and variation probability P m, the change of crossover probability and variation probability is along with fitness changes automatically, and expression formula is as follows:

$P_c=\left\{\begin{array}{cc}{P}_{c1}-\frac{(P_{c1}-P_{c2})(f_c-f_p)}{f_{\max }-f_p},& (f_c\&GreaterEqual;f_p)\\ P_{c1}& (f_c<f_p)\end{array}\right.---\left(7\right);$

$P_m=\left\{\begin{array}{cc}{P}_{m1}-\frac{(P_{m1}-P_{m2})(f_m-f_p)}{f_{\max }-f_p},& (f_m\&GreaterEqual;f_p)\\ P_{m1}& (f_m<f_p)\end{array}\right.---\left(8\right);$

In formula: f _maxfor the maximum adaptation degree value in population; f _pfor the average fitness value in population; f _cfor intersecting individual fitness value; f _mfor the individual fitness value that makes a variation; P _c1for produced first crossover probability value, P _c2for produced second crossover probability value, P _m1for produced the 1st variation probable value, P _m2for the 2nd produced variation probable value.

8. electric network reactive-load capacity configuration optimizing method as claimed in claim 1, it is characterized in that, in described step S7, described speed of convergence is according to the mode of taboo question blank, set up chromosome question blank, to per generation population carry out chromosome inquiry, storage, prevent that phase homologous chromosomes from carrying out repeated power flow calculating, for improving speed of convergence; Adopt average fitness value determining method as stop technology criterion, guarantee just to stop finishing after output global optimum, that is:

$|{\stackrel{\&OverBar;}{F}}^{\left(n\right)}-{\stackrel{\&OverBar;}{F}}^{(n-1)}|<\mathrm{\&epsiv;}---\left(9\right);$

Wherein, ε determines according to the fitness value of practical problems; with the average fitness value that represents respectively n generation and n-1 generation, n is natural number.