CN105825002A - Method for modeling dynamic equivalence of wind power farm based on dynamic grey-relevancy analysis method - Google Patents

Abstract

The invention relates to a method for modeling dynamic equivalence of a wind power farm based on a dynamic grey-relevancy analysis method. The method comprises the following steps: surveying relevancy between any time data value of actually measured operating data and historical data during actually measured operating data preprocessing by adopting autocorrelation analysis, mining an implicit rule in the data, selecting reasonable time span, and further determining small data samples at K moments in the span to serve as simulation model input according to different wind speed scales, so that the influence of simulation model precision and analysis and calculation duration due to too few data samples or multifarious redundancy is reduced. Moreover, an association degree matrix is established to serve as a clustering index by adopting the dynamic grey-relevancy analysis method, and complex relation and grey property in an operating state of each wind generation set are considered into the fleet clustering process, so that the fleet clustering result is reasonable, and the accuracy of a dynamic equivalence model of the wind power farm is greatly improved.

Description

A kind of wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis

Technical field

The present invention relates to regenerative resource interconnection technology field, more particularly, it relates to a kind of wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis.

Background technology

Wind-power electricity generation is big owing to having installed capacity growth space, and the advantages such as cost declines fast, safety, the energy never exhaustion are increasingly paid attention to by countries in the world.But wind energy has randomness, intermittent and instable feature, along with Wind turbines single-machine capacity and the continuous expansion of wind energy turbine set scale, wind-electricity integration is the most notable on the impact of stability of power system.According to incompletely statistics, within only 2011, just there occurs 193 Wind turbines off-grid accidents, wherein 2.24 largest Jiuquan large-scale wind power off-grid accidents of China's generation in 2011, cause 598 typhoon group of motors off-grids altogether, loss is exerted oneself and is reached 840.43MW, northwest major network frequency the most as little as 49.854Hz.

Therefore, in order to accurately analyze and evaluate the interaction between high-capacity wind power plant and power system and impact, study and to seek suitable wind energy turbine set Dynamic Equivalence significant to the power system dynamic stability simulation analysis containing wind energy turbine set.

In prior art, wind energy turbine set Dynamic Equivalence is broadly divided into unit method of equivalents and multimachine method of equivalents.

Wherein, unit method of equivalents modeling process is simple; usually assume that the input wind speed of all Wind turbines is identical; it is that a typhoon group of motors just may be used by whole wind energy turbine set equivalence; but for large-scale wind energy turbine set, due to topography and geomorphology and wake effect and the impact of time lag, wind speed skewness in wind energy turbine set; the wind speed of Wind turbines differs greatly, and uses unit equivalence method to usually there will be bigger error.

In multimachine equivalence, the most frequently used and simplest method is based on arrangement position and wind energy turbine set carries out classification polymerization, often will be equivalent to a wind energy conversion system with exhaust blower, take into account the difference of fan operation situation between different row, reduce the error that wind energy turbine set unit dynamic equivalent model brings, but in actual wind energy turbine set, even if also likely to be present bigger wind speed difference with exhaust blower.

In prior art, relevant Equivalent Model is to all data sample equivalent processes, and input vector is many, the numerous and diverse redundancy of data sample, be unfavorable for analyzing, and simulation accuracy, simulation analysis time all can be caused certain harmful effect by this；On the other hand, the relevant Equivalent Model of prior art typically carries out unit equivalence division using certain quantity of state independent of each typhoon group of motors or whole monitoring project amounts as Clustering Model input, relatedness between each running of wind generating set situation and grey majorized model are lacked and considers, and the running status of each typhoon group of motors is different and non-deterministic by the influence degree of the many factors such as weather, grid operating conditions, temperature, is a complicated non-linear process.

Thus, it is comprehensive not with substantially that the dependency of running status between Wind turbines is considered by above-mentioned unit equivalence division methods, can affect the reasonability of cluster result and reduce simulation accuracy.Therefore, how the relatedness between each running of wind generating set situation and grey majorized model are taken into account based on actual measurement service data and combine clustering algorithm the modeling of wind energy turbine set dynamic equivalent in be the problem that need to solve further.

Summary of the invention

It is an object of the invention to overcome the deficiencies in the prior art, a kind of wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis is provided, reduce that factor data sample is very few or numerous and diverse redundancy, reduce the very few of factor data sample or numerous and diverse redundancy to phantom precision and the impact of analytical calculation duration；Use dynamic Gray Association Analysis to set up degree of association matrix as clustering target, relatedness complicated between each running of wind generating set situation and grey majorized model are taken into account in group of planes cluster process so that group of planes cluster result is more reasonable.

Technical scheme is as follows:

A kind of wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis, step is as follows:

Step 1) according to wind speed scale, determine the small data sample in wind energy turbine set actual operating data sample；

Step 2) set up the wind-powered electricity generation group of planes running status degree of association based on dynamic Gray Association Analysis matrix, and as clustering target；

Step 3) use K-means clustering algorithm to carry out group of planes cluster, obtain group of planes cluster result；

Step 4) use the typhoon group of motors in a group of planes that a group of planes is carried out equivalence, set up the dynamic equivalent model of wind energy turbine set.

As preferably, step 1) in, use autocorrelation analysis that wind energy turbine set actual measurement service data is carried out pretreatment, particularly as follows:

1.1) having Wind turbines altogether n platform in setting wind energy turbine set, choose actual measurement service data that wind energy turbine set is incorporated into the power networks in time period T as data sample, sampling number is i, and sample frequency is f；

1.2) wind speed in time period T is surveyed service data as time series, the autocorrelation coefficient ρ of the sequence obtained after calculating this time series and its Δ t translation in time_Δt:

${\mathrm{\ρ}}_{\mathrm{\Δ}t}=\frac{\mathrm{cov}(y_t,y_{t+\mathrm{\Δ}t})}{\sqrt{\mathrm{var}\left(y_t\right)\·\mathrm{var}\left(y_{t+\mathrm{\Δ}t}\right)}};$

Wherein, y_tFor the wind series of t, y_t+ΔtThe sequence obtained after translating for t wind series Δ t in time；

1.3) with selected Δ t as time span, in intercepting time period T, wind speed changes maximum actual measurement service data sample in default wind speed range, according to default wind speed scale, in the data sample with Δ t as time span intercepted, choose the small data sample in K moment.

As preferably, step 1.2) in, after more than Δ t time span, the autocorrelation coefficient of wind series drops to below predetermined coefficient.

As preferably, step 2) in, set up the wind-powered electricity generation group of planes running status degree of association based on dynamic Gray Association Analysis matrix particularly as follows:

2.1) set wind energy turbine set and include m typhoon group of motors, gather n continuous quantity monitoring project data, particularly as follows:

{x_ij∈(S_i, X_j, T_t)}_m*n*N；

Wherein, i=1,2 ..., m；J=1,2 ..., n；T=1,2 ..., N；x_ijT () represents the jth continuous quantity monitoring project data of the t moment the i-th Fans；S_iRepresent the operation conditions of the i-th typhoon group of motors, X in certain moment wind energy turbine set_jRepresent selected jth continuous quantity monitoring project, T_tRepresent the t selected moment；N is moment number；

2.2) definition A₁,A₂,...,A_nFor the continuous quantity monitoring project information matrix of running of wind generating set situation the most in the same time, the data comprised in matrix are small data sample selected in step 1, particularly as follows:

$A_1=\left(\begin{array}{cccc}{x}_{11}\left(1\right)& x_{12}\left(1\right)& ...& x_{1n}\left(1\right)\\ x_{21}\left(1\right)& x_{22}\left(1\right)& ...& x_{2n}\left(1\right)\\ ...& ...& ...& ...\\ x_{m1}\left(1\right)& x_{m2}\left(1\right)& ...& x_{mn}\left(1\right)\end{array}\right)$

$A_2=\left(\begin{array}{cccc}{x}_{11}\left(2\right)& x_{12}\left(2\right)& ...& x_{1n}\left(2\right)\\ x_{21}\left(2\right)& x_{22}\left(2\right)& ...& x_{2n}\left(2\right)\\ ...& ...& ...& ...\\ x_{m1}\left(2\right)& x_{m2}\left(2\right)& ...& x_{mn}\left(2\right)\end{array}\right)$

...

$A_N=\left(\begin{array}{cccc}{x}_{11}\left(N\right)& x_{12}\left(N\right)& ...& x_{1n}\left(N\right)\\ x_{21}\left(N\right)& x_{22}\left(N\right)& ...& x_{2n}\left(N\right)\\ ...& ...& ...& ...\\ x_{m1}\left(N\right)& x_{m2}\left(N\right)& ...& x_{mn}\left(N\right)\end{array}\right);$

2.3) Largest Mean calculated on the continuous quantity monitoring project time dimension in each moment is:

$M_0=\[m_1,m_2,\mathrm{...},m_n\]=\[\underset{t}{\max }\left(\underset{i}{mean}\right(x_{i1}\left(t\right)\left)\right),\underset{t}{\max }\left(\underset{i}{mean}\right(x_{i2}\left(t\right)\left)\right),\mathrm{...},\underset{t}{\max }\left(\underset{i}{mean}\right(x_{in}\left(t\right)\left)\right)\];$

2.4) to small data sample selected in step 1 without guiding principle quantification treatment, particularly as follows:

${\left\{x_{ij}^{\′}\left(t\right)\right\}}_{m*n*N}=\[x_{i1}^{\′}\left(t\right),x_{i2}^{\′}\left(t\right),\mathrm{...},x_{in}^{\′}\left(t\right)\]=\[\frac{x_{i1}\left(t\right)}{m_1},\frac{x_{i2}\left(t\right)}{m_2},\mathrm{...},\frac{x_{in}\left(t\right)}{m_n}\];$

2.5) comparative sequences of certain moment separate unit running of wind generating set situation is:

S_i(t)={ x '_ij(t)}_m*n*N=[x '_i1(t),x′_i2(t),...,x′_in(t)]；

2.6) the dynamic reference sequence of all Wind turbines integrated operation situations is:

$s_0\left(T\right)=\[x_1\left(T\right),x_2\left(T\right),\mathrm{..},x_n\left(T\right)\]=\[\underset{t}{\max }\underset{i}{\max }\left(x_{i1}^{\′}\right(t\left)\right),\underset{t}{\max }\underset{i}{\max }\left(x_{i2}^{\′}\right(t\left)\right),\mathrm{...},\underset{t}{\max }\underset{i}{\max }\left(x_{in}^{\′}\right(t\left)\right)\];$

2.7) expression each moment, the comparative sequences S of each running of wind generating set situation are calculated_i(t) and reference system S₀(T) grey relational grade between is:

${\mathrm{\γ}}_{0i}=\frac{1}{n}\underset{j=1}{\overset{n}{\Σ}}{\mathrm{\γ}}_{0i}\left(j\right);$

Wherein, γ_0iJ () is grey incidence coefficient, as follows:

${\mathrm{\γ}}_{0i}\left(j\right)=\frac{m+\mathrm{\ρ}M}{{\mathrm{\Δ}}_i\left(j\right)+\mathrm{\ρ}M};$

Wherein, two-stage lowest differenceTwo-stage maximum differenceρ is resolution ratio, Δ_0i(j)=| x_j(T)-x′_ij(t)|。

2.8) grey relational grade matrix is:

$Y=\left(\begin{array}{cccc}{y}_1& y_2& ...& y_i...y_m\end{array}\right)={\left({\mathrm{\γ}}_{ti}\right)}_{N\×m}=\left[\begin{array}{cccc}{\mathrm{\γ}}_{11}& {\mathrm{\γ}}_{12}& ...& {\mathrm{\γ}}_{1m}\\ {\mathrm{\γ}}_{21}& {\mathrm{\γ}}_{22}& ...& {\mathrm{\γ}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{\γ}}_{N1}& {\mathrm{\γ}}_{N2}& ...& {\mathrm{\γ}}_{Nm}\end{array}\right];$

Wherein, i represents the i-th typhoon group of motors, and t represents the t time cross-section, γ_tiRepresent the gray relation grades between t the i-th typhoon group of motors operation conditions and all Wind turbines integrated operation situations, y_iRepresent the gray relation grades change sequence of the i-th typhoon group of motors operation conditions with wind speed change of rank in Δ t time span.

As preferably, step 3) particularly as follows:

3.1) using grey relational grade matrix as the input data of K-means Clustering Model, wherein the gray relation grades change sequence of every typhoon group of motors operation conditions is an input vector:

Y=(y₁y₂...y_i...y_m)=(γ_ti)_N×m；

3.2) h sample is chosen as initial clustering point；

3.3) center of gravity of each class is calculated, as cluster centre:

$d_{ij}={\[\underset{t=1}{\overset{N}{\Σ}}{({\mathrm{\γ}}_{ti}-{\mathrm{\γ}}_{tj})}^2\]}^{1/2};$

3.4) other samples are included into the class minimum with its distance；

3.5) repeat step 3.2), 3.3), till all samples all can not distribute；

3.6) profile value S (i) is calculated, particularly as follows:

$S\left(i\right)=\frac{min\left(b\right)-a}{max\[a,min\left(b\right)\]},t=1,2,...,n;$

Wherein: a is the average distance between sample i and other samples of same bunch；B is a vector, its element be sample i from different bunches bunch in average distance between sample；

The threshold value of predetermined S (i), if S (i) can not meet condition, the most again chooses initial clustering point and clusters, until S (i) meets condition；If all of initial clustering point all can not meet, then re-enter h value, cluster.

As preferably, step 4) particularly as follows: be a typhoon group of motors by the Wind turbines equivalence in same group, and use capacity weighting method to calculate the parameter of equivalent Wind turbines, set up the dynamic equivalent model of wind energy turbine set.

As preferably, step 1) in, actual measurement service data obtains connection amount monitoring project data by SCADA system output.

As preferably, SCADA system connection amount monitoring project data include: A phase voltage surveyed by generator drive end end bearing temperature, gearbox input shaft temperature, cabin cabinet temperature, net, A phase current surveyed by electromotor anti-drive end end bearing temperature, gearbox output shaft temperature, tower base cabinet temperature, net, generator windings temperature, gear-box oil temperature, wind speed, active power, electromotor cooling air temperature, gear-box base bearing temperature, mean wind speed in 5 minutes, mains frequency, generator speed, cabin temperature, ambient temperature one or more.

Beneficial effects of the present invention is as follows:

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis of the present invention, when surveying service data pretreatment, use autocorrelation analysis that any time data value of actual measurement service data is investigated with the relatedness of historical data, excavate out the implicit rule in data, select rational time span, and then in span, determine according to different wind speed scales that the small data sample in K moment as phantom input, decreases the very few of factor data sample or numerous and diverse redundancy to phantom precision and the impact of analytical calculation duration.

On the other hand, the present invention uses dynamic Gray Association Analysis to set up degree of association matrix as clustering target, relatedness complicated between each running of wind generating set situation and grey majorized model are taken into account in group of planes cluster process, make group of planes cluster result more reasonable, substantially increase the accuracy of wind energy turbine set dynamic equivalent model.

Accompanying drawing explanation

Fig. 1 is wind energy turbine set dynamic equivalent modeling method flow chart based on dynamic Gray Association Analysis；

Fig. 2 is certain detailed phantom of actual wind energy turbine set of embodiment；

Fig. 3 is the IEEE39 node system of embodiment；

When Fig. 4 is gust disturbances, active power dynamic response curve comparison diagram when wind energy turbine set uses unit Equivalent Model, present invention wind energy turbine set based on dynamic Gray Association Analysis dynamic equivalent model and detailed model；

When Fig. 5 is gust disturbances, reactive power dynamic response curve comparison diagram when wind energy turbine set uses unit Equivalent Model, present invention wind energy turbine set based on dynamic Gray Association Analysis dynamic equivalent model and detailed model；

When Fig. 6 is system side generation three-phase ground short trouble, active power dynamic response curve comparison diagram when wind energy turbine set uses unit Equivalent Model, present invention wind energy turbine set based on dynamic Gray Association Analysis dynamic equivalent model and detailed model；

When Fig. 7 is system side generation three-phase ground short trouble, reactive power dynamic response curve comparison diagram when wind energy turbine set uses unit Equivalent Model, present invention wind energy turbine set based on dynamic Gray Association Analysis dynamic equivalent model and detailed model.

Detailed description of the invention

Below in conjunction with drawings and Examples, the present invention is described in further detail.

The present invention is to solve all deficiencies that prior art exists, it is provided that a kind of wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis, as it is shown in figure 1, step is as follows:

Step 1) according to wind speed scale, determine the small data sample in wind energy turbine set actual operating data sample；

Step 2) set up the wind-powered electricity generation group of planes running status degree of association based on dynamic Gray Association Analysis matrix, and as clustering target；

Step 3) use K-means clustering algorithm to carry out group of planes cluster, obtain group of planes cluster result；

Step 4) use the typhoon group of motors in a group of planes that a group of planes is carried out equivalence, set up the dynamic equivalent model of wind energy turbine set.

Step 1) in, use autocorrelation analysis that wind energy turbine set actual measurement service data is carried out pretreatment, particularly as follows:

1.1) having Wind turbines altogether n platform in setting wind energy turbine set, choose actual measurement service data that wind energy turbine set is incorporated into the power networks in certain time period T as sample, sampling number is i, and sample frequency is 1 beat/min.Actual measurement service data includes the SCADA system continuous quantity monitoring project of 19 energy general performance wind energy turbine set real-time operating conditions, as shown in table 1.

The SCADA system continuous quantity monitoring project of table 1:19 item energy general performance wind energy turbine set real-time operating conditions

1.2) consider that wind speed change is the basic reason of fan parameter change, therefore use wind speed as the foundation of preprocessing process actual measurement service data by stages.Wind speed in time period T is surveyed service data as time series, the autocorrelation coefficient ρ of the sequence obtained after calculating this time series and its Δ t translation in time_Δt:

${\mathrm{\ρ}}_{\mathrm{\Δ}t}=\frac{\mathrm{cov}(y_t,y_{t+\mathrm{\Δ}t})}{\sqrt{\mathrm{var}\left(y_t\right)\·\mathrm{var}\left(y_{t+\mathrm{\Δ}t}\right)}};$

Wherein, y_tFor the wind series of t, y_t+ΔtThe sequence obtained after translating for t wind series Δ t in time；

Owing to it is generally acknowledged that before current data and Δ t, the dependency of data variation is more weak when autocorrelation coefficient is less than 0.3.Therefore, taking certain selected Δ t is that wind energy turbine set actual measurement service data is analyzed by time span data intercept sample.The condition that this Δ t should meet is: after more than this Δ t time span, the autocorrelation coefficient of wind series drops to less than 0.3；

1.3) with selected Δ t as time span, in intercepting time period T, wind speed changes maximum actual measurement service data sample in default wind speed range (3m/s～21m/s), according to default wind speed scale, in the data sample with Δ t as time span intercepted, choose the small data sample in K moment.In the present embodiment, the division of wind speed scale is particularly as follows: mild wind (3.4m/s～5.4m/s) and wind (5.5m/s～7.9m/s), strong wind (8.0m/s～10.7m/s), high wind (10.8m/s～13.8m/s), strong wind (13.9m/s～17.1m/s) and strong wind (17.2m/s～20.7m/s).Then in the small data sample in this K moment, the air speed data of every typhoon group of motors should at least include front 4 wind speed scales.

Step 2) in, set up the wind-powered electricity generation group of planes running status degree of association based on dynamic Gray Association Analysis matrix particularly as follows:

2.1) set wind energy turbine set and include m typhoon group of motors as a multisystem, multiple index evaluation system, i.e. wind energy turbine set, gather n continuous quantity monitoring project data, particularly as follows:

{x_ij∈(S_i, X_j, T_t)}_m*n*N；

Wherein, i=1,2 ..., m；J=1,2 ..., n；T=1,2 ..., N；x_ijT () represents the jth continuous quantity monitoring project data of the t moment the i-th Fans；S_iRepresent the operation conditions of the i-th typhoon group of motors, X in certain moment wind energy turbine set_jRepresent step 1) in selected jth continuous quantity monitoring project, T_tRepresent step 1) in selected the t moment；N is moment number；

2.2) definition A₁,A₂,...,A_nFor the SCADA system continuous quantity monitoring project information matrix of running of wind generating set situation the most in the same time, the data comprised in matrix are small data sample selected in step 1, particularly as follows:

$A_1=\left(\begin{array}{cccc}{x}_{11}\left(1\right)& x_{12}\left(1\right)& ...& x_{1n}\left(1\right)\\ x_{21}\left(1\right)& x_{22}\left(1\right)& ...& x_{2n}\left(1\right)\\ ...& ...& ...& ...\\ x_{m1}\left(1\right)& x_{m2}\left(1\right)& ...& x_{mn}\left(1\right)\end{array}\right)$

$A_2=\left(\begin{array}{cccc}{x}_{11}\left(2\right)& x_{12}\left(2\right)& ...& x_{1n}\left(2\right)\\ x_{21}\left(2\right)& x_{22}\left(2\right)& ...& x_{2n}\left(2\right)\\ ...& ...& ...& ...\\ x_{m1}\left(2\right)& x_{m2}\left(2\right)& ...& x_{mn}\left(2\right)\end{array}\right)$

...

$A_N=\left(\begin{array}{cccc}{x}_{11}\left(N\right)& x_{12}\left(N\right)& ...& x_{1n}\left(N\right)\\ x_{21}\left(N\right)& x_{22}\left(N\right)& ...& x_{2n}\left(N\right)\\ ...& ...& ...& ...\\ x_{m1}\left(N\right)& x_{m2}\left(N\right)& ...& x_{mn}\left(N\right)\end{array}\right);$

2.3) Largest Mean calculated on the SCADA system continuous quantity monitoring project time dimension in each moment is:

$M_0=\[m_1,m_2,\mathrm{...},m_n\]=\[\underset{t}{\max }\left(\underset{i}{mean}\right(x_{i1}\left(t\right)\left)\right),\underset{t}{\max }\left(\underset{i}{mean}\right(x_{i2}\left(t\right)\left)\right),\mathrm{...},\underset{t}{\max }\left(\underset{i}{mean}\right(x_{in}\left(t\right)\left)\right)\];$

2.4) to small data sample selected in step 1 without guiding principle quantification treatment, particularly as follows:

${\left\{x_{ij}^{\′}\left(t\right)\right\}}_{m*n*N}=\[x_{i1}^{\′}\left(t\right),x_{i2}^{\′}\left(t\right),\mathrm{...},x_{in}^{\′}\left(t\right)\]=\[\frac{x_{i1}\left(t\right)}{m_1},\frac{x_{i2}\left(t\right)}{m_2},\mathrm{...},\frac{x_{in}\left(t\right)}{m_n}\];$

2.5) comparative sequences of certain moment separate unit running of wind generating set situation is:

S_i(t)={ x '_ij(t)}_m*n*N=| x '_i1(t),x′_i2(t),...,x′_in(t)]；

2.6) the dynamic reference sequence of all Wind turbines integrated operation situations is:

$s_0\left(T\right)=\[x_1\left(T\right),x_2\left(T\right),\mathrm{..},x_n\left(T\right)\]=\[\underset{t}{\max }\underset{i}{\max }\left(x_{i1}^{\′}\right(t\left)\right),\underset{t}{\max }\underset{i}{\max }\left(x_{i2}^{\′}\right(t\left)\right),\mathrm{...},\underset{t}{\max }\underset{i}{\max }\left(x_{in}^{\′}\right(t\left)\right)\];$

2.7) expression each moment, the comparative sequences S of each running of wind generating set situation are calculated_i(t) and reference system S₀(T) grey relational grade between is:

${\mathrm{\γ}}_{0i}=\frac{1}{n}\underset{j=1}{\overset{n}{\Σ}}{\mathrm{\γ}}_{0i}\left(j\right);$

Wherein, γ_0iJ () is grey incidence coefficient, as follows:

${\mathrm{\γ}}_{0i}\left(j\right)=\frac{m+\mathrm{\ρ}M}{{\mathrm{\Δ}}_i\left(j\right)+\mathrm{\ρ}M};$

Wherein, two-stage lowest differenceTwo-stage maximum differenceρ is resolution ratio, in the present embodiment, ρ=0.5, Δ_0i(j)=| x_j(T)-x′_ij(t)|。

2.8) grey relational grade matrix is:

$Y=\left(\begin{array}{cccc}{y}_1& y_2& ...& y_i...y_m\end{array}\right)={\left({\mathrm{\γ}}_{ti}\right)}_{N\×m}=\left[\begin{array}{cccc}{\mathrm{\γ}}_{11}& {\mathrm{\γ}}_{12}& ...& {\mathrm{\γ}}_{1m}\\ {\mathrm{\γ}}_{21}& {\mathrm{\γ}}_{22}& ...& {\mathrm{\γ}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{\γ}}_{N1}& {\mathrm{\γ}}_{N2}& ...& {\mathrm{\γ}}_{Nm}\end{array}\right];$

Wherein, i represents the i-th typhoon group of motors, and t represents the t time cross-section, γ_tiRepresent the gray relation grades between t the i-th typhoon group of motors operation conditions and all Wind turbines integrated operation situations, y_iRepresent the gray relation grades change sequence of the i-th typhoon group of motors operation conditions with wind speed change of rank in Δ t time span.

Step 3) particularly as follows:

3.1) the grey relational grade matrix calculated in selecting step 2 is as the input data of K-means Clustering Model, and wherein the gray relation grades change sequence of every typhoon group of motors operation conditions is an input vector:

Y=(y₁y₂...y_i...y_m)=(γ_ti)_N×m；

3.2) h sample is chosen as initial clustering point；

3.3) center of gravity of each class is calculated, as cluster centre:

$d_{ij}={\[\underset{t=1}{\overset{N}{\Σ}}{({\mathrm{\γ}}_{ti}-{\mathrm{\γ}}_{tj})}^2\]}^{1/2};$

3.4) other samples are included into the class minimum with its distance；

3.5) repeat step 3.2), 3.3), till all samples all can not distribute；

3.6) profile value S (i) is calculated, particularly as follows:

$S\left(i\right)=\frac{min\left(b\right)-a}{max\[a,min\left(b\right)\]},t=1,2,...,n;$

Wherein: a is the average distance between sample i and other samples of same bunch；B is a vector, its element be sample i from different bunches bunch in average distance between sample；

The threshold value of predetermined S (i), if S (i) can not meet condition, the most again chooses initial clustering point and clusters, until S (i) meets condition；If all of initial clustering point all can not meet, then re-enter h value, cluster.

Step 4) particularly as follows: be a typhoon group of motors by the Wind turbines equivalence in same group, and use capacity weighting method to calculate the parameter of equivalent Wind turbines, set up the dynamic equivalent model of wind energy turbine set.

As a example by being below the actual enforcement of actual wind energy turbine set:

On power system full digital trigger technique device (ADPSS) platform, wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis to the present invention carries out emulating explanation.As in figure 2 it is shown, have 22 typhoon group of motors in this wind energy turbine set, the type of unit is GE1.5MW, and wind energy turbine set total installation of generating capacity is 33MW.

19 the selected SDADA system continuous quantity monitoring project actual measurement service datas chosen in two weeks on October 15,1 day to 2014 October in 2014 are analyzed.Wind turbines in field is used wind energy turbine set dynamic equivalent based on the dynamic Gray Association Analysis modeling of the present invention to be modeled this wind energy turbine set by the actual measurement service data using this period.When analyzing gust disturbances with system side generation three-phase ground short trouble, unit Equivalent Model, present invention wind energy turbine set based on dynamic Gray Association Analysis dynamic equivalent model and the dynamic response characteristic of detailed model.

Gust disturbances: setting the fitful wind 4s startup met to wind energy turbine set, 8s terminates, and fitful wind maximum is 3m/s, as shown in Figure 4, Figure 5.Can find under wind speed disturbance by contrast, the wind energy turbine set dynamic equivalent model that the present invention sets up is closer to reactive power dynamic response curve with the active power of detailed model.

System side three-phase ground short trouble: use IEEE39 node system (as shown in Figure 3) as wind farm grid-connected system, 35 nodes of wind energy turbine set interventional systems in emulation experiment.If IEEE39 node system is when t=1.00s, there is three-phase ground short trouble in bus 24, Failure elimination during t=1.12s.As shown in Figure 6, Figure 7, can find under system side generation three-phase ground short trouble by contrast, the wind energy turbine set dynamic equivalent model that the present invention sets up is closer to reactive power dynamic response curve with the active power of detailed model.

Wind speed disturbance and system side three-phase ground short trouble disturbance simulation result explanation, the small data sample using the data preprocessing method of the present invention to obtain is reasonable and effective, the data sample of numerous and diverse redundancy can be reduced, less on the dynamic response characteristic impact before and after modeling.Additionally, the dynamic Gray Association Analysis using the present invention to propose carries out group of planes cluster, grey majorized model complicated between running of wind generating set situation and relatedness are taken into account in clustering target all sidedly, with the accuracy that the contrast of unit Equivalent Model can effectively improve the modeling of wind energy turbine set dynamic equivalent.

Above-described embodiment is intended merely to the present invention is described, and is not used as limitation of the invention.As long as according to the technical spirit of the present invention, be changed above-described embodiment, modification etc. all will fall in the range of the claim of the present invention.

Claims (8)

1. a wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis, it is characterised in that step is as follows:

Step 1) according to wind speed scale, determine the small data sample in wind energy turbine set actual operating data sample；

Step 2) set up the wind-powered electricity generation group of planes running status degree of association based on dynamic Gray Association Analysis matrix, and as clustering target；

Step 3) use K-means clustering algorithm to carry out group of planes cluster, obtain group of planes cluster result；

Step 4) use the typhoon group of motors in a group of planes that a group of planes is carried out equivalence, set up the dynamic equivalent model of wind energy turbine set.

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis the most according to claim 1, it is characterised in that step 1) in, use autocorrelation analysis that wind energy turbine set actual measurement service data is carried out pretreatment, particularly as follows:

1.1) having Wind turbines altogether n platform in setting wind energy turbine set, choose actual measurement service data that wind energy turbine set is incorporated into the power networks in time period T as data sample, sampling number is i, and sample frequency is f；

1.2) wind speed in time period T is surveyed service data as time series, the autocorrelation coefficient ρ of the sequence obtained after calculating this time series and its Δ t translation in time_Δt:

${\mathrm{\ρ}}_{\mathrm{\Δ}t}=\frac{\mathrm{cov}(y_t,y_{t+\mathrm{\Δ}t})}{\sqrt{\mathrm{var}\left(y_t\right)\·\mathrm{var}\left(y_{t+\mathrm{\Δ}t}\right)}};$

Wherein, y_tFor the wind series of t, y_t+ΔtThe sequence obtained after translating for t wind series Δ t in time；

1.3) with selected Δ t as time span, in intercepting time period T, wind speed changes maximum actual measurement service data sample in default wind speed range, according to default wind speed scale, in the data sample with Δ t as time span intercepted, choose the small data sample in K moment.

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis the most according to claim 2, it is characterised in that step 1.2) in, after more than Δ t time span, the autocorrelation coefficient of wind series drops to below predetermined coefficient.

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis the most according to claim 1, it is characterised in that step 2) in, set up the wind-powered electricity generation group of planes running status degree of association based on dynamic Gray Association Analysis matrix particularly as follows:

2.1) set wind energy turbine set and include m typhoon group of motors, gather n continuous quantity monitoring project data, particularly as follows:

{x_ij∈(S_i,X_j,T_t)}_m*n*N；

Wherein, i=1,2 ..., m；J=1,2 ..., n；T=1,2 ..., N；x_ijT () represents the jth continuous quantity monitoring project data of the t moment the i-th Fans；S_iRepresent the operation conditions of the i-th typhoon group of motors, X in certain moment wind energy turbine set_jRepresent selected jth continuous quantity monitoring project, T_tRepresent the t selected moment；N is moment number；

2.2) definition A₁,A₂,...,A_nFor the continuous quantity monitoring project information matrix of running of wind generating set situation the most in the same time, the data comprised in matrix are small data sample selected in step 1, particularly as follows:

$A_1=\left(\begin{array}{cccc}{x}_{11}\left(1\right)& x_{12}\left(1\right)& ...& x_{1n}\left(1\right)\\ x_{21}\left(1\right)& x_{22}\left(1\right)& ...& x_{2n}\left(1\right)\\ ...& ...& ...& ...\\ x_{m1}\left(1\right)& x_{m2}\left(1\right)& ...& x_{mn}\left(1\right)\end{array}\right)$

$A_2=\left(\begin{array}{cccc}{x}_{11}\left(2\right)& x_{12}\left(2\right)& ...& x_{1n}\left(2\right)\\ x_{21}\left(2\right)& x_{22}\left(2\right)& ...& x_{2n}\left(2\right)\\ ...& ...& ...& ...\\ x_{m1}\left(2\right)& x_{m2}\left(2\right)& ...& x_{mn}\left(2\right)\end{array}\right)$

…

$A_N=\left(\begin{array}{cccc}{x}_{11}\left(N\right)& x_{12}\left(N\right)& ...& x_{1n}\left(N\right)\\ x_{21}\left(N\right)& x_{22}\left(N\right)& ...& x_{2n}\left(N\right)\\ ...& ...& ...& ...\\ x_{m1}\left(N\right)& x_{m2}\left(N\right)& ...& x_{mn}\left(N\right)\end{array}\right);$

2.3) Largest Mean calculated on the continuous quantity monitoring project time dimension in each moment is:

$M_0=\[m_1,m_2,\mathrm{...},m_n\]=\[\underset{t}{\max }\left(\underset{i}{mean}\left(x_{i1}\left(t\right)\right)\right),\underset{t}{\max }\left(\underset{i}{mean}\left(x_{i2}\left(t\right)\right)\right),\mathrm{...},\underset{t}{\max }\left(\underset{i}{mean}\left(x_{in}\left(t\right)\right)\right)\];$

2.4) to small data sample selected in step 1 without guiding principle quantification treatment, particularly as follows:

${\left\{x_{ij}^{\′}\left(t\right)\right\}}_{m*n*N}=\[x_{i1}^{\′}\left(t\right),x_{i2}^{\′}\left(t\right),\mathrm{...},x_{in}^{\′}\left(t\right)\]=\[\frac{x_{i1}\left(t\right)}{m_1},\frac{x_{i2}\left(t\right)}{m_2},\mathrm{...},\frac{x_{in}\left(t\right)}{m_n}\];$

2.5) comparative sequences of certain moment separate unit running of wind generating set situation is:

$S_i\left(t\right)={\left\{x_{ij}^{\′}\left(t\right)\right\}}_{m*n*N}=\[x_{i1}^{\′}\left(t\right),x_{i2}^{\′}\left(t\right),\mathrm{...},x_{in}^{\′}\left(t\right)\];$

2.6) the dynamic reference sequence of all Wind turbines integrated operation situations is:

$S_0\left(T\right)=\[x_1\left(T\right),x_2\left(T\right),\mathrm{...},x_n\left(T\right)\]=\[\underset{t}{\max }\underset{i}{max}\left(x_{i1}^{\′}\left(t\right)\right),\underset{t}{\max }\underset{i}{max}\left(x_{i2}^{\′}\left(t\right)\right),\mathrm{...},\underset{t}{\max }\underset{i}{max}\left(x_{in}^{\′}\left(t\right)\right)\];$

2.7) expression each moment, the comparative sequences S of each running of wind generating set situation are calculated_i(t) and reference system S₀(T) grey relational grade between is:

${\mathrm{\γ}}_{0i}=\frac{1}{n}\underset{j=1}{\overset{n}{\Σ}}{\mathrm{\γ}}_{0i}\left(j\right);$

Wherein, γ_0iJ () is grey incidence coefficient, as follows:

${\mathrm{\γ}}_{0i}\left(j\right)=\frac{m+\mathrm{\ρ}M}{{\mathrm{\Δ}}_i\left(j\right)+\mathrm{\ρ}M};$

Wherein, two-stage lowest differenceTwo-stage maximum differenceρ is resolution ratio, Δ_0i(j)=| x_j(T)-x′_ij(t)|。

2.8) grey relational grade matrix is:

$Y=\left(\begin{array}{cccc}{y}_1& y_2& \mathrm{...}& y_i...y_m\end{array}\right)={\left({\mathrm{\γ}}_{ti}\right)}_{N\×m}=\left[\begin{array}{cccc}{\mathrm{\γ}}_{11}& {\mathrm{\γ}}_{12}& ...& {\mathrm{\γ}}_{1m}\\ {\mathrm{\γ}}_{21}& {\mathrm{\γ}}_{22}& ...& {\mathrm{\γ}}_{2m}\\ ...& ...& ...& ...\\ {\mathrm{\γ}}_{N1}& {\mathrm{\γ}}_{N2}& ...& {\mathrm{\γ}}_{Nm}\end{array}\right];$

Wherein, i represents the i-th typhoon group of motors, and t represents the t time cross-section, γ_tiRepresent the gray relation grades between t the i-th typhoon group of motors operation conditions and all Wind turbines integrated operation situations, y_iRepresent the gray relation grades change sequence of the i-th typhoon group of motors operation conditions with wind speed change of rank in Δ t time span.

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis the most according to claim 4, it is characterised in that step 3) particularly as follows:

3.1) using grey relational grade matrix as the input data of K-means Clustering Model, wherein the gray relation grades change sequence of every typhoon group of motors operation conditions is an input vector:

Y=(y₁y₂…y_i…y_m)=(γ_ti)_N×m；

3.2) h sample is chosen as initial clustering point；

3.3) center of gravity of each class is calculated, as cluster centre:

$d_{ij}={\[\underset{t=1}{\overset{N}{\Σ}}{({\mathrm{\γ}}_{ti}-{\mathrm{\γ}}_{tj})}^2\]}^{1/2};$

3.4) other samples are included into the class minimum with its distance；

3.5) repeat step 3.2), 3.3), till all samples all can not distribute；

3.6) profile value S (i) is calculated, particularly as follows:

$S\left(i\right)=\frac{min\left(b\right)-a}{max\[a,min\left(b\right)\]},t=1,2,...,n;$

Wherein: a is the average distance between sample i and other samples of same bunch；B is a vector, its element be sample i from different bunches bunch in average distance between sample；

The threshold value of predetermined S (i), if S (i) can not meet condition, the most again chooses initial clustering point and clusters, until S (i) meets condition；If all of initial clustering point all can not meet, then re-enter h value, cluster.

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis the most according to claim 1, it is characterized in that, step 4) particularly as follows: be a typhoon group of motors by the Wind turbines equivalence in same group, and use capacity weighting method to calculate the parameter of equivalent Wind turbines, set up the dynamic equivalent model of wind energy turbine set.

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis the most according to claim 2, it is characterised in that step 1) in, actual measurement service data obtains connection amount monitoring project data by SCADA system output.

Wind energy turbine set dynamic equivalent modeling method based on dynamic Gray Association Analysis the most according to claim 7, it is characterized in that, SCADA system connection amount monitoring project data include: generator drive end end bearing temperature, gearbox input shaft temperature, cabin cabinet temperature, A phase voltage surveyed by net, electromotor anti-drive end end bearing temperature, gearbox output shaft temperature, tower base cabinet temperature, A phase current surveyed by net, generator windings temperature, gear-box oil temperature, wind speed, active power, electromotor cooling air temperature, gear-box base bearing temperature, mean wind speed in 5 minutes, mains frequency, generator speed, cabin temperature, one or more of ambient temperature.