Method of calculating available output power of wind farm

Abstract

A method of calculating available output power of wind farm includes following steps. A space vector Vk is obtained by decomposing a power sequence of benchmarking wind turbines in a wind farm based on empirical orthogonal function. A typical power sequence of the benchmarking wind turbines is calculated by restoring the space vector Vk. A total power Ptotal of a feeder on which the benchmarking wind turbines is operated is obtained by enlarging a typical power of each benchmarking wind turbine in proportion according to the quantity of the benchmarking wind turbines operated on the feeder. An output power Pestimate of the wind farm is obtained by accumulating the total power Ptotal of all the benchmarking wind turbines.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims all benefits accruing under 35 U.S.C. §119 from China Patent Application 201410231145.2, filed on May 29, 2014 in the China Intellectual Property Office, the disclosure of which is incorporated herein by reference.

BACKGROUND

1. Technical Field

The present disclosure relates to a method of calculating available output power of the wind farm, especially for a method of calculating the available output power of large-scale wind farm (above 30 MW) based on the empirical orthogonal function (EOF) in the natural state.

2. Description of the Related Art

With the rapid development of wind farm industry, the installed wind power capacity in the power network operation has reached 100 million kilowatts. There is an obvious case of brownouts with the limit of power transmission capacity and the ability of consumptive. Currently, the amount of abandoned wind power is generally calculated by annual generating capacity of wind turbines in theory, and the available output power is calculated based on the amount of abandoned wind power. However, the result is very different from the actual.

What is needed, therefore, is a method of calculating available output power of the wind farm that can overcome the above-described shortcomings.

BRIEF DESCRIPTION OF THE DRAWINGS

Many aspects of the embodiments can be better understood with reference to the following drawings. The components in the drawings are not necessarily drawn to scale, the emphasis instead being placed upon clearly illustrating the principles of the embodiments. Moreover, in the drawings, like reference numerals designate corresponding parts throughout the several views.

FIG. 1 shows a flow chart of one embodiment of a method of calculating available wind power of a wind farm based on EOF.

DETAILED DESCRIPTION

The disclosure is illustrated by way of example and not by way of limitation in the figures of the accompanying drawings in which like references indicate similar elements. It should be noted that references to “an” or “one” embodiment in this disclosure are not necessarily to the same embodiment, and such references mean at least one.

Referring to FIGURE, a method of calculating available output power of the wind farm comprises following steps:

step S10, obtaining a space vector V_k by decomposing a power sequence of benchmarking wind turbines in a wind farm based on empirical orthogonal function;

step S20, calculating a typical power sequence P of benchmarking wind turbines by restoring the space vector V_k;

step S30, getting a total power P_total of a feeder on which the benchmarking wind turbines is operated by enlarging a typical power of each benchmarking wind turbine in proportion according to the number of the benchmarking wind turbines operated on the feeder;

step S40, obtaining an output power P_estimate of the wind farm by accumulating the total power P_total of all the benchmarking wind turbines.

In step S10, the power sequence of benchmarking wind turbines in the wind farm can be obtained through the available output power of the wind farm in one year. The power sequence is defined through the number of benchmarking wind turbines and a time sequence. The number of benchmarking wind turbines is set for m, and the time sequence of each of the benchmarking wind turbines is set for n based on the available output power of the wind farm in one year. According to EOF, the wind farm can be abstracted as a space field with m-dimensional random variable X, and a plurality of samples with a capacity of n are obtained in the space field. The plurality of samples are defined as X_i(1≦i≦n). X_i is an m-dimensional vector representing the benchmarking wind turbines. The X_i is denoted as follows:X_t=(x_1t,x_2t, . . . ,x_mt)^T, t=1,2, . . . ,n;  (1)

According to EOF decomposition form, formula (1) can be expressed as follows:

$\begin{array}{cc}{X}_t=\sum _{k=1}^K{\alpha }_k\left(t\right)V_k+{\varepsilon }_t;& \left(2\right)\end{array}$

wherein V_k is the unknown m dimensional space vector, ε_t is an m dimension error vector corresponding to the V_k, α_k(t) is a time factor, which is a weighting factor while the k^th space vector V_k represents X_t.

The space vector V_k can be obtained by following substeps:

Step S11, obtaining a first space vector V₁, wherein a residual error sum of squares E₁ of the formula:X_t=α₁(t)V₁+ε_t  (3)is smallest.

The residual error sum of squares E₁ can be expressed as:

$\begin{array}{cc}\begin{array}{c}{E}_1=\frac{1}{n}\sum _{t=1}^n\sum _{i=1}^m{\varepsilon }_{\mathrm{it}}^2=\frac{1}{n}\sum _{t=1}^n{\varepsilon }_t^T{\varepsilon }_t\\ =\langle {\varepsilon }_t^T{\varepsilon }_t\rangle =\langle {\left(X_t-{\alpha }_1\left(t\right)V_1\right)}^T\left(X_t-{\alpha }_1\left(t\right)V_1\right)\rangle \end{array}& \left(4\right)\end{array}$

wherein α₁(t)=X_t^TV₁ or α₁(t)=V₁^TX_t, α₁(t)=0, and V₁^TV₁=1.

The residual error sum of squares E₁ can further be expressed as:

$\begin{array}{cc}\begin{array}{c}{E}_1=\langle X_t^T{X}_t-{\alpha }_1\left(t\right)X_t^T{V}_1-{\alpha }_1\left(t\right)V_1^T{X}_t+{\alpha }_1^2\left(t\right)V_1^T{V}_1\rangle \\ =\langle X_t^T{X}_t\rangle -\langle {\alpha }_1\left(t\right){\alpha }_1\left(t\right)\rangle -\langle {\alpha }_1\left(t\right){\alpha }_1\left(t\right)\rangle +\langle {\alpha }_1^2\left(t\right)\rangle \\ =\langle X_t^T{X}_t\rangle -\langle {\alpha }_1^2\left(t\right)\rangle \end{array}& \left(5\right)\end{array}$

wherein

$\langle X_t^T{X}_t\rangle =\frac{1}{n}\sum _{t=1}^n\sum _{i=1}^m{x}_{\mathrm{it}}^2=\sum _{i=1}^m\frac{1}{n}\sum _{t=1}^n{x}_{\mathrm{it}}^2,$ is a total variance expressed as VarX which is depended on the spatial field of space research field and independent of V₁ and α₁(t), then there are:E₁=VarX−Varα₁  (6)

The Varα₁ can be represented with X and V₁, thus:Varα₁=α₁(t)α₁(t)=V₁^TX_tX_t^TV₁=V₁^TX_tX_t^TV₁  (7)

wherein, the element on the i line and the j column of the X_tX_t^T can be expressed as x_itx_jt is a covariance of i and j's grid points sequence. The X_tX_t^T is a covariance matrix:Σ=(X_tX_t^T)  (8)

Furthermore, the residual error sum of squares E₁ can be expressed as:E₁=VarX−V₁^TΣV₁  (9)

Step S12, calculating the minimum value of E₁ and the V₁ corresponding to the minimum value of E₁ under the condition of V₁^TV₁=1.

The minimum value of E₁ and the V₁ can be obtained through Lagrangian method, which is expressed as:F(v₁₁,v₂₁, . . . ,v_m1)=F(V₁)=VarX−V₁^TΣV₁+λ(V₁^TV₁₋1)  (10).

Through the derivation function, then:

$\begin{array}{cc}\frac{\partial F}{\partial V_1}=-2\sum V_1+2\lambda V_1.& \left(11\right)\end{array}$

Making it to zero vector, then:ΣV₁=λV₁  (12).

Thus V₁ is an eigenvector of the covariance matrix Σ, and λ is a Lagrange multiplier, which is an eigenvalues corresponding to the eigenvector V₁.

Taking the formula (12) into the formula (9):

$\begin{array}{cc}\begin{array}{c}{E}_1=\mathrm{Var}X-V_1^T\sum V_1\\ =\mathrm{Var}X-V_1^T\lambda V_1\\ =\mathrm{Var}X-\lambda V_1^T{V}_1\\ =\mathrm{Var}X-\lambda .\end{array}& \left(13\right)\end{array}$

In order to get the minimum E₁, a maximum eigenvalue λ₁ of the eigenvalue λ is selected, which is expressed as λ=λ₁. Thus V₁ is the eigenvector corresponding to the maximum eigenvalue λ₁.

Then:

$\begin{array}{cc}\begin{array}{c}\mathrm{Var}{\alpha }_1=\langle {\alpha }_1\left(t\right){\alpha }_1\left(t\right)\rangle =\langle V_1^T{X}_t{X}_t^T{V}_1\rangle =V_1^T\sum V_1\\ =V_1^T\lambda V_1={\lambda }_1{V}_1^T{V}_1=\lambda .\end{array}& \left(14\right)\end{array}$

Step S13, calculating variance contribution rate η.

A variance contribution Q_k of the space vector V_k is a total number of the error variance which is reduced in the space field after the space vector V_k is added into the expansion of formula (9). The variance contribution Q_k can be obtained:

$\begin{array}{cc}{Q}_k=E_{k-1}-E_k=\left(\mathrm{Var}X-\sum _{l=1}^{k-1}{\lambda }_l\right)-\left(\mathrm{Var}X-\sum _{l=1}^k{\lambda }_l\right)={\lambda }_k.& \left(15\right)\end{array}$

Furthermore, because the total variance of the space field is expressed as

$\mathrm{Var}X=\sum _{k=1}^m{\lambda }_k,$ thus the variance contribution rate η is:

$\begin{array}{cc}\eta =\frac{Q_k}{\mathrm{Var}X}=\frac{{\lambda }_k}{\sum _{k=1}^m{\lambda }_k}\times 100\%.& \left(16\right)\end{array}$

The variance contribution rate η reflects the space-based capabilities to the original description of the space field, the greater the variance contribution rate η, the closer between the space filed which is reduced from the space vector and the original space field. A typical space vector V can be determined by the variance contribution rate η.

In step S20, the threshold of the variance contribution rate η is defined as 95% to obtain the typical space vector V. The typical space vector V under the variance contribution rate η and the time factor α can be expressed as:V=(b₁,b₂,b₃, . . . b_m)  (17).

Thus the typical power sequence P of each benchmark wind turbine is:P=α(b₁,b₂,b₃, . . . ,b_m)  (18).

In step S30, a feeder series of the benchmarking wind turbines which are switched on in the number of m are defined as c₁, c₂, c₃, . . . , c_m, thus a transposed number of the feeder series is defined as C:C=(c₁,c₂,c₃, . . . ,c_m)^T  (19).

Then a total power of the feeder P_total on which the benchmarking wind turbines is operated is expressed as:

$\begin{array}{cc}\begin{array}{c}{P}_{\mathrm{total}}=\mathrm{PC}=\alpha \left(b_1,b_2,b_3,\dots ,b_m\right){\left(c_1,c_2,c_3,\dots ,c_m\right)}^T\\ =\alpha \left(b_1{c}_1,b_2{c}_2,b_3{c}_3,\dots ,b_m{c}_m\right)\\ =\left(\alpha b_1{c}_1,\alpha b_2{c}_2,\alpha b_3,c_3,\dots ,\alpha b_m{c}_m\right).\end{array}& \left(20\right)\end{array}$

In step S40, the output power P_estimate of the whole wind farm can be obtained by summing the all components of the total power of the feeder P_total in the formula (20):P_estimate=αb₁c₁+αb₂c₂+αb₃c₃+ . . . +αb_mc_m  (21).

The method of calculating available output power of wind farm has following advantages. First, the method of calculating available output power takes advantages of the EOF in mining characteristics and generality of each benchmarking turbines, and the random sequences in the decomposition can be destructed and superposed. Second, the large calculation error in calculating the theoretical generating capacity of the wind farm in the natural state can be overcome, and the calculation accuracy of the theoretical generating capacity can be improved. Third, the method can provide an important reference to assess the economic losses of brownouts or overhaul.

Depending on the embodiment, certain of the steps of methods described may be removed, others may be added, and that order of steps may be altered. It is also to be understood that the description and the claims drawn to a method may include some indication in reference to certain steps. However, the indication used is only to be viewed for identification purposes and not as a suggestion as to an order for the steps.

It is to be understood that the above-described embodiments are intended to illustrate rather than limit the disclosure. Variations may be made to the embodiments without departing from the spirit of the disclosure as claimed. It is understood that any element of any one embodiment is considered to be disclosed to be incorporated with any other embodiment. The above-described embodiments illustrate the scope of the disclosure but do not restrict the scope of the disclosure.

Claims

1. A method of calculating available output power of wind farm comprising: obtaining a space vector Vk by decomposing a power sequence of benchmarking wind turbines in a wind farm based on empirical orthogonal function;calculating a typical power sequence P of the benchmarking wind turbines by restoring the space vector Vk;getting a total power Ptotal of a feeder on which the benchmarking wind turbines is operated by enlarging a typical power of each benchmarking wind turbine in proportion according to a quantity of the benchmarking wind turbines operated on the feeder;accumulating the total power Ptotal of all the benchmarking wind turbines.

2. The method of claim 1, wherein the power sequence of benchmarking wind turbines in the wind farm is obtained through an output power of the wind farm in past one year.

3. The method of claim 2, wherein the power sequence is defined through a number of benchmarking wind turbines and a time sequence, the number of benchmarking wind turbines is set for m, and a time sequence of each of the benchmarking wind turbines is set for n.

4. The method of claim 3, wherein the wind farm is abstracted as a space field with m-dimensional random variable X, and a plurality of samples with a capacity of n is obtained in the space field.

5. The method of claim 4, wherein the plurality of samples are defined as Xi(1≦i≦n), and Xi is an m-dimensional vector representing the benchmarking wind turbines denoted as:

6. The method of claim 5, wherein the space vector Vk is obtained by: obtaining a first space vector V1, wherein a residual error sum of squares E1 in the formula: E1=VarX−V1TΣV1  (3)

7. The method of claim 6, wherein the variance contribution Qk is a total number of the error variance reduced in the space field after the space vector Vk is added into expansion of the formula of the residual error sum of squares E1.

8. The method of claim 7, wherein the variance contribution Qk is obtained by:

9. The method of claim 6, wherein a typical space vector V is obtained by setting the variance contribution rate η as 95%, and the typical space vector V under the variance contribution rate η and a time factor α is expressed as: V=(b1,b2,b3, . . . ,bm).

10. The method of claim 9, wherein the typical power sequence P of each benchmark wind turbine is: P=α(b1,b2,b3, . . . ,bm).

11. The method of claim 10, wherein a feeder series of the benchmarking wind turbines which are switched on are defined as c1, c2, c3 . . . , cm, and a transposed number of the feeder series is defined as C: C=(c1,c2,c3, . . . ,cm)T.

12. The method of claim 11, wherein the total power Ptotal of the feeder on which the benchmarking wind turbines is operated is:

13. The method of claim 12, wherein accumulating the total power Ptotal of all the benchmarking wind turbines is performed by summing the all components of the total power of the feeder Ptotal, and the output power Pestimate is: Pestimate=αb1c1+αb2c2+αb3c3+ . . . +αbmcm.