CN112035783A - Wind power characteristic evaluation method based on time-frequency analysis - Google Patents

Abstract

The invention discloses a wind power characteristic evaluation method based on time-frequency analysis, which comprises the following steps: 1. carrying out discrete S transformation on the wind power; 2. establishing a time-frequency characteristic index considering the volatility, randomness and intermittence of the wind power on a wind power time-frequency domain; 3. and establishing a wind power characteristic evaluation model based on time-frequency analysis by utilizing an analytic hierarchy process according to the wind power time-frequency characteristic index. The method can solve the problem that the wind power characteristics cannot be comprehensively evaluated from a time domain or a frequency domain, and make up for the deficiency of the research on the wind power intermittency and randomness indexes, thereby providing a certain reference for the wind power grid-connected scheduling.

Description

A wind power feature evaluation method based on time-frequency analysis

technical field

The invention relates to a wind power feature evaluation method based on time-frequency analysis, and belongs to the technical field of electrical engineering.

Background technique

With the increasing depletion of traditional energy, people pay more and more attention to the development of renewable energy. Among the many renewable energy sources, wind energy is developing rapidly as a kind of green and environment-friendly renewable energy. However, there are fluctuations, randomness and intermittency in wind energy, which lead to the volatility, randomness and intermittency of wind power output from wind farms. As the penetration rate of wind power in the power grid increases, the grid-connected wind power with volatility, randomness and intermittency also increases the dispatching impact on the power grid. Therefore, it is necessary to analyze and evaluate the characteristics of wind power to clarify the degree of influence of wind power on grid dispatch.

Existing research mainly analyzes the characteristics of wind power from the perspective of time domain or frequency domain. By analyzing the characteristics of wind power from the time domain, the output of wind power in each period can be obtained, and the variation trend and fluctuation characteristics of wind power can be grasped, but the frequency composition of wind power fluctuation components and the energy of each fluctuation component cannot be known. The output of generator sets with different response speeds cannot be reasonably arranged in wind power grid-connected scheduling. By analyzing the characteristics of wind power from the frequency domain, the frequency of wind power fluctuation and its energy information can be obtained, but the time information of each fluctuation component cannot be obtained, so it is impossible to accurately determine the generator set in different time periods in wind power grid-connected scheduling. output. To sum up, it is very one-sided to analyze the characteristics of wind power only from the perspective of time domain or frequency domain. Therefore, it is urgent to analyze the characteristics of wind power from the time domain and frequency domain at the same time, that is, to perform time-frequency analysis of wind power to obtain the characteristic indicators of wind power in the time-frequency domain, so as to lay the foundation for more reasonable wind power grid-connected scheduling.

In addition, the feature analysis of wind power in the existing literature mainly starts from the volatility of wind power, but the research on the inherent randomness and intermittency of wind power itself is insufficient. The randomness of wind power is represented by the uncertainty and unpredictability of wind power output; the intermittent performance of wind power is that the output of wind power is zero or extremely small in a certain period of time. The randomness and intermittency of wind power will cause great adverse effects on grid-connected wind power dispatching: the strong randomness of wind power will increase the forecast error of wind power, thereby affecting the rational formulation of power generation dispatching plans; the intermittency of wind power It will threaten the power balance of the power grid and affect the frequency stability of the power grid.

To sum up, the traditional wind power feature analysis mainly adopts the time domain and frequency domain methods, and the analysis results are one-sided, and it is difficult to provide comprehensive and complete information for wind power feature evaluation and wind power scheduling.

SUMMARY OF THE INVENTION

In order to solve the above-mentioned shortcomings of the prior art, the present invention proposes a method for evaluating the characteristics of wind power based on time-frequency analysis, in order to extract the fluctuation, intermittent and intermittent characteristics of wind power by performing time-frequency analysis on wind power and extracting the frequency spectrum from the wind power. In order to solve the problem that the wind power characteristics cannot be comprehensively evaluated only from the time domain or frequency domain, and make up for the intermittent and The lack of random index research provides a certain reference for wind power grid-connected scheduling.

In order to achieve the above object, the technical scheme adopted in the present invention is:

The characteristics of a wind power feature evaluation method based on time-frequency analysis of the present invention are performed according to the following steps:

Step 1. It is known that the wind power sequence {P(j)} with a sampling interval of ΔT in a wind farm, j=0,1,...,N-1, use formula (1) to perform discrete S transform on it to obtain the time sampling The frequency spectrum S[j,n] of wind power with point j and frequency sampling point n:

In formula (1), j represents the time sampling point, j=0,1,...,N-1; n represents the frequency sampling point, n=0,1,...,N/2; i is the imaginary unit; m is the frequency translation amount; N represents the total number of sampling points, and N is an even number; P(j) represents the wind power when the time sampling point is j; exp represents the exponential function with the natural constant e as the base;

Step 2. In view of the volatility of wind power, use equations (2) and (3) to define the time-frequency domain mean m _{(t, f)} and the time-frequency domain variance in the time-frequency domain of wind power respectively.

Step 3. Aiming at the randomness of wind power, use formula (4) to define the time-spectrum entropy ε _(t,f) of wind power in the time-frequency domain:

Step 4. For the intermittency of wind power, define the total intermittent times n _sum of wind power and the average intermittent duration t _mean of wind power in the time-frequency domain;

Step 4.1. Use formula (5) to calculate the intermittent state quantity C(j) of wind power when the time sampling point is j on the time spectrum:

In formula (5), S _min is the time-frequency intermittent value of wind power;

Step 4.2. Use equation (6) to calculate the intermittent state transition c(j) of wind power when the time sampling point is j:

In formula (6), C(j+1) represents the intermittent state quantity of wind power when the time sampling point on the time spectrum is j+1;

Step 4.3. Use formula (7) to calculate the total intermittent times n _sum of wind power:

Step 4.4. Use formula (8) to calculate the average intermittent duration t _mean of wind power:

In formula (8), ΔT is the sampling interval of wind power;

Step 4.5. Use formula (9) to form the characteristic index vector x of wind power:

x=[x ₁ ,x ₂ ,..., _xi ,...,x ₅ ] (9)

时频谱熵ε_(t,f)、间歇次数n_sum、间歇平均时间t_mean；i＝1,2,…,5；In formula (9), x _i represents the i-th characteristic index value, x ₁ , x ₂ , x ₃ , x ₄ , and x ₅ are the time-frequency domain mean m _{(t, f)} and time-frequency domain variance of wind power, respectively.

Time spectrum entropy ε _(t,f) , interval times n _sum , interval average time t _mean ; i=1,2,...,5;

Step 5. Construct the index matrix X of D wind farms;

Step 5.1. Calculate the characteristic index of wind power according to steps 1 to 4 for each wind farm, and use formula (10) to construct the index vector {x _d ,d=1,2,...,D} of each wind farm:

x _d =[x _d1 ,x _d2 ,...,x _di ,...,x _d5 ] (10)

In formula (9), x _di represents the i-th characteristic index value of the d-th wind farm; x _d1 , x _d2 , x _d3 , x _d4 , and x _d5 are the time-frequency domain mean values of the wind power of the d-th wind farm, respectively , time-frequency domain variance, time-spectrum entropy, interval times, interval average time;

Step 5.2. Use formula (11) to obtain the index matrix X of D wind farms:

Step 6. Use AHP to establish a wind power characteristic evaluation model based on time-frequency analysis, and perform characteristic evaluation on the grid-connected wind power of each of the D wind farms;

Step 6.1. Construct judgment matrix A;

Use formula (12) to construct the judgment matrix A of the index:

In equation (12), a _ij represents the importance scale of the ith index x _i compared with the j th index x _j in the feature index vector x shown in equation (9); when i=j, let a _ij =1, when i≠j, let a _ij =1/a _ji ;

Step 6.2. Use formula (13) to construct a standard judgment matrix

表示式(9)所示的特征指标向量x中第i个指标x_i和第j个指标x_j相比的重要性标度标准值；In formula (13),

The importance scale standard value of the ith index x _i compared with the jth index x _j in the feature index vector x shown in expression (9);

Step 6.3. Use formula (14) to calculate the index weight vector w:

In formula (14), the superscript T represents the transpose of the vector;

Step 6.4. AHP consistency check;

Use formula (15) to get the maximum characteristic root λ _max :

In formula (15), (Aw) _i is the i-th element in the vector Aw formed by the product of the judgment matrix A and the indicator weight vector w, and w _i represents the i-th element in the indicator weight vector w;

Use formula (16) to obtain the consistency index C _I of the judgment matrix A:

Use formula (17) to obtain the consistency ratio _CR of the judgment matrix A:

In formula (17), R _I is the average random consistency index;

When C _R <δ, it means that the judgment matrix A satisfies the consistency; otherwise, the judgment matrix A is re-adjusted to meet the consistency; δ indicates the set consistency ratio threshold;

Step 6.5. Standardize the indicator matrix X;

Standardize the index matrix X of the D wind farms in step 5, and use the formula (18) to obtain the standard index matrix

表示第d个风电场风电功率的第i个特征指标标准值；In formula (18),

Represents the standard value of the i-th characteristic index of the wind power of the d-th wind farm;

Step 6.6. Obtain the wind power characteristic evaluation results of D wind farms;

利用式(19)得到D个风电场的风电功率特征评价向量v：According to the indicator weight vector w and the standard indicator matrix

Using Equation (19), the wind power characteristic evaluation vector v of D wind farms is obtained:

Compared with the prior art, the beneficial effects of the present invention are reflected in:

The invention solves the problem that the wind power characteristics cannot be comprehensively evaluated only from the time domain or the frequency domain, and makes up for the deficiency of the research on the randomness and intermittent index of the wind power. Through time-frequency analysis of wind power, characteristic indicators are extracted from the time spectrum, and a wind power characteristic evaluation model is established, so as to obtain more accurate and effective wind power characteristic evaluation results, and provide a certain reference for wind power grid-connected scheduling. The specific effects are reflected in the following aspects:

1. The present invention uses the discrete S transform shown in step 1 to perform time-frequency analysis on wind power to obtain the time spectrum of wind power, so that the frequency components and energy changes of wind power at each moment can be accurately described, which are the characteristics of wind power in the future. The evaluation provides more specific and comprehensive information on the characteristics of wind power;

2. The present invention uses the time spectrum entropy ε _{(t, f)} shown in step 3 to reflect the randomness of the signal to define the randomness of wind power on the time spectrum, and uses the total intermittent times of wind power shown in step 4. n _sum and the average intermittent duration t _mean of wind power to define the intermittent of wind power in time spectrum, thus making up for the lack of research on wind power randomness and intermittent index;

3. The present invention adopts the analytic hierarchy process shown in step 6 to establish an evaluation model of wind power characteristics based on time-frequency analysis, and the obtained evaluation results can more comprehensively and effectively reflect the characteristics of wind power, thereby providing a more effective method for grid-connected dispatching of wind power. refer to.

Description of drawings

FIG. 1 is a flow chart of the method for evaluating the characteristics of wind power based on time-frequency analysis according to the present invention.

Detailed ways

In this embodiment, as shown in FIG. 1 , a method for evaluating wind power characteristics based on time-frequency analysis is performed according to the following steps:

Step 1. It is known that the wind power sequence {P(j)}, j=0, 1, . Spectrum S[j,n] of wind power with point j and frequency sampling point n:

In formula (1), j represents the time sampling point, j=0, 1, ..., N-1; n represents the frequency sampling point, n=0, 1, ..., N/2; i is the imaginary unit; m is the frequency translation amount; N represents the total number of sampling points, and N is an even number; P(j) represents the wind power when the time sampling point is j; exp represents the exponential function with the natural constant e as the base;

Step 2. In view of the volatility of wind power, use equations (2) and (3) to define the time-frequency domain mean m _{(t, f)} and the time-frequency domain variance in the time-frequency domain of wind power respectively.

表示风电功率在时频域的波动程度，时频域方差

越大，风电功率波动越剧烈，对电网的影响也越大；The larger the mean value m _{(t, f)} in the time-frequency domain, the greater the average energy of wind power, and the greater the impact of its grid-connected fluctuations on the power grid; the variance in the time-frequency domain is greater.

Indicates the degree of wind power fluctuation in the time-frequency domain, and the variance in the time-frequency domain

The larger the value, the more severe the wind power fluctuation, and the greater the impact on the power grid;

Step 3. In view of the randomness of wind power, use formula (4) to define the time-spectrum entropy ε _{(t, f)} of wind power in the time-frequency domain

The time-spectral entropy ε _{(t, f)} represents the randomness of wind power in the time-frequency domain. The larger the value of the time-spectral entropy ε _{(t, f)} , the stronger the randomness and the weaker the predictability of wind power, which affects the power generation. Reasonable formulation of scheduling plans;

Step 4. For the intermittency of wind power, define the total intermittent times n _sum of wind power and the average intermittent duration t _mean of wind power in the time-frequency domain;

Step 4.1. Use formula (5) to calculate the intermittent state quantity C(j) of wind power when the time sampling point is j on the time spectrum:

In formula (5), S _min is the time-frequency intermittent value of wind power, which is taken as 5%×m _{(t, f)} in this embodiment; when C(j)=1, it means that the wind power at j is an intermittent state , when C(j)=0, it means that the wind power at j is a non-intermittent state;

Step 4.2. Use equation (6) to calculate the intermittent state transition c(j) of wind power when the time sampling point is j:

In formula (6), C(j+1) represents the intermittent state quantity of the wind power when the time sampling point is j+1 on the time spectrum; when c(j)=1, it represents that the wind power at j is changed from the non-intermittent state. into an intermittent state;

Step 4.3. Use formula (7) to calculate the total intermittent times n _sum of wind power:

Step 4.4. Use formula (8) to calculate the average intermittent duration t _mean of wind power:

In formula (8), ΔT is the sampling interval of wind power;

The total intermittent times n _sum and the average intermittent duration t _mean are used to _represent the intermittency of _wind power. The greater the balance threat is, the easier it is to affect the frequency stability of the power grid;

Step 4.5. Put the above five indicators together, and use the formula (9) to form the characteristic index vector x of wind power:

x=[x ₁ , x ₂ ,..., _xi ,...,x ₅ ] (9)

时频谱熵ε_(t，f)、间歇次数n_sum、间歇平均时间t_mean；i＝1，2，…，5；In formula (9), x _i represents the ith characteristic index value; x ₁ , x ₂ , x ₃ , x ₄ , and x ₅ are the time-frequency domain mean m _{(t, f)} and time-frequency domain variance of wind power, respectively

Time spectrum entropy ε _{(t, f)} , interval times n _sum , interval average time t _mean ; i=1, 2,..., 5;

Step 5. Construct the index matrix X of D wind farms;

Step 5.1. Calculate the characteristic index of wind power according to steps 1 to 4 for each wind farm, and use formula (10) to construct the index vector {x _d , d=1, 2, ..., D} of each wind farm:

x _d =[x _d1 , x _d2 , ..., x _di , ..., x _d5 ] (10)

In formula (9), x _di represents the i-th characteristic index value of the d-th wind farm; x _d1 , x _d2 , x _d3 , x _d4 , and x _d5 are the time-frequency domain mean values of the wind power of the d-th wind farm, respectively , time-frequency domain variance, time-spectrum entropy, interval times, interval average time;

Step 5.2. Use formula (11) to obtain the index matrix X of D wind farms:

Step 6. Use AHP to establish a wind power characteristic evaluation model based on time-frequency analysis, and perform characteristic evaluation on the grid-connected wind power of each of the D wind farms;

Step 6.1. Construct judgment matrix A;

Use formula (12) to construct the judgment matrix A of the index:

In equation (12), a _ij represents the importance scale of the ith index x _i compared with the j th index x _j in the feature index vector x shown in equation (9); when i=j, let a _ij =1, when i≠j, let a _ij =1/a _ji ;

间歇次数n_sum、间歇平均时间t_mean四个特征指标两两之间的重要性程度相同，风电功率的时频谱熵ε_(t，f)与其他四个指标相比都稍微重要，因此判断矩阵A可取式(1*)中的值：In this embodiment, the time-frequency domain mean value m _{(t, f)} and the time-frequency domain variance of the wind power are taken

Intermittent times n _sum , intermittent average time t _mean four characteristic indicators have the same degree of importance, and the time-spectrum entropy ε _{(t, f)} of wind power is slightly more important than the other four indicators, so the judgment matrix A can take the value in formula (1*):

Step 6.2. Use formula (13) to construct a standard judgment matrix

表示式(9)所示的特征指标向量x中第i个指标x_i和第j个指标x_j相比的重要性标度标准值；In formula (13),

The importance scale standard value of the ith index x _i compared with the jth index x _j in the feature index vector x shown in expression (9);

Step 6.3. Use formula (14) to calculate the index weight vector w:

In formula (14), the superscript T represents the transpose of the vector;

Step 6.4. AHP consistency check;

Use formula (15) to get the maximum characteristic root λ _max :

In formula (15), (Aw) _i is the i-th element in the vector Aw formed by the product of the judgment matrix A and the indicator weight vector w, and w _i represents the i-th element in the indicator weight vector w;

Use formula (16) to obtain the consistency index C _I of the judgment matrix A:

Use formula (17) to obtain the consistency ratio _CR of the judgment matrix A:

In formula (17), R _I is the average random consistency index, which is only related to the order of the judgment matrix. The order of the judgment matrix in the present invention is 5, and δ=1.12;

When C _R < δ, it means that the judgment matrix A satisfies the consistency; otherwise, the judgment matrix A is re-adjusted to meet the consistency; δ represents the set consistency ratio threshold, and the default δ=0.1;

Step 6.5. Standardize the indicator matrix X;

Standardize the index matrix X of the D wind farms in step 5, and use the formula (18) to obtain the standard index matrix

表示第d个风电场风电功率的第i个特征指标标准值；In formula (18),

Represents the standard value of the i-th characteristic index of the wind power of the d-th wind farm;

Step 6.6. Obtain the wind power characteristic evaluation results of D wind farms;

利用式(19)得到D个风电场的风电功率特征评价向量v：According to the indicator weight vector w and the standard indicator matrix

Using Equation (19), the wind power characteristic evaluation vector v of D wind farms is obtained:

The d-th element v _d in the influence degree vector v represents the influence degree of the wind power of the d-th wind farm connected to the grid on the grid dispatching. The larger the value, the greater the influence on the grid dispatching.

Claims (1)

1. A wind power characteristic evaluation method based on time-frequency analysis is characterized by comprising the following steps:

step 1, knowing a wind power sequence { p (j) }, j is 0,1, …, N-1 of a certain wind farm with a sampling interval Δ T, performing discrete S transformation on the wind power sequence by using a formula (1) to obtain a wind power time frequency spectrum S [ j, N ] with a time sampling point j and a frequency sampling point N:

in formula (1), j represents a time sampling point, and j is 0,1, …, N-1; n represents a frequency sampling point, N is 0,1, …, N/2; i is an imaginary unit; m is the frequency translation amount; n represents the total number of sampling points and is an even number; p (j) represents the wind power when the time sampling point is j; exp represents an exponential function with a natural constant e as the base;

step 2, aiming at the volatility of the wind power, respectively defining a time-frequency domain mean value m on a wind power time-frequency domain by using the formula (2) and the formula (3)_(t,f)Sum time-frequency domain variance

And 3, aiming at the randomness of the wind power, defining the time-frequency spectrum entropy of the wind power on a time-frequency domain by using the formula (4)_(t,f)：

Step 4, aiming at the intermittence of the wind power, defining the total intermittence times n of the wind power on a time-frequency domain_sumAnd the average intermittent duration t of the wind power_mean；

Step 4.1, calculating the intermittent state quantity C (j) when the time sampling point on the time spectrum of the wind power is j by using the formula (5):

in the formula (5), S_minThe time-frequency interval value of the wind power is obtained;

and 4.2, calculating an intermittent state transition variable c (j) of the wind power when a time sampling point is j by using the formula (6):

in the formula (6), C (j +1) represents an intermittent state quantity when a time sampling point on a time frequency spectrum of the wind power is j + 1;

step 4.3, calculating the total intermittent times n of the wind power by using the formula (7)_sum：

Step 4.4, calculating the average intermittent duration t of the wind power by using the formula (8)_mean：

In the formula (8), Δ T is a sampling interval of the wind power;

and 4.5, forming a characteristic index vector x of the wind power by using a formula (9):

x＝[x₁,x₂,…,x_i,…,x₅] (9)

in the formula (9), x_iIndicates the i-th characteristic index value, x₁,x₂,x₃,x₄,x₅Time-frequency domain mean values m of wind power respectively_(t,f)Time-frequency domain variance

Entropy of time-frequency spectrum_(t,f)Intermittent number of times n_sumIntermittent average time t_mean；i＝1,2,…,5；

Step 5, constructing index matrixes X of the D wind power plants;

step 5.1, calculating the characteristic index of the wind power of each wind power plant according to the steps 1 to 4, and constructing an index vector { x ] of each wind power plant by using the formula (10)_d,d＝1,2,…,D}：

x_d＝[x_d1,x_d2,…,x_di,…,x_d5] (10)

In the formula (9), x_diRepresenting the ith characteristic index value of the d wind power plant; x is the number of_d1,x_d2,x_d3,x_d4,x_d5Respectively representing the mean value of the time-frequency domain, the variance of the time-frequency domain, the entropy of the time-frequency spectrum, the intermittent times and the intermittent average time of the wind power of the d-th wind power plant;

and 5.2, obtaining index matrixes X of the D wind power plants by using the formula (11):

step 6, establishing a wind power characteristic evaluation model based on time-frequency analysis by using an analytic hierarchy process, and performing characteristic evaluation on the grid-connected wind power of each wind power plant in the D wind power plants;

step 6.1, constructing a judgment matrix A;

a judgment matrix A of the index is constructed by using the formula (12):

in the formula (12), a_ijThe ith index x in the feature index vector x represented by the expression (9)_iAnd the jth index x_jScale of importance of the comparison; when i is j, let a_ijWhen i ≠ j, let a_ij＝1/a_ji；

Step 6.2, constructing a standard judgment matrix by using the formula (13)

In the formula (13), the reaction mixture is,

the ith index x in the feature index vector x represented by the expression (9)_iAnd the jth index x_jSignificance scale criteria values for comparisons;

and 6.3, calculating an index weight vector w by using the formula (14):

in the formula (14), the superscript T represents the transpose of the vector;

step 6.4, checking consistency of an analytic hierarchy process;

obtaining maximum characteristic root lambda by using formula (15)_max：

In formula (15), (Aw)_iIs used for judging the ith element, w, in the vector Aw formed by the product of the matrix A and the index weight vector w_iRepresenting the ith element in the indicator weight vector w;

obtaining a consistency index C of the judgment matrix A by using the formula (16)_I：

Obtained by the formula (17)Determining the consistency ratio C of the matrix A_R：

In the formula (17), R_IIs an average random consistency index;

when C is present_RIf yes, the judgment matrix A meets the consistency; otherwise, readjusting the judgment matrix A to meet the consistency; indicating a set threshold of the consistency ratio;

6.5, standardizing the index matrix X;

standardizing the index matrixes X of the D wind power plants in the step 5, and obtaining a standard index matrix by using a formula (18)

In the formula (18), the reaction mixture,

the ith characteristic index standard value of the wind power of the d wind power plant is represented;

6.6, obtaining wind power characteristic evaluation results of the D wind power plants;

according to the index weight vector w and the standard index matrix

Obtaining wind power characteristic evaluation vectors v of D wind power plants by using an equation (19):

Images