CN103296701B - Active power control method in wind power plant - Google Patents

Abstract

The invention provides an active power control method in a wind power plant. The active power control method in the wind power plant comprises the steps of computing the active power prediction errors of all predicted points in the wind power plant, establishing a wind power plant active power prediction error distribution function, establishing a wind power plant active power prediction error distribution module, a wind power plant active power prediction confidence degree module and a wind power plant active power control module in sequence, optimizing the wind power plant active power control module, obtaining an optimization confidence weight of all unit sets, and controlling the active power of the wind power plant. The active power control method in the wind power plant is based on historical data analysis. By means of the active power control method in the wind power plant, the power-outputting error of the wind power plant is reduced, the reliability and accuracy of the wind power plant are guaranteed, and a powerful guarantee is supplied to the wind power plant in the grid-connected operation.

Description

A kind of wind power station active power control method

Technical field

The invention belongs to Power System Planning and running technology field, be specifically related to a kind of wind power station active power control method.

Background technology

According to country's " New Energy Industry development planning " draft, the year two thousand twenty China wind-powered electricity generation total installation of generating capacity will reach 1.5 hundred million kilowatts, be 5 times that within 2007, issue " planning of regenerative resource Long-and Medium-term Development " target.According to planning, China will in Gansu, Xinjiang, Hebei, Jilin, the Inner Mongol, provinces and regions, six, Jiangsu make 7 ten million multikilowatt wind power base.Jiuquan ten million multikilowatt wind power base construction plan total installation of generating capacity is 3,565 ten thousand kilowatts; Hami plans 2,000 ten thousand kilowatts; Inner Mongol planning construction 5,000 ten thousand kilowatts, wherein covers 2,000 ten thousand kilowatts, west, covers 3,000 ten thousand kilowatts, east; Hebei planning is built together in coastal and northern territory and is established 1,000 ten thousand kilowatts; Jiangsu planning construction 1,000 ten thousand kilowatts, wherein 7,000,000 kilowatts, coastal waters; West Area of Jilin Province, mainly in cities such as loose source and Baicheng, is planned for the year two thousand twenty and reaches 2,300 ten thousand kilowatts.By the end of the year 2009, China's installed capacity of wind-driven power will have reached 2,500 ten thousand kilowatts, leap to global second.

China's wind-power electricity generation development is swift and violent, but faces the multiple challenge of fault in the primary stage.There is fluctuation and intermittence in wind-resources, therefore wind-power electricity generation is a kind of insecure forms of electricity generation.Wind power resources is by effectively including the Forecasting Methodologies such as ultra-short term prediction, short-term forecast in electric power system dispatching system, the impact of wind-powered electricity generation fluctuation on electric power system dispatching can be reduced to a certain extent, but analyze existing Forecasting Methodology, what large metering method obtained is all the power prediction sequence determined, only provides one group of wind power prediction sequence determined.

Therefore, there is certain error in the result obtained according to existing wind-powered electricity generation Forecasting Methodology.The distribution probability and probabilistic forecasting result that obtain wind power prediction error if can analyze, then can improve the precision of wind-powered electricity generation scheduling, reduces the system cost increase that wind-powered electricity generation predicated error is brought.

The prediction of wind energy turbine set ultra-short term can predict the wind power in following 4h, because existing Forecasting Methodology all exists certain error, if ultra-short term predicted power is directly applied to scheduling, and can to power system economy negative effect.

Summary of the invention

In order to overcome above-mentioned the deficiencies in the prior art, the invention provides a kind of wind power station active power control method, based on historical data analysis, output of wind electric field error can be reduced, ensure that reliability and the accuracy of wind energy turbine set, for wind farm grid-connected operation provides powerful guarantee.

In order to realize foregoing invention object, the present invention takes following technical scheme:

A kind of wind power station active power control method is provided, said method comprising the steps of:

Step 1: calculate wind energy turbine set each future position active power predicated error, and set up active power of wind power field predicated error distribution function;

Step 2: based on active power of wind power field predicated error distribution function, sets up active power of wind power field predicated error distributed model;

Step 3: based on active power of wind power field predicated error distributed model, sets up active power of wind power field forecast confidence model;

Step 4: comprehensive active power of wind power field predicated error distributed model and active power of wind power field forecast confidence model, sets up active power of wind power field Controlling model;

Step 5: be optimized described active power of wind power field Controlling model, obtains the optimization confidence weight of each machine group, and then controls active power of wind power field.

Described step 1 comprises the following steps:

Step 1-1: input wind energy turbine set historical forecast power data and history actual power data, selects wind energy turbine set rated power to calculate wind energy turbine set each future position active power predicated error as a reference value, has:

$e_i=\frac{P_{\mathrm{predict}\_i}-P_{\mathrm{actual}\_i}}{P_{\mathrm{farm}}}---\left(1\right)$

Wherein: e _ifor the active power predicated error of wind energy turbine set i-th time point; P _{predict_i}for wind energy turbine set i-th time point predicted power; P _{actual_i}for wind energy turbine set i-th time point actual power; P _farmfor the specified installed capacity of wind energy turbine set;

Step 1-2: judge active power of wind power field predicated error is as which kind of distribution by hypothesis testing in mathematical statistics, and set up active power of wind power field predicated error distribution function.

Described step 1-2 comprises the following steps:

Whether step 1-2-1: suppose that active power of wind power field predicated error is normal distribution, check this hypothesis to set up:

Utilize maximum-likelihood method estimation error average and variance, calculate interval frequency and the statistic test value in each interval if meet test value in acceptance region, then determine that active power of wind power field predicated error is normal distribution;

Step 1-2-2: set up active power of wind power field predicated error distribution function;

Active power of wind power field predicated error distribution function is described as:

$f\left(e_i\right)=\frac{1}{\mathrm{\σ}\sqrt{2\mathrm{\π}}}{e}^{-\frac{{(e_i-\mathrm{\μ})}^2}{{2\mathrm{\σ}}^2}}---\left(2\right)$

Wherein: μ is wind energy turbine set anticipation error, σ is the standard deviation of active power predicated error;

Then at time point t place, the active power predicated error desired value E of wind energy turbine set _i(t) be:

E _i(t)＝e(t) （3）

Wherein: t is current time, e (t) is t active power of wind power field predicated error.

In described step 2, active power of wind power field predicated error Normal Distribution is normal distribution, active power predicated error average and the variance of wind energy turbine set is can be calculated by historical data, because each future position of wind energy turbine set is discrete, calculate the active power predicated error of each time point of wind energy turbine set according to active power of wind power field predicated error distribution function, obtain active power of wind power field predicated error distributed model thus.

In described step 3, for active power of wind power field predicated error, select confidence level, and then the wind energy turbine set calculated under this confidence level expects confidential interval, is expressed as:

P{e _min＜e＜e _max}≥1-α （4）

Wherein: e _minfor lower limit of confidence interval, e _maxfor the confidential interval upper limit, 1-α is confidence level, (e _min, e _max) be confidential interval.

In described step 4, for the wind energy turbine set comprising N platform wind power generating set, Wind turbines is divided into M group by the running status according to unit each in wind energy turbine set;

The distribution of comprehensive active power of wind power field predicated error and confidence level, the choosing of confidential interval, set up active power of wind power field Controlling model, object function is:

${\min f}_i=|\underset{j=1}{\overset{M}{\mathrm{\Σ}}}({\mathrm{\ξ}}_{\mathrm{ij}}+\frac{{\mathrm{\λ}}_{\mathrm{ij}}}{2})P_{\mathrm{group}\_\mathrm{ij}}-P_{\mathrm{predict}\_i}\·E_i|---\left(5\right)$

Wherein: f _ibe that the i-th time point wind energy turbine set desired output power and history error distribute and expect difference, ξ _ijfor the weight coefficient of a jth group wind-powered electricity generation group of planes i-th time point; λ _ijfor the fluctuation weight coefficient of a jth group wind-powered electricity generation group of planes i-th time point; P _{group_ij}for the prediction generated output of jth group wind-powered electricity generation group of planes time point; P _{predict_i}it is the wind energy turbine set predicted power of the i-th time point; E _ifor the active power predicated error desired value of each time point of wind energy turbine set.

Statistical history data, were divided into n interval, calculate the active power predicated error desired value of each time point of wind energy turbine set, thus obtain each interval error e by one day _idistribution;

The active power predicated error desired value E that integration obtains each time point of wind energy turbine set is asked for each interval _i, have

$E_i={\∫}_{t_{i-1}}^{t_i}{e}_i\mathrm{dt}---\left(6\right)$

Ask for f _iduring minimum of a value, generated output is divided into by variable bound to retrain and weight coefficient constraint;

Generated output is constrained to:

$\left\{\begin{array}{c}\underset{j=1}{\overset{M}{\mathrm{\Σ}}}({\mathrm{\ξ}}_{\mathrm{ij}}+{\mathrm{\λ}}_{\mathrm{ij}})P_{\mathrm{group}\_\mathrm{ij}}=P_{\mathrm{predict}\_i}\×(1+e_{+})\\ \underset{j=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{\ξ}}_{\mathrm{ij}}{P}_{\mathrm{group}\_\mathrm{ij}}=P_{\mathrm{predict}\_i}\×(1-e_{-})\\ ({\mathrm{\ξ}}_{\mathrm{ij}}+{\mathrm{\λ}}_{\mathrm{ij}})P_{\mathrm{group}\_\mathrm{ij}}\≤\underset{j=1}{\overset{M}{\mathrm{\Σ}}}{Q}_{\mathrm{ij}}{P}_{\mathrm{gen}}\end{array}\right.,i=\mathrm{1,2,3}...,N;j=\mathrm{1,2},...,M---\left(7\right)$

Wherein: e ₊for predicted power confidence upper limit; e _-for predicted power confidence lower limit; Q _ijfor a jth group of planes i-th moment unit number; P _genfor Wind turbines rated power;

In formula (7), the equality constraint equal sign left side represents the general power of each wind turbine cohort, represents the power bound of wind energy turbine set on the right of equal sign; The inequality constraints equal sign left side represents the general power of each wind turbine cohort, represents the power upper limit of each machine group on the right of equal sign;

Weight coefficient is constrained to:

0≤ξ _ij＜1，i=1,2,3...,N;j=1,2,...,M（8）

0≤λ _ij＜1，i=1,2,3...,N;j=1,2,...,M（9）。

In described step 5, according to determined object function and variable bound, described active power of wind power field Controlling model is optimized, obtains the optimization confidence weight of each machine group, and then control active power of wind power field;

Optimizing process is expressed as Mathematical Modeling, has:

$\left.\begin{array}{c}s.t.\min f(u,x)\\ h(u,x)=0\\ g(u,x)\≥0\end{array}\right\}---\left(10\right)$

Wherein: u and x is state variable, and state variable x comprises the weight coefficient of each machine group; F is that wind energy turbine set desired output power and history error distribute and expect difference, h and g is respectively equality constraints functions and inequality constraints function.

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

(1) machine group predicted power weight reassignment method in wind energy turbine set is proposed, overcome the scheduling error that predicated error in conventional method causes, by the introducing of historical data analysis and weight coefficient, find suitable mathematical optimization object function, and then set up Optimized model, for wind farm grid-connected control provides suitable method;

(2) contemplated by the invention various factors in wind energy turbine set, more accurately rationally, reduce wind farm grid-connected loss, method is simple to operation for gained active power controller scheme, is convenient to promote.

(3) based on historical data analysis, output of wind electric field error can be reduced, ensure that reliability and the accuracy of wind energy turbine set, for wind farm grid-connected operation provides powerful guarantee.

Accompanying drawing explanation

Fig. 1 is wind power station active power control method flow chart.

Detailed description of the invention

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

As Fig. 1, consider that the active power of wind power field optimal control method of confidence level comprises the following steps:

(1) input historical forecast power data and the actual power data of wind energy turbine set, calculate active power of wind power field predicated error distribution function;

(2) wind energy turbine set predicated error is analyzed, select suitable confidence level, calculate wind energy turbine set predicted power confidential interval;

(3) according to the realtime power data of the wind turbine cohort collected, Wind turbines is hived off;

(4) active set algorithm optimization is utilized to find the optimum confidence level weight of each machine group;

(5) based on the confidence level weight calculated, control is optimized to active power of wind power field.

Be below an example of calculation of the inventive method, based on the ultra-short term prediction data of Gansu wind energy turbine set, realize each machine group power weight reallocation in wind energy turbine set and regulate.Comprising 15 single-machine capacities in wind energy turbine set is the Wind turbines of 3MW, and unit is all controlled, and temporal resolution is 30 minutes, after each group of planes receives dispatch command, can be adjusted to dispatch value before future time point.And according to analysis needs, configuration scheduling instruction.Before and after optimizing, the difference of wind power vacancy is as table 1:

Table 1

Time t/h	Difference power/MW
		0.5	0.7138
1	0.7138
		1.5	1.9201
2	1.3843
		2.5	2.0061
3	1.7158
		3.5	1.7506
4	1.6988
		6.5	2.1053
7	1.4473
		7.5	1.2804
8	0.5268
		11	0.6836

11.5	0.8016
		12	0.3652
22	0.3316
		22.5	0.5528
23	1.8566
		23.5	3.0329
24	2.4334

As shown in Table 1, after being optimized by said method, when wind energy turbine set generated output is greater than schedule power, largest optimization amount is about 2MW; When wind energy turbine set generated output is less than schedule power, maximumly reduce about 3MW power shortage.Therefore for this wind energy turbine set, this active power optimal control method has important investment value, and example also demonstrates the feasibility of method proposed by the invention simultaneously, is the effectively method solving active power of wind power field control problem.

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 with reference to above-described embodiment to invention has been detailed description, those of ordinary skill in the field are to be understood that: still can modify to the specific embodiment of the present invention or equivalent replacement, and not departing from any amendment of spirit and scope of the invention or equivalent replacement, it all should be encompassed in the middle of right of the present invention.

Claims (1)

1. a wind power station active power control method, is characterized in that: said method comprising the steps of:

Step 1: calculate wind energy turbine set each future position active power predicated error, and set up active power of wind power field predicated error distribution function;

Step 2: based on active power of wind power field predicated error distribution function, sets up active power of wind power field predicated error distributed model;

Step 3: based on active power of wind power field predicated error distributed model, sets up active power of wind power field forecast confidence model;

Step 4: comprehensive active power of wind power field predicated error distributed model and active power of wind power field forecast confidence model, sets up active power of wind power field Controlling model;

Step 5: be optimized described active power of wind power field Controlling model, obtains the optimization confidence weight of each machine group, and then controls active power of wind power field;

Described step 1 comprises the following steps:

Step 1 ?1: input wind energy turbine set historical forecast power data and history actual power data, select wind energy turbine set rated power to calculate wind energy turbine set each future position active power predicated error as a reference value, have:

$e_i=\frac{P_{\mathrm{predict}\_i}-P_{\mathrm{actual}\_i}}{P_{\mathrm{farm}}}---\left(1\right)$

Wherein: e _ifor the active power predicated error of wind energy turbine set i-th time point; P _{predict_i}for wind energy turbine set i-th time point predicted power; P _{actual_i}for wind energy turbine set i-th time point actual power; P _farmfor the specified installed capacity of wind energy turbine set;

Step 1 ?2: judge active power of wind power field predicated error is as which kind of distribution by hypothesis testing in mathematical statistics, and set up active power of wind power field predicated error distribution function;

Described step 1 ?2 to comprise the following steps:

Whether step 1 ?2 ?1: suppose that active power of wind power field predicated error is normal distribution, check this hypothesis to set up:

Utilize maximum-likelihood method estimation error average and variance, calculate interval frequency and the statistic test value in each interval if meet test value in acceptance region, then determine that active power of wind power field predicated error is normal distribution;

Step 1 ?2 ?2: set up active power of wind power field predicated error distribution function;

Active power of wind power field predicated error distribution function is described as:

$f\left(e_i\right)=\frac{1}{\mathrm{\σ}\sqrt{2\mathrm{\π}}}{e}^{-\frac{{(e_i-\mathrm{\μ})}^2}{2{\mathrm{\ρ}}^2}}---\left(2\right)$

Wherein: μ is wind energy turbine set anticipation error, σ is the standard deviation of active power predicated error;

Then at time point t place, the active power predicated error desired value E of wind energy turbine set _i(t) be:

E _i(t)＝e(t) (3)

Wherein: t is current time, e (t) is t active power of wind power field predicated error;

In described step 2, active power of wind power field predicated error Normal Distribution is normal distribution, active power predicated error average and the variance of wind energy turbine set is can be calculated by historical data, because each future position of wind energy turbine set is discrete, calculate the active power predicated error of each time point of wind energy turbine set according to active power of wind power field predicated error distribution function, obtain active power of wind power field predicated error distributed model thus;

In described step 3, for active power of wind power field predicated error, select confidence level, and then the wind energy turbine set calculated under this confidence level expects confidential interval, is expressed as:

P{e _min＜e＜e _max}≥1-α (4)

Wherein: e _minfor lower limit of confidence interval, e _maxfor the confidential interval upper limit, 1-α is confidence level, (e _min, e _max) be confidential interval;

In described step 4, for the wind energy turbine set comprising N platform wind power generating set, Wind turbines is divided into M group by the running status according to unit each in wind energy turbine set;

The distribution of comprehensive active power of wind power field predicated error and confidence level, the choosing of confidential interval, set up active power of wind power field Controlling model, object function is:

$\min f_i=|\underset{j=1}{\overset{M}{\mathrm{\Σ}}}({\mathrm{\ξ}}_{\mathrm{ij}}+\frac{{\mathrm{\λ}}_{\mathrm{ij}}}{2})P_{\mathrm{group}\_\mathrm{ij}}-P_{\mathrm{predict}\_i}\·E_i|---\left(5\right)$

Wherein: f _ibe that the i-th time point wind energy turbine set desired output power and history error distribute and expect difference, ξ _ijfor the weight coefficient of a jth group wind-powered electricity generation group of planes i-th time point; λ _ijfor the fluctuation weight coefficient of a jth group wind-powered electricity generation group of planes i-th time point; P _{group_ij}for the prediction generated output of a jth group wind-powered electricity generation group of planes i-th time point; P _{predict_i}it is the wind energy turbine set predicted power of the i-th time point; E _ifor the active power predicated error desired value of each time point of wind energy turbine set;

Statistical history data, were divided into n interval, calculate the active power predicated error desired value of each time point of wind energy turbine set, thus obtain each interval error e by one day _idistribution;

The active power predicated error desired value E that integration obtains each time point of wind energy turbine set is asked for each interval _i, have

$E_i={\∫}_{t_{i-1}}^{t_i}{e}_i\mathrm{dt}---\left(6\right)$

Ask for f _iduring minimum of a value, generated output is divided into by variable bound to retrain and weight coefficient constraint;

Generated output is constrained to:

$\left\{\begin{array}{c}\underset{j=1}{\overset{M}{\mathrm{\Σ}}}({\mathrm{\ξ}}_{\mathrm{ij}}+{\mathrm{\λ}}_{\mathrm{ij}})P_{\mathrm{group}\_\mathrm{ij}}=P_{\mathrm{predict}\_i}\×(1+e_{+})\\ \underset{j=1}{\overset{M}{\mathrm{\Σ}}}{\mathrm{\ξ}}_{\mathrm{ij}}{P}_{\mathrm{group}\_\mathrm{ij}}=P_{\mathrm{predict}\_i}\×(1-e_{-})\\ ({\mathrm{\ξ}}_{\mathrm{ij}}+{\mathrm{\λ}}_{\mathrm{ij}})P_{\mathrm{group}\_\mathrm{ij}}\≤\underset{j=1}{\overset{M}{\mathrm{\Σ}}}{Q}_{\mathrm{ij}}{P}_{\mathrm{gen}}\end{array}\right.,i=1,\mathrm{2,3}...,N;j=\mathrm{1,2},...,M---\left(7\right)$

Wherein: e ₊for predicted power confidence upper limit; e _-for predicted power confidence lower limit; Q _ijfor a jth group of planes i-th moment unit number; P _genfor Wind turbines rated power;

In formula (7), the equality constraint equal sign left side represents the general power of each wind turbine cohort, represents the power bound of wind energy turbine set on the right of equal sign; The inequality constraints equal sign left side represents the general power of each wind turbine cohort, represents the power upper limit of each machine group on the right of equal sign;

Weight coefficient is constrained to:

0≤ξ _ij＜1，i＝1,2,3...,N；j＝1,2,...,M (8)

0≤λ _ij＜1，i＝1,2,3...,N；j＝1,2,...,M (9)；

In described step 5, according to determined object function and variable bound, described active power of wind power field Controlling model is optimized, obtains the optimization confidence weight of each machine group, and then control active power of wind power field;

Optimizing process is expressed as Mathematical Modeling, has:

$\left.\begin{array}{c}s.t.\min f(u,x)\\ h(u,x)=0\\ g(u,x)\≥0\end{array}\right\}---\left(10\right)$

Wherein: u and x is state variable, and state variable x comprises the weight coefficient of each machine group; F is that wind energy turbine set desired output power and history error distribute and expect difference, h and g is respectively equality constraints functions and inequality constraints function.