CN106655190A - Method for solving P-OPF (Probabilistic-Optimal Power Flow) of wind power stations - Google Patents

Abstract

The invention relates to a method for solving a P-OPF of wind power stations. The method comprise that a P-OPF model is established, a random variable X is input, and the random variable X represents the wind speeds vi of the different wind power stations and uncertain components P<Li> and Q<Li> of load power; an orthogonal transformation technology is used to linearly transform the random variable X into an uncorrelated random variable Z, and an orthogonal transformation matrix B is obtained; a mean value muZ, a variance sigmaZ, a deviation parameter lambda<z,3>=( lambda<z1,3>, lambda<z2,3>,... lambda<zn,3>)<T> and a kurtosis parameter lambda<z,4>=( lambda<z1,4>, lambda<z2,4>,... lambda<zn,4>)<T> of the uncorrelated random variable Z are calculated, a position parameter xi<k,i> and a probability p<k,i> of different estimation points of the uncorrelated random variable Z are determined according to a point estimation method, and m estimation points Z<k,i> are constructed; the m estimation points Z<k,i> of the uncorrelated random variable Z are transformed inversely to estimation points of the random variable X; and according to each estimation point of the random variable X, the P-OPF model is solved in a modern interior point method. The method for solving the P-OPF of wind power stations maintains higher precision and consumes less time.

Description

A kind of method for solving wind energy turbine set probability optimal load flow

Technical field

The present invention relates to wind energy turbine set probability optimal load flow technical field, more particularly to a kind of optimum tide of solution wind energy turbine set probability The method of stream.

Background technology

Trend optimum is widely used in Power System Planning and real-time operation, however, a large amount of accesses of wind-powered electricity generation New requirement is proposed to traditional optimal load flow technology, the optimum that two classes consider power system enchancement factor has thus been developed Power Flow Problem：Probability optimal load flow and random optimum Power Flow Problem.Random optimum Power Flow Problem is that (probability is about by chance constraint Beam) enchancement factor is considered into optimal load flow model, it is clear that and enchancement factor will affect calculating process and final result.Probability is most Excellent Power Flow Problem is then the probability-distribution function according to stochastic variable in power system, obtains optimal load flow some state variables The probability distribution of (conventional power plants are exerted oneself, branch power, node voltage amplitude and phase angle etc.).Wind power output power has fluctuation With intermittent feature, with the continuous expansion of wind-electricity integration scale, the impact that electrical network is subject to also increasingly increases.Wind energy turbine set is ultrashort Phase power prediction provides the reference of unit output ability for active power of wind power field control, helps electrical network to formulate rational scheduling meter Draw, enable wind energy turbine set to receive the United Dispatching of electrical network.

In correlation technique, domestic ultra-short term power forecasting method will mainly pass by several points (such as nearest a hour) Actual wind data of surveying is contrasted with the numerical weather forecast wind speed and direction of several points in the past, and with this number of following several points is corrected Value weather forecast wind speed and direction, then calculates power using revised numerical weather forecast wind speed and direction.Due to the method not Using actual power, can preferably adapt to abandon wind field scape, at home using relatively broad.

But, because China is with a varied topography and weather condition difference is larger, ripe forecast system cannot be applied directly, be surpassed Short-term wind-electricity power precision of prediction is relatively low, therefore, research is adapted to the wind energy turbine set ultra-short term work(of China's topography and geomorphology and weather conditions Rate Predicting Technique is significant to China's extensive development wind-powered electricity generation.

The content of the invention

To overcome problem present in correlation technique, the present invention to provide a kind of side for solving wind energy turbine set probability optimal load flow Method, including：

Probability optimal load flow model is set up, stochastic variable X is input into, stochastic variable X is each wind farm wind velocity v_iAnd it is negative The uncertain component P of lotus power_LiAnd Q_Li；

Using orthogonal transformation technology, it is determined that the stochastic variable x-ray is transformed into uncorrelated random variables Z, and obtain Orthogonal transform matrix B；

Calculate the mean μ of uncorrelated random variables Z_Z, variances sigma_Z, straggling parameterKurtosis Parameter

Determine the location parameter ξ of each estimation points of uncorrelated random variables Z according to point estimations_k,iAnd Probability p_k,i, construct m Individual estimation point Z_k,i；

By the m estimation point Z of uncorrelated random variables Z_k,iIt is inversely transformed into the estimation point of stochastic variable X；

According to each estimation point of stochastic variable X, using modern interior-point method optimal load flow model is solved.

Preferably, the utilization orthogonal transformation technology, it is determined that the stochastic variable x-ray is transformed into uncorrelated random change Amount Z, and orthogonal transform matrix B is obtained, including：

Stochastic inputs variable X=(x₁,x₂,...,x_n), mean μ_x=(μ₁,μ₂,...,μ_n)^T, variances sigma_x=(σ₁,σ₂,..., σ_n)^TIf, x in stochastic variable X_k1And x_k2Between coefficient correlation be ρ_k1k2, then covariance matrix：

The degree of bias parameter of stochastic variable X isKurtosis parameter is

Stochastic variable X is converted into by uncorrelated random variables Z by Z=BX；

Using square-root method to covariance matrix C_XCarry out decomposition and obtain C_X=LL^T；

Transformation matrix B is derived according to Z=BX：

Preferably, the location parameter ξ that each estimation points of uncorrelated random variables Z are determined according to point estimations_k,iAnd probability p_k,i, construct m estimation point Z_k,i, including：

According to arbitrary model F, stochastic variable y to be asked and stochastic variable X=(x₁,x₂,...,x_n) mapping relations can represent For：

Y=F (X)=F (x₁,x₂,...,x_n), wherein, each stochastic variable mean μ=(μ₁,μ₂,...,μ_n)^T, variances sigma= (σ₁,σ₂,...,σ_n)^T；

Each stochastic variable x_k(k=1 ..., n) on take m point x_k,i(k=1 ..., n；I=1 ..., m) construct m Individual estimation point x_k,iFor：

x_k,i=μ_k+ξ_k,iσ_k, wherein ξ_k,iFor location parameter, if it is p that stochastic variable is taken at the probability of each estimation point_k,i(k= 1,...,n；I=1 ..., m), then：

Make λ_k,jFor stochastic variable jth rank central moment M_j(x_k) and standard deviation sigma_kJ powers ratio, i.e.,：

F (x in formula_k) it is stochastic variable x_kProbability density function, by formulaCan Know, λ_k,1=0, λ_k,2=1, and λ_k,3And λ_k,4It is respectively stochastic variable x_kDeviation and kurtosis；

By y=F (X) in x_kMean μ_kPlace's Taylor series expansion, uses successively λ_k,jY is estimated on m point, can be obtained ArriveWherein, j=1,2 ..., 2m-1, k=1,2 ..., n.

Preferably, described each estimation point according to stochastic variable X, using modern interior-point method optimum tide is solved Flow model, including：

The deterministic models of each estimation point are calculated, values y (k, i) of the y in each estimation point is obtained, and then obtains y Each rank square estimate：

Preferably, it is described to set up probability optimal load flow model, including：

With the minimum object function of generating expense, it is considered to which power flow equation equality constraint and system operation, physical limit are not Equality constraint, and count Large Scale Wind Farm Integration and exert oneself at random, setting up the Probabilistic optimal load flow model containing wind power integration is：

In formula, S_BFor node set, S_GFor generator node set, S_RFor reactive source set, S_WTo access the section of wind energy turbine set Point is combined, P_GiFor the active power that node i normal power supplies send, Q_RiFor the reactive power that all kinds of reactive sources of node i send, P_Wi And Q_WiActive, reactive power, P are sent for node i wind energy turbine set_LiAnd Q_LiFor node i load is active, reactive power；

V_iAnd δ_iFor node voltage amplitude and phase angle, Y_ijFor bus admittance matrix element, a_ijIt is corresponding for bus admittance matrix Element phase angle, δ_ij=δ_i-δ_j-a_ij；

P _GiFor P_GiCorrespondence bound, P _WiFor P_WiCorrespondence bound, Q _RiFor Q_RiCorrespondence bound, V _iIt is node i voltage magnitude bound.

Preferably, each wind farm wind velocity v_iSolved according to the probability density function of node i wind farm wind velocity：

Wherein, field gas velocity v_iIt is according to the probability density function of node i wind farm wind velocity：

In formula, K is the form parameter of Weibull distribution, and C is scale parameter.

Preferably due to the error of the aspect such as power system measuring, estimation, real system part of nodes load power it is pre- Indeterminacy is true, and with randomness；

Node i load component P_LiAnd Q_LiIt is uncertain, and meet with ground state load power μ_PLiFor average, with σ₁For standard deviation Normal distribution, then P_LiProbability density function be：

Load power factor is constant, the uncertain component Q of node i load_LiBy P_LiIt is determined that.

Preferably, the P_LiAnd Q_LiSolution include：

The power output of wind energy turbine set depends on the power output of each typhoon group of motors in wind energy turbine set, and the generating of Wind turbines Power changes with the fluctuation of wind speed, and the relation between the generated output and wind speed of the Wind turbines is：

In formula, v wind speed, v_inTo cut wind speed, v_outFor cut-out wind speed, P_rFor wind-powered electricity generation The rated output power of unit, P_WgFor Wind turbines real output；

WithFor constant；

The Power Output for Wind Power Field of node i is：P_Wi=N_WiP_Wgi, in formula, N_WiIt is the Wind turbines platform of node i wind energy turbine set Number.

The technical scheme that embodiments of the invention are provided can include following beneficial effect：

The method for solving wind energy turbine set probability optimal load flow provided in an embodiment of the present invention, establishes the wind for considering that wind speed is related The Probabilistic optimal load flow computation model that electric field is accessed, proposes using the stochastic variable of orthogonal transformation technical finesse, shape Into the point estimations based on orthogonal transformation.Probability optimal power flow problems are converted into deterministic optimization by point estimations Problem, is then solved using modern interior-point method.Compared with prior art, the optimum tide of solution wind energy turbine set probability disclosed by the invention While degree of precision is kept, elapsed time is less, has broad application prospects for stream method.

It should be appreciated that the general description of the above and detailed description hereinafter are only exemplary and explanatory, not The present invention can be limited.

Description of the drawings

Accompanying drawing herein is merged in specification and constitutes the part of this specification, shows the enforcement for meeting the present invention Example, and be used to explain the principle of the present invention together with specification.

In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, for those of ordinary skill in the art Speech, without having to pay creative labor, can be with according to these other accompanying drawings of accompanying drawings acquisition.

Fig. 1 is a kind of method flow schematic diagram for solving wind energy turbine set probability optimal load flow provided in an embodiment of the present invention；

Fig. 2 is replicated respectively for provided in an embodiment of the present invention with point estimations and the calculated voltage of monte carlo method Probability density curve；

The active power cumulative probability that Fig. 3 point estimations and monte carlo method provided in an embodiment of the present invention are obtained is bent Line.

Specific embodiment

Here exemplary embodiment will be illustrated in detail, its example is illustrated in the accompanying drawings.Explained below is related to During accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represent same or analogous key element.Following exemplary embodiment Described in embodiment do not represent and the consistent all embodiments of the present invention.Conversely, they be only with it is such as appended The example of the consistent apparatus and method of some aspects described in detail in claims, the present invention.

Probability optimal power flow problems of the present invention for Large Scale Wind Farm Integration after grid-connected, establish the wind-powered electricity generation for considering that wind speed is related The Probabilistic optimal load flow computation model that field is accessed.Propose using the related stochastic variable of orthogonal transformation technical finesse, Define based on the point estimations of orthogonal transformation.Probability optimal power flow problems are converted into deterministic optimum by point estimations Change problem, is then solved using modern interior-point method.

A kind of method flow schematic diagram of solution wind energy turbine set probability optimal load flow that Fig. 1 is provided for inventive embodiments.

Before carrying out solving wind energy turbine set probability optimal load flow, the probability optimal load flow model containing wind energy turbine set is initially set up. In the step s 100, probability optimal load flow model is set up, stochastic variable X is input into, stochastic variable X is each wind farm wind velocity v_i And the uncertain component P of load power_LiAnd Q_Li。

Set up probability optimal load flow model process as follows：

With the minimum object function of generating expense, it is considered to which power flow equation equality constraint and system operation, physical limit are not Equality constraint, and count Large Scale Wind Farm Integration and exert oneself at random, set up the Probabilistic optimal load flow model containing wind power integration and join See formula (1) and formula (2).

Formula (1)：

Formula (2)

In formula, S_BFor node set, S_GFor generator node set, S_RFor reactive source set, S_WTo access the section of wind energy turbine set Point is combined, P_GiFor the active power that node i normal power supplies send, Q_RiFor the reactive power that all kinds of reactive sources of node i send, P_Wi And Q_WiActive, reactive power, P are sent for node i wind energy turbine set_LiAnd Q_LiFor node i load is active, reactive power；

V_iAnd δ_iFor node voltage amplitude and phase angle, Y_ijFor bus admittance matrix element, a_ijIt is corresponding for bus admittance matrix Element phase angle, δ_ij=δ_i-δ_j-a_ij；

P _GiFor P_GiCorrespondence bound, P _WiFor P_WiCorrespondence bound, Q _RiFor Q_RiCorrespondence bound, V _iIt is node i voltage magnitude bound.

Wind energy turbine set stochastic model is set up, process is as follows：

The power output of wind energy turbine set depends on the power output of each typhoon group of motors in wind energy turbine set, and the generating of Wind turbines Power changes with the fluctuation of wind speed, and the relation between Wind turbines and wind speed is referring to formula 3.

Formula (3)：

In formula (3), v wind speed, v_inTo cut wind speed, v_outFor cut-out wind speed, P_rFor the rated output power of Wind turbines, P_WgFor Wind turbines real output,WithFor constant；

The Power Output for Wind Power Field of node i is referring to formula (4)：

P_Wi=N_WiP_Wgi；

In formula (4), N_WiIt is the Wind turbines number of units of node i wind energy turbine set.

At present, double-fed asynchronous generator or direct drive generator is adopted to be trends of the times in Large Scale Wind Farm Integration, therefore this Bright method is applied to the probability optimal power flow problems of such Wind turbines access system.It is assumed that Wind turbines with firm power because Number mode is run, and node i wind energy turbine set absorbing reactive power is referring to formula (5).

Formula (5)：Q_Wi=P_Witanθ_i；

In formula, θ_iFor the power of fan factor angle of node i wind energy turbine set.

A large amount of measured datas show that the wind speed in an area approximately obeys Two-parameter Weibull Distribution, node i wind energy turbine set Probability gathers function referring to formula (6)：

Formula (6)：

In formula, K is the form parameter of Weibull distribution, and C is scale parameter.

Each wind farm wind velocity v_iCan obtain from above-mentioned wind energy turbine set stochastic model, each data for getting are stochastic variable X In parameter.

In the probability optimal load flow model of foundation, in addition it is also necessary to load power stochastic variable.Load power stochastic variable includes The uncertain component P of load power_LiAnd Q_Li, the uncertain component P of load power_LiAnd Q_LiCan obtain from load power stochastic model.

Due to the error of the aspects such as power system measuring, estimation, the forecasting inaccuracy of real system part of nodes load power Really, with randomness.Using node load power as stochastic variable, and its randomness can be answered with normal distribution approximate reverse.

Node i load air quantity P_LiAnd Q_LiIt is uncertain, and meet with ground state load power μ_PLiFor average, with σ₁For standard deviation Normal distribution, then P_LiProbability density function be：

Formula (7)：

Load power factor is constant, the uncertain component Q of node i load_LiBy P_LiIt is determined that.

In step s 200, using orthogonal transformation technology, it is determined that the stochastic variable x-ray is transformed to uncorrelated random Variable Z, and obtain orthogonal transform matrix B.

Point estimations cannot directly process the related situation of stochastic variable, if stochastic variable X is not separate, can adopt Orthogonal transformation technology, by original related stochastic variable x-ray uncorrelated random variables is transformed to.

Stochastic inputs variable X=(x₁,x₂,...,x_n), mean μ_x=(μ₁,μ₂,...,μ_n)^T, variances sigma_x=(σ₁,σ₂,..., σ_n)^TIf, x in stochastic variable X_k1And x_k2Between coefficient correlation be ρ_k1k2, then covariance matrix is referring to formula (8)：

The degree of bias parameter of stochastic variable X is

Kurtosis parameter is

C_XFor symmetrical matrix, there is orthogonal transform matrix B, by formula (9)：Z=BX, stochastic variable X is converted into Uncorrelated random variables Z；

In Practical Project, C_XGenerally symmetric positive definite matrix, using square-root method to covariance matrix C_XCarry out decomposition to obtain C_X=LL^T；

Transformation matrix B is derived according to Z=BX：

Stochastic variable Z after conversion has following features：

Stochastic variable Z mean μ_Z=B μ_X, the covariance matrix C of Z_ZFor unit matrix, i.e.,：C_Z=I.

The straggling parameter and kurtosis parameter of Z can be obtained by formula (10), formula (10)：

In formula, B_lrThe element of representing matrix B l rows r row.

In step S300, the mean μ of uncorrelated random variables Z is calculated_Z, variances sigma_Z；

Straggling parameterKurtosis parameter

According to μ_Z=B μ_XAnd C_Z=I calculates mean μ_Z, variances sigma_Z, straggling parameterPeak Degree parameter

In step S400, according to point estimations the location parameter ξ of each estimation points of uncorrelated random variables Z is determined_k,iWith it is general Rate p_k,i, construct m estimation point Z_k,i。

The thought of point estimations is in known input stochastic variable probability characteristics parameter (average, standard deviation, kurtosis, the degree of bias Deng) under conditions of, the estimate of each rank square for obtaining stochastic variable to be asked is calculated by the deterministic models of limited number of time, to determine The probability characteristics parameter of stochastic variable to be asked, so as to its stochastic behaviour of quantitative analysis.

Step S401, according to arbitrary model F, stochastic variable y to be asked and stochastic inputs variable X=(x₁,x₂,...,x_n) reflect The relation of penetrating is represented by：

Formula (11)：Y=F (X)=F (x₁,x₂,...,x_n), wherein, each stochastic variable mean μ=(μ₁,μ₂,...,μ_n)^T, Variances sigma=(σ₁,σ₂,...,σ_n)^T；

Step S402, each stochastic variable x_k(k=1 ..., n) on take m point x_k,i(k=1 ..., n；I=1 ..., M) m estimation point x is constructed_k,iFor：

Formula (12)：x_k,i=μ_k+ξ_k,iσ_k, wherein ξ_k,iFor location parameter, if the probability that stochastic variable is taken at each estimation point is p_k,i(k=1 ..., n；I=1 ..., m), then, and formula (13)：

Make λ_k,jFor stochastic variable jth rank central moment M_j(x_k) and standard deviation sigma_kJ powers ratio, i.e.,：

Formula (14)：

F (x in formula_k) it is stochastic variable x_kProbability density function, from formula formula (14), λ_k,1=0, λ_k,2=1, and λ_k,3 And λ_k,4It is respectively stochastic variable x_kDeviation and kurtosis；

Step S403, by y=F (X) in x_kMean μ_kPlace's Taylor series expansion, uses successively λ_k,jY is carried out on m point Estimate, be obtainedWherein, j=1,2 ..., 2m-1, k=1,2 ..., n.

According to formula (14), λ can be directly calculated_k,j, by simultaneous formula (123) and formula (14), obtain stochastic variable and respectively estimating The probability P of enumeration value_k,i, and the location parameter ξ of each estimation point_k,i, m estimation point x is constructed by formula (12)_k,i。

In step S500, by the m estimation point Z of uncorrelated random variables Z_k,iIt is inversely transformed into estimating for stochastic variable X Enumeration.

According to formula (9) by the m estimation point Z of uncorrelated random variables Z_k,iIt is inversely transformed into the estimation of stochastic variable X Point.

In step S600, according to each estimation point of stochastic variable X, solve using modern interior-point method optimum Tide model.

The deterministic models of each estimation point are calculated, values y (k, i) of the y in each estimation point is obtained, and then obtains y Each rank square estimate, referring to formula (15)：

In formula (15), during l=1, E (y) is the mean μ of stochastic variable y to be asked_y, work as l=2, the standard deviation of y can be calculated

Estimate that points m is higher, treat and ask the estimated accuracy of stochastic variable y higher, but can so bring solution difficulty, calculate The low problem of efficiency.In actual solution, m=2 or m=3 two schemes are typically chosen.During using m=2, location parameter ξ is solved_k,i And Probability p_k,iOnly make use of λ_k,3Degree of bias parameter information, calculation error is larger, therefore selects m in method provided in an embodiment of the present invention =3 schemes, one of estimation point is taken as stochastic variable average, the ξ of calculated each estimation point_k,iAnd p_k,iIt is as follows：

Formula (16)：

Formula (17)：

Although constructing 3n estimation point, wherein n estimation point is identical, so only needing to carry out 2n+1 model meter Calculate.

Seen from the above description, method disclosed by the invention considers the Probabilistic that the related wind energy turbine set of wind speed is accessed Optimal load flow computation model, and propose, using the related stochastic variable of orthogonal transformation technical finesse, to define based on orthogonal transformation Point estimations.Probability optimal power flow problems are converted into deterministic optimization problem by point estimations, then using interior point Send out and solved.

The following is the validation verification that method is disclosed the present invention, using Yunnan somewhere network system, the contrast present invention Disclosed method and monte carlo method.

1st, system outline

In Yunnan somewhere, the node 23 and 39 of network system accesses two wind energy turbine sets, with method disclosed by the invention and illiteracy Special chucking method calculates probability optimal load flow, and the checking present invention discloses the validity of method.

The wind energy turbine set basic parameter for being accessed such as table 1.The load power of 50 nodes is random in hypothesis system, obeys with original First given load power is average, the normal distribution that standard deviation is 5% average.

The wind-powered electricity generation field parameters of table 1

Wind energy turbine set	Blower fan number of units	Rated capacity	Incision wind speed	Cut-out wind speed	Rated wind speed	K	C
										MW	(m/s)	(m/s)	(m/s)		(m/s)
1	100	0.75	4.0	25	15.0	1.4	6.0
								2	50	1.50	3.0	30	14.0	1.8	7.0

2nd, point estimations validation verification

5000 Monte Carlo simulations standard as a comparison is selected, verifies that point estimations calculate the effective of probability optimal load flow Property.The result of calculation for making point estimations and monte carlo method is respectively H_PEAnd H_MC, ε is the relative error of the two, i.e.,：

ε=| H_MC-H_PE|/H_MC× 100%

It is the average and standard deviation of the cost of electricity-generating that two methods of point estimation and Monte Carlo are obtained in table 2, it is clear that point is estimated The computational accuracy of meter method is very high.It is assumed here that the wind speed of the wind energy turbine set for accessing is uncorrelated.Situation contrast, wind-powered electricity generation are not accessed with wind-powered electricity generation Coal consumption cost, the pressure that alleviating energy crisis bring can be reduced after access.Table 3 is that point estimations are calculated with monte carlo method As a result mean error, the error of average is less than 0.5%, and the error of standard deviation is less than 4%.

Table 2：The average and standard deviation of object function

The mean error of table 3

Shown in Fig. 2 is that Yunnan somewhere grid nodes system is calculated respectively with point estimations and monte carlo method The probability density curve of the voltage magnitude of node 15 for arriving.Obviously, point estimations result is fine with Monte Carlo simulation degree of fitting, very The probability distribution of good reflection node voltage.With the point estimation method and monte carlo method obtain branch road 34-37's respectively in Fig. 3 The cumulative probability curve of active power.It can be seen that, cumulative probability curve is obtained with higher precision, Ke Yiman with the point estimation method The demand of sufficient practical application.

3rd, impact of the wind speed correlation to probability optimal load flow

It is the average and standard deviation of the cost of electricity-generating that two methods of point estimation and Monte Carlo are obtained in table 4.Wind energy turbine set wind Under fast 3 kinds of degrees of correlation, the computational accuracy of point estimations is all satisfactory.It is point estimations and monte carlo method in table 5 The mean error of result of calculation, with the increase of wind speed coefficient correlation, the degree of correlation of wind farm wind velocity increases, the optimum tide of probability The average and standard deviation of each state variable of stream all increased, but as a result be still acceptable.

The average and standard deviation of the object function of table 4

The randomness of two wind energy turbine sets and 50 node loads is considered in the probability optimal load flow model of this paper, 52 are had Individual stochastic variable.The sampling number needed for the point estimations of this paper 2n+1 patterns is adopted for 105, the required calculating time is 6.319s.The calculating time of 5000 monte carlo methods is 306.905s, be only the time required to point estimations its 2.059%.

The mean error of table 5

From above-mentioned contrast verification, the access of large-scale wind power field brings a large amount of uncertain factors to power system, Launch research for the probability optimal power flow problems containing Large Scale Wind Farm Integration, establish the electricity for considering that the related wind energy turbine set of wind speed is accessed Force system probability optimal load flow computation model.Using the related stochastic variable of orthogonal transformation technical finesse, it is proposed that based on orthogonal The point estimations of conversion.Probability optimal power flow problems are converted into by deterministic optimization problem by the point estimation method, then Solved using interior point method.The result of calculation of Yunnan somewhere network system shows, compared with monte carlo method, the method While higher computational accuracy is kept, less calculating time is consumed, with good application prospect.

Those skilled in the art will readily occur to its of the present invention after considering specification and putting into practice disclosure of the invention here Its embodiment.The application is intended to any modification of the present invention, purposes or adaptations, these modifications, purposes or Person's adaptations follow the general principle of the present invention and including the undocumented common knowledge in the art of the present invention Or conventional techniques.Description and embodiments are considered only as exemplary, and true scope and spirit of the invention are by following Claim is pointed out.

It should be appreciated that the precision architecture for being described above and being shown in the drawings is the invention is not limited in, and And can without departing from the scope carry out various modifications and changes.The scope of the present invention is only limited by appended claim.

Claims (8)

1. it is a kind of solve wind energy turbine set probability optimal load flow method, it is characterised in that include：

Probability optimal load flow model is set up, stochastic variable X is input into, stochastic variable X is each wind farm wind velocity v_iAnd load power Uncertain component P_LiAnd Q_Li；

Using orthogonal transformation technology, it is determined that the stochastic variable x-ray is transformed into uncorrelated random variables Z, and obtain orthogonal Transformation matrix B；

Calculate the mean μ of the uncorrelated random variables Z_Z, variances sigma_Z, straggling parameterKurtosis Parameter

Determine the location parameter ξ of each estimation points of the uncorrelated random variables Z according to point estimations_k,iAnd Probability p_k,i, construct m Individual estimation point Z_k,i；

By the m estimation point Z of the uncorrelated random variables Z_k,iIt is inversely transformed into the estimation point of stochastic variable X；

According to each estimation point of stochastic variable X, using modern interior-point method the probability optimal load flow model is solved.

2. it is according to claim 1 solve wind energy turbine set probability optimal load flow method, it is characterised in that it is described using orthogonal Converter technique, it is determined that the stochastic variable x-ray is transformed into uncorrelated random variables Z, and obtains orthogonal transform matrix B, bag Include：

The random input variable X=(x₁,x₂,...,x_n), mean μ_x=(μ₁,μ₂,...,μ_n)^T, variances sigma_x=(σ₁,σ₂,..., σ_n)^TIf, x in stochastic variable X_k1And x_k2Between coefficient correlation be ρ_k1k2, then the covariance matrix：

$C_X=\left[\begin{array}{cccc}{\mathrm{\&sigma;}}_1^2& {\mathrm{\&rho;}}_{12}{\mathrm{\&sigma;}}_1{\mathrm{\&sigma;}}_2& ...& {\mathrm{\&rho;}}_{1n}{\mathrm{\&sigma;}}_1{\mathrm{\&sigma;}}_n\\ {\mathrm{\&rho;}}_{12}{\mathrm{\&sigma;}}_2{\mathrm{\&sigma;}}_1& {\mathrm{\&sigma;}}_2^2& ...& {\mathrm{\&rho;}}_{2n}{\mathrm{\&sigma;}}_2{\mathrm{\&sigma;}}_n\\ ...& ...& ...& ...\\ {\mathrm{\&rho;}}_{1n}{\mathrm{\&sigma;}}_n{\mathrm{\&sigma;}}_1& {\mathrm{\&rho;}}_{2n}{\mathrm{\&sigma;}}_n{\mathrm{\&sigma;}}_2& ...& {\mathrm{\&sigma;}}_n^2\end{array}\right];$

The degree of bias parameter of stochastic variable X isKurtosis parameter is

Stochastic variable X is converted into by uncorrelated random variables Z by Z=BX；

Using square-root method to the covariance matrix C_XCarry out decomposition and obtain C_X=LL^T；

Transformation matrix B is derived according to the Z=BX：

$C_Z=\mathrm{cov}\left(Z,Z^T\right)=\mathrm{cov}\left(BX,X^T{B}^T\right)=B\mathrm{cov}\left(X,X^T\right)B^T={\mathrm{BC}}_X{B}^T=B\left({\mathrm{LL}}^T\right)B^T=I\&DoubleRightArrow;B=L^{-1}.$

3. it is according to claim 1 solve wind energy turbine set probability optimal load flow method, it is characterised in that it is described to be estimated according to point Meter method determines the location parameter ξ of each estimation points of the uncorrelated random variables Z_k,iAnd Probability p_k,i, construct m estimation point Z_k,i, Including：

According to arbitrary model F, stochastic variable y to be asked and stochastic variable X=(x₁,x₂,...,x_n) mapping relations are represented by：

Y=F (X)=F (x₁,x₂,...,x_n), wherein, each stochastic variable mean μ=(μ₁,μ₂,...,μ_n)^T, variances sigma=(σ₁, σ₂,...,σ_n)^T；

Each stochastic variable x_k(k=1 ..., n) on take m point x_k,i(k=1 ..., n；I=1 ..., m) construct m estimation Point x_k,iFor：

x_k,i=μ_k+ξ_k,iσ_k, wherein ξ_k,iFor location parameter, if it is p that stochastic variable is taken at the probability of each estimation point_k,i(k= 1,...,n；I=1 ..., m), then：

Make λ_k,jFor stochastic variable jth rank central moment M_j(x_k) and standard deviation sigma_kJ powers ratio, i.e.,：

$\left\{\begin{array}{c}{M}_j\left(x_k\right)={\&Integral;}_{-\mathrm{\&infin;}}^{+\mathrm{\&infin;}}{\left(x_k-{\mathrm{\&mu;}}_k\right)}^{j}f\left(x_k\right){\mathrm{dx}}_k\\ {\mathrm{\&lambda;}}_{k,j}=\frac{M_j\left(x_k\right)}{{\left({\mathrm{\&sigma;}}_k\right)}^j},j=1,2,3,\mathrm{...}\end{array}\right.;$

F (x in formula_k) it is stochastic variable x_kProbability density function, by formulaUnderstand, λ_k,1=0, λ_k,2=1, and λ_k,3And λ_k,4It is respectively stochastic variable x_kDeviation and kurtosis；

By y=F (X) in x_kMean μ_kPlace's Taylor series expansion, uses successively λ_k,jY is estimated on m point, is obtainedWherein, j=1,2 ..., 2m-1, k=1,2 ..., n.

4. it is according to claim 3 solve wind energy turbine set probability optimal load flow method, it is characterised in that described in the basis Each estimation point of stochastic variable X, using modern interior-point method optimal load flow model is solved, including：

The deterministic models of each estimation point are calculated, values y (k, i) of the y in each estimation point is obtained, and then obtains each of y The estimate of rank square：

5. it is according to claim 1 solve wind energy turbine set probability optimal load flow method, it is characterised in that it is described to set up probability Optimal load flow model, including：

With the minimum object function of generating expense, it is considered to power flow equation equality constraint and system operation, the inequality of physical limit Constrain, and count Large Scale Wind Farm Integration and exert oneself at random, setting up the Probabilistic optimal load flow model containing wind power integration is：

$\begin{array}{cc}\min & F_{Cost}=\underset{i\&Element;S_G}{\mathrm{\&Sigma;}}\left(a_{0,i}+a_{1,i}{P}_{Gi}+a_{2,i}{P}_{Gi}^2\right)\end{array};$

$s.t\left\{\begin{array}{c}\begin{array}{cc}{U}_i\underset{j\&Element;i}{\mathrm{\&Sigma;}}{Y}_{ij}{U}_j{\mathrm{cos\&delta;}}_{ij}+P_{Li}-P_{Gi}-P_{Wi}=0& \\ U_i\underset{j\&Element;i}{\mathrm{\&Sigma;}}{Y}_{ij}{U}_j{\mathrm{cos\&delta;}}_{ij}+Q_{Li}-Q_{Gi}-Q_{Wi}=0& i,j\&Element;S_B\end{array}\\ \begin{array}{cc}{\underset{\&OverBar;}{P}}_{Gi}\&le;P_{Gi}\&le;{\stackrel{\&OverBar;}{P}}_{Gi}& i\&Element;S_G\\ {\underset{\&OverBar;}{P}}_{Wi}\&le;P_{Wi}\&le;{\stackrel{\&OverBar;}{P}}_{Wi}& i\&Element;S_W\\ {\underset{\&OverBar;}{Q}}_{Ri}\&le;Q_{Ri}\&le;{\stackrel{\&OverBar;}{Q}}_{Ri}& i\&Element;S_R\\ {\underset{\&OverBar;}{V}}_i\&le;V_i\&le;{\stackrel{\&OverBar;}{V}}_i& i\&Element;S_B\end{array}\end{array}\right.;$

In formula, S_BFor node set, S_GFor generator node set, S_RFor reactive source set, S_WTo access the node knot of wind energy turbine set Close, P_GiFor the active power that node i normal power supplies send, Q_RiFor the reactive power that all kinds of reactive sources of node i send, P_WiAnd Q_Wi Active, reactive power, P are sent for node i wind energy turbine set_LiAnd Q_LiFor node i load is active, reactive power；

V_iAnd δ_iFor node voltage amplitude and phase angle, Y_ijFor bus admittance matrix element, a_ijFor bus admittance matrix respective element Phase angle, δ_ij=δ_i-δ_j-a_ij；

P _GiFor P_GiCorrespondence bound, P _WiFor P_WiCorrespondence bound, Q _RiFor Q_RiCorrespondence bound, V _iIt is Node i voltage magnitude bound.

6. it is according to claim 1 solve wind energy turbine set probability optimal load flow method, it is characterised in that each wind energy turbine set Wind speed v_iSolved according to the probability density function of node i wind farm wind velocity：

Wherein, field gas velocity v_iIt is according to the probability density function of node i wind farm wind velocity：

$f\left(v_i\right)=\frac{K}{C}{\left(\frac{v_i}{K}\right)}^{K-1}\exp \left\{-{\left(\frac{v_i}{C}\right)}^K\right\};$

In formula, K is the form parameter of Weibull distribution, and C is scale parameter.

7. it is according to claim 1 solve wind energy turbine set probability optimal load flow method, it is characterised in that due to power system The error of the aspects such as measurement, estimation, the forecasting inaccuracy of real system part of nodes load power is true, and with randomness；

Node i load component P_LiAnd Q_LiIt is uncertain, and meet with ground state load power μ_PLiFor average, with σ₁For standard deviation just State is distributed, then P_LiProbability density function be：

$f\left(P_{Li}\right)=\frac{1}{\sqrt{2\mathrm{\&pi;}}{\mathrm{\&sigma;}}_i}\exp \left(-\frac{{\left(P_{Li}-{\mathrm{\&mu;}}_{PLi}\right)}^2}{2{\mathrm{\&sigma;}}_i^2}\right);$

Load power factor is constant, the uncertain component Q of node i load_LiBy P_LiIt is determined that.

8. it is according to claim 5 solve wind energy turbine set probability optimal load flow method, it is characterised in that the P_LiAnd Q_Li's Solution includes：

The power output of wind energy turbine set depends on the power output of each typhoon group of motors in wind energy turbine set, and the generated output of Wind turbines Change with the fluctuation of wind speed, the relation between the generated output and wind speed of the Wind turbines is：

In formula, v wind speed, v_inTo cut wind speed, v_outFor cut-out wind speed, P_rFor Wind turbines Rated output power, P_WgFor Wind turbines real output；WithFor constant；

The Power Output for Wind Power Field of node i is：P_Wi=N_WiP_Wgi, in formula, N_WiIt is the Wind turbines number of units of node i wind energy turbine set.